Ningbo Univ 1 and 2

NINGBO DAXUE XUEBAO LIGONG BAN
25
2012
10
13
20
24
27
3
,
PETROV Y V, BRATOV V
35
41
BRATOV V, PETROV Y V, VOLKOV G

,
GREKOV M, MOROZOV N
,
48
53
60
64
BRAGOV A M, BALANDIN V V,
KONSTANTINOV A YU, KONSTANTINOVA YU V, LOMUNOV A K, FILIPPOV A R
70
GRUZDKOV A, PETROV Y
,
TiNi
,
RAZOV A
,
SMIRNOV I, PETROV Y, SUDENKOV Y
Radon-based ULM
,
KCaA
CO2
CN 33-1134/N 1988 q
A4
132
zh en
74
78
83
89
94
98
103
108
113
119
125
128
P 10.00
1000 25 2012-01
JOURNAL OF NINGBO UNIVERSITY

NATURAL SCIENCE AND ENGINEERING EDITION
Vol.25 No.1
Jan. 10, 2012
CONTENTS
Marine Aquatic Production & Biologic Technology
Effects of Dietary Supplementation with Compound Bacillus on Growth and Digestive Enzyme Activity in Portunus
Trituberculatus
DING Xue-yan, ZHOU Fan, HUANG Fu-yong, HE Feng, MENG Qing-hui ( 1 )
Analysis on Temperature Characteristics of the Litopenaeus vannamei Culture Ponds and Water Temperature Prediction
CEN Bo-ming, HU Zhou, CHEN Han-chun ( 7 )
Comparison of Nutritional Quality in Muscle of Oujiang Colour Carp (Cypinus carpio var. color) of 3 Different Sizes
XU Ru-wei, SHENTU Ji-kang, JIANG Jin-po, HUANG Fu-you, YAO Zi-liang ( 13 )
Comparison and Optimization of Total Lipid Detection Methods for Microalgae Chlorella
CONG Feng, SUN Xue, XU Nian-jun ( 20 )
Effects of Light Condition and Low Temperature on Survivability in Three Kinds of Ulva
HUANG Xian-jun, LUO Qi-jun ( 24 )
Mechanics Research (Column for Sino-Russia Workshop)

Rate-dependent Damage Evolution and its Influence on Dynamic Behavior of Materials at High Strain Rates
WANG Li-li, ZHOU Feng-hua, SUN Zi-jian, HUANG De-jin ( 27 )
Spatio-temporal Model of Multiscale Fracture in Brittle Solids
PETROV Y V, BRATOV V ( 35 )
Hierarchic Nature of Dynamic Deformation and Fracture of Rock Mass

QI Cheng-zhi, WANG Ming-yang, CHEN Jian-jie, QIAN Qi-hu ( 41 )
Existence of Optimal Energy Saving Parameters for Different Industrial Processes
BRATOV V, PETROV Y V, VOLKOV G ( 48 )
Dynamic Collapse of Honeycombs under In-plane Compressions
YU Tong-xi, GAO Zi-yang, HU Ling-ling ( 53 )
Surface Effects and Problems of Nanomechanics
GREKOV M, MOROZOV N ( 60 )
A Three-dimensional Viscoelastic Constitutive Model of Finite Deformation
SHEN Li-jun, BIAN Zhong-jing ( 64 )
Identification and Verification of Some Plasticity Models for Structural Materials by Using Pressure Bar Technique
BRAGOV A M, BALANDIN V V, KONSTANTINOV A YU, KONSTANTINOVA YU V, LOMUNOV A K, FILIPPOV A R ( 70 )
Stamp Collapse under Compression
WU Jian, HUANG Ke-zhi, HUANG Yong-gang ( 74 )
Incubation-time Based Approach for Dynamic Yielding of Metals
GRUZDKOV A, PETROV Y ( 78 )
An Interrupted Dynamic Tensile Testing for Pure Copper Bars

MA Dong-fang, CHEN Da-nian, WU Shan-xing, WANG Huan-ran, CAI Can-yuan ( 83 )
Effect of High Strain Rate on TiNi Shape Memory Alloys
RAZOV A ( 89 )
Hypervelocity Impact of the Icy Droplet on Al Shell at Nanoscale: A Molecular Dynamics Probe
YUAN Quan-zi, ZHAO Ya-pu ( 94 )
Threshold Characteristics of Pulse Loads Causing Fracture on an Example of Concrete and Rocks
SMIRNOV I, PETROV Y, SUDENKOV Y ( 98 )
Communication Engineering & Computer Application Technology

Research on Radon-based ULM Image Blur Direction Detection Algorithm
Thesaurus-based Ontological Construction Using Ocean Domain Ontology
HUANG Wen, ZHANG Xue-yan ( 103 )

ZHOU Jun-gen, LIU Bai-song ( 108 )
The Embedded Remote Detection System for Bridge Cable Force Based on ERA
ZHOU Hong-qiong, MIN Jie, ZHOU Yu, WANG Xiao-dong ( 113 )
Order Allocation Driven by Manufacturing Resources in Industrial Clusters

LIU Lan-lan, ZHANG Sheng, YE Fei-fan, LI Guo-fu ( 119 )
Development in Biology Research on Periplaneta fuliginosa
LIU Jian-fa, WANG Shi-ping, JIANG Wen-wen, ZHANG Jie-nan, LIN Lei, ZHOU Xi-wu, ZHOU Fei, HU Qi-feng ( 125 )
Synthesis of KCaA Molecular Sieve and its CO2 Adsorption Capacity
FAN Hai-yan, KONG Chun-long, XIA Hai-ping, CHEN Liang ( 128 )
25
, 2012 1
Vol.25 No.1, Jan. 2012
JOURNAL OF NINGBO UNIVERSITY ( NSEE )
310012
(Portunus trituberculatus Miers)

:
(T1)
21
(T2)
(T0);
; T1
15.38
T2
18.80 .
(D1)
, D1
D2
(D2)
14.4 ; D2
D1
3.4 .
.
;
: S963.73 9
;
:A
: 1001-5132
2012
01-0001-06
,
,
,
(Bacillus
sp.)
;
,
1
1.1
1.1.1
,
[1]
.
(Bacillus subtilis Natto)
(Portunus trituberculatus Miers),
(Bacillus subtilis Cohn),
,
,
[2]
10
110 cfug-1.
1.1.2
.
[3-6]
,
(
31 g
).
,
2011 08 16.
1966
25~35 g,
: http://3xb.nbu.edu.cn
,
2007BAD43B08
,
,
;
,
2007C22038 .
. E-mail: dingxy_sc@sina.com
2012
25000
2h
1.1.3
,1
, 65
,
1.
(
, T0
) 1.0
(T1
.
3~5 mm
2.0
(T2
80
,
(D1
(D0)
(D1)
,
43.7
6.3 .
0.33 hm2
0.33 hm2,
(D2). 3
667 m2
:
,
(D0
D2
4600~4800.
3 cm
2 ~7 .
),
,
30
23
39.92
1.3
6.35
1.3.1
12
7.47
23
<15
2~3 d
50 .
-80
:
,
7.37
, 10 000 rmin
4.95
.
;
1.35
-1
3 min
9
, 2 500 rmin-1
1.2
10 min,
1.2.1
1.3.2
3
6
70 cm,
0.8 m (
120 cm,
60 cm)
,
20
(31.180.89) g,
1
.
4
3
2
(T0 T1
l8:00,
, DO
pH
8.0, 37
.
,
:
0.003
22
24 h
29~32
(T0)
,
T2)
6:30
50
:
10
0.5
10 min,
,
,
16.4
13
, 11
21
1
1
29
0.1
4~5 d
30 ~ 50 ,
0.4 hm2
3 cm
10 ~15 ;
1.2.2
, 65
-20
1 d,
19
,4d
,
1 mol
-1
4.0 mgL .
37
,
37
10 mg
30 min,
;
37
1 g
18.80 ,
(P>0.05).
2
.
.
T0
T1
T2
/g
31.140.42
31.261.08
31.150.79
/g
43.750.86
44.311.27
43.990.80
72.507.50
80.005.00
82.502.50
69.402.82
72.301.99
71.204.53
40.510.88
41.800.81
41.231.03
2.440.12
2.330.17
2.230.08
2.340.28
1.980.12
1.900.11
1.4
( )=100
( )=2
/(
+
),
( )=100(
)/
,
)/(
),
( )=100(
2)/(
)).
16.0
2.2
SPSS
(T0)
(P<0.05),
, Tukeys
,
P<0.05.
(T1)
(T2)
(P>0.05). T0
(P<0.05), T1
T2
(P>0.05).
2.1
,
(P>0.05).
T2
2. 4
, 1
(T1)
(P<0.05),
(T2)
T0
T1
(P>0.05).
2.3
4.
(P>0.05).
,
0.05).
(P>
(D2
2
667 m
59.5 kg,
(D0)
58.5 kg;
15.38
,
(D1)
Umg-1
3
T0
T1
T2
9 171.51353.25a
3 747.792449.62b
6 227.98168.86b
7.590.25b
10.900.91a
12.534.15a
0.560.30
1.230.79
0.600.33
1.870.15b
3.702.03
:
(P<0.05).
5.433.25
.
3
D0 (
ab
)
)
D1 (
D2 (
+2
/hm
0.33
0.40
0.33
/kg
292.5
312.0
297.5
/kg
/cm
58.5
52.0
59.5
8.49
8.18
8.46
12.27
3.86
3.89
2012
14.4 .
, D2
(8.46 cm)
D1
D0
(8.18 cm)
3.3 ,
(8.49 cm)
,
43.6
119.0
46.8 ;
,
65.1
7.5 ,
[18]
(Carassais auratus gibebio)
,
[19]
(Pelodiscus sinensis)[20]
(Scophthalmus maximus)[21]
:
[7]
,
;
,
,
,
.
(Penaeus monodon)
,
[8-9]
[22]
(Macrobrachium rosenbergii)
(Oreochromis niloticus O.
[10]
(Litopenaeus
4
vannamei)
110 cfug
-1
aureus)
-1
510 cfug
[23]
.
0.2
[11]
(Homarus gammarus L.)
(Bacillus
licheniformi)
(Lates calcarifer)
,
,
[12]
[24]
(Bacillus licheniformis)
(Hyriopsis cumingi)
,
pH
,
6
,
,
1.010 cfumL
pH
30 d
,
[13]
15 d
,
. Mohapatra
(Labeo rohita)
,
,
VC VB
-1
[25]
VK
,
,
,
[14-15]
,
;
,
,
.
[16-17]
,
,
5
Effects of a probiotic bacterium on black tiger shrimp
Penaeus monodon survival and growth[J]. Aquaculture,
.
2
1998, 167:301-313.
[9] Boonthai T, Vuthiphandchai V, Nimrat S. Probiotic
[26]
.
idella)[27]
bacteria effects on growth and bacterial composition of
(Ctenopharyngodon
black tiger shrimp (Penaeus monodon)[J]. Aquaculture
(Sparus latus)[28]
Nutrition, 2011, doi: 10.1111/j.1365-2095.2011.00865.x.
[10] Keysami M A, Saad C R, Sijam K, et al. Effect of

Bacillus subtilis on growth development and survival of
larvae Macrobrachium rosenbergii (de Man)[J]. Aqua-
culture Nutrition, 2007, 13:131-136.
[11] Shen W Y, Fu L L, Li W F, et al. Effect of dietary

supplementation with Bacillus subtilis on the growth,
,
.
performance, immune response and antioxidant activities
of the shrimp (Litopenaeus vannamei)[J]. Aquaculture
Research, 2010, 41:1691-1698.

[12] Daniels C L, Merrield D L, Boothroyd D P, et al. Effect
.
,
of dietary Bacillus spp. and mannan oligosaccharides
(MOS) on European lobster (Homarus gammarus L.)
larvae growth performance, gut morphology and gut

microbiota[J]. Aquaculture, 2010, 304:49-57.
[13] Moriarty D. The role of microorganisms in aquaculture

ponds[J]. Aquaculture, 1997, 151:333-349.
[14] Gatesoupe F J. The use of probiotics in aquaculture[J].
Aquaculture, 1999, 180:147-165.
[1] Ring E, Olsen R E, GifstadI T , et al. Prebiotics in

aquaculture: A review[J]. Aquaculture Nutrition, 16:117-
[15] Ziaei-Nejad S, Rezaei M H, Takami G A, et al. The effect

of Bacillus spp. bacteria used as probiotics on digestive
136.
[2] Cui Z X, Liu Y, Luan W S, et al. Molecular cloning and
enzyme activity, survival and growth in the Indian white
characterization of a heat shock protein 70 gene in
shrimp Fenneropenaeus indicus[J]. Aquaculture, 2006,
swimming crab (Portunus trituberculatus)[J]. Fish &
252:516-524.
[16] Sakata T. Microora in the digestive tract of sh and
Shellsh Immunology, 2010, 28:56-64.

[3]
[4]
[5]
shellsh[C]//Lesel R. Microbiology in Poecilotherms.
.
[J].
[J].
Amsterdam: Elsevier, 1990:171-176.
, 1997, 21:246-251
[17] Prieur G, Nicolas J L, Plusquellec A, et al. Interactions
between bivalves molluscs and bacteria in the marine
, 2003, 25:24-26
,
environment[J]. Oceanography and Marine Biology: An
Annual Review, 1990, 28:227-352.
[J].
[18]
, 2009, 1:26-29.
[6]
the growth and digestive enzyme activity of white pacific
, 2010, 22:492-497.
shrimp (Litopenaeus vannamei)[J]. Oceanic and Coastal

Sea Research, 2008, 7:2152-2218.
Improvement of the probiotic effect of micro-organisms

combination
, 2002, 26:4451-4482
[19] Gomez R, Geovanny D, Ma S. Influence of probiotics on
[7] Bomba A, Nemcov R, Gancarckov S, et al.

their
,
[J].
.
[J].
by
with
maltodextrins,
fructo-
[20]
oligosaccharides and polyunsaturated fatty acids[J].
[J]
British Journal of Nutrition, 2002, 88(S1):95-99.

[8] Rengpipat S, Phianphak W, Piyatiratitivorakul S, et al.
, 2010, 22:235-240
[21]
2012
[J].
[22]
rohita fingerlings effects on growth, nutrient digestibility
, 2011, 35:10-16
and retention, digestive enzyme activities and intestinal
microflora[J]. Aquaculture Nutrition, 2011, doi: 10.1111/
[J].
j.1365-2095.2011.00866.x.
, 2008, 38:30-35
[23]
[26]
.
[J].
,
[J]
[27]
43-48
[24]
, 2011, 30:
,
, 2006, 27:54-58
,
[J]
,
[J]
[28]
, 2010, 10:17-19
,
, 2009, 21:95-100
[25] Mohapatra S, Chakraborty T, Prusty A K, et al. Use of
De
[J]
, 2010, 6:1-6
different microbial probiotics in the diet of rohu, Labeo
Effects of Dietary Supplementation with Compound Bacillus on Growth and

Digestive Enzyme Activity in Portunus Trituberculatus
DING Xue-yan, ZHOU Fan, HUANG Fu-yong, HE Feng, MENG Qing-hui
( Zhejiang Fisheries Technical Extension Station, Hangzhou 310012, China )
Abstract: The study is conducted to investigate the effects of dietary supplementation with compound bacillus
(Bacillus subtilis Natto and Bacillus subtilis Cohn) on growth condition, digestive enzyme activity and
cultivation efficiency in Portunus trituberculatus (Miers, 1876). The results from the indoor feed trial indicate
that the colonization rate, survival rate and weight gain in 1 (T1) and 2 (T2) bacillus supplementation
groups are higher than that of the control group. Feed intake presents a decline trend with dietary bacillus level.
Feed conversion rate in T1 and T2 are reduced by 15.38 and 18.80 when compared with the control diet,
respectively. Dietary compound bacillus shows no significant effect on amylase activity among the groups, but it
influences the other digestive enzyme activities. The highest average product in the outdoor pond feeding trail is
observed in P. trituberculatus fed with commercial feed with 2
compound bacillus (D2). The final breadth in
D2 group is found to be 3.3 higher than that of in D1(commercial feed) group. Feed conversion rates are found
very similar in D1 and D2 groups. It suggests that 2 compound bacillus supplemented in diet is suitable for P.
trituberculatus.
Key words: Portunus trituberculatus; compound bacillus; growth performance; digestive enzyme; cultivation
efficiency
25
, 2012 1
Vol.25 No.1, Jan. 2012
1.
1*
315300; 2.
315300
:
.
9.47
,
,
30~33
Durbin-h
,
,
5.0
,
.
.
;
: S915; S914
:A
: 1001-5132
2012
01-0007-06
(Litopenaeus vannamei),
.
,
[1-2]
DB 3302/T-2007
,
2
3 333.3 hm ,
40 m 150 m,
1.5 m.
,
.
,
,
.
[3]
[4]
[5]
DB 330282/T040-2009
[6]
[7-10]
,
4.5 m,
1.2
YDJ-1T
(
),
,
0.1
30 cm
0~100
,
250 cm
120 cm 3
1
1.1
,
2010
28
26
2011 07 20.
1959
1981
,
24
24
,
,
,
,
,
200801 .
,
,
/
:
,
. E-mail: cbm@eastagri.com
. E-mail: hilfreich@126.com
2012
.
2010
2.1.2
19
,
2010
2010
18
26
2(a)
2010
29
,
.
1.3
2(b)
2010
.
5
13
SPSS 17.0
.
.
Durbin-Watson d
Durbin-h
,
,
,
.
,
,
2.1
2.1.1
,
(
1
1).
,
,
;
,
32
.
,
(a)
32.3
33.7
9.47
12.4
,
,
.
(1.24
(2.86
),
)
,
.
(b)
2
2.1.3
,
,
2(a)),
15:00 13:00
1
1~3 h,
,
.
.
3
,
,
,
:
(R>0.96).
Yt
,
,
b1 b2 X t
, Yt
(1)
; Xt
;
; b1, b2
.
,
,
Durbin-Watson d
[11]
,
.
Yt
b1
b2 X t
b3Yt
, Yt
; Yt
; Xt
ut ,
(2)
; b1, b2, b3
ut
2.1.4
Yt
(2)
, d
.
.
, Durbin
;
,
30~33
,
,
[12]
1,
(R>0.8),
Durbin-Watson d
, d<dl,
;
(R<0.6).
,
1
d
2.2
(
: 1)
d1)
0.805
108
0.737
0.873
186
0.765
0.319
5.65
0.293
0.536
20.20
0.314
60,
50. n=
60, k=1, dl=1.503, du=1.585; n=50, k=1, dl=1.549, du=1.616.
(R>0.9),
,
2.
,
.
,
,
Durbin-h
,
,
10
2012
2
Yt=3.79+0.458Xt +
0.402Yt-1
Yt=1.74+0.432Xt +
0.545Yt-1
Yt=3.18+0.144Xt +
0.802Yt-1
Yt=1.98+0.122Xt +
2)
:*
=0.05
0.846Yt-1
/2=1.96;
h*
0.876
94.2
3.87
0.947
247
0.923
0.895
98.6
2.89
0.909
111
1.35
Yt*=3.00+0.53Xt*+
0.279Yt-1*
0.826
61.1
1.16
5.0
3.9
Yt*=2.70+0.177Xt*+
0.808
0.745Yt-1*
45.9
0.768
2.9
2.0
2)
3
/
0.70
0.978
3.1
0.71
0.986
2.1
0.95
1.019
2.6
0.95
1.020
2.2
0.85
0.977
3.1
0.81
0.988
2.1
0.92
1.025
2.8
0.87
1.029
3.3
.
,
5.0
2.3
,
,
2011
,
,
,
,
,
,
[7]
3):
.
,
2.1 ~3.3 ,
0.70~0.95,
0.977~1.029,
(1.0)
)[13].
2.4
.
18~35
30~33
,
18
9
Klemetson
,
[1]
2.8~4.4
[13]
. Little
[14]
9~10
,
,
11
32.3
9.47
Little
30~33
18
4),
2).
;
20~25
.
,
3~4
,
,
9.47
30~33
,
,
.
;
[15]
,
.
Durbin-h
,
,
.
[13]
Klemetson
[10]
,
(
5.0
MAPT
[10]
. Zhu
,
[16]
. Sarkar B
,
:
[1]
,
[18]
[2]
[J].
,
[4]
.
, 2005, 27(5):1-8.
.
[6]
, 2008, 33(6):117-121.
,
[J].
, 1998(9):14-20.
,
[J].
, 2006,
43(3):614-617.
[7]
[J].
,
,
, 2006, 30(1):63-67.
,
[J].
[5]
[7]
[3]
[9]
[J].
,
,
2001, 6(4):296-300.
,
[J].
[17]
[8]
[12]
, 2004, 23(4):52-55.
,
.
[J].
, 2010, 29(6):
12
2012
in a greenhouse covered aquaculture pond: a progress
674-677.
[9]
report[J]. American Society of Agricultural Engineers,
.
[J].
1987, 87:4022-4033.
,
[15]
2010, 26 (1):60-64.
,
[J].
[10] Zhu S, Deltour J, Wang S. Modeling the thermal characteristics of greenhouse pond systems[J]. Aquacultural
,
,
[J].
studies of a greenhouse fish pond in the central Himalayan

region[J]. Energy Conversion and Management, 2006,
47(18-19):3174-3184.
, 2008, 19(1):47-52.
[12] Gujarati D N, Porter D C. Essentials of Econometrics[M].

4th ed.
, 2004, 25(2):62-65.
[16] Sarkar B, Tiwari G N. Thermal modeling and parametric
Engineering, 1998, 18(3):201-217.

[11]
[17]
.
[J].
, 2010.
4(3):191-208.
, 2003, 29(2):175-180.
[13] Klemetson S L, Rogers G L. Aquaculture pond

temperature modeling[J]. Aquacultural Engineering, 1985,
(Scylla serrata)
[18]
[J].
, 2002, 18(1):18-22.
[14] Little M A, Wheaton F W. Water temperature prediction
Analysis on Temperature Characteristics of the Litopenaeus vannamei Culture

Ponds and Water Temperature Prediction
CEN Bo-ming1, HU Zhou1*, CHEN Han-chun2
( 1.Cixi Ocean and Fishery Bureau, Cixi 315300, China; 2.Cixi Fisheries Technical Extension Center, Cixi 315300, China )
Abstract: Temperature characteristics and water temperature prediction are studied in this paper using statistical
analytical approach with the on-site measured data of water and air temperature in the open-air and greenhouse
ponds designated for Litopenaeus vannamei culture. The results show that daily fluctuation of water temperature
of pond is relatively small and occurring time of daily maximum water temperature lags behind that of the air
temperature. Daily change of water and air temperature under clear weather is more frequent than that under
overcast one. The temperature variation and time lagging effect are not obvious on the upper and bottom layers
of pond due to oxygen aerator operation. The variation of water temperature in the greenhouse ponds is found to
be insignificant than that in the open-air ponds, and less affected by weather conditions. On average, a 9.47
increase in water temperature can be achieved by adding a greenhouse rooftop. Greenhouse pond maintains
water temperature at 3 33 , which proves optimal for Litopenaeus vannamei development. Correlation
analysis between water temperature and air temperature shows that water temperature is closely correlated with
air temperature. Based on time lagging effect of water temperature, autoregressive models of daily maximum
and minimum water temperature are established in the open-air and greenhouse ponds, respectively.
Autocorrelation in the time series are checked by Durbin-h test and removed using generalized difference
method. Verifications of these models are statistically effective and their average relative errors do not exceed
5.0 . The verifying results indicate that the developed models with wide applicability and good reliability can
be applied to forecast water temperature of Litopenaeus vannamei culture pond.
Key words: Litopenaeus vannamei; greenhouse; pond; temperature characteristics; prediction
25
, 2012 1
Vol.25 No.1, Jan. 2012
3
1
1.
,
,
315211; 2.
3.
315012
323000
:
,
(54.762.21)
(40.992.83) ; 3
44 ,
, 4
29
91 ~95 ;
;
,
.
;
: S917
:A
: 1001-5132 2012 01-0013-07

,
(Cypinus carpio var. color),

,
,
[1]
,
,
1.1
2010
,
1200
,
;
(53.5 g)
.
10
(170.8 g)
(308.5 g) 3
3d
.
[2-4]
1.2
500 g
[5-6]
1.2.1
,
.
,
;
3
,
0.0l g
,
100
,
2011 08 01.
1957
).
.
:
. E-mail: 214015@nbip.net
,
(
2009C12021
,
,
,
/
14
2012
1.2.2
5009.5-2003);
,
100
).
[10]
550
1.2.3
[9]
(GB/T 5009.4-
2003).
,
1.3.2
100 ).
GB/T18246-2000
1.2.4
GB/T15339-1994
,
(
) 4
40 g,
105
(NBFMS/XZ043-2005).
1.4
SPSS 11.5
(ANOVA)
.
, P 0.05
1.2.5
2
, 9
2.1
;
WHO (
2.1.1
/FAO (
,
(54.76 ),
[7]
(CS)
(51.63 ),
(40.99 ),
(AAS)
(P
(EAAI)
[7-8]
0.05).
,
1.3
.
1.3.1
(105
) (GB/T 5009.3-
2003);
2.1.2
(GB/T
1
/cm
12.750.56a
18.600.93b
23.130.38c
/g
53.524.59a
168.2722.56b
308.5026.97c
40.992.83a
51.631.92b
54.762.21b
2.740.09a
1.720.22b
1.180.26b
10.020.19a
7.690.48b
6.130.67b
79.050.15
77.500.90
78.051.15
15.950.45b
14.400.80
2.020.00
1.860.04
2.320.23
1.920.49
1.641.64
2.220.45
2.610.00a
1.470.03c
/
3
17.701.50
1.990.03
Turkey
(P<0.05);
:3
2.74
1.72
1.18
6.13
(P
(P
15
10.02 ,
7.69
(P
0.05) (
1).
2.2.3
0.05);
0.05).
;
,
1),
(17.7 ),
(P
((14.40.8)
0.05);
(2.32 ),
2.2
2.02
2.2.1
(P
1
0.05).
1.86 , 3
,
,
(79.5 ),
(77.50.9)
(78.051.15) ;
(P
0.05).
2.2.2
2.2.4
(1.47 ~
1.92
1.64
2.22 ,
2.61 ),
(P
0.05),
+
*
7.140.01
Asp
3.220.02
Thr
7.400.11
7.320.07
3.160.02b
3.330.03
Ser
2.970.01
3.040.04
2.970.06
Glu
12.530.02
12.830.20
12.360.19
Gly
4.710.03
4.380.09
4.780.12
Ala
5.020.02
5.090.08
4.970.06
Cys
1.020.06
1.010.01
1.040.04
Val
3.410.03
3.570.05
3.380.02
2.220.02b
3.750.07
3.460.04
6.920.09b
*
*
2.300.03
Met
ab
3.560.03
Ile
7.160.04
Ieu
ab
2.380.00
7.470.12
Tyr
2.200.01
2.310.03
2.150.02
2.840.00
2.990.04
2.810.03
6.400.03b
Phe
Lys
6.090.01
His
2.150.04a
2.340.03b
1.850.01c
Arg
4.340.03
4.410.07
4.500.06
Pro
2.150.01b
72.890.16
74.781.07
72.430.82
31.790.05
33.170.43
31.540.24
6.490.03
6.750.10
6.340.07
34.610.08
34.860.54
34.550.51
29.400.06
29.700.47
29.430.44
0.440.00
0.440.00
0.440.00
0.910.01a
2.240.01
WEAA/WTAA
WEAA/WNEAA
:*
ab
;#
0.920.00
;+
6.370.08
2.120.02
0.950.00
16
2012
(2.61 ) (
1).
2.3
, WEAA
WDAA
, WEAA/WNEAA
,
17
9
2),
(EAA)
(Val, Met, Ile, Leu, Tyr, Phe, Lys, Thr, Cys)

(HEAA) 2
(NEAA) 6
(Asp, Ser, Glu, Gly, Ala, Pro).

Glu
12.36 );
7.47
2.4
(His, Arg)
3
12.83
EAA
6.92 ), Cys
(12.53
Ilu
10
1.01
WTAA
44
(WDAA)
,
(PUFA)6
1.04 ).
;4
(Glu, Asp, Gly, Ala)
0.15)
(28.270.00
0.05).
((38.160.15) ).
,
(AA)
;
FAO/WHO
1.12
1.17
1.18
(P
(DHA)
1.61 ~3.78 ,
3
,
-6
-3
0.42~0.12,
; DHA/EPA
1.9~4.76,
(P
0.05).
UFA
((46.730.10)
(46.640.10)
3. 3
((68.19
(Lys)
SFA
);
(EPA)+
(P
(Val);
(71.750.21) ); MUFA
29.40 ~29.70
(61.790.28) ,
37.88 ; WEAA/WNEAA (
91 ~95 ; WEAA/
(38.220.03) ,
(WTAA)
)
(MUFA) 4
((31.850.11)
(WHEAA)
39.92
(SFA) 6
(7.16
(1.02
16
(WEAA)
38.28
0.05).
UFA
MUFA
PUFA
AAS
CS
EAAI
FAO
*
AAS
2.5
0.89
0.94
0.87
4.4
1.02
1.06
0.98
3.4
1.12
1.17
1.18
2.5
0.80
0.83
0.79
3.1
0.68#
+
*
+
*
CS
EAAI
:
0.72
2.2
0.94
0.96
0.93
3.8
0.83
0.87
0.82
3.31
0.67
0.71
0.65
5.34
0.84
0.87
0.81
4.41
0.86
0.90
0.91
2.92
0.69
0.71
0.68
0.69
0.54
0.52 #
4.10
0.52
3.86
0.54
0.55
0.53
5.65
0.56
0.59
0.55
65.64
68.28
65.07
:3
17
4
C14:0-CH3
0.680.05a
1.000.05ab
0.750.03b
C15:0-CH3
0.37a
0.340.06a
0.140.02b
C16:0-CH3
23.270.62a
20.600.16ab
18.860.25b
ab
0.540.03
0.320.05b
C17:0-CH3
1.010.16
C18:0-CH3
11.430.11a
7.670.02b
6.080.25c
C20:0
1.650.14
1.710.07
2.140.005
SFA
38.220.03
31.850.11
28.270.00c
C16:1n7-CH3
4.350.34a
5.510.36ab
2.620.16b
C18:1n9-CH3
32.130.06a
40.030.16b
43.570.25c
0.300.005b
C20:1n9-CH3
0.650.05
0.510.04a
C20:3
1.030.08a
0.680.06b
0.160.005c
MUFA
38.160.15a
46.730.10b
46.640.10b
C18:2n6-CH3
11.410.33a
13.560.69a
20.540.30b
C18:3n3
2.510.18a
2.900.18b
0.910.07b
C20:2n6-CH3
1.430.07
1.050.09
1.160.11
C20:4n6-CH3
4.510.19
0.900.03c
C20:5n3-CH3
1.310.07
0.610.09
0.280.01b
C22:6n3-CH3
2.470.18
1.430.3
PUFA
23.630.13
1.890.04
1.330.21
21.430.05
25.110.11c
7.320.15a
5.610.62a
2.670.27b
17.350.07a
16.500.57a
22.590.16b
3/ 6
0.420.01a
0.340.05a
0.120.01b
1.900.23
2.340.17
4.760.90
2.50.22b
DHA/EPA
AA+EPA+DHA
8.290.31
EPA+DHA
3.780.12a
3.930.43
2.040.39b
1.610.20b
.
215 g)
56.57
[7]
54.3
[7]
168~
[8]
(240~330 g) 51.94
(265 g) 55.71
[9]
:
,
[10]
.
.
,
.
3.2
,
.
,
.
,
53.52 g 168.27 g
308.50 g
40.99
54.76 .
51.63
3
(14.40.8)
(17.71.5)
(15.950.45)
.
500 g
700 g
18
2012
(18.040.50)
[6]
0.58)
[5]
(18.06
UFA,
PUFA
168.27 g
308.5 g
3.4
;
,
,
.
,3
,
.
,
,4
29 ,
150~300 g
,
,
[10-11]
WEAA/WTAA
44 , WEAA/WNEAA
.
,
91 ~95 .
.
FAO/WHO
WEAA/WTAA
WNEAA
40
, WEAA/
[12]
60
(1)
3.3
(2)
.
PUFA
;
,
[13]
53.52~308.5 g
.
,
. SFA
((38.22
0.03) ~(28.270.00)
(3) 3
, 3
16
150 g
(4)
(UFA)
,
,
((61.790.28) ~(71.750.21)
150~300 g.
MUFA
((38.160.15) ~(46.730.10)
PUFA
((23.630.13) ~(25.110.11)
[6]
700 g)
(25.080.44) , UFA (74.920.87)

PUFA
(41.890.65)
.
150 g
,
SFA
(MUFA
(33.030.22)
:
[1]
[2]
.
[J].
, 2001, 28(2):
56-63.
[3]
[M].
, 1991:109-110
,
.
,
[J].
.
, 2002, 30(6):858-
:3
19
[9]
860
[4]
[J].
,
,
[J].
[8]
[10]
[11]
[J].
, 1998(3):16-19.
[J].
, 2005, 29(5):502-506.
[12]
[13]
, 2009, 2(2):154-156.
.
, 1999, 15(2):69-72.
, 2010, 22(2):505-510.
[J].
, 2003, 30(5):198-200.
,
[J].
[7]
, 1996,
9(1):26-27.
,
[J].
[6]
.
[J].
2003(4):42-43.
[5]
[J].
, 1988(4):41-43.
.
[J].
, 2000, 17(1):l6-19.
Comparison of Nutritional Quality in Muscle of Oujiang Colour Carp

(Cypinus carpio var. color) of 3 Different Sizes
XU Ru-wei1, SHENTU Ji-kang2, JIANG Jin-po1, HUANG Fu-you3, YAO Zi-liang3
( 1.School of Marine Science, Ningbo University, Ningbo 315211, China; 2.Ningbo Academy of Ocean and Fishery, Ningbo
315012, China; 3. Fisheries Technical Extension Station of Lishui City, Lishui 323000, China )
Abstract: The rate of flesh content, ratio of viscera, body, general nutritional composition, amino acids and fatty
acids of Oujiang colour carp (Cypinus carpio var. color) of 3 different sizes are analyzed in this paper. The carp
nutritional values are compared and estimated. The results show that the rate of flesh content of the adult fish
group (54.762.21)
is significantly higher than that of small size fish (40.992.83) ; in the total content of
amino acids in muscle of 3 different sizes of Oujiang colour carp, the essential amino acids (WEAA) is 44 , of
which the first limiting amino acid is valine (Val) for all fish of 3 different sizes, and 4 amino acids account for
more than 29 . The WEAA/WNEAA (total nonessential amino acids) is ranging 91 ~95 . There are
more polyunsaturated fatty acids (PUFA) in adult fish group, and more eicosapntemacnioc acid (EPA) and
docase hexaeno acid (DHA) in small size group. The above results indicate that Oujiang colour carp is
characterized by high nutrition and better composition of amino acids and fatty acids in muscle, which is a
choice animal protein and suitable for fish farming, and dried fish processing as well.
Key words: Oujiang colour carp (Cypinus carpio var.color); muscle; nutritional composition; amino acid; fatty
acid
25
, 2012 1
Vol.25 No.1, Jan. 2012
1,2
1,2
1,2*
1.
2.
315211;
315211
Bligh-Dyer
:
,
:3
100 mg;
,
,
10 mg,
0.5~3.0 mg.
-1
200 L
0.058 8~0.588 0 m mL
.
;
: Q543
:A
,
: 1001-5132 2012 01-0020-04
[1]
1.1
,
[2]
.
.
,
[3]
(0.5 m mL-1)
CO2.
,
[4]
-20
1.2
.
,
1.2.1
10~
5 000 mL OD460
15
[5-6]
3 000 rmin
(Chlorella)
-1
[9]
,
/
.
(Chlorella pyrenoidosa)
.
[7]
1987
1973
,
,
,
,
:
:
8 min,
125 mL
4 mL
2009R50012-6
-1
7.6 mL
,
5 min, 3 000 rmin

1
2011 07 11.
(1:2:0.8),
125 mL
/H2O
,
,
100 mg
10 mL
2 min,
2010C33066 ;
2010C10022 .
,
,
/
24 h
1.2.2
[7]
, -20
Bligh-Dyer
[8]
8 min
,
.
0.542
2010C10022
. E-mail: fenger19871221@126.com
. E-mail: xunianjun@nbu.edu.cn
2:2:0.8,
/
/H2O
30 s,
2:2:1.8,
21
0.588 0 m mL-1
4 mL
30 s,
2h
0.294 mgmL
96
,
(0.1 mgmL ),
540 nm
100 mg
30 mL
10 mL
50 mL
,
.
.
.
2.1
[8]
Bligh-Dyer
0.1 m mL
2.0 m mL
-1
0.2 m mL
-1
5.0 m mL
0.5 m mL
2.5 mL,
1.0 m mL
0.1 mL
, 100
1.978 m mL
-1
-1
0.5 mL
,
100 mg
0.0130 g 0.012 6 g 0.0134 g
,
-1
580 nm
1.2.4
7 min,
2:1 (v /v )
,
200 L
-1
2 L
1.2.3
-1
2 h,
-1
13.0
13.4 ,
10 min,
(13.00.4) .
2.2
-1
528 nm
,
12.5
3
[10]
10 mg
100 mg
0.012 5 g 0.012 4 g 0.013 0 g,
OD528
12.6
12.4
13.0 ,
(12.6300.321) .
2.3
1.2.5
0.5 mg 1.0 mg 1.5 mg 2.0 mg 2.5 mg

3.0 mg
1.0 mL
10 min,
5 mL,
.
, 100
1.978 mgmL
2 h,
:
2
0.986 6.
528 nm
OD528
0.591 0.623 0.634,

(2.2870.090) m mL-1,
1.8 mg
(12.58 0.495) .
2.4
[10]
1.2.6
13.0
0.2 mgmL
1.0 mgmL
0.248 4 x 0.048, R
-1
OD528
3
1)
-1
2.0 mgmL
-1
,
200 L
5.0 mgmL
3
.
0.5 mgmL
-1
96
2.
-1
,x
(mg), y
OD528
OD528
1.402 1.429 1.384,
(0.224 80.004) mg,
(12.490.215) .
2.5
13
.
3.
1.2.7
-1
0.058 8 mgmL
0.235 2 mgmL
3
-1
-1
y = 5.824 6x+0.095 7, R = 0.997 0,
540 nm
-1
(0.1 mgmL
7 min,
-1
10.0 mgmL
1.5 mgmL
2 L
580 nm
-1
-1
3.528 0 mgmL
-1
y = 0.310 2x+0.119 8, R =0.793 9,
-1
,x
0.117 6 mgmL
0.470 4 mgmL
,
:
(m mL-1), y
22
2012
1
1
5.0
1.2510.072
/(m mL-1)
OD528
OD
2
2.0
0.6520.075
3
1.0
0.2960.011
4
0.5
0.1480.012
5
0.2
0.0780.019
6
0.1
0.0490.004
OD
/mg
0.5
1.0
1.5
2.0
2.5
3.0
/mg
0.065
0.130
0.195
0.260
0.325
0.390
0.5240.077
0.8200.078
1.1820.012
1.6350.077
1.9710.063
2.3930.119
OD528
3
-1
/(m mL )
-1
/(m mL )
OD580
OD
0.2
0.5
1.0
2.0
5.0
10.0
0.026
0.065
0.130
0.260
0.650
1.30
0.027 10.012
0.109 40.034
0.170 70.023
0.297 90.009
0.401 70.020
0.465 90.053
4
1
-1
/(m mL )
/(m mL-1)
OD580
OD
2
0.058 8
0.117 6
0.235 2
0.352 8
0.470 4
0.588 0
0.007 644
0.015 288
0.030 576
0.045 864
0.061 152
0.076 440
0.085 70.004
0.108 50.024
0.131 90.009
0.153 90.010
0.164 00.019
0.185 90.025
(100 mg
0.198 5 0.226 6 0.259 8,

-1
(0.349 80.099) m mL ,
(23.326.595) .
,
(P
),
(10 mg
0.05).
2.6
4
:
2
y = 1.369 6x+0.084 2, R =0.971 2,

(m mL-1), y
,x
.
,
3
,
0.131 5 0.138 7 0.133 6,

-1
(0.036 80.003) mgmL ,

-1
0.294 mgmL ,
(12.520.921) .
(0.5~3.0 mg).
3
,
Bligh-Dyer
,
, 2
13.0
(P
0.01),
12.6 ,
.
,
,
. Liu
[10]
480 nm
23
microalgae[J]. Applied Microbiology and Biotechnology,
, 580 nm
2008, 79(5):707-718.
.
.
,
[4] Courchesne N M D, Parisien A, Wang B, et a1.

Enhancement of lipid production using biochemical,
genetic and transcription factor engineering approaches
[J]. Journal of Biotechnology, 2009, 141:31-41.
,
(0.058 8~0.588 0 mgmL-1)
[5] Wijffels R H, Barbosa M J. An outlook on microalgal

biofuels[J]. Science, 2010, 329(13):796-799.
[6] Williams P. Biofuel: Microalgae cut the social and
ecological costs[J]. Nature, 2007, 450:478.
,
,
[7] Bligh E G, Dyer W J. A rapid method of total lipid
extraction and purification[J]. Canadian Journal of

Biochemistry and Physiology, 1959, 37:911-917.
[8] Gunnel A, Leonardo M. Lipid analysis of freshwater
microalgae: A method study[J]. Arch Hydrobiol, 1991,

121(3):295-306.
[9] Cooksey K E, Guckert J B, Williams S A, et a1.
:
[1] Chisti Y. Biodiesel from microalgae[J]. Biotechnology
microalgal cells using nile-red[J]. J Microbiol Methods,
Advances, 2007, 25(3):294-306.

[2]
Fluorometric determination of the neutral lipid content of

1987, 6:333-345.
.
[J].
, 2009, 27(3):
[10] Liu Z Y, Wang G C, Zhou B C. Effect of iron on growth

and lipid accumulation in Chlorella vulgaris[J]. Bio-
65-69.
[3] Wang B, Li Y, Wu N, et a1. CO2 bio-mitigation using
resource Technology, 2008, 99(11):4717-4722.
Comparison and Optimization of Total Lipid Detection Methods for Microalgae Chlorella
CONG Feng1,2, SUN Xue1,2, XU Nian-jun1,2*
( 1.Key Laboratory of Applied Marine Biotechnology, Ministry of Education, Ningbo University, Ningbo 315211, China;
2.Key Laboratory of Marine Biotechnology, School of Marine Science, Ningbo University, Ningbo 315211, China )
Abstract: Three methods for detecting the total lipid content of the microalgae Chlorella pyrenoidosa, that is,
bligh-dyer method, vanillin colorimetric method and nile-red method, are compared and analyzed for their
efficiency and sensitivity, and the optimization method is also introduced. Test results indicate that the
gravimetric method is of generic and the results from using it are accurate, but the gravimetric method is not
adequately sensitive in the detection process of trace samples because of the minimum detectable amount being
100 mg. The vanillin colorimetric method and nile-red method both are simple and sensitive. The detection using
the vanillin colorimetric method is limited to 10 mg, and the detection range of the optimized vanillin
colorimetric method is 0.5 3.0 mg. The detection using nile-red method is limited to the amount ranging from
0.058 8 mgmL-1 to 0.588 mgmL-1 algal solution. In conclusion, the lipid content of the microalgae Chlorella
pyrenoidosa can be detected by the optimized vanillin and combined with the nile-red method, which turns out to
be easier and more sensitive than the gravimetric method and can thus be used for the small amount of samples.
The study provides a technical approach for the rapid detection of the lipid content of biofuel microalgae.
Key words: Chlorella; lipid detection; vanillin colorimetric method; nile-red method
25
, 2012 1
Vol.25 No.1, Jan. 2012
3
*
315211
,
,
24 h, -24
:
3d
100 ;
3~6 d,
: Q142
:A
: 1001-5132 2012 01-0024-03
2 000 lx,
1.2
.
[1]
88
1.2.1
0.01 L -1
.
[2]
(Ulva prolifera)
[3]
80
[4]
60
2h
[5]
,
[8]
24 h,
1.2.2
[6]
20~22
25
1 000~3 000 lx;
.
50 lx.
,
,
(Ulva intestinalis)
4 h 8 h 12 h 24 h 36 h 48 h
72 h
(Ulva prolifera),
18
1.2.3
(Ulva
[7]
compressa) .
,
,
3
-4
-12
-24
0.5 d 1 d 2 d 4 d 8 d 15 d
.
1
2
1.1
2009
10
2010
3
72 h
10 cm
20
2011 07 14.
1986
1965
100 .
,
,
201105023
,
,
;
,
,
.
. E-mail: huangxianjunhdm@163.com
. E-mail: luoqijun @nbu.edu.cn
2010C10024
,
,
:
:
1)
25
1
0h
U.p
4h
2.3 1.16
8h
12 h
24 h
36 h
48 h
72 h
6.731.62
21.144.32
38.455.34
49.687.01
71.239.36
100
U.i
2.591.33
7.352.19
23.695.10
47.126.17
53.138.19
79.237.53
100
U.c
4.511.52
9.412.58
29.776.12
47.085.28
58.428.85
85.238.12
100
U.p
1.580.38
3.250.46
13.051.47
21.927.28
33.623.36
79.755.00
100
U.i
1.721.71
5.042.61
21.424.23
32.839.17
40.138.19
77.398.17
100
U.c
2.491.35
7.872.92
23.182.86
30.943.36
47.428.85
87.479.28
100
: U.p
; U.i
; U.c
Ulva prolifera
0h
4h
8h
12 h
24 h
36 h
48 h
72 h
28.921.72
46.300.25
70.520.08
86.030.16
94.050.63
93.480.20
93.580.20
22.251.73
41.530.65
71.810.33
83.210.44
92.280.15
92.710.38
92.870.20
U. Prolifera
0.5 d
1d
2d
4d
8d
15 d
13.750.41
32.271.04
69.754.39
88.755.72
98.427.58
100
-4
15.000.57
50.363.83
74.468.26
89.757.40
99.429.23
100
-12
22.621.48
67.576.57
83.683.30
92.621.63
100
100
-24
30.171.16
71.404.96
85.382.83
98.481.52
100
100
4
0.5 d
1d
2d
4d
8d
15 d
53.75+0.42
62.35+1.04
79.75+4.39
88.75+5.72
94.42+7.58
95.58+9.95
-4
55.00+0.57
60.35+3.83
74.75+8.26
89.75+7.40
91.42+9.23
94.58+12.27
-12
42.62+1.48
67.23+6.57
73.75+3.30
88.62+1.63
90.71+8.07
95.14+7.18
-24
40.17+1.16
71.40+4.96
75.38+2.83
83.48+1.52
91.52+4.92
95.01+6.94
72 h
48 h
70
48 h,
.
1.
,
,
72 h
4d
48 h
72 h
,
8d
,
,
,
.
,
(12 h
),
24 h
, 36 h
,
,
,
10
2),
48 h
100 .
, 48 h
3.
-4
-12
-24
.
,
,
,
26
2012
,
,
95
5
,
90
100 ,
2),
[1] Martins I, Marques J C. A model for the growth of
72h
100%.
opportunistic macroalgae (Enteromorpha sp.) in tidal
estuaries[J]. Coastal and Science, 2002, 55:247-257.
[2]
(Enteromorpha profilera)
[J].
[3]
[4]
.
,
10 ~15
-20
.
, 2008, 29(5):
[5]
10 ~ 15
30 d
[J].
[6]
, 2002, 19(1):31.
.
[J].
[7]
[9]
[J].
99 ~100
, 2009, 37(20):9566-9567.
[9]
130-136.
8d
99
[J].
.
,
, 2009, 40(1):
68-71.
-12~-24
36 h
3d
prolifera)
, 2009, 22(4):57-62.
.
(Enteromorpha
[J].
, 2007, 26(5):
111-116.
[8]
,
,
[J].
, 2001, 21(2):30-34.
[9]
,
[J].
.
, 2007, 19(2):131-136.
Effects of Light Condition and Low Temperature on

Survivability in Three Kinds of Ulva
HUANG Xian-jun, LUO Qi-jun*
( Key Laboratory of Applied Marine Biotechnology, Ministry of Education, Ningbo University, Ningbo 315211, China )
Abstract: In this thesis three kinds of Ulva from nori net are selected as experimental samples. Under the variant
conditions of light and temperature, we use drying and freezing methods to probe Ulva in an effort to study how
the mortality of Ulva responses to these methods and explore the optimal conditions to physically rid the Ulva on
breeding net. The results show that the response is different from one kind of Ulva to another. In the drying
experiment, the mortality under light condition turns out to be higher than in non-light condition, and the
mortality of Ulva under lit condition reaches 100 in three days. The Ulva mortality rate through freezing at low
temperature is found to be very high. Therefore, dehydration for 24 h and refrigeration in -24
low
temperature for three to six days is recommended as the optimal configuration.
Key words: Ulva; drying; freezing
25
, 2012 1
Vol.25 No.1, Jan. 2012
Rate-dependent Damage Evolution and its

Influence on Dynamic Behavior of Materials
at High Strain Rates
WANG Li-li, ZHOU Feng-hua, SUN Zi-jian, HUANG De-jin
( Mechanics and Materials Science Research Center, Ningbo University, Ningbo 315211, China )
Abstract: Both experimental and numerical results reveal that dynamic damage evolution depends on
both strain and strain rate. Based on the thermal activated mechanism, a new damage evolution law and a
new fracture criterion are proposed. For determining the damage parameters of material, two approaches are
developed: the DM-ZWT approach and the SHPB+BP neural network approach. The predictions by both
approaches are in good agreement with the experimental data.
Key words: damage evolution; dynamic behavior of materials; high strain rate; adiabatic shearing
CLC number: O346.5
Document code: A
The study of the rate-dependent deformation and
Article ID: 1001-5132 2012 01-0027-08
is given in the present paper.
fracture of materials by taking into account the damage

evolution has become one of the research frontiers,
receiving more and more attention by scientists of both
Experimental and numerical observation on dynamic damage evolution
mechanics and materials science.

The main difficulty lies in the interaction between
the flow/deformation process and the damage evolution
[1]
1.1 Experimental observation of macro-damage

evolution
process . In fact, on the one hand, the damage
Since there is still no way to directly measure the
evolution is a stress/strain-assisted process. On the other
evolution of internal damage during a deformation
hand, the damage evolution inevitably influences the
process, particularly at high strain rates, the change of
macro-response of materials. Such kind of damage-
elastic modulus of the specimen having experienced a
deformation interaction becomes more complicated
certain impact deformation (after a SHPB test) is used to
under explosion/impact condition due to the coupled
approximately characterize the macro-damage DE , as
inertia (wave) effects and strain-rate effects of materials.
DE
1 ED / E0 ,
(1)
In the authors lab, both the dynamic behavior of
where ED is the elastic modulus of the damaged
materials and the damage evolution were experimentally
material measured after an impact test and E0 the
and theoretically studied in a range of strain rates from
initial elastic modulus of the undamaged material.
-5
-1
10 10 s . A brief overview of those research works
For a polypropylene-polyamide (PP/PA) polymer
Received date: 2011 10 30.

JOURNAL OF NINGBO UNIVERSITY ( NSEE ): http://3xb.nbu.edu.cn
Foundation items: Supported by the National Natural Science Foundation of China (11032001); Russian Foundation for Basic Research (11-01-91217);
K. C. Wong Magna Fund in Ningbo University.
The first author: WANG Li-li (1934-), male, Ningbo Zhejiang, professor, research domain: impact dynamics. E-mail: wanglili@nbu.edu.cn
28
2012
blend, the typical results obtained by such experimental

method are given in Fig 1(a), in the form of DE versus
for different final strains (from 4
For a -phase titanium alloy (Ti-8Cr-5Ni-5V-3Al),
to 12 ) and in
our experimental results (Fig. 2) shown that the evo-
for different
lution of adiabatic shear band is also dependent on both
Fig 1(b), in the form of DE versus

strain rates, respectively
tive instability [3].
[2]
strain rate and strain [4].
(a) Micrographs of specimens at room temperature: showing the

development of shear band with increase of strain under
Fig. 1(a)
DE - curves for PP/PA blend at different strains
Fig. 1(b) DE - curves for PP/PA blend at different strain-rates
From these figures it is clear that (a) under a
=1.5103 s-1;
8.8 ,
16.3 ,
27.4 ,
42.3
(intersected shear bands with crack)
(b) Experimental results at room temperature ( : no shear band, :

deformed band;
transformed band, *: cracking; Theoretical predict-
certain loading condition, a deformation process is
tion by Wangs criterion [3]: a-A =1.153 6, b-A =1.138 2, c-A =1.114 3)
accompanied with a damage evolution process, and a
Fig. 2 Dependence of shear band evolution on strain and
threshold strain
th
strain rate for a -Titanium alloy
for the beginning of damage
evolution exists, which can be approximately estimated,

for the PP/PA blend, in the range of 4 <
th
<6
depending on strain rate; and (b) the evolution of
1.3 Numerical investigation of micro-damage

evolution
By using an improved fuse network model and
macro-damage DE is dependent on not only the strain
Monte-Carlo
but also the strain-rate, i.e. DE
features of damage evolution were numerically studied
DE ( , ) .
1.2 Experimental observation of meso-damage

evolution
Adiabatic shear bands are usually regarded as one
random
method,
the
rate-dependent
at different constant strain rates [5].

For
three
strain-rates (
different
0
constant
dimensionless
10 , 10 , 10 ), the configurations of the
of elemental forms of damage in meso-lever. The
micro-crack distribution at different time are given in
adiabatic shear banding at high strain rates is generally
Fig. 3. As can be seen, the damage in the form of
attributed to the so-called thermo-viscoplastic constitu-
micro-cracks evolves stochastically within the network
29
WANG Li-li, et al: Rate-dependent Damage Evolution and its Influence on Dynamic Behavior of Materials at High Strain Rates
at early stage, then tends to develop in an intensively
depends on the
localizing part at later stage and finally the network is
Since the experimental and numerical results
destroyed. Such a process of damage evolution is
mentioned above reveal that the damage evolution is
influenced strongly by strain-rate, and thus the failure of
also a rate process similar to the rate process for
network depends on both strain rate and strain (
viscoplastic deformation, it is not unreasonable to
consider the damage evolution as a thermal activated
approximately).
process. In other words, if the macro-damage, according

to Kachanov (1958) [6], is defined as
D
app
t = 7.5
t = 1.6
where
app
1, or,
(3)
(1 D) ,
app
t = 48.4
)/ , o
is the stress for material without damage and
is the apparent stress for material with damage,
then similar to Eq. (2), the rate-dependent damage

evolution may be expressed by:
D
t = 63.1
t = 0.7
t = 1.9
D/ t
D0 exp( U D / kT ),
(4)
where D is the rate of damage evolution, D0 is the

frequency coefficient and U D is the activation energy
for damage evolution, respectively. Instead of analyzing
the relation between the U D and the applied stress
t = 73.9
(a)
100
Fig. 3
in detail, assume that for the same material
t = 12.8
(b)
101
(c)
10
concerned a proportional relation exists between the
Damage distribution within network at different

times under different constant
U D in Eq. (4) and the U s in Eq. (2), that is

UD
Thus, either from the experimental or the

numerical investigation, it is clear that the damage
evolution is dependent on both strain and strain rate,
where
(5)
Us ,
is a material parameter. Then, after some
mathematics calculations, we have

D / D0
regardless whether the form of damage is.
KD
( /
) , or, D
D0 /
KD
(6)
or by integrating,
Dynamic damage evolution law
KD
( ),
( )
t
t0
dt ,
(7)
which formulates a rate dependent, dynamic damage

It is well-known that the macroscopic ratedependent viscoplastic behavior of materials can be mi-
evolution law in a simple form.

In the case of constant strain rate, by integrating Eq.
croscopically explained by the thermal activated motion
(7) and recognizing that a strain-threshold,
of dislocation and described by the following Arrhenius
damage evolution, Eq. (7) reduces to
type equation:
0
where
exp( U s ( ) / kT ) ,
(2)
is the viscoplastic strain rate,
KD
th
), for
th
th,
exists for
(8)
moreover, considering that D may depend on
is a
nonlinearly in more general cases, a generalized form of
frequency coefficient, k is the Boltzmanns constant, T is
Eq. (8) can be proposed by introducing a material
the absolute temperature, and the activation energy U s
parameter
is a function of the applied stress
. In the thermal
activation theory, a key problem is how the U s
KD
th
) , if
th
(9)
It is worthwhile to note that the adiabatic shearing
30
2012
at high strain rates is usually regarded as one of
tutive response for a variety of plastics, polymer- matrix
elemental forms of damage, which is generally
composites and even concretes can be well described by
attributed to the so-called thermo-viscoplastic constitu-
the
tive instability
[9]
. The thermal softening induced by the
following
nonlinear
viscoelastic
equation proposed by Zhu, Wang and Tang (ZWT) [7] in
adiabatic heating during the adiabatic shearing process,
the case of
actually, can be considered as a weakening induced by a
depending on different polymers:
kind of generalized damage. In fact, the adiabatic
temperature rise dT and the corresponding stress

decrease d
d
dT , where
/ T )dT
d / ( CV ), and
is the Taylor-
Quinney coefficient of viscoplastic work converted to

heat,
have
( )
, which is e.g. around 5 -7
th
( ) E1
E2
during the adiabatic shearing can be
expressed, respectively, as dT
constitutive
( ) exp( (t
E0
( )
( ) exp( (t
)/
3
[1 exp(
) / 1 )d
)d ,
(10a)
(10b)
i
(m ) / i )] .
(10c)
i 1
Where the first term
describes the
( )
the material density, CV the specific heat. If
nonlinear elastic equilibrium response which can be
/ T (>0) can be regard as a constant, then we
described by Eq. (10a) or Eq. (10b); the next integral
dD
d /
/ CV )d ,
or,
term describes the linear viscoelastic response at low
( CV )) , which is actually a special case of Eq. (8)
strain rates (E1 and
when K D
2 . Thus, the
relaxation time, respectively), and the last integral term
adiabatic temperature rise can be indeed regarded as a
describes the linear viscoelastic response at high strain
kind of generalized damage.
rates (E2 and
/ ( CV ), ,
0 and
Now the main problem is how to experimentally

determine the material parameters .
and
th
contained
are the elastic constant and
are the elastic constant and relaxation
time, respectively). In Eq. (10a) E0,
and
denote the
initial elastic modulus and nonlinear elastic constants
in the dynamic damage evolution law (Eq. (9)). The
respectively; while in Eq. (10a)
difficulty lies in that there is practically no available
totic maximum, m the ratio of E0 and
technique to directly observe and measure the damage
tive integer n is a material parameter characterizing the
evolution during an impact process.
initial linearity.
Two indirect approaches are developed in our lab
In case
th
denotes the asympm,
and the posi-
, in principle, a rate-dependent
to distinguish the dynamic constitutive response and the
constitutive relation taking into account the damage
damage response, and then determine the macro-damage
evolution can be obtained by combining Eq. (10) and Eq.
evolution in a
(9), briefly called the damage-modified ZWT (MD-
- process of materials. One is the
approach named damage modified ZWT equation
ZWT) equation and expressed as
(DM-ZWT approach), and another is an approach by
(1 D)[
combining the Split Hopkinson Pressure Bar (SHPB)
E2
technique with the Back-Propagation neural network

technique. The details of them will be discussed in the
following.
KD
t
0
( ) E1
t
0
( ) exp( (t
1
th
) , if
( ) exp( (t
)/
)d ] ,
th
) / 1 )d
(11a)
(11b)
the material parameters in Eqs. (11) can be determined

by fitting the experimental data obtained in a series of
Determination of coupled damageconstitutive parameters
INSTRON and SHPB tests using the Genetic Algorithm

method [8-9].
A typical example is shown below. The test
3.1 DM-ZWT approach

It was found that the nonlinear dynamic consti-
material is a polypropylene-polyamide (PP/PA) polymer

blend compatibilized with the polypropylene grafted
Tab. 1
31
Parameters of DM-ZWT constitutive equation fitted for the PP/PA polymer blend tested
m/MPa
E1/MPa
44.1
39.4
278
1/s
E2/GPa
7.22
2.29
2/
D0
0.107
0.639
th
1.13
0.617
Eq. (12)
with maleic anhydride (PP-g-MAH). The typical stress-
strain rate. Moreover, it is evident that the strain-
strain curves experimentally determined at different
softening appearing in Fig. 4 at large deformation does
strain rates by SHPB and INSTRON tests are shown by
be induced by damage evolution.
solid lines in Fig. 4. The strain-softening phenomenon

appearing at large deformation and high strain-rate is
attributed to the internal damage evolution, as will be
further discussed below.
Fitting the experimental data obtained, the material
parameters of the DM-ZWT equation (Eqs. (11a) and
(11b)) are determined by using the Genetic Algorithm
method, and given in Tab. 1.
The curves predicted by those parameters in Table
1 are shown in Fig. 4 too, denoted by dashed lines. As
can be seen, the theoretical curves predicted by the
DM-ZWT model are in good agreement with the
D- curves at different for PP/PA blend
Fig. 5
th
3.2 Approach by combining SHPB and BP neural

network technique
This approach combines the SHPB technique with
in Eq. (11b) is dependent on strain rate too, which can
the Back-Propagation (BP) artificial neural network
be described by Eq. (12)
technique
experimental curves in a wide range of strain rate. Note

that from the experimental data, the threshold strain
th
0.075 1.02 10
(12)
[10-11]
, wherein no pre-assumption of constitu-
tive model for test material is required. However, the

threshold strain
th
for damage evolution should be pre-
estimated approximately. Then the SHPB experimental

data in both the case of
th
(Case A) and
th
(Case B) are used as appropriate input/ output to train

the BP neural network and both the dynamic constitutive
response of undamaged material (Case A), and the
dynamic constitutive response taking account of damage
evolution (Case B) can be predicted by the trained BP
neural network. Comparing those two sets of results, the
influence of the dynamic damage evolution on the
Fig. 4
The
- curves at different for PP/PA blend
curves
showing
rate-dependent
damage
dynamic constitutive behavior for the material tested can

be obtained.
evolution law D=D( , ) predicted by Eq. (11) and Eq.
A typical example is shown below for the same
(12) are plotted in Fig. 5. As can be seen, strain rate
polypropylene-polyamide (PP/PA) discussed in previous
plays a significant role in damage evolution. Particularly,
sub-section, for which the experimental stress-strain
the damage develops much quickly with increasing
curves were shown by solid lines in Fig. 4.
strain at impact condition than at quasi-static condition,

while the threshold strain
th
decreases with increasing
In the Case A, the BP neural network is trained

from the experimental data determined in SHPB tests by
32
2012
(a) in Case A
(b) in Case B
Comparison of the prediction (dashed curve) with the measurement (solid curve) at strain rate of 1.22103 s-1
Fig. 6
taking the strain (t) and strain rate
(t) as the input
and the stress (t) as the output. Taking

3
th
as 4.8 , the
-1
- curve in strain rate of 1.2210 s predicted by the
evolution D
D( , ) can be finally obtained, as
plotted in Fig. 7 in form of a set of D- curves at

different constant strain rate.
trained BP neural network is shown in Fig. 6(a) by
To compare the results obtained from two different
dashed line. In the same figure, the solid line denotes the
approaches, the D- curves at the same strain rate of
experimental curve. Note that a normalized stress

is the maximum stress
are plotted in the same figure (Fig. 8), as denoted by the
appeared among all the tests. As can be seen, the
curve I and the curve II respectively. As can be seen,
predicted
curve (actual output) is in good agree-
both curves consistently reveal that the dynamic damage
ment with the experimental curve (expected output) in
evolution depends on both strain and strain rate. The
. The increasing deviation between
existence of some difference between two curves could
is attributed to the damage
be understood, since for the curve II the damage
is introduced where
2.0103 s-1 given in Fig. 5 and that given in Fig. 7
the range of
th
two curves when
th
evolution, which plays a weakening role as expected.

In the Case B, the BP neural network in principle
test with BP neural network without any pre-assumption
(t ) and D (t ) as
of constitutive model, while for the curve I the damage
(t ) as the output. However, no success-
evolution is determined by the DM-ZWT viscoelastic
should be trained by taking

the input and
evolution is determined by the method combining SHPB
(t ) ,
ful technique could be used to directly measure D (t )
constitutive model, which will provide an underestima-
during the SHPB test. Since t ( D ) could be regarded as

an inverse function of damage evolution, t
g 1 ( D)
t ( D) , the BP neural network in the Case B thus can be
trained by taking
(t ) ,
(t ) and t ( D ) as the input
(t ) as the output. The - curve predicted by the
and
BP neural network trained in such way is shown by the

dashed line in Fig. 6(b), which is in good agreement
with the experimental curve (expected output) in whole
strain range, as shown by the solid line in the same
figure.
Comparing the results predicted in Case A and
Case B, the dynamic rate-dependent law of damage
Fig. 7
D-e curves at different
predicted by BP
neural network for PP/PA blend
n 4, m 22.3, E1 0.897 GPa,

GPa,
95.4
and Df 10.2
s, KD 1.82,
33
15.3 s, E2 3.07
1.17,
1,
th
0.06
. As can be seen, the rate-dependent
damage evolution should be taken into account and the

theoretical predictions are in good agreement with the
experimental measurements.
Fig. 8
Comparison of results for two approaches

(I: DM-ZWT, II: SHPB+BP neural network)
tion of damage evolution if the test material experiences

additional visco-plastic deformation, which was in fact
observed at larger deformation.
(a) Dynamic - curves at different strain rates
Dynamic fracture based on

damage evolution
If the dynamic fracture of material is dominated by
damage evolution, the critical condition for dynamic

fracture could be expressed as D
D f , where Df is the
critical damage corresponding to the critical state of

dynamic fracture. Then from Eq. (9), the critical
condition for dynamic fracture
rewritten in terms of and
KD
th
as
can be
Df
[1]
Df .
(13)
Obviously, there exist three possible cases:

(1) If
fracture strain
1 , then with increasing strain-rate, the

f
decreases, which corresponds to the
case of impact embrittlement;

(2) If
fracture strain
Comparisons between the experimental data and

theoretical predictions given by (a) Eq. (11)
and (b) Eq. (13) for PMMA
Conclusion
1 , then with increasing strain-rate, the

f
increases, which corresponds to the
1 , then Eq. (13) is reduced to a critical
A typical example is shown below for a

[12]
. The experimental
-4
(1) Both experimental and numerical investigations

reveal a significant fact that dynamic damage evolution
strain criterion and independent to strain-rate.

polymethylmethacrylate (PMMA)
From above results the following conclusions can

be drawn.
case of impact toughening;

(3) If
(b) Dynamic fracture at different strain rates
Fig. 9
-1
is a rate-dependent process, depending on both strain

and strain rate.
stress-strain curves at strain rates from 10 to 10 s are
(2) Considering the damage evolution as a thermal
given in Fig. 9. The material parameters in Eq. (11)
activated process, a new dynamic damage evolution law
91.8 MPa,
is proposed, describing how the damage evolution
determined by DM-ZWT approach are
34
2012
ature[J]. Journal de Physique, 1988, 49(C3):207-214.
depends on both strain and strain rate.
[5] Huang Dejin, Shi Shaoqiu, Wang Lili. Studies on
(3) For determining the material parameters con-
rate-dependent evolution of damage and its effects on
tained in the dynamic damage evolution law, two
dynamic constitutive response by using a random fuse
approaches are developed: the DM-ZWT approach and
network model[M]//Chiba A, Tanimura S, Hokamoto K.
the approach by combining the SHPB technique with
Impact engineering and application, proc 4th int symp
the BP neural network technique. The predictions by
impact engineering. Tokyo: Elsevier Science Ltd, 2001:
both approaches are in good agreement with the
743-748.
[6] Kachanov L M. On the creep rupture time[J]. Izv AN
experimental data.
SSSR Otd Tekhn Nauk, 1958, 8:26-31.
(4) Corresponding to the damage evolution law
[7] Wang Lili, Yang Liming. The nonlinear viscoelastic
proposed, a new critical condition for dynamic fracture
constitutive relation of solid polymeric materials[M]//
is further derived, showing the dynamic fracture is
Wang Lili, Yu Tongxi, Li Yongchi. Progress in impact
dependent on both strain and strain rate too. The
dynamic. Hefei: The Press of the University of Science
theoretical predictions are in good agreement with the
and Technology of China, 1992:88-116.
experimental measurements.
[8] Holland J H. Adaption in natural and artificial systems

[M]. Ann Arbor: Univ Michigan Press, 1975.
References:
[9] Jiang Houman, Zhang Ruoqi, Zhang Shouqi. Applying
[1] Wang Lili, Jiang Zhaobiao, Chen Jingying. Studies on
genetic algorithm to determine parameters in equation of
rheological relation of materials by taking account of
state[J]. Chinese J High Pressure Physics, 1998, 12(1):47-
rate-dependent evolution of internal defects at high strain
53.
rates[M]//Wang Ren. IUTAM Sym: Rheology of bodies
[10] Xu Mingqiao, Wang Lili. A new method for studying the
with defects. Dordercht: Kluwer Academic Publishers,
dynamic response and damage evolution of polymers at
1999:167-178.
high strain rates[J]. Mechanics of Materials, 2006, 38(1/
[2] Sun Zijian. Studies on impact constitutive behavior and
2):68-75.
damage evolution for two PP/PA polymer blends at large
[11] Li Xueqiao, Ma Lili. Neural Network Engineering[M].
deformation[D]. Ningbo: Ningbo University, 2005.
Chongqing: Chongqing University Publishing House,
[3] Wang Lili. A criterion of thermo-viscoplastic instability
1996.
for adiabatic shearing[M]//Zheng Zhemin. Proc int symp
[12] Zhou Fenghua, Wang Lili, Hu Shisheng. A damage
on intense dynamic loading and its effects. Beijing:
modified nonlinear viscoelastic constitutive relation and
Science Press, 1986:787-792.
failure criterion of PMMA at high strain rates[J].
[4] Wang Lili, Bao Hesheng, Lu Weixian. The dependence of
Explosion and Shock Waves, 1992, 12(4):333-342.
adiabatic shear banding on strain-rate, strain and temper-
,
,
315211
:
.
.
:
.
;
: DM-ZWT
BP
25
, 2012 1
Vol.25 No.1, Jan. 2012
Spatio-temporal Model of Multiscale Fracture in Brittle Solids

PETROV Y V1,2, BRATOV V1,2
( 1.Saint Petersburg State University, Saint Petersburg 198504, Russia; 2. Inst Probl Mech Engng of
the Russian Academy of Sciences, Saint Petersburg 199178, Russia )
Abstract: We analyze examples illustrating typical dynamic effects inherent in dynamic fracture process,
and we propose a unified interpretation for fracture of solids utilizing structural-temporal approach based on
the concept of the fracture incubation time. Corresponding generalized model accounting for fracture scale
level will also be presented. The model will be used to predict fracture of quasi-brittle heterogeneous
materials on different scale levels. It will be demonstrated that can give a possibility to predict fracture on a
higher (real) scale level and having experimental data obtained on a lower (laboratory) scale level.
Key words: incubation time criterion; dynamic fracture; dynamic strength; fracture toughness; spatiotemporal discretization; scale levels; crack propagation; pipelines; FEM modelling.
CLC number: O342
Document code: A
Article ID: 1001-5132 2012 01-0035-06
Experiments on the dynamic fracture of solids
we analyze examples illustrating typical dynamic effects
caused by intense dynamic loads or focused energy
inherent in dynamic fracture of brittle media. We
fluxes reveal a number of effects indicating fundamental
propose a unified interpretation for the fracture of solids
differences between dynamic rupture (breakdown) of
using the structural-temporal approach based on the
materials and a similar process under slow quasistatic
concept of the fracture incubation time [1-4]. This paper is
loads. For example, one of the basic problems in testing
also presenting structural-temporal approach for analysis
of dynamic-strength properties of materials is associated
of multiscale nature of brittle fracture. Problem of
with the dependence of the limiting characteristics on
experimental determination of a fixed scale level is
the duration, amplitude, and rate of an external load, as
discussed. Possible interconnections of this scale level
well as on a number of other factors. Whereas a critical
with a higher and a lower scale levels are discussed.
value is a constant for a material in the static case,

experimentally determined critical characteristics in
Incubation-time criterion
dynamics are strongly unstable, and as a result, their

behavior becomes unpredictable. The indicated (some
The basic cause of difficulties in modeling the
other) features of the behavior of materials subjected to
aforementioned effects of mechanical strength is the
pulsed loads are common for a number of seemingly
absence of an adequate limiting condition that determines
quite different physical processes, such as dynamic
the instant of rupture or breakdown. This problem can
fracture (crack initiation and scabbing), cavitation in
be solved using both the structural macromechanics of
liquids, and electrical breakdown in solids. In this paper,
fracture and the concept of the fracture incubation time,

Foundation items: Supported by the National Natural Science Foundation of China (11032001); Russian Foundation for Basic Research (11-01-91217).
The first author: PETROV Y V (1957- ), male, Correspondence Academician of the Russian Academy of Sciences, professor,
reaserch domain: dynamic fracture. E-mail: yp@YP1004.spb.edu
36
2012
which represents the kinetic processes of the formation

[1-3]
practice, the break of a material (sample) is a process
. The above effects become
proceeding in time, and the function f (t ) describes
essential for loads with periods comparable to the scale
the micro-scale level kinetics of the passage from a
determined by the fracture incubation time associated
conditionally defect-free state ( f (0)
with preparatory relaxation processes of microdefects
pletely broken state at the given point ( f (0) 1) that
development in the material structure.
can be associated with the macro-fracture event (Fig. 1).
of macroscopic breaks
0) to the com-
The criterion for the fracture incubation time
On the other hand, applying fracture criterion (1) to
proposed in [1-4] makes it possible to calculate effects
macro-scale level situation (F (t ) PH (t )) , we arrive at
of the unstable behavior of dynamic-strength charac-
the relation for time to fracture t*
for P
Fc .
teristics. These effects are observed in experiments on

the fracture of solids. This criterion can be generalized
in the form of the condition:
t
1
t
F (t )
Fc
dt
1,
(1)
where F (t ) is the intensity of a local force field causing the fracture of the medium, Fc is the static limit of
the local force field, and
is the incubation time
Fig. 1
Schematic micro-scale level kinetics of the

fracture of a sample at the break point
associated with the dynamics of a relaxation process

preparing the break. The fracture time t* is defined as
In other words, the incubation time introduced
the time at which condition (1) becomes an equality. The
above is equal to the duration of the pre-fracture process
parameter
after the stress in the material has reached the static
characterizes the sensitivity of a material
strength on the given scale level. This duration can be
to the intensity of the force field causing fracture.

By example of the mechanical break of a material,
measured in experiments on quasi-static fracture of
we now consider one of the possible methods of
samples measuring the time of the decrease of stress at
interpreting and determining the parameter
. We
the unloading wave front, which can determined by
assume that a standard sample made of a given material
interferometric (visar-based or photoelasticity-based)
under tension is broken into two parts under the stress
method using the velocity profile of points on the
P arising at a certain time t
sample. We analyze examples of the actual utilization of
0 : F (t )
PH (t ) , where
H (t ) is the heaviside step function. In the case of
criterion (1) in various physicomechanical problems.
quasi-brittle fracture, the material is unloaded, and the

local stress at the break point decreases rapidly (but not
instantaneously) from P
Fracture of solids
to 0. In this case, the
corresponding unloading wave is generated, propagates
A typical example illustrating the complicated
over the sample, and can be detected by well-known
behavior of the dynamic strength of solids is the time
(e.g., interferometric) methods. The stress variation at
dependence of strength observed under scabbing
the break point can be conditionally represented by the
conditions
dependence
P Pf (t ) , where the function f (t )
time t* on the critical pulse amplitude P* for different
varies from 0 to 1 within a certain time interval T . The
pulse durations shows that the dynamic strength is not a
case f (t )
H (t ) corresponds to the classical theory of
material constant but is depending on the time-to-
strength. In other words, according to the classical
fracture (i.e. sample lifetime). The criterion of critical
approach, break occurs instantaneously ( T
stress
(t )
0 ). In
[5]
(t )
(Fig. 2). This dependence of the fracture
, where
is the static strength, is
PETROV Y V, et al: Spatio-temporal Model of Multiscale Fracture in Brittle Solids
able to predict long-term quasistatic fracture caused by

long-duration wave pulses
the amplitude and
(t )
P (t ) , where P
is
(t ) is the load time profile
1
t
where
1
dx
37
(x',t ) dx'dt
(3)
is the incubation time of a fracture process (or
function. However, in the case of short-duration pulses,
fracture microstructural time): a parameter characteri-
fracture time weakly depends on the threshold pulse
zing the response of the material to applied dynamic
amplitude, and this dependence has a certain asymptote.
loads (i.e. is constant for a given material and does not
This effect is called the phenomenon of the mechanical
depend on problem geometry, the way a load is applied,
branch of the strength time dependence. Neither the
the shape of a load pulse or its amplitude). d is the
conventional theory of strength nor known time criteria
characteristic size of a fracture process zone and is
explains this phenomenon.
constant for the given material and chosen spatial scale.
Time dependence of strength can be obtained on
is normal stress at a point, changing with time and
the basis of the incubation time criterion (1). For the
is its critical value (ultimate stress or critical tensile
scabbing fracture under consideration, this criterion
stress evaluated in quasistatic conditions). x* and t*
takes the form of the limiting condition previously
are local coordinate and time.
proposed in [1-4]:
t
1
(t ) dt
Assuming
c
(2)
(2 / )( K Ic2 /
2
c
),
(4)
where K Ic is a critical stress intensity factor for mode I
(t ) is the time dependence of the local stress
loading (mode I fracture toughness), measured in
at the break point. The scheme for the application of
quasistatic experimental conditions, it can be shown that
criterion (2) to split problems is given by Petrov and
within the framework of linear fracture mechanics for
where
Morozov
[1,3-4]
. An example of a calculation utilizing
criterion (2) in order to predict time dependence of the

strength of aluminum (
0.75 s,
103 MPa) for
triangular pulses realized in the experiments reported by

Zlatin et al [5] is represented in Fig. 2 by solid line.
case of fracture initiation in the tip of an existing mode I

loaded crack, Eq. (3) is equivalent to:
t
1
K I (t ) dt
K IC
(5)
Condition (2) arises from the requirement that (1)

is equivalent to Irwins criterion K I (t )
tatic conditions ( t
K Ic in quasis-
). This means that a certain size
typical for fractured material appears. This size should

be associated with a size of a failure cell on the current
spatial scale-all rupture sized essentially less than d
cannot be called fracture on the current scale level.
Thus, by introduction of
and d time-spatial
domain is discretized. Once material and scale one is

working on are chosen, gives a time, such, that energy,
Fig. 2
Logarithm of the fracture-process duration t* vs.
accumulated during this time can be released by rupture
the threshold amplitude P* of a stress pulse that causes
of the cell that accumulated it. Linear size d assigns
scabbing in an aluminum sample (Zlatin et al [5]).
dimensions for the cell. Introduction of temporal and
Incubation time fracture criterion, originally propo-
spatial domain discretization is a very important step. To
sed to predict crack initiation in dynamic conditions,
our belief, a correct description of high loading rate
was formulated in [1-4]. This criterion for fracture at a
effects is not possible if this time-spatial discreteness is
point x, at time t, reads:
not accounted somehow. Advantage of incubation time
38
2012
approach is that one can stay within the framework of
the set of material parameters of the criterion
continuum linear elasticity, utilizing all the consequent
(or
advantages and accounting discreetness of the problem
We suppose that any given scale level is characterized
only inside critical fracture condition.
by two characteristic lengths representing its lower and
As it was shown in multiple publications
[4]
, crite-
, d,
, K Ic , ) is associated with the given scale level.
upper limits:
rion (3) can be successfully used to predict fracture
Lower boundary:
initiation in brittle solids. For slow loading rates and,
hence, times to fracture that are essentially bigger than ,
Objects sized essentially less then the lower
condition (3) for crack initiation gives the same
boundary cannot display fracture on the given scale
predictions as classical Irwins criterion of the critical
level. Samples sized essentially less that this boundary
stress intensity factor. For high loading rates and times
cannot be used in order to experimentally evaluate
to fracture comparable with
fracture parameters on the given scale level.
all the variety of effects
experimentally observed in dynamical experiments
[6-8]
2
c
).
(6)
Upper boundary:
can be received using (3) both qualitatively and quanti[2-3]
(2 / )( K Ic2 /
c ,
(7)
. Application of condition (3) to description
where c is the speed of the energy transport. Thus,
of real experiments or usage of (3) as a critical fracture
biggest possible volume, the energy able to produce
condition in finite element numerical analysis gives a
fracture can occupy within the time equal to incubation
possibility for better understanding of fracture dyna-
time is introduced. Objects sized essentially more than
tatively
micss nature
[9-10]
and even prediction of new effects
typical for dynamical processes
[11]
the upper boundary cannt correctly display fracture

parameters on the given scale level. Samples sized
Rate dependences K Id of the dynamic fracture

toughness, which were observed in experiments, are
essentially more that boundary cannt be used in order to

evaluate fracture parameters on the given scale level.
characterized by a strong instability and can noticeably
Permissible characteristic sizes L of tested
change when varying the duration of the load rise stage,
specimens for the given scale level (object sizes for
the shape of the time profile of a loading pulse, sample
which any fracture theory based on strength properties
geometry, and the method of load application
[6-8]
. The
calculations based on the concept of the incubation time

corresponding to the conditions of a number of experiments were carried out by Petrov and Morozov
[2]
. The
results show that the dynamic fracture toughness is not
, K Ic , ) measured on this scale level is valid) should
stay within approximate range: d
Thereby it is supposed that any i-th scale level is

characterized by a pair of linear sizes { di , Di } and the
range of permissible object dimensions:
di
employment of both the criterion of the critical intensity
At the same time we suppose that
K Id and the characteristic K Id as
a material parameter defining the dynamic fracture

(in analogy to the static parameter K Ic ) are incorrect.
di
Di .
an intrinsic characteristic of a material. Therefore, the

coefficient K I (t )
D.
Di ,
(8)
(9)
and therefore
i
di 1 / c .
(10)
Thus, knowing strength properties (including
Interconnection of rupture processes on different scale levels
fracture incubation time) on some scale level it is

possible to estimate the upper boundary for smaller
scale level, the lower boundary for the larger scale level
Consider fracture process that is determined and
and the incubation time for the smaller scale level.
controlled by the incubation time criterion. Assume that
Modified incubation time fracture criterion is giving a
PETROV Y V, et al: Spatio-temporal Model of Multiscale Fracture in Brittle Solids
39
possibility to establish interconnections between fracture
FEM modeling was performed for three different
on different scale levels. It also provides a possibility to
steels (X80, X90, X100) [14]. Fig. 3 represents crack ex-
assign fracture scale level for different experiments on
tension histories for pipelines made from three different
the same material.
steels and gives an overview of the propagating crack.
The abovementioned ideas were applied while

modeling propagation of dynamic crack in trunk gas
pipeline
[12]
. In these experiments a section of gas pipe
was loaded by internal pressure close to operational

pressure inside the gas pipeline. A furrow was made in
the part of the pipe parallel to its central axis. The
furrow was filled with an explosive substance. When the
explosive is blasted the crack starts to propagate from
Fig. 3
the furrow. Pipelines made of several different pipe
Crack extension histories for pipelines

made of different steels
steels were tested.

The main difficulty was connected with a fact that
It was found that in the modeled situation the speed
the experimental data about strength properties of steel
of the crack is close to the speed of the acoustic wave in
were evaluated on laboratory scale level (samples sized
gas that determines the speed of the front of the pressure
several centimeters) while gas pipeline with diameter of
drop. This leads to a conclusion about instability of
1.5 m is a steel shell and almost 10 m wide and infinite
crack propagation regimes in the modeled situation that
in another direction was to be modeled. Laboratory tests
a small change in properties of the pipeline material can
of relatively small specimens of that particular pipe steel
result in qualitative change it crack propagation regime,
showed large plastic zones and big ductility while large
and should the speed of the crack be higher than the
scale tests of the pipes discovered long distant crack
speed of acoustic signal in gas, the crack will never
propagations that mostly could be interpreted as quasi-
arrest. It was found that length of the resulting crack
brittle behavior on the large scale level. FEM model
does strongly depend on material of the pipeline and the
predicting propagation of crack in the pipe-line was
length of the resulting crack does vary significantly
developed. Utilization of material properties received on
(from 3 to 300 m) though all steels had very similar
laboratory scale as input parameters for the FEM model
properties. The origin of this instability was understood
resulted in crack propagation histories drastically diff-
due to numerical analysis presented above. As a result,
ering from the ones observed in reality. The abovemen-
received crack extension histories were very close to the
tioned approach based on the criterion introduced above
ones measured experimentally [12-13].
was applied in order to modify material properties.

The model was created for a section of pipeline
with length of 9 m. Pipeline diameter is 1.22 m
[12-13]
Conclusions
Considering large scale fracture of a trunk pipe we used
Thus, the examples considered above show the
quasi-brittle approach based on the assumption that the
fundamental importance of the incubation processes
size of an element on the crack path is equal to D
c1 ,
preparing abrupt structural changes (fracture and crack
where c1 is the speed of the longitudinal wave in steel
propagation) in continuum under intense pulsed loads.
(pipeline material) and
is the fracture incubation
The fracture incubation time is evidently a universal
time for steel measured on laboratory scale were the
basic characteristic of the dynamic strength and must
fracture process was accompanied by large plastic zone.
become one of the main material parameters to be
40
2012
experimentally determined (measured). The above
gation into dynamic fracture: Crack initiation and arrest
results show that the incubation-time approach is funda-
[J]. Int J Fract, 1984, 25:247-262.
mental and makes it possible to adequately represent the

dynamics of fracture in solids on various scale levels. A
corresponding generalized model accounting for fracture
[7] Kalthoff J F. Fracture behaviour under high rates of

loading[J]. Engng Fract Mech, 1986, 23:289-298.
[8] Dally J W, Barker D B. Dynamic measurements of
initiation toughness at high loading rates[J]. Exp Mech,
scale level is also presented. It was shown that this can
1988, 28:298-303.
give a possibility to predict fracture on a higher (real)
[9] Bratov V, Petrov Y. Application of incubation time
scale level having experimental data obtained on a lower
approach to simulate dynamic crack propagation[J]. Int J
(laboratory) scale. This possibility is of extreme
Fract, 2007, 146:53-60.
importance for many applications where the possibility
[10] Bratov V, Petrov Y. Optimizing energy input for fracture

by analysis of the energy required to initiate dynamic
to evaluate material strength properties on real structure
mode I crack growth[J]. International Journal of Solids
scale level does not exist (ex. geological objects, big

concrete structures, trunk pipelines, etc.)
and Structures, 2007, 44:2371-2380.

[11] Petrov Y, Sitnikova E. Temperature dependence of spall
strength and the effect of anomalous melting tempera-
References:
tures in shock-wave loading[J]. Technical Physics, 2005,
[1] Petrov, Yu V. On quantum nature of dynamic fracture in
50:1034-1037.
brittle solids[J]. Sov Phys Dokl, 1991, 36:802-804.
[12] Abakumov A I. An experimental study of buckling of
[2] Petrov Y V, Morozov N F. On the modeling of fracture of
cylindrical shells subjected to static and dynamic axial
brittle solids[J]. J Appl Mech, 1994, 61:710-712.
impact[J]. International Journal of Modern Physics B,
[3] Petrov Y V, Morozov N F, Smirnov V I. Structural macromechanics approach in dynamics of fracture[J]. Fatigue
2008, 22(9/11):1369-1376.
[13] Igi S, Akiyama, T. Multiscale fracture model for quasi-
Fract Engng Mater Struct, 2003, 26:363-372.
brittle materials[J]. Applied Mechanics and Mateirals,
[4] Morozov N, Petrov Y. Dynamics of Fracture[M]. BerlinHeidelberg, New-York: Springer-Verlag, 2000.
2011, 82:160-165.
[14] Petrov Y, Bratov V. Multiscale Fracture Model for Quasi-
[5] Zlatin N A, Mochalov S M, Pugachev G S, et al.Temporal
brittle Materials[J]. Applied Mechanics and Materials
features of fracture in metals under pulsed intense actions
2011, 82:160-165.
[J]. Sov Phys Solid State, 1974, 16:1137-1140.

[6] Ravi-Chandar K, Knauss W G. An experimental investi-
PETROV Y V1,2, BRATOV V1,2

1.
198504; 2.
199178
.
:
25
, 2012 1
Vol.25 No.1, Jan. 2012
Hierarchic Nature of Dynamic Deformation and

Fracture of Rock Mass
QI Cheng-zhi1, WANG Ming-yang2, CHEN Jian-jie2, QIAN Qi-hu2
( 1.Beijing Research Center of Engineering Structures and New Materials, Beijing University of Civil Engineering and
Architecture, Beijing 100044, China; 2.Engineering Institute of Cops of Engineers,
PLA University of Science and Technology, Nanjing 210007, China )
Abstract: Dynamic deformation and fracture process of rock mass are investigated from the perspective of
structural hierarchy. The relationship between spatial scales of deformation and fracture with strain rate is
explored in the framework of relaxation model. The velocity of crack propagation in dependence on loading
intensity is discussed, and the hierarchic nature of deformation and fracture are examined. At last the
fracture criteria are discussed from the aspects concerning structural hierarchy of rock mass. The
investigation shows that dynamic deformation and fracture process of rock mass possess hierarchic nature
which depends on the spatial and temporal characteristics of external loading, structural hierarchy of rock
mass and finiteness of the velocity of the fracture process. The temporal criterion and limit strain criterion
may serve as fracture criteria adequately well.
Key words: hierarchy of internal structure; rock mass; dynamic deformation and fracture; hierarchic nature
CLC number: TU452
Document code: A
Introduction
Article ID: 1001-5132 2012 01-0041-07
of geo-blocks exists:
Generally in continuum mechanics solid is

represented
as
ideal
continuum
medium
whose
deformation under external forces is completely
where
( 2)
(1)
,
6
2.5 10 m is the radius of Earths core; i
is positive integers.
According to the investigation in [2], the ratio of
reversible if the internal stresses in solid dont exceed
openings of cracks
their limit values characterizing the strength of solid.
separated by these cracks at i-th scale level is stable, and
But real materials have complex internal structure

which has decisive effect on mechanical behavior of
to linear size of blocks
can be described by the next relation that has a normal

statistical distribution
materials. For rock mass an important peculiarity of
( )
such discreteness is the similarity of the internal
10 2 ,
(2)
structure in a wide range of sizes. Investigation showed
where
that [1], a fundamental canonical series for the sizes
2, and parameter
is a coefficient changing in the interval 1/2

is termed as geo-mechanical

Foundation items: Supported by the National Natural Science Foundation of China (11032001, 51174012, 50825403) and by Russian Foundation for Basic
Research (11-01-91217); 973 Key State Research Program (010CB732003); The Innovation School Foundation (51021001); Beijing
Natural Science Foundation (KZ200810016007); Scientific School of Modeling and Analysis of Nonlinear Systems (PHR201107123).
The first author: QI Cheng-zhi (1965 ), male, Taian, Shandong, doctor/professor, research domain: shock dynamics, geo-mechanics, structural dynamics.
E-mail: qczbicea@yahoo.com.cn
42
2012
invariant in [2].
mechanical behavior of rock mass. If the strength of
The structural hierarchy of rocks specifies the
crystals with ideal regular lattices is their theoretical
hierarchy of deformation and fracture processes. Slow
strength, then the strength of real materials is about 2
processes, such as energy storage between earthquakes
3 orders lower than their theoretical strength. Obviously,
at the geotectonic levels as well as the diffusion stress
the complex hierarchic internal structure of real
waves, tectonic solitary waves with different periods,
materials will causes the stress concentration and strain
rotary waves, seismicity caused by the oil and gas
localization which are responsible for lowering of real
production, and the construction of petroleum storage
material strength. Accordingly the stresses in hetero-
reservoirs etc., commonly occur in the hierarchical
geneous solids may be looked as being composed of two
[3-4]
levels with geo-blocks of great sizes
components[6]: elastic stresses induced by the reversible
The fast processes, for instance, the deform[5]
volume and shear deformations, and the local inelastic
ation and fracture of materials under shock loading , are
stresses in heterogeneities which are responsible for the
related to the mesoscopic or microscopic structural levels.
irreversible deformations. The elastic stresses are related
The interim processes, including micro-earthquakes,
to the reversible deformations linearly. As to the residual
rock-bursts, and so on, occur at levels between the
stresses (inelastic stresses), they arise at definite strain
[6]
tectonic and meso-microscopic levels .
rate, and relax with time. The evolution equation for the
It has been from the conceptions of the structural

hierarchy that scientists have been studying the laws of
deformation and fracture of geo-medium at different
structural levels since the 1970s. For example, the
relations between the sizes of geo-blocks and the
fracture times were examined in [3,7]. The comparisons
between the stiffness of inter-block layers and the
thickness of geo-block layers were analyzed in [8-9].
The relations between the structural hierarchy, the
viscosity, and the strength of geo-blocks were
investigated in [10-11]. However, the hierarchic nature
of deformation and fracture of rock mass are still not
fully understood, the essence of the relations between
spatial and temporal scales is not revealed completely,
the deformation and fracture processes of rock mass
from the viewpoint of structural hierarchy is not fully
clarified. The present paper is devoted to the
investigation of the above mentioned problems in order
to reveal the hierarchic nature of deformation and
fracture of rock mass.
residual stress deviator
The relationship between temporal

and spatial scales of deformation
and fracture of rock mass
The internal structure has decisive impact on
sijl in heterogeneities with
characteristic linear size l may be described by Maxwell

model
d sijl
dt
2 cs2 eij
sijl
l
(3)
sijl is the residual stress deviator components
where
in heterogeneities with characteristic scale l ; eij is the

residual strain rate deviator components;
is the
density of the medium; v is the relaxation velocity;

cs is the elastic shear wave propagation velocity. Here
we suppose that all residual stress deviator components

relax with the same relaxation time.
The main feature of this model is that, the
relaxation rate of the residual deviator stresses in
heterogeneities is proportional to the magnitude of the
residual stresses, and inversely proportional to the size
of the heterogeneities.
The solution of equation (3) has the following
form:
sijl
l
2 cs2 eij (1 e
v
vt / l
),
(4)
with the time the residual deviator stresses approach

their stationary values
l
sijl 2 cs2eij .
v
(5)
QI Cheng-zhi, et al: Hierarchic Nature of Dynamic Deformation and Fracture of Rock Mass
Substituting equation (5) into the expression

l
ij
l
ij
3 s s / 2, for intensity of residual stress
deviator, we obtain
where
2eij eij / 3,
(6)
is the residual strain rate
intensity.
It can be seen from equations (5) and (6) that at
given strain rate the greater the size of the
heterogeneities is, the greater the residual stresses are. If
the size of the body is infinite, then we can always find
large enough heterogeneities that their residual stresses
are large enough to cause the fracture of the body. If we
denote the limit residual stress causing fracture of the
body as
3 cs2
(8)
v,
which denotes the propagation distance of crack at the
l
3 cs2 eij ,
v
43
, then we can determine the minimal size of
moment of failure.
Propagation velocity v of crack depends on loading conditions. Experimental observations of fracture
propagation[12-13] indicate that crack may grow for
energy lower than the critical limit of fracture. At microscale, the tensile failure due to the sub-critical propagation of cracks may represent the main micromechanism of creep at the macro-scale. The dependence
of fracture propagation rate on the stress intensity factor
in mode I may be approximated by tri-modal behavior,
which is shown in Fig. 1.
heterogeneities corresponding to the limit residual stress

:
l
v
2
s
3 c
(7)
In this way, at constant strain rate among the

parameters of solid a parameter with dimension of
length arises.
It can be seen from equation (7) that, for any body
we can find one strain rate that does not make the body
Fig. 1
The dependence of crack propagation velocity on

stress intensity factor
fractured. In the frame of this model such a behavior of
In region I, the rate of stress corrosion reaction
solid corresponds to creep.
controls the velocity of crack growth. Region II is
Hierarchic natures of deformation

and fracture of rock mass
mainly determined by the rate of transport of reactive

species to crack tips. In region III, the velocity of crack
growth increases drastically up to failure and is
Now let us analyze equation (7). For plastically

non-compressible rock mass Poissons ratio
Youngs elastic modulus is E
2
s
term
2(1
0, and
)G
2(1 0.5)
2
s
3 c . Therefore in one-dimensional case the

2
s
/ (3 c ) on the right of equation (7) represents
the magnitude
of deformation corresponding to
2
s
/ (3 c ) denotes
, and the term
strength limit
the time for strain to reach deformation

strain rate
2
s
/ (3 c )
at constant
. v may be
looked at as the propagation velocity of crack. Therefore

equation (7) may be rewritten as
relatively independent of the chemical environment and

is controlled by mechanical rupture.
Considering crack growth as a succession of water
vapor enhanced rupture of small material element
immediately adjacent to the crack tips, Salganik et al[14]
showed that the three regimes of crack propagation can
be predicted, respectively, by three following equations:
vI
U
v0 0n
exp( 01
nmol
vII
b0 p
,
nmol N A 2 mkT
K1 / 2 d m
),
kT
(9)
(10)
44
2012
vIII
dm
exp(
U 03
K1 / 2 d m
),
kT
(11)
phase that induces dynamic response of materials. The

shorter the rise time tr is, and the greater the stress
is the typical
magnitude is, the higher is the strain rate. If we denote
period of atomic fluctuations; v0 is the mean velocity of
the wave propagation velocity in material as c , the more
diffusion;
is the relative concentration of water in
the strain rate is, the less is the region ctr covered by
the gas next to the crack tips; nmol is the number of
the rise phase of the wave, and the higher is the strength
molecules of water required for the water assisted
of material.
where T is the absolute temperature;
rupture of a single bridging bonding; d m is the length of
From other hand, in the region ctr covered by the
the material structure; b0 is the bridging bond length;
rise phase of stress wave, because of the finiteness of the
N A is the Avogadro number; m is the molecular mass
wave propagation velocity v
of water; p is the partial pressure of water vapor; U 01
elements with maximum size D0 ~ ctr , the strength of
and U 03 are the zero stress activation energy in regions I
structural surface of material is
and III, respectively; Q1 and Q3 are stress sensitivity
structural surface to open is t0
factors in regions I and III, respectively.
the growth of loading is so faster that before the
vC , for 0-th level
N0
, the entire time for
N0
c 2 ) D0 / v . If
/(
Experiments show that the limit crack growth
complete failure of 0-th level elements the magnitude of
velocity is about 0.4 of Reyleigh wave speed CR [15].
stress reaches the strength of the 1-th level element with
Recently the relationship between limit crack growth
size D1
velocity vc
surfaces of 1-th level elements start to crack. The time
and the mechanical characteristics of
necessary for 1-th level inter- element surfaces to failure
material has been obtained theoretically in [16]

vc ( R ) ((1
) / (4(1
)((
ln(2 R / a) 3 / 8 3/ 32
( E / )1/2 ,
D0 , in this way the inter-element structural
completely is t1
3 /2 7 / 8)
N1
/(
c 2 ) D1 / v .
If before the complete failure of 0-th and 1-th level
a 2 /(8 R 2))))1/2
(12)
elements loading continues to grow, then the initiation

of cracking of inter-element surfaces of the next smaller
where R is the size of region of the material involved in
scale level elements is activated and so on. Thus the
the crack propagation;
is the Poissons ratio; E is
strength of inter-element structural surfaces of the least
is the material density;
scale elements at moment of macroscopic failure is the
Youngs elastic modulus;
a is the half-length of the crack. For example, for

R 2.6a ,
0.25 , equation (21) gives vc 0.38 E /
dynamic strength of material.
Therefore there are two characteristic crack

propagation velocities: the crack growth velocity
corresponding to plateau II vsub and limit crack growth
In this way we can clearly see the intrinsic

relationship between size and strain rate effects. The
fracture time is the minimum of the quantity
(
velocity vC
the thickness of the weakened surfaces is proportional to
c ) Di / v for all involved structural levels, i.e.

frac
According to equation (2), at every structural level
Ni
min(
Ni
c2
(13)
Di / v ).
It is clear that the term
Ni
/(
c 2 ) in the
the characteristic size of the elements at given structural
right-hand side of equation (13) is related to the elastic
level. It is natural to assume that the smaller the
behavior of material, and the second term Di / v
thickness of the structural surfaces is, i.e. the smaller the
represents the relaxation time at i-th scale level.
character size of this structural level is, the higher the

mechanical characteristics of rock mass at this level are.
On fracture criteria of rock mass
The magnitude of strain rates is related to the stress

wave profile. The rise phase of stress wave is the main
In static loading conditions or at low strain rates
45
cracks at the largest scale level have time to develop,
displacement of particles reaches critical value, fracture
and the strength of the structural surfaces at this level
occurs.
controls the static strength
of rock mass. Therefore
Displacement is the macroscopic measure of
force criteria are good enough for the description of
deformation of materials. If the characteristic dimension
the fracture. At high strain rates the situation becomes
covered by shock wave is Lshock , then displacement can
much complicated. If the strain rates and the maximum
be expressed by deformation
c0
as
(18)
attained stresses are high enough, then before the
fracture of rock mass at the largest scale level
Therefore formula (17) becomes
deformation and fracture processes at smaller scale

levels will be activated. Thus force criteria are not
enough for describing the deformation and fracture, they
are only necessary conditions for fracture of rock mass.
In order that fracture takes place at i-th scale level the
loading time must exceeds the time Di / v necessary
for the cracks to propagate through out the elements at
this level. From this viewpoint the concept of incubation
time for fracture proposed by Morozov N F and Petrov
i
Y V[17] is appropriate. Obviously, incubation time tinc
i
for fracture at i-th scale level is tinc
i i
i inc
of the strength
ci
Di / v , the product
and incubation time tinc at
i-th scale level constitutes the critical value of impulse

for the fracture of rock mass.
Nikiforovsky-Shemyakins
impulse
criterion[18]
supposes that, fracture occurs when the integral of local

stress with time, i.e. when the local stress
(t ) impulse
Lshock .
t
0
(t )dt
DLshock
i.e. when deformation of solids reaches critical value,

fracture occurs.
But with the increase of strain rate more scale
levels are involved into deformation and fracture
processes, the critical strain magnitude should increase.
For example if shock wave covers the j-th scale level
element, and if only the structural surfaces at this level
are fractured, then the limit strain at j-th scale level
should be close to geo-mechanical invariant
(20)
jcr
[6]
Experimental data conform this conclusion .

If the next smaller scale level, i.e. (j-1)-th level, is
also activated, and
(t )dt
(14)
Jc.
by Morozov N F and Petrov Y V, equation (14) should

be rewritten as
0
(t )dt
(16)
is the density of materials; v is the velocity
of particle of materials; and D is the velocity of shock

wave. Hence for pulse criterion (14), we have
0
(t )dt
(1
jcr
(21)
If the next two smaller scale levels: (j-1)-th and

(j-2)-th levels are activated, then the limit strain will be
jcr
(1
(22)
and so on.
t
0
Dvdt
Du
increases with the strain rate[19]. The rock-like material,

concrete, also shows the same strain rate sensitivity of
Dv,
1 , then the limit
Experiments show that the limit failure strain really
(15)
c 0 tinc .
In solids the next relation holds

where
According the concept of incubation time proposed
j 1
strain will be
exceeds one critical value Jc, i.e.

0
(19)
Jc ,
Jc ,
limit failure strain[20].

With more and more scale levels being activated,
the limit strain will approach to the following limit
jcr
(17)
where u is particle displacement.

Formula (17) shows that, when macroscopic
(23)
For example, if
1 / 2 , then
jcr
3.4
which agrees with the experimental data in [19-20].

The relative stable nature of limit failure strain
46
2012
allows us to use limit strain criterion as an alternative

fracture criterion for rock mass together with temporal
fracture criteria. The essence of temporal criteria and
limit strain criterion is that the material must has enough
time to develop deformation before limit failure strain is
attained.
34(5):16-33.
[2] Kurlenia M V, Oparin V N, Eremenko A A. On ratio of
linear sizes of blocks to openings of cracks in structural
hierarchy of rock mass[J]. Journal of Mining Science,
1993, 29(2):6-33.
[3] Sadovsky M A, Volkhovitinov L G, Pisarenko V F.
Deformation of geophysical medium and seismic
process[M]. Moscow: Nauka, 1987.
Conclusions
[4] Nikolaevsky V N. Geomechanics and fluid dynamics[M].

Moscow: Nedra, 1996.
Rock mass has complex internal structural

hierarchy, at every scale level mechanical properties of
materials are different. Such structural hierarchy
[5] Meyers M A. Dynamic behavior of materials[M]. New

York: Wiley, 1994.
[6] Radionov V N, Sizov I A, Tsvetkov V M. Fundamental
of geomechanics[M]. Moscow: Nedra, 1986.
specifies the hierarchic nature of deformation and
[7] Kuksenko V S. Physical and methodological fundamental
fracture of rock mass. According to the temporal and
of forecasting of rock-bursts[J]. Journal of Mining
spatial properties of loading different scale levels are

activated in the processes of deformation and fracture of
rock mass. In the present paper the relationship between
spatial and temporal scales of deformation and fracture
is studied from the viewpoint of structural hierarchy. It
Science, 1987, 23(1):9-21.

[8] Kostjuchenko V N, Kocharyan G G, Pavlov D V.
Deformation characteristics of layers between blocks at
different scale levels[J]. Physical Mesomechanics, 2002,
5(5):23-42.
[9] Kocharyan G G, Kuljukin
. Study of collapse of
is shown that the relationship between spatial scales and
underground opening in rock mass with block structure
temporal scales of deformation and fracture of rock
under dynamic Loading, Part II[J]. Journal of Mining
mass is determined by the structural hierarchy of rock

mass and the limitness of crack propagation. The
essence of strain rate effect of strength is that because of
the limitness of crack propagation velocity the increase
Science, 1994, 30(5):27-37.

[10] Qi Chengzhi, Wang Mingyang, Qian Qihu, et al.
Structural hierarchy and mechanical properties of rock
mass, part II[J]. Physical Mesomechanics, 2006, 9(6):4152.
of loading activates the deformation and fracture
[11] Qi Chengzhi, Wang Mingyang, Qian Qihu, et al.
processes at smaller scale levels before the macro-
Structural hierarchy and mechanical properties of rock
fracture of the body, the dynamic strength of materials is
mass, part I[J]. Physical Mesomechanics, 2006, 9(6):29-
the structural surface strength of the smallest activated

structural elements before the macro- fracture of the
sample. It is also shown that temporal criteria and limit
strain criterion may serve as favorable fracture criteria
for rock mass, The essence of temporal criteria and limit
39.
[12] Anderson O, Grew P. Stress corrosion theory of crack
propagation with application to geophysics[J]. Reviews
of Geophysics and Space Physics, 1977, 15:77-104.
[13] Atkinson B, Meredith P. The theory of subcritical crack
growth with application to minerals and rocks[C]//
strain criterion is that the material must has enough time
Fracture Mechanics of Rock. New York: Academy Press,
to develop deformation before limit failure strain is
1987:111-166.
attained.
[14] Salganik R, Rapoport L, Gotlib V. Effect of structure on

environmentally assisted subcritical crack growth in
References:
brittle materials[J]. International Journal of Fracture,
[1] Oparin V N, Jushkin V F, Akinin A A, et al. On new scale
1997, 87:21-46.
of structural hierarchy presentation as basic characteristic

of geological objects[J]. Journal of Mining Science, 1998,
[15] Fineberg J, Marder M. Instability in dynamic fracture[J].

Physical Reports, 1999, 313:1-108.
47
[16] Chekunaev N I, Kaplan A M. Limiting velocity of crack
amphibolites under impact compressive loading[J].
propagation in elastic materials[J]. Journal of Appllied
Chinese Journal of Rock Mechanics and Engineering,

2009, 28(1):2113-2120.
Mechanics and Technical Physics, 2009, 50(4):677-683.
[20] Shi Shaoqiu, Wang Yongzhong. The modified Johnson-
[17] Morozov N F, Petrov Y V. Dynamics of fracture[M].
Cook strength model for C30 concrete with consideration
Berlin: Springer-Verlag, 2000.
of strain rate sensitive micro-damage evolution[C]//Bai
[18] Nikifrovsky B S, Shemyakin E I. Dynamical fracture of
Yilong. Dynamic Response of Materials and Structures,
solid[M]. Novosibirsk: Nauka, 1979.
Hefei: Chinese Science and Technology University Press,
[19] Liu Junzhong, Xu Jinyu, Liu Xiaocong, et al.
2005:98-108.
Experimental study on dynamic mechanical properties of
1.
,
,
2.
100044;
210007
.
,
,
.
,
.
.
:
25
, 2012 1
Vol.25 No.1, Jan. 2012
Existence of Optimal Energy Saving Parameters for

Different Industrial Processes
BRATOV V 1,2, PETROV Y V1,2, VOLKOV G2
( Saint-Petersburg State University, Saint-Petersburg 198504, Russia; 2.Institute of Problems of Mechanical Engineering,
Russian Academy of Sciences, Saint-Petersburg 198504, Russia )
Abstract: It i demonstrated that energy input for fracture in many industrial processes can be optimised so
that the energy cost of the process is minimised. Using a simple example of central crack it is shown that for
a certain shape of the load pulse energy transmitted to the sample in order to initiate the crack has a strongly
marked minimum. Received results indicate a possibility to optimise energy consumption of different
industrial processes connected with fracture. Possible applications include drilling or rock pounding where
energy input often accounts for the largest part of the process cost. Using this approach it will be possible to
predict optimal operational parameters for bores, grinding machines, etc. and hence significantly reduce the
process cost. In the second part of the paper the behaviour of energy input for initiation of fracture in
conditions of contact interaction is studied. It is considered that a spherical particle is impacting the
half-space. Stress field created as a result of the interaction can be estimated using the Hertz solution.
Threshold particle velocity (and, hence, threshold kinetic energy) corresponding to initiation of rupture in
the half-space can be found once the fracture criterion is defined. It will be shown that the value of this
energy does significantly depend on load duration and has a marked minimum. Existence of energetically
optimal modes of dynamic impact is claimed.
Key words: energy saving; optimal operational mode; incubation time; brittle fracture; rocks
CLC number: O34
Document code: A
The possibility of optimizing the amount of energy
Article ID: 1001-5132 2012 01-0048-05
ation becomes evident.
required to fracture materials is of great interest in

connection with many applications. Energy inputs for
fracture induced by short impulse loadings are of the
Estimation of energy required to

initiate central crack in a plate
major importance in such areas as percussive, explosive,

hydraulic, electro-impulse and other means of mining,
In [2] an attempt to estimate the amount of energy
drilling, pounding etc. In these cases energy input
sufficient to initiate mode I loaded central crack in a
usually accounts for the largest part of the process cost
plate subjected to plane strain deformation was made.
(e.g. [1]). Taking into consideration the fact that the
The study involves analysis of interaction of the wave
efficiency of the mentioned processes rarely exceeds a
approaching from infinity with an existing central crack
few percent, the importance of energy
in a plane. The existing crack is oriented parallel to the
inputs
optimiz-

Foundation items: Supported by the National Natural Science Foundation of China (11032001) and by Russian Foundation for Basic Research (11-01-91217).
The first author: BRATOV V (1978 ), male, doctor/senior researcher, research domain: dynamic fracture. E-mail: vladimir@bratov.com
BRATOV V, et al: Existence of Optimal Energy Saving Parameters for Different Industrial Processes
49
front of the wave. The process was analyzed utilizing

the finite element method. The ABAQUS finite element
package was used to solve the problem. Computations
were performed for granite ( E = 96.5 GPa,
-3
kgm ,
and
= 2 810
= 0.29, where E is the elasticity modulus

is the Poissons ratio). The results of the
investigation will qualitatively hold for a large variety of

quasi-brittle materials.
Fracture criterion fulfilment was checked for
different load amplitudes and durations. Dependence of
*
Fig. 1
Filled area corresponds to a set of possible pulses
time-to-fracture T on the amplitude of the load applied
leading to crack initiation (at T =72 s energy needed to
was investigated. Time-to-fracture is the time from the
advance the crack is minimized)
beginning of interaction between the wave package and
With the majority of non-explosive methods used
the crack to fracture initiation. Incubation time criterion
to fracture materials (drilling, grinding etc.), it is
[3-4]
was adopted. A similar approach to be
possible to control amplitude and frequency of impacts
used in the case of short cracks is given in [5]. Using the
from the side of a rupture machine. The performed
incubation time criterion, the dependence of time-to-
modeling shows that at a certain load duration (at impact
fracture on the amplitude of the load pulse applied was
fracture of large volumes of material, impulse duration
studied. Values of K
72 s
is connected to the frequency of the machine impacts)
typical for the granite were used. Integration of the
energy input for crack propagation has a marked
temporary dependence of the stress intensity factor was
minimum. Analogously to Fig. 1, it is possible to plot
done numerically. Energy transmitted to the sample by a
the limiting curve for the set of energy values leading to
virtual loading device in the process of impact can be
propagation of a crack in the sample at different load
evaluated analytically. A specific (per unit of length)
amplitudes. Thus, it is possible to establish ranges of
energy transmitted to the stripe can be calculated using
amplitudes and frequencies of load, at which energy
the solution for the uniformly distributed load acting on
costs for fracture of the material are minimized. These
a half plane. This problem can be easily solved utilizing
ranges are dependent on parameters of fractured
the D Alembert method. Fig. 1 is giving a limiting
material, predominant length of existing material cracks
curve for a set of energies that, being transmitted to the
and the way the load is applied.
of fracture
2.4 MPa m and
sample, cause the crack propagation. The minimum
Dependence of the optimal load parameters on the
energy able to increment the crack length (172e6 J) is
crack length was also studied. It was found that the
reached at load pulses with duration of 78 s . As it is
duration of the load, that minimizes energy, and
evident from Fig. 1, the minimum energy, required to
momentum
propagate the crack by impacts with durations differing
dependent on the existing crack length. With the
greatly from the optimal one, significantly exceeds the
disappearing crack length, the duration of the load
minimum possible value. Thus the minimum energy,
minimizing momentum needed for crack propagation
initiating the crack for the load with duration of 92 s
approaches zero. At the same time, the duration optimal
(at this impact duration crack propagation is possible
for the energy inputs tends to the micro-structural time
with the impact of threshold amplitude), will exceed the
of the fracture process. The maximum possible time-to-
minimum energy possible by 10 , and at duration
fracture also tends to the micro-structural time of
of 40 s it will be more than two times bigger.
fracture.
inputs
are
linearly
or
quasi-linearly
50
2012
Minimization of fracture energy

in the case of contact interaction
of loading rates and load durations. In this regard the

incubation time fracture criterion provides a unique
possibility to achieve correct estimations of conditions
While analyzing dynamic strength of materials,
leading to fracture for the complex problem of rocks
one is facing a contradiction between available
spudding. Suppose that the shape of a loading pulse can
experimental results and classical quasi-static appro-
be approximated by a smooth function:
aches in fracture. Numerous experiments demonstrate
exp(1 / (1 (2t / t0 1) 2 ),
that under high-rate dynamic loads, materials are able to
(t )
| t t0 / 2 |
endure loads significantly exceeding fracture loads in

static (quasi-static) conditions. At the same time, in
some of the experiments fracture in dynamic conditions
is initiated at a moment when local stresses at a rupture
point are significantly less as compared to stresses
leading to fracture initiation in static (quasi-static)
conditions. These obvious contradictions led to attempts
to correct and generalize classical fracture criteria in
order to make it applicable in the case of high-rate loads.
This led to the appearance of a concept of dynamic
t0 / 2,
0, | t t0 / 2 | t0 / 2,
where t0 is for the load duration. Then the load is

given by
(t )
where
max
(t ) ,
(1)
is the load amplitude. Substituting
max
equation (1) into incubation time fracture criterion, one

can obtain the critical amplitude
max
leading to
fracture and corresponding to execution of the

incubation time criterion:
*
strength, depending not only on the material properties,
/ max
(s )ds.
t 0; t0
but also on the loading rate and even the time-shape of
As an option for the way the energy is delivered to
the load pulse[6-7]. Practical utilization of this approach
the fracture zone, consider a problem of impact
is rather complicated and often impossible, as there is no
interaction. In [8], Petrov et al analyzed a problem for a
possibility of evaluating dynamic strength for all variety
spherical particle having radius R and velocity V
of loading rates and load shapes.
impacting an elastic half-space using the classical Hertz
Most researchers dealing with problems of dynamic
contact scheme. The maximal stresses appearing in the
fracture are using fracture criteria based on extrapolation
half-space and the duration of interaction between the
of quasi-static fracture criteria to dynamic conditions.
particle and the half-space were calculated. According to
Though they normally account for inertia and temporal
the Hertz hypothesis, the contact force P arising
characteristics of the load applied, temporal character-
between the particle and the half space can be presented
istics of the fracture process are usually not taken into
as
consideration. Utilizing this kind of approach, it is

impossible to predict a critical situation, leading to
fracture, applicable to both dynamic (high-rate loads)
P (t ) = kh3/ 2 ,
k
4/3 R
E
2
(1
(2)
and quasi-static cases. In this section, an incubation time
where h is a particle penetration and
fracture criterion is used in order to predict fracture in
Poissons ratio of the elastic media. The maximal
the case of contact interactions. Employing this
penetration h0 can be found as
approach one does not need to worry about the time
h0
(5mV 2 / (4k )) 2/5 ,
is the
(3)
scale of the problem: the criterion gives correct
where m is the mass of a particle. The impact duration
predictions in a wide range of loading rates from static
can be presented as
problems to the extremely dynamic ones. For the

present analysis, we need to consider a wide range
t0
2h0
V
1
0
5/2
2.94
h0
,
V
(4)
BRATOV V, et al: Existence of Optimal Energy Saving Parameters for Different Industrial Processes
51
the dependence of time on penetration h(t ) can be
versus impact duration in Fig. 2. The properties of the
approximated by
material are taken to be equal to the properties of
h(t )
(5)
h0sin( t / t0 ),
time-dependent maximum tensile stress generated in the

impacted media can be estimated by
1 2
2
(V , R, t )
P (t )
,
a 2 (t )
(6)
where the radius of the contact area a(t ) is given by

2
a(t ) (3P (t )(1
R 1/3
) .
4E
gabbro-diabase[9] (E
6.2109 Nm2,
Nm2, v 0.26.and
440 s ).
44.04106
Conclusions
The results received indicate a possibility of
(7)
optimising the energy consumption of different fractureconnected industrial processes (e.g., pounding, drilling).
Knowing the duration and amplitude of the applied
It is shown that the energy cost of crack propagation
load, the mass and velocity of the impacting particle can
strongly depends on the amplitude and duration of the
be found from equations (2) (7):
load applied. For example, in the studied problem when
R
V
t0 6
max
(
)1/2 ,
2.94 5 (1 2 )
(
where
5
4
(1
E
the duration of the load differs from the optimal one by

10 , the energy cost of initiating the crack is exceeding
6
max
)5/2 ,
5 (1 2 )
the minimum value by more than 10 . The obtained
is a parameter of load intensity, having a
pulse on the existing crack length can help in predicting
)2 (
dimension of mass density, and
dependencies of the optimal characteristics of a load
is the maximum
energy-saving parameters for the fracture processes by
stress (i.e. load amplitude). Evaluating the initial kinetic
investigating the predominant crack size in a fractured
energy of the spherical particle, one can estimate the
material. Knowing the fracture incubation time for the
energy required in order to create fracture in the
particular material we can select the most energetically
impacted media
favorable mode of treatment. In particular, adjusting the
3 13/2
0 max
3/2 4
max
duration of impacts, we can optimize the operation of
rupture devices of the impact type. Similarly, it is
where
possible to choose the vibration modes for decreasing

2
5(1
(
3 (2.94)3
4
)4 (
6
)13/2
5(1 2 )
the energy losses during processing of various materials.

Thus, it was demonstrated that the incubation time
is a dimensionless coefficient. This energy, cor-
approach is providing a possibility to predict the
2 400 kgm-3, is plotted
strength of rocks in a wide range of loading rates as well
responding to the value
as to optimize the energy input needed to create rupture

of rock media.
References:
[1] Royal Dutch Petroleum Company Annual Report, 2003
[EB/OL]. [2004-05-28]. http://peakoil.com/ publicpolicy/
[2] Bratov V, Petrov Y. Optimizing energy input for fracture
by analysis of the energy required to initiate dynamic
mode I crack growth[J]. Int J Solid Struct, 2007, 44:
2371-2380.
Fig. 2
Energy necessary to fracture versus load

pulse duration for gabbro-diabase
[3] Petrov Y, Morozov N. On the modeling of fracture of

brittle solids[J]. ASME J Appl Mech, 1994, 61:710-712.
52
2012
Teplotechnik Press, 2007.
[4] Morozov N, Petrov Y. Dynamics of fracture[M].
[8] Petrov Y V, Morozov N F, Smirnov V I. Structural
Berlin-Hidelberg-New York: Springer-Velrag, 2000.

[5] Petrov Y, Taraban V. On double-criterion models of the
macromechanics approach in dynamics of fracture[J].
fracture of brittle materials[J]. St.-Petersburg University
Fatigue Fract Engng Mater Struct, 2003, 26:363-372.
Mechanics Bulletin, 1997, 2:78-81.
[9] Petrov Y V, Smirnov V I, Krivosheev S I, et al. Impact
[6] Protasov Y I. The fracture of rocks[M]. Moscow:
loading of rocks[C]//Fedorovsky G D, Utkin A A.
Moscow State Mining University Press, 2002.

[7] Latyshev
G.
Fracture
of
Proceedings of International Conference on Shock Waves
rocks[M].
Moscow,
in Condensed Matter, St.-Petersburg, 2004:17-19.
BRATOV V1,2, PETROV Y V1,2, VOLKOV G2

1.
198504; 2.
198504
.
,
.
.
.
:
,
;
.
;
25
, 2012 1
Vol.25 No.1, Jan. 2012
Dynamic Collapse of Honeycombs under In-plane Compressions

YU Tong-xi1, GAO Zi-yang2, HU Ling-ling3
( 1.Mechanics and Material Science Research Center, Ningbo University, Ningbo 315211, China;
2. Hong Kong Applied Science and Technology Research Institute, Hong Kong 999077, China;
3.School of Engineering, Sun Yatsen University, Guangzhou 510275, China )
Abstract: Based on our experimental and numerical studies, the present paper mainly describes our
findings in the quasi-static and dynamic collapse of honeycombs under in-plane compression. In quasi-static
experiments, in-plane equi-biaxial compression was applied to a square block of circular honeycomb, while
the onset and evolution of inhomogeneous deformation were observed. To explore the effects of the loading
speed, a finite element simulation on the collapse process of a circular honeycomb block under biaxial
dynamic compression is carried out. The simulation reveals the transition from a global collapse mode to a
progressive collapse mode (i.e. a shock mode). The critical loading velocity, which marks this mode
transition, is further studied numerically by various models and compared with the prediction from the wave
trapping theory.
Key words: cellular materials; circular honeycombs; in-plane compression; dynamic collapse mode; critical
velocity
CLC number: O344
Owing
to
their
Document code: A
light-weight
and
Article ID: 1001-5132 2012 01-0053-07
excellent
transition from a global collapse mode to the progre-
energy-absorption capacity, cellular materials, such as
ssive (shock) mode of a circular honeycomb under in-
honeycombs, foams and lattices, have rapidly extended
plane bi-axial dynamic compression, and the relevant
their engineering applications in automobiles, aircraft
critical velocity will be analyzed and discussed.
and space structures, as well as in various kinds of

energy
absorbers
protection
[1-2]
for
transportation
and
human
. Since the energy absorbing structures
containing cellular materials are often subjected to
Quasi-static experiments of circular honeycombs under in-plane

compression
impact and explosive loading, their dynamic behavior is

a major concern in the material selection and in the
design of sandwich structures.
Among various types of cellular materials,

honeycombs have been firstly and widely used in
In this paper, we will briefly review the studies on
lightweight sandwich panels and in protective structures.
the quasi-static and dynamic behaviors of honeycombs,
In order to improve the design of these structures, it is
and then summarize our systematic studies on honey-
important to understand the deformation mechanism of
combs under in-plane compression. With an emphasis
honeycombs subjected to multi-axial loadings. Exten-
placed on the dynamic behavior of honeycombs, the
sive experimental, analytical and numerical studies on

Foundation items: Supported by the National Natural Science Foundation of China (11032001, 11172335); Russian Foundation for Basic Research (11-01-91217).
The first author: YU Tong-xi (1941-), male, Jiangling Hubei, professor, research domain: impact dynamics and engineering plasticity.
E-mail: metxyu@ust.hk
54
2012
the mechanical behavior of honeycombs have been
lower arms remained stationary while the upper two
conducted, while many of these studies focused on
arms, which were rigidly connected, could move down-
honeycombs under uniaxial compression. Expressions
wards to compress a square honeycomb block. By using
for the mechanical properties, such as the elastic
this test rig, equal remote in-plane strains were assigned
modulus and the collapse strength, have been introduced
to the honeycomb block. Nevertheless, the deformation
for both plastic and brittle base materials
[1]
. Papka and
inside the square block was not homogeneous and
studied the load-displacement response
various patterns of the cells deformation could be
of hexagonal-cell aluminum honeycombs as well as
distinguished. The cells attached to or close to the
circular polycarbonate honeycombs under in-plane
corners where the stationary and the moving arms met in
uniaxial loading. They observed various deformation
the loading devise experienced larger deformation in
patterns (modes), which were related to the particular
comparison to the cells close to the other two corners, as
ratio between the components of the applied displace-
shown in Fig. 3(b).
Kyriakides
[3-4]
The normalized stress-strain curve is plotted in Fig.
ments or forces.
[5]
had
3(a), whilst the deformation process of the sample is
circular cells in a hexagonal close packed arrangement
illustrated in Fig. 3(b) with a series of photographs taken
(Fig. 1) and it is made from polycarbonate extruded
at various overall area strains of the sample correspond-
The honeycomb used in our experiments
tubes with radius R = 3.47 mm and an average wall

thickness h = 0.151 mm forming a block with a depth of
25.4 mm. The density of the polycarbonate is
= 1 191
-3
kg m . The biaxial compressions were conducted in the

Impact Lab of the Hong Kong University of Science and
Technology (HKUST), with a particularly designed test
rig (Fig. 2), whose four arms had equal lengths. Two
Fig. 3(a) The normalized load vs. the diagonal
Strain relation of a honeycomb sample under
quasi-static equi-biaxial compression
Loading direction
Fig. 1
Fig. 2
A honeycomb sample
A 3-D sketch of the equi-biaxial compression rig
[1]
= 0.021
[2]
= 0.074
[3]
= 0.277
[4]
= 0.433
[5]
= 0.535
[6]
= 0.636
Fig. 3(b) A series of photographs showing the deformed

configurations of a honeycomb sample under quasi-static
equi-biaxial compression (marked by [1], [2],
)
correspond to the points marked on Fig. 3(a)
YU Tong-xi, et al: Dynamic Collapse of Honeycombs under In-plane Compressions
55
ing to the equilibrium states identified on the response
revealed the effects of cell wall thickness and impact
in Fig. 3(a). In a square sample under equi-biaxial
velocity on the mode of localized deformation and the
compression, the overall area strain of the sample is
plateau stress. The impact velocity of the rigid plate
related to its diagonal strain,
varied from 3.5 m s-1 to 280 m s-1 in order to study the
dia
, as:
2
dia
(1)
effect of loading rate. As shown in Fig. 4(a), for v =
From this experimental study, we have seen the
3.5 m s-1, the first localization bands formed X shapes
1 (1
)(1
dia
distinguishing characteristics of honeycombs under in-
and
plane bi-axial loading; that is, the nominal stress-strain
honeycomb was completely crushed. When the impact
curve of the circular honeycomb under quasi-static
velocity was even higher, for example at v = 70 m s-1,
compression exhibits three stages, which are typical for
only a localized transverse band perpendicular to the
many kinds of cellular materials under uni-axial
impact (called the I mode, as shown in Fig. 4(b)) was
compression, the bi-axial deformation of honeycombs is
observed at the loading edge and it continued to
notably inhomogeneous with some localization zones;
propagate, layer by layer, to the fixed edge. This
whilst a few interesting and repeatable deformation
deformation was in the manner of plane wave
patterns initiate and develop within the bi-axially
propagation.
developed
into
rhombus
shapes,
until
the
compressed sample (refer to [5] for details). Evidently,

these characteristics are quite different from the
deformation behavior of uniform materials (e.g. a metal
block) under compression.
Dynamic behavior of honeycombs

under uni-axial compression
Fig. 4(a)
The X mode, v=3.5 m s-1, =18.2 mm (t=5.2 ms)
Fig. 4(b)
The I mode, v=70 m s-1, =40.3 mm (t =0.58 ms)
Regarding the dynamic response of hexagonal

honeycombs, Stronge and Shim [6] proposed one-dimensional models to account for the strain-softening within
each cell and strain rate effect. Dynamic tests were also
performed for hexagonal honeycombs, e.g. Zhao et al [7]
studied the in-plane crushing behavior of aluminum
honeycombs by using the split Hopkinson pressure bar.
Using FE simulation, Ruan et al
[8]
studied the in-plane
dynamic behavior of hexagonal aluminum honeycombs.

After that, Zheng et al
[9]
At intermediate impact velocities (not shown here),
investigated the influences of
a number of cells near the right edge of the model were
cell irregularity and impact velocity on the deformation
slightly crushed to form a V shaped band. Thus, all
mode and plateau crush pressure of 2D cellular
the observed deformation modes under the uni-axial
structures.
compression may be classified into three types (Fig.
The dynamic behavior of other types of cellular

materials, such as on 1D cellular chains
hollow spheres
[11-12]
, planar lattice materials
[10]
, metal
[13- 14]
, have
also been studied by the authors group at HKUST.

By employing FE simulation, Ruan et al
5(a)), and summarized into a Mode Classification Map,

Fig. 5(b). The critical velocity at which the deformation
mode switches from one type to another depends on the
cell wall thickness.
[8]
Regarding the plateau stress under impact loading,
56
2012
travels at an increasing speed, leading to a shock wave

front. It is now clear that almost all of the cellular
materials considered earlier have such a stress-strain
curve and hence shock wave theory should apply here
too
[16-17]
, provided that the impact velocity is suffi-
ciently high and a plane plastic wave front develops.

Based on a rigid-perfectly-plastic-localization (RFig. 5(a)
Definitions of the three deformation modes
P-P-L) idealization of the materials, a relationship

between the dynamic stress
stress
d,
the quasi-static plateau
and the applied impact velocity v can be
derived as follows [16-18],

d
)v 2 .
(3)
The second term on the right-hand side represents

a dynamic enhancement due to impact loading. It is seen
that Eq. (3) is very similar to Eq. (2) in the form,
although Eq. (2) comes from data fitting of FE
simulations
Fig. 5(b)
[8]
in
terms of the cell wall thickness and velocity,

k
where
Av 2 ,
(2)
is the dynamic plateau stress,
statically collapse strength, k
consider
micro-structures
of
honeycombs and lattices, whilst Eq. (3) is derived from
Mode classification map
an empirical formula was derived by Ruan et al
which
is the
2 / 3 1.15 is the
an idealized R-P-P-L theoretical model.
Dynamic behavior of circular

honeycombs under in-plane
biaxial compression
coefficient between plane strain and plane stress

conditions, A is a function of thickness ratio (h / l ) .
Finite element analysis by employing commercial
In the case of very low impact velocity v , the
code ABAQUS 6.4 is carried out to simulate the
expression could be degenerated to the static collapse
equi-biaxial
strength
In the study of dynamic behavior of lattice

[14]
process
of
circular
honeycomb sample at large strains. The sample size and
derived from static analysis [15].
materials under impact loading, Qiu et al
compression
found that
material properties are taken as the same as those

employed in our experiment
[5]
. The FE model is first
the dynamic plateau stress can also be expressed in the
verified by comparing the stress-strain curves obtained
form of Eq. (2). A recent study made by Hu et al (to be
from the experiment and from the FE simulation. It is
published) on the crushing behavior of honeycombs
found that when the model contains a connection angle
with various inner angles also confirmed the validity of
of 13 between circular cells and a friction co-efficient
Eq. (2).
of 0.13, it will give the best approximation to the

experimental result.
Rigid, perfectly-plastic shock theory
With the increase of the impact velocity, two types

of collapse modes are observed from the simulation as
As known from the stress wave theory in
shown in Fig. 6(a) and Fig. 6(b), respectively. When the
non-linear solids, when the stress-strain curve is convex
impact velocity is low, crush bands initiate at the two
about the strain axis, the subsequent plastic stress wave
diagonal corners and then produce complex inhomo-
57
geneous deformation in the central region. In this case,
which took place when the impact velocity was larger
the inhomogeneous deformation spreads to the whole
than a critical value.
sample, so this represents a global collapse mode (Fig.
Honig and Stronge
[20]
pointed out that this critical
6(a)). When the impact velocity is much higher, no more
impact velocity can be predicted by the theory of wave
crush bands appear inside the sample, whilst the
trapping. In fact, according to the stress wave theory, the
collapse of cells starts from the impact ends and
speed of the stress wave propagated in non-linear solids
propagates towards the distal ends (Fig. 6(b)). Hence, it
is determined by
is called the progressive collapse mode or shock mode,

which is also typical in cellular materials.
'( )
c( )
1 d
d
(4)
Therefore, if the uni-axial stress-strain curve of a

non-linear material has the characteristics of monotonically increasing until softening or plateau begins,
localization at the wave front will occur and the wave
Fig. 6(a)
Global collapse mode: crush bands occurred
will be trapped when the wave speed c becomes zero,
from the two diagonal corners
and the particle velocity reaches the critical value:

cr
vcr
where
0
cr
c( ) d ,
(5)
is the critical strain, in which the tangent modu-
lus of the stress-strain curve, d / d , becomes zero.

When applying this wave trapping theory to our
Fig. 6(b)
Progressive collapse mode (shock mode): crush
is found that when the model only contains one row of
bands occurred at the impact ends
Between these two collapse modes, therefore, there

must exist a critical impact velocity. Observed from the
collapse patterns of a honeycomb block under different
impact velocities ranging from 7.5 m s-1 to 30 m s-1, the
collapse mode transition happens between V0
10 m s-1
and 12.5 m s-1, so that the critical impact velocity, Vcr,

should be within this range.
bi-axial loading case, different models are tested and it
Prediction of critical impact velocity
cells on each side that directly contact with the

respective loading arm, the row model provides the
best prediction of the critical impact velocity in
comparison to the other models, see Fig. 7. In fact, this
row model well represents the local crush bands when
the impact velocity is high. The deformed shape of the
row model is similar to the local deformed shape of the
block model when the critical impact velocity is
imparted.
Generally speaking, for a cellular solid subjected to

a uni-axial impact loading, a critical impact velocity
exists and divides the collapse modes into two
categories as described above. In fact, this phenomenon
is even more obvious for a cellular solid whose cell
exhibits a softening behavior, e.g., in a type II chain [19].
In the numerical simulation of hexagonal honeycombs
under uniaxial dynamic loading, the I-mode found by
Ruan et al
[8]
is indeed a progressive collapse mode,
Fig. 7 Comparison of the critical impact velocities

predicted by using different models with
the observation of simulation results
58
2012
Effect of the initial impact velocity

Fig. 8 depicts the effect of the initial impact
velocity. When it is smaller than the critical impact

velocity (Vcr=12.5 m s-1 in this case), i.e., the sample
deforms in a global collapse mode, the increase of the
initial impact velocity only results in an increasing
initial peak stress, whereas the plateau stress doesnt
change much. However, when it is larger than Vcr, i.e.
the sample deforms in a progressive collapse mode, both
the initial peak stress and the plateau stress increase
Fig. 8 Effect of the initial impact velocity,

Vcr= 12.5 m s-1, friction co-efficient = 0.13
notably, showing the energy absorbing capacity of the

material is greatly enhanced.
Based on the stress-strain data shown in Fig. 8, Fig.
9 depicts the effect of the impact velocity on the specific
energy absorption capacity of the circular honeycomb
block under in-plane biaxial compression, in which the
absorbed energy is calculated from the corresponding
stress-strain curve within a nominal strain range of
00.4 and then divided by the weight of the sample.
It is clear that the energy absorption curve displays a
Fig. 9 Effect of the impact velocity on the

specific energy absorption capacity
quicker increase after the impact velocity exceeds the
energy absorption capacity of the cellular materials in
critical one, implying that the energy absorption
their applications.
capacity of the honeycomb block is greatly enhanced in
To explore the effects of the loading speed, a finite
the progressive collapse mode. It is also noted the
element simulation on the collapse process of the
specific energy absorption capacity of the circular
circular honeycomb under in-plane bi-axial compression
honeycomb block (in the range of 0.43 J g-1 when the

impact velocity is below 40 m s-1) is relatively low in
is performed. Besides the load-displacement curve and

deformation patterns, the simulation also reveals the
comparison with most of energy absorbing materials
transition from a global collapse mode (despite it is
because the honeycomb block is compressed within its
inhomogeneous) to a progressive collapse mode (i.e. a
plane instead of out-of-plane compression usually
shock mode). The velocity at this transition, termed as
utilized in practical applications.
the critical loading velocity, is further studied numeri-
Concluding remarks
In the experiments of circular honeycomb samples
under in-plane equi-biaxial compression, the evolution

of localization was observed and described. As a matter
cally by various models and found good agreement with

the prediction from the wave trapping theory.
References:
[1] Gibson L J, Ashby M F. Cellular solids: Structure and
properties[M]. 2nd ed. Cambridge: Cambridge University
Press, 1997.
of fact, inhomogeneous deformation is a genetic feature
[2] Lu Guoxing, Yu Tongxi. Energy absorption of structures
of cellular materials. This will not only create complex
and materials[M]. Cambridge: Woodhead Publishing Ltd,
and interesting deformation patterns, but also affect the
2003.
59
simulation on the mechanical properties of MHS
[3] Papka S D, Kyriakides S. In-plane compressive response
materials[J]. Acta Mechanica Sinica, 2006, 23:65-75.
and crushing of honeycomb[J]. J Mech Phys Solids, 1994,
[13] Qiu Xinming, Zhang J, Yu Tongxi. Collapse strength of
42:1499-1532.
[4] Papka S D, Kyriakides S. In-plane crushing of a polycar-
periodic planar lattices under uniaxial compression, Part I:
bonate honeycomb[J]. Int J Solids Sruct, 1998, 35:239-
Quasi-static strength predicted by limit analysis[J]. Int J
267.
Impact Engng, 2009, 36(10/11):1223-1230.
[5] Hu Lingling, Yu Tongxi, Gao Ziyang, et al. The inhomo-
[14] Qiu Xinming, Zhang J, Yu Tongxi. Collapse strength of
geneous deformation of polycarbonate circular honey-
periodic planar lattices under uniaxial compression, Part
combs under in-plane compression[J]. Int J Mech Sci,
II: Dynamic crushing based on finite element simula-
2008, 7:1224-1236.
tion[J]. Int J Impact Engng, 2009, 36(10/11):1231-1241.
[6] Stronge W J, Shim V P W. Microdynamics of crushing in
[15] Zhang J, Ashby M F. Buckling of honeycombs under
cellular solids[J]. J Eng Mater Technol. 1988, 110:185-190.
in-plane biaxial stresses[J]. Int J Mech Sci, 1992, 34(6):

491-509.
[7] Zhao H, Gary G. Crushing behavior of aluminum honeycombs under impact loading[J]. Int J Impact Engng, 1998,
[16] Reid S R, Peng C. Dynamic uniaxial crushing of wood[J].

Int J Impact Engng, 1997, 19:531-570.
21:827-836.
[17] Ashby M F, Evans A, Fleck N A, et al. Metal foams: A
[8] Ruan D, Lu G, Wang B, et al. In-plane dynamic crushing
design guide[M]. Boston: Buttterworth Heinemann, 2000.
of honeycombs-a finite element study[J]. Int J Impact
[18] Tan P J, Reid S R, Harrigan J J, et al. Dynamic
Engng, 2003, 28(2):161-182.

[9] Zheng Zhijun, Yu Jilin, Li Jianrong. Dynamic crushing of
compressive strength properties of aluminium foams. Part
2D cellular structures: A finite element study[J]. Int J
II: Shock theory and comparison with experimental data
Impact Engng, 2005, 32:650-664.
and numerical models[J]. J Mech Phys Solids, 2005, 53:
[10] Gao Ziyang, Yu Tongxi. One-dimensional analysis on the
2206-2230.
dynamic response of cellular chains to pulse loading[J].
[19] Gao Ziyang, Yu Tongxi X, Lu Guoxing. A study on type
Journal of Mechanical Engineering Science, 2006, 220
II structures, Part II: Dynamic behavior of a chain of
(5C):679-689.
pre-bent plates[J]. Int J Impact Engng, 2005, 7:911-926.
[11] Gao Ziyang, Yu Tongxi, Zhao H. Mechanical behavior of
[20] Honig A, Stronge W J. In-plane dynamic crushing of
metallic hollow sphere (MHS) materials: An experimental
honeycomb, Part I: Crush band initiation and wave
study[J]. ASCE J Aerospace Eng, 2008, 21(4):206-216.
trapping[J]. Int J Mech Sci, 2002, 44:1665-1696.
[12] Gao Ziyang, Yu Tongxi, Karagiozova D. Finite element
1.
315211; 2.
999077; 3.
,
510275
(
.
25
, 2012 1
Vol.25 No.1, Jan. 2012
Surface Effects and Problems of Nanomechanics

GREKOV M, MOROZOV N*
( Saint-Petersburg State University, Saint-Petersburg 198504, Russia )
Abstract: A boundary value problem on a circular nanometer hole in an elastic plane loaded at the
boundary and infinity is solved. It is assumed that complementary surface stresses are acting at the
boundary of the hole. Based on Goursat-Kolosovs complex potentials and Muskhelishvilis technique, the
solution of the problem is reduced to a hypersingular integral equation in an unknown surface stress. The
solution of the problem shows that, due to an existence of the surface stresses, the stress concentration at the
boundary depends on the elastic properties of a surface and bulk material, and also on the radius of the hole.
Key words: nanometer circular hole; surface stress; hypersingular integral equation; stress concentration
CLC number: O341
Document code: A
The theory of elasticity which takes into con[1-4]
sideration the action of stresses at a surface of a body
Article ID: 1001-5132 2012 01-0060-04

boundary conditions for the circular hole of radius a
abnormal properties of nano-samples in comparison with
in a plane are given as follows

s
s
1
prr , r
rr
a
a
s
here
is the surface stress;
macro-samples of the same mamerial. In particular, the
the classical stresses in the polar coordinates r , , with
surface stresses are directly related to the scale effect
the center coinciding with that of the circular hole;
is recently applied to the problems of nanotechnology.

The surface stresses are supposed to be responsible for
that means the material properties of a spesimen depend

on its size[5-6]. Furthermore, unexpected effects not
prr , pr
(1)
pr ,
rr
respectively.
The conditions at infinity are
apparent. For example, Youngs modulus of a cylindrical
lim
s jk , lim
jk
(2)
0,
specimen increases significantly, when the cylinder
where
diameter becames very small[7]. From these positions the
coordinates x1 and x2 ( x1 r cos , x2

is a turning angle of the material particle.
circular hole loaded at the boundary and infinity will be

considered. It is assumed that complementary surface
stresses act along the boundary of the hole.
are
are the external normal and tangential loads
corresponding to our traditional representations become
classical problem concerning an elastic plane with a
and
jk
( j , k 1, 2) are the stresses in the Cartesian

r sin ),
The constitutive equations for the surface[1,4] and

volume linear elasticity in the case of plane strain are
reduced to the following
s
s
(2 s
s
0
s
33
The statement of the problem
s
0
s
s
(3)
and
According to the Laplace-Young law
[1,4]
, the
(2
rr
rr
(2
rr

The first author: GREKOV M (1946 ), doctor/professor, research domain: mechanics of surface and near surface defects. E-mail: magrekov@mail.ru
Corresponding author: MOROZOV N (1932 ), doctor/professor, research domain: fracture mechanics and nanomechanics.
E-mail:Morozov@nm1016.spb.edu
GREKOV M, et al: Surface Effects and Problems of Nanomechanics
33
rr
s
In equations (3) and (4),

rr
s
and
r
s
(4)
).
is the surface strain;
are the strains in the bulk material;
are the modules of the surface elasticity, similar
to the Lam constants
of the bulk material;
conditions (3) we obtain from (8)

( )
( )
t s ( ) p( ),
(9)
( ) are the limiting values of function
here
The solution of the boundary problem (9) is given
the residual surface stress under unstrained conditions.
in [8] and can be written in the form

( z)
I ( z ) I p ( z ) S ( z ), | z | 1,
where
Basic relations
1
2 i|
I ( z)
First, we construct a solution for the hole of unit

radius, therefore, introduce a 1 in equation (1). Then
a, p
prr
n, t rotated anticlockwise with respect to the system
by the angle
x1 , x2
2 . Then following
Muskhelishvilis technique[8], one can derive the joint

expression for the traction
where G
(
u1 iu2
dz
( z)
(z
dz
( z)
for
nn
and the
nt
),
( s11 s22 ) 4, c2
c1
2 i (1
)|
( z)
( z )),
(6)
( s22
p( )d .
| 1
us ( )
where
is the displacement vector of the

. From (12) follows the same for the
and the deformation on the
volume deformations
s
, i.e.,
s
(z)
( ), |
z
| z| 1
The relations (11) (13) result in the equation for
derivative with respect to the argument. The increment
( ), |
d z is taken in the direction of the axis t . Thus in
where M s
dz | dz | e , dz
[8]
Muskhelishvili
dz . According to
we define function
( z)
(z ) z
e dr when
'
(z ) z
( z ), | z | 1. (7)
following equation
.
, we find
i
from equation (8) for dz

i
Using equality (7), we derive from (6) the

G( z z )
( z)
dz 1
( ( ( z)
dz z 2
| z | 1,
(14)
| 1,
To obtain expression for
holomorphic
for r 1 by the formula
(13)
| 1.
the surface stress

s
( )
Ms
s
equation (6),
(12)
z
| z| 1
morphic for r 1 . A quantity with the bar denotes
s11 2is12 ) 2,
boundary
lim u ( z ) u s ( ), | | 1,
are functions holo-
complex conjugation and the prime denotes the
(11)
placements vector under passing from the volume to the
boundary
lim
2 du dz for
p( )
d ,
z
1
We impose the continuity constraint on the dis-
boundary point
1 and G
3 )/(
[9].
displacement vector u
G( z z )
1
2 i|
and
ipr ,
Introduce a local orthogonal system of coordinates
ts ( )
d , I p ( z)
z
1
(10)
S ( z ) c c1 / z c2 / z 2 ,
in the complex writing, the conditions (1) take the form

i r t s ( ) p ( ), | | 1,
(5)
rr
s
where t s
i s /
i is the imaginary unit.
( z)
1 0.
on the circumference of unit radius when | z |
is
61
ire d
rr
and
and dz
1 and then use two first equations
(4). This way yields the equation

2
( )
Re(
( )
( )).
(15)
Introducing (15) into (14) and taking into account

( z)
1
( )) ( z
z
equations (10) and (11), we arrive at the following equation

1
)
z
( z )),
(8)
ei in equation (8) and direct

we take the limit z
axis n towards the center z 0 . Since in this case
3 2 and dz
i | dz | ei , by virtue of
M Re( I ( ) I ( )) M (
1)
( S ( ) S ( )) M Re( I p ( ) I p ( )),
where M
i
Ms / 2 .
s
Let
s
s
(16)
s
s
. Since
( ) , the Sokhotskii-Plemelj formulas for the
62
2012
Cauchy type integrals I ( z ) and I p ( z ) acquire the forms

s
I ( )
1
2 i|
( )
2
s( )
( )
2
s( )
Ip ( )
1
2 i|
p( )
which is free from the external load ( p

(17)
d ,
Cauchy principal value.

Introducing (17) into (16) and taking into account
1
2
( ), d
( )
( ),
( )
d , we obtain the following singular
d0
1))
1
2 i|
2 i|
2a
( )
( )
( )
s
r a
a M
(2 3H 2 )( s22
1)( S ( ) S ( ))
(18)
In (21) and (22) we denote

1,
a,
It is clear that
s
0
from equation (18) that surface stress
s
0
at the
s
0
, we reduce equation (19) to the hypersingular
(2a M (
(
| | 1
M ( 1)
2 i
2
/ ) ( )
d
Ma (
(
)2
(24)
(1) Equibiaxial tension/compression s11 s22 p,

s12
0,
s
r a
a M
M (1 )
p 2 p.
2(a M )
(2) Uniaxial tension (the Kirsch problem) s11

s22
s12
a M
(1
3M (1 )
) p1 cos 2 .
2a M (3 )
(3) Simple shear s12 q, s11 s22 0,
s
r a
( S ( ) S ( )) 2Ma Re( I p ( ) I p ( )),
p1 ,
M (1 )
) p1
4(a M )
(2
1)
(25)
0,
r a
1)) ( )
| 1.
s11 ) cos 2
Particular cases
( ),
integral equation in the unknown function
(1 H1 )( s11 s22 )
and Rajapakse[5] obtained by another way.
. We derive
boundary of the circular hole in the unloaded plane is

a
s
(19)
0
s.
a M
Using the regularization formula[10] and denoting
s
(23)
This formula coincides with the solution of Tian
const if the external loads
are absence. Let s jk 0, p 0 and
ei dr and
1
( )), | 1 | a.
2(2 3H 2 ) s12 sin 2 .
are points on the circumference of radius a .

s
(21)
we obtain from equation (23)
d )
| 1
Ma (
Taking into account (10), (11), (19), (21) and (22),
2 Ma Re( I p ( ) I p ( )) .
where
d2
from equation (8) under dz

( 1 ) Re(3 ( )
1)
| 1
1
The expression for the hoop stresses we derive
( ) M(
( )
d2
where
d 0 aH1 ( s11 s22 ), 2d 2 aH 2 ( s22 s11 2is12 ), (22)
and
M (1 )
M (1 )
, H2
.
H1
4(a M )
2a M (3 )
integro-differential equation
(2a M (
0 ), the
solution of the equation (20) has the simple form
| 1
where the integral is understood in the sense of the
the relations
Solution for the circular hole with

free boundary
For the case of a circular hole the boundary of
d ,
| 1
p( )
2
a M
2(2
(26)
3M (1 )
) q sin 2 . (27)
2a M (3 )
(20)
The equalities (25) (27) show that the presence
It is important to note that according to the cons-
of the surface stress decreases the stress concentration if
0 . The residual stress
produces the same
truction of the equation (20), the homogeneous equation.
corresponding to the equation (20) does not have
effect. Besides, the stress concentration depends on the
non-trivial solution.
radius of the hole (scale effect). According to theoretical
GREKOV M, et al: Surface Effects and Problems of Nanomechanics
calculations for cubic metals[11] and estimates produced

in [6], the parameter M
the residual stress
M
(10
10
(2
s
10 9 ) m, and
1 Nm-1. In this case if
material surfaces[J]. Arch Rational Mech Anal, 1975, 57:

146-147.
) / (2 ) in which
is negligible has an order
a 10 nm, the first member in (25) (27) has the order
[2] Murdoch A I. Some fundamental aspects of surface

modeling[J]. J Elasticity, 2005, 80(1-3):291-323.
[3] Podstrigach Ya S, Povstenko Yu Z. An introduction to the
mechanics of surface phenomena in deformable solids[M].
108 Nm-2 =100 MPa. So, for the values of remote loads
p, p1 , q up to 100 MPa, the influence of the residual
stress
Kiev: Naukova Dumka, 1985.

[4] Duan H L, Wang J, Karihaloo L. Theory of elasticity at
the nanoscale[J]. Adv Appl Mechanics, 2009, 42:1-68.
on the stress distribution at the boundary and
especially on the stress concentration is comparable with
63
[5] Tian L, Rajapakse R K D. Analytical solution for sizedependent
the influence of remote loads. Furthermore, the effect of
elastic
field
of
nanoscale
circular
inhomogeneity[J]. Trans ASME J Appl Mech, 2007, 74
the residual surface stress surpasses that of surface
(5):568-574.
elasticity because the rest members in (25) (27) con-
[6] Goldstein P V, Gorodtsov V A, Ustinov K V. Effect of
taining parameter M are less than the first one. This
residual stress and surface elasticity on deformation of
phenomenon was recently ascertained by Goldstein et al[6].
nanometer spherical inclusions in an elastic matrix[J].

Phys Mesomechanics, 2010, 13(5-6):318-328.
[7] Eremeyev V A, Morozov N F. The effective stiffness of a
Conclusion
nanoporous rod[J]. Dokl Phys, 2010, 55(6):279-282.
The general analytical solution of the 2D problem
[8] Muskhelishvili N I. Some basic problems of the

mathematical
on a circular nanometer hole in an elastic plane is
theory
of
elasticity[M].
Groningen:
Noordhoff, 1963.
constructed and reduced to the hypersingular integral
[9] Grekov M A. A singular plane problem in the theory of
equation in the surface stress. For the case of remote
elasticity[M]. St. Petersburg: Izdatelstvo Sankt Peter-
loading, the solution of this equation is obtained in a
burgskogo Universiteta, 2001.
closed form. The effect of the surface stress and residual
[10] Linkov A M. Boundary integral equations in elasticity
surface stress on the stress concentration is analyzed for
theory[M]. The Netherlands: Kluwer Academic Publisher,

2002.
the simplest kinds of loading.
[11] Shenoy V B. Atomic calculations of elastic properties of

metallic fcc crystal surfaces[J]. Phys Rev B, 2005, 71(9):
References:
94-104.
[1] Gurtin M E, Murdoch A I. A continuum theory of elastic
GREKOV M, MOROZOV N*
,
198504
:
.
Muskhelishvilis
.
,
:
.
;
25
, 2012 1
Vol.25 No.1, Jan. 2012
A Three-dimensional Viscoelastic Constitutive

Model of Finite Deformation
SHEN Li-jun, BIAN Zhong-jing
( Key Laboratory of Impact and Safety Engineering Affiliated to Ministry of Education, Ningbo University, Ningbo 315211, China )
Abstract: Hereditary integral formulations and differential formulations identified with spring and dashpot
constructions can express the viscoelastic behaviors of polymeric materials at small deformations. In this
paper, the models of small deformation serve as a starting point for the development of the viscoelastic
constitutive models of finite deformation. A process of finite deformation is decomposed into a series of
sub-processes of small deformation. The rotations of stress in sub-processes are determined by elastic
constitutive equation. The changes in the principal stress are calculated using the spring and dashpot
constructions. Then, a viscoelastic constitutive model, which satisfies the principle of objectivity, is
presented. Such form of constitutive model in principle can be suitable for a range of strain-rates, e.g. either
for quasi-static loading or for impact loading, with different material parameters in different strain-rates. As
an application example, the simple shear deformation is computed to show that the proposed model can
adequately well describe the viscoelastic behavior for polymers at finite deformations.
Key words: viscoelastic; finite deformation; the principle of objectivity
CLC number: O341
Document code: A
Article ID: 1001-5132 2012 01-0064-06
The constitutive theories of viscoelastic material
derivative of Cauchys stress, many authors developed
have attracted substantial attention during the past half
the rate-form rheological models for finite deformations
century. The viscoelastic constitutive models at small
(e.g. [5]). However, there are infinite kinds of objective
deformations can be described with hereditary integral
rates of stress. We dont know which kind of objective
[1-2]
. The rate-
rate is suitable for the constitutive equations. Many
form rheological models are commonly used. However,
authors chose an objective rate for the rate-form
much less progress has been made for the constitutive
constitutive equation without any suitable reason. There
models at finite deformations. Several authors have
are many disputes on how to choose the objective rate.
developed the viscoelastic constitutive model of finite
The definition of elastic and inelastic strain rates is
deformation on the basis of the second law of thermo-
another issue in developing the constitutive equations of
dynamics (e.g. [3]). It is known that the constitutive
finite deformation. Several authors have defined elastic
models for finite deformations must satisfy the principle
and inelastic deformation rates by using the decomposi-
formulation or differential formulations
[4]
of objectivity . By replacing the strain rate with the
tion of the deformation gradient as a product of elastic
stretching (the deformation rate) and the material time
and inelastic parts[6]. However, the multiplicative decom-
derivative of Cauchys stress with an objective
position of deformation gradient is not consistent with

The first author: SHEN Li-jun (1962 ), male, associate professor, research domain: plasticity, constitutive model. E-mail: shenlijun@nbu.edu.cn
SHEN Li-jun, et al: A Three-dimensional Viscoelastic Constitutive Model of Finite Deformation
the additive decomposition of strain rate into the elastic

[7]
and inelastic parts. Lion
pushed forward the Greens
65
equation (2) satisfies the principle of objectivity.
1.2 The Maxwell model for finite deformations
strain to intermediate configuration and separated the
The Maxwell model consists of a linear spring and
strain into elastic and inelastic parts to develop the finite
a dashpot in series. For small deformation, the linear
deformation constitutive relation consistent with the
spring and the dashpot are expressed by
[8]
second law of thermodynamics. Huber and Tsakmakis
separated the transformed Almansi strain into elastic and
) / E)
((1
(3a)
( /E )tr( ) I ,
(3b)
/ ,
inelastic parts and defined the associated stress to extend
respectively, where the sub indeces e and v imply elastic
constitutive laws of viscoelastic materials for small
and viscoelastic respectively,
deformations to these for finite deformations. Drozdov
Youngs modulus and the Poissons ratio. Assume that
[9]
et al
made the objective inelastic stretching by means
of a rotation transformation and proposed that the

Shen
presented a decomposition of deformation into
and
are the
the strain rate is equal to the sum of elastic and

viscoelastic strain rates, i.e.,
inelastic stretching is proportional to the total stretching.

[10]
(4)
thus, we have
elastic and inelastic parts and proposed a constitutive
((1
) / E)
( /E )tr( ) I
/ .
(5)
model for elastic-plastic materials at finite deformations.
For finite deformations, the strain rate may be
This paper will develop a constitutive model for
replaced by the deformation rate, the deformation rate is
viscoelastic materials at finite deformations through a
expressed as the sum of elastic and viscoelastic parts
new approach. Also it will calculate the stress of simple
shear to compare the constitutive equations proposed
The material time derivative is replaced by the
here with the existing constitutive equations.
Dv .
The constitutive models for finite

deformations
where
((1
) / E)
1.1 The Kelvin model for finite deformations
( /E )tr( ) I
/ ,
(7)
stress, which is commonly expressed in the form

o
where
The Kelvin model for small deformation is
is the objective derivative of the Cauchys
expressed by
I tr( ) 2
(
(1 / 3) tr ( ) I )
(1/ 3) tr( ) I ,
(6)
objective derivative. Thus, we obtain

D
De
(8)
is a rotational rate (spin). However, it is
difficult to choose the spin in Formulate (8). The

rotational rate is commonly obtained from the total
deformation gradient and is believed to be a geometric
(1)
quantity. The material rotational rate, the relative
where super dot denotes the material time derivative,
rotational rate and the Euler frame rotational rate are
and
are Lames constants,
are the
obtained from the total deformation gradient. We think
deviatoric and volumetric viscosities respectively, may
that the rotational rate applicable to the formulate (8) is
be taken as a function of stress or strain rate. For
dependent on not only the total deformation but also the
simplification, let
viscoelastic part of the deformation. In addition, the
equal
and
. In the case of finite
deformation, the model may be expressed by
additive decomposition of deformation rate (6) is not

(2)
consistent with the multiplicative decomposition of
where D is the stretching (the deformation rate), V is
deformation gradient. Shen[10] presented an approach
the left stretch tensor. In extending the limits of equation
through which the constitutive relations are extended to
(1), the strain is defined by the logarithmic of the left
include the case of finite deformation. In this study, we
stretch tensor, the strain rate is replaced by the stretching.
develop the Maxwell model for finite using this
I tr ( ln V ) 2 ln V
D,
66
2012
approach. Consider a deformation from time t0 , t1 up
RE Da RET . Hence, if the elastic deformation vanishes,
to tn : I
the strain rate RE Db RET consequent upon the increase
F1
Fn . The deformation gradient
F2
at time t1 is expressed in the forms

F1 = RE1V 1 RLT1 = RE1V 1eV 1v RLT1.
RET . If the
RE
of RL is independent of the stress

(9)
deformation is elastic, RL does not affect the stress
In Cartesien coordinate system, RE and RL are
either. Hence, we assume that in the process from t1 to
orthotropic matrices, V is diagram matrix. Sub indecs
t2 , RL 2 b does not affect the stress at time t2 . Thus, the
e and v implies elastic and viscoelastic. The process
principal direction of the stress is RE 2 b and the stress
from t0 to t1 is a small deformation. When the
may be expressed by
T
L1
configuration at time t0 undergoes the rotation R ,
RE 2 b
t2
RET 2 b .
(16)
the deformation gradient at time t1 becomes RE1V 1 .
The principal stresses may be obtained as follows
Hence, for isotropic materials, RL1 does not affect the
ln V
stress at the time t1 , and the principal direction of this

stress may be expressed by RE1 . The stress at the
time t1 is expressed in the form
T
t 1 = RE 1 1 RE 1.
(11)
ln V 1e ln V 1v .
The principal values of the stress at t1 can be

obtain as follows
ln V 1
((1
) / E)
( / E)tr(
)I
1
/ , (12)
1
O) /
(t1 0) . The elastic deformation V 1e may be obtained
as follows
ln V 1e
) ( / E )tr(
)I
/ , (17)
ln V
(ln V
2b
2b
ln V 1e ) / (t2 t1 ),
) / (t2 t1 ).
In a similar way, we can obtain the stresses from

elastic
deformation,
configurations
at
the
stress-free
current
times
intermediate
are
the
initial
configuration. We may obtain the constitutive equation

of elastic deformation
(ln V 1 ln I ) / (t1 0) ,
where, ln V 1
the time t3 up to the time tn . In the case of pure
position of logarithmic strain rate

ln V 1
)/ E
where
(10)
From the decomposition (9), we have the decom-
((1
2b
((1
) / E)
( / E )tr(
)I.
(13)
Taking the free-stress intermediate configuration as
REi
ti
ln V i
((1
T
REi
, ( i 1, 2, 3,
) / E)
n ),
( /E )tr(
(18a)
i
(18b)
)I.
The equation (18) may be rewritten as

RE ln V RET
((1
) / E)
( / E )tr( ) I ,
where the objective rate
( RE RET )
(19)
( RE RET )
reference configuration, the deformation gradients at t1
and is Euler frame corotational rate. The constitutive
and t2 are respectively expressed by

F1b RE1V 1e ,
equation (19) may be expressed by

(14a)
and
((1
) / E)
where the objective rate

F2 b
T
L1
F2 (V 1v R )
RE 2bV
2b
T
L2b
(14b)
The deformation rate is decomposed in the following form

(15a)
D = RE ( Da + Db ) R ,
Db
(20)
corotational
log
[11]
rate
log
If
and is the
the
elastic
deformation vanishes, the stress-free intermediate constitutive relation becomes

D
-1
(15b)
VV ,
1
T
L
(1 / 2)(V R RLV
T
L
V R RLV ),
(15c)
where Da is a diagonal matrix, Db is a symmetric

matrix whose diagonal components are all zero. The
stress
figurations are the current configurations, the conT

E
where
Da
logarithmic
( / E )tr( ) I ,
RE
T
E
yields the viscous-strain rate
/ .
(21)
The constitutive equations (19 and 21) are consistent with the existing constitutive equations.
1.3 The Burgers constitutive model for finite

deformations
Burgers model can describe the main characters of
67
some viscoelastic materials. Burgers model implies that
deformation from time t0 to t1 , we obtain the deviatoric
the deformation can be decomposed into three parts.
principal stress from the following equations
For small deformations, the constitutive equation is
s1
2 1 (ln V 1e (1) (1 / 3) I tr (ln V 1e (1) )),
expressed by
s1
2 2 (ln V 1e (2) (1/ 3) I tr (ln V 1e (2) ))
e e1 e2 e3 ,
2 1e1 , s
s
p
K1 1 , p
(1/3)
2 2 e2
K2
3 3
e, s
e,
2 2
(1 / 3)
3 3
2 2
s1
,
(22a-h)
where s is the deviatoric stress, e the deviatoric

strain, p Hydrostatic stress,
and,
the volumetric strain
K is bulk modulus. We easily obtain the
constitutive equation of p and
. From equations
(22a) (22e), we obtain the rate-form constitutive

equation of deviatoric stress and deviatoric strain
(27a)
(ln V 1e (2) (1 / 3) I tr (ln V 1e (2) )),
(27b)
(ln V 1v (1 / 3) I tr (lnV 1v )).
(27c)
Equations (27a) (27c) are the elastic element, the

Kelvin element and the viscous-element constitutive
relations respectively. The equation (27b) implies that
the Kelvin element consists of spring parallel with
dashpot.
It
is
supposed
that
the
deforming
body
intermediately unloaded and again loaded generates the

same stress as this continuously loaded body does.
(23)
Hence, we can take the free-stress intermediate
is
configuration as reference configuration. Thus the
easily obtained. By replacing the material time
deformation gradients at t1 and t2 are respectively
derivative s by the objective derivative s , the two-
expressed by
order derivative s by the two-order objective derivative
F1b
RE1V 1e (1) V 1e (2) ,
deformation rate D , the two-order derivative e by
F2 b
F2 (V 1v RLT1 )
the objective derivative D of the deformation rate,
As in the process from time t0 to t1 , the stress at
we can obtain a rate-form constitutive equation for finite
t2 can be obtained. The RL 2 b does not influence the
deformations
stress at t2 . Thus, the principal direction of stress at t2
p0 s
p1 s
p2 s q1e q2e.
The constitutive equation of the p and the

o
oo
s , the deviatoric strain rate e by the deviatoric
and
oo
(28a)
RE 2bV
2b
RLT2 b .
(28b)
(24)
is RE 2 b . From the equation similar to (27a) (27c), we
However, it is difficult to choose the objective
can obtain the principal stress at t2 . Thus, the Burgers
p0 s
p1 s p2 s q1 D
q2 D .
derivatives for the constitutive equation. As in the case

of the Maxwell model, the equation (23) is extended to
include the case of finite deformation. Consider the
deformation from t0 to t1 up to tn : I
F1
Fn . The deformation gradient at time t1 is
F2
expressed in the forms

F1
RE1V 1 RLT1.
(25)
We have the decomposition

T
L1
model for finite deformations is developed. In this study,

we propose an approach through which the constitutive
relation of small deformation is extended to include the
case of finite deformation.
The simple shear deformation

We calculate the stresses of simple shear deforma-
tion by using the proposed Burgers constitutive model

(26)
and the existing Burgers constitutive model (24) respec-
where sub indexes e(1), e(2) and v imply the elastic

element, the Kelvin element and the dashpot element
tively. The motion in the simple shear is expressed as

x1 X 1 kX 2 , x2 X 2 , x3 X 3 ,
(29)
respectively. Like the Maxwell model, the principal
where
direction of the stress at time t1 is RE1 . For the small
ordinates of the current and the initial configuration
F1
RE1V 1e (1)V 1e (2)V 1v R ,
xi
and
Xi
are rectangular Cartesian co-
68
2012
respectively. We have the deformation gradient F

1 k
(30)
F
.
0 1
material rotational rate, the relative rotational rate, the

Euler frame rotational rate, the logarithmic rotational
rate and zero.
Here the matrix of tensor is presented in a 2 2

truncated matrix form with all other components that
have an index equal to 3 being identically zero. The
deformation rate (stretching) D
and the material
rotational rate w are respectively

k 0 1
D
,
2 1 0
and
k 0 1
w
.
2 1 0
(31)
The deformation gradient F is expressed in the

following form
REV RLT
RRLV RLT .
(32)
The Euler rotational rate and the relative rotational

rate are respectively
E
RE RET
0 1
,
4 1 0
(33a)
and
RR
2k
0 1
.
4 1 0
Fig.1 Principal stress in simple shear resulting from
Burgers constitutive model where the shear strain rate

k =10 s-1 material parameters used 1 = 1 500 MPa ,
1 000 MPa, 2 100 MPas, 3 180 MPas.
2
Curve (1) is based on the proposed Burgers constitutive model. Curves (2 6) are based on the
existing Burgers constitutive model (24) where the
rotational rate is the material rotational rate, the
relative rotational rate, the Euler frame rotational
rate, the logarithmic rotational rate and zero
(33b)
The logarithmic rotational rate is

Log
k 4
(
4 4 k2
k
2
4 k sh (k /2)
0 1
. (34)
1 0
The stresses are shown in Fig. 1 and Fig. 2. For the

principal stress, the proposed constitutive model is
between the existing constitutive models (24) where the
objective rate is the material corotational rate and the
relative corotational rate respectively. For the principal
direction, the proposed constitutive model is near to the
existing constitutive model (24) where the objective rate
is the material corotational rate.
Fig. 1 principal stress in simple shear resulting
from Burgers constitutive model where the shear strain
rate k 10 s 1 , material parameters used
2
1 000 MPa,
100 MPa s,
1 500 MPa,
Fig.2 Principal direction angle in simple shear

resulting from Burgers constitutive model where the
shear strain rate k =10 s-1, material parameters used
1 000 MPa, 2 100 MPas,
1 = 1 500 MPa,
2
180
MPas.
Curve
(1) is based on the proposed
3
Burgers constitutive model. Curves (2 6) are based
on the existing Burgers constitutive model (24) where
the rotational rate is the material rotational rate, the
relative rotational rate, the Euler frame rotational
rate, the logarithmic rotational rate and zero
180 MPa s . Curve
(1) is based on the proposed Burgers constitutive model.
Conclusion
Curves (2 6) are based on the existing Burgers constitutive model (24) where the rotational rate is the
The constitutive model of viscoelastic material at
69
small deformation can be expressed by the elastic
[3] Holzapfel G A, Simo J C. A new viscoelastic constitutive
element and the dashpot in parallel and series. The
model for continuous media at finite thermomechanical

changes[J]. Int J Solids and Structures, 1996, 33:3019-
constitutive model at finite deformation must obey the
3034.
principle of objectivity. In this study, the constitutive
[4] Truesdell C, Noll W. The non-linear field theories of
model of small deformation is extended to make the
mechanics[M]. New York: Springer-Verlag, 1965.
model applicable in finite deformation through a new
[5] Bardenhagen S G, Stout M G, Gray G T. Three-
approach. The deformation gradient is decomposed into
dimensional, finite deformation, viscoplastic constitutive
elastic and inelastic parts, and thus strain rate can be
models for polymeric materials[J]. Mechanics of Materials,
expressed as the sum of elastic and inelastic parts. The
1997, 25:235-253.
[6] Lubliner T. A model of ruber vicoelasticity[J]. Mech Res
decomposition of deformation gradient is consistent
Commun, 1985, 12:93-99.
with that of strain rate. The rate-form constitutive
[7] Lion A. A physically based method to represent the
relations proposed here obey the principle of objectivity
thermo-mechanical behavior of elastomers[J]. Acta
though we need not make the choice of the objective
Mechanica, 1997, 123:1-25.
corotational rate of stress which is arbitrarily made by
[8] Huber N, Tsakmakis C. Finite deformation viscoelasticity

laws[J]. Mechanics of Materials, 2000, 32:1-18.
many authors. The simple shear deformation has been
[9] Drozdov A D, Al-Mulla A, Gupta R K. Finite visco-
worked out by using the existing Burgers model of finite

deformation
and
the
proposed
model
of
plasticity of polycarbonate reinforced with short glass
finite
fibers[J]. Mechanics of Materials, 2005, 37:473- 491.
deformation. The proposed model can describe the
[10] Shen L. Constitutive relations for isotropic or kinematic
viscoelastic constitutive relation of finite deformation.
hardening at finite elastic-plastic deformations[J]. Int J

Solids and Structures, 2006, 43:5613-5627.
References:
[11] Xiao H, Bruhns O T, Meyers A. Hypo-elasticity model
[1] Ferry J D. Viscoelastic properties of polymers[M]. 3rd ed.
based upon the logarithmic stress rate[J]. Journal of
New York: John Wiley, 1980.
Elasticity, 1997, 47:61-68.
[2] Flggle W. Viscoelasticity[M]. 2nd ed. New York:

Springer-Verlag, 1975.
,
,
315211
:
.
,
.
,
.
25
, 2012 1
Vol.25 No.1, Jan. 2012
Identification and Verification of Some Plasticity Models for

Structural Materials by Using Pressure Bar Technique
BRAGOV A M1, BALANDIN V V1, KONSTANTINOV A YU1,
KONSTANTINOVA YU V2, LOMUNOV A K1,3, FILIPPOV A R1
( 1.Research Institute of Mechanics, Lobachevsky State University, Nizhny Novgorod 603950, Russia;
2.Research Institute of Mechanical Engineering, Dzerzhinsk Nizhegorodsky 603004, Russia;
3.Nizhny Novgorod State University of Architecture and Civil Engineering, Nizhny Novgorod 603950, Russia )
Abstract: The paper proposed methodology for complex research of properties of materials, including
obtaining of demanded mechanical properties of materials in a wide range of strain rate and temperatures,
definition (identification) on their basis of necessary parameters of models of deformation and destruction,
and also checking of adequacy (verification) of these models by means of special natural and numerical
experiments. The possibility of this methodology is illustrated for some structural metals and alloys.
Key words: mechanical properties; plasticity models; identification; verification; dynamic experiment;
pressure bar technique; numerical simulation
CLC number: O343
Document code: A
Article ID: 1001-5132 2012 01-0070-04
At designing military and industrial objects,
these factors on mechanical characteristics of a material
nuclear stations, large public buildings and also contain-
is very important at calculation of a stress-strained
ners for transportation of explosive, toxic and radio-
condition and strength of constructions. The major
active substances the main problem is devoted to an
element of such calculation is the defining relationships
estimation of their vulnerability in case of occurrence of
(models) adequately describing elastic-plastic behaviour
emergencies: acts of terrorism, the natural and man-
of a material in a wide range of strain rates and
caused accidents, accompanied intensive dynamic
temperatures. There is a necessity of realization of
influences. The use of numerical simulation in the early
complex
design stages can significantly speed up and reduce
include obtaining of a spectrum of mechanical charac-
price of the developing process. However, it should be
teristics of a material, a choice of defining relationships
understood that the accuracy of the numerical experi-
and their equipping by necessary parameters, and also
ment (the efficiency of its use) is largely determined by
their subsequent verification on the basis of comparison
the accuracy of material models involved.
results of a set of numerical and natural verification
experimental-numerical
researches
which
Now it is well-known, that mechanical properties
experiments. On an example of some metals the
of materials change depending on speed of deformation
methodology of such complex research of dynamic
and temperature. Taking into account an influence of
properties of materials is presented.

Foundation items: Supported by the National Natural Science Foundation of China (11032001); Russian Foundation for Basic Research (11-01-91217);
RFBR (10-01-00585- ); Limited Company Strela, Snezhinsk, Russia.
The first author: BRAGOV A M (1947- ), male, doctor/professor, research domain: dynamical mechanics. E-mail: bragov@mech.unn.ru
BRAGOV A M, et al: Identification and Verification of Some Plasticity Models for Structural Materials by Using Pressure Bar Technique
Experimental methods
and dynamic indentation
[4]
71
of indenters with a head-tip
both in the form of a cone and a hemisphere. In modiDetermination of dynamic stress-strain curves under
[1]
fied Taylors test, the accelerated specimen hits not on
rigid semi-infinite barrier, but on the pressure bar equip-
To determine strength and deformation failure properties
ed by strain gauges. In each of the specified methods for
of materials under tension the Nicholas scheme was
checking adequacy of parameters of models is used,
compression the traditional SHPB scheme were used
used
[2]
. Static and dynamic experiments were realized
except the analysis of the residual form of the specimen,
a comparison of experimentally registered and numeri-
Dynamic properties at compression and tension were
cally simulated strain pulse in a pressure bar. In modified
in the range of temperatures from 70
up to +300
2
3 -1
Taylor test also high-speed film-registration of change
obtained in the range of strain rates 510 510 s .

Software package LS-DYNA is very widely using
of specimens form during impact has been done which
all over the world for numerical simulation. It has the
has allowed estimating a reliability of models more
big library of models (material functions) and criteria of
precisely. It is necessary to notice that modes (a kind of
destruction. For a few popular models of plasticity
the stress-strain state, a strain-rate range) of these
(Johnson-Cook, Zerilli-Armstrong, Cowper-Symonds
experiments differ from one-dimensional experiments on
etc.) necessary parameters have been defined. The
the Kolsky method in which parameters of models have
problem of definition of an optimum set of parameters in
been obtained. Thus possibilities of models are checked
this case is identical to a problem of approximation of
at other modes of a loading.
experimental data in four-dimensional space deformation-strain rate-temperature-yield stress. The decision of

an optimizing problem was done numerically by a
Experimental results and some

plasticity models
method of the conjugate gradients with use of mathematical packages Maple and MathCAD. Then the estima-
Dynamic properties at compression and tension for
tion of adequacy (verification) of the received models
some structural materials were obtained by using the
was carried out.
SHPB and some original modifications and then
For verification of defined plasticity models, the
parameters of popular plasticity models from LS-DYNA
authors used a set of original verification experiments
library were determined. Possibilities of the specified
(in natural and numerical realization) using pressure bar
methodology are illustrated on an example of research of
[3]
dynamic properties of Russian aluminium AMg6 alloy
technique: modified Taylor test, direct impact method
Fig. 1
Natural (dotted lines) and simulated (solid lines) stress-strain curves of AMg6 alloy for
(a) Zerilli- Armstrong model and (b) power law plasticity model
72
2012
and English industrial high-pure C101 copper in

condition both as-received and after annealing.
Tab. 1 Parameters of some reductive models for C101 copper

Johnson-Cook model
n
T=(A+B p )
Cowper-Symonds model
=k n
constitutive equations were equipped by necessary para-
As-received After annealing
As-received After annealing
meters: Zerilli-Armstrong model (c1=143.032; c2=
A = 356 MPa
A = 35 MPa
k = 462 MPa
k = 627 MPa
B = 129 MPa
B = 603 MPa
n = 0.055
n = 0.484
n = 0.451
n = 0.547
For AMg6 alloy in as-received condition two
5.565 8; c3=1.4510-3; c4=1.4710-3; c5=380; n= 0.462 42)

and power law plasticity (Cowper-Symonds) model (k=
365; n=0.14; C=17120; p=2.338). Also the tabular input
of stress-strain diagrams (Model 24 Piecewise linear
plasticity) was used.
Comparison of the stress-strain curves experimental and regenerated by using specified models is shown
in Fig.1.
Very strong influence of an initial state of C101
copper onto the dynamic diagram was marked. On the
other hand, in the investigated range of strain rate, sensitivity of the material to this parameter is not observed.
Experiments with strain-rate jumps have shown that influence of strain-rate history can be neglected [5]. Absence
Fig. 2
Natural (solid lines) and simulated (dotted lines)

stress-strain curves of C101 copper for
as-received state and after annealing
of influence both of strain-rate value and its history onto
In Fig. 4, a comparison of experimental and simu-
mechanical properties of copper has allowed using for
lated residual profiles of copper specimens both in a
its modeling simple defining relationships (the state equa-
condition as-received (a) and after an annealing (b) is
tions), not considering high-speed hardening (Tab. 1).
shown. It is well visible, that after an annealing the zone
Experimental stress-strain curves (markers) in

comparison with the models are presented in Fig. 2.
of plastic deformation of the specimen is much more,

though the degree of radial deformation is less.
Some results of verification for obtained models of
For two specified models an equally good
AMg6 alloy and C101 copper are shown below. A com-
conformity of experimental and simulated residual
parison of natural and simulated strain pulses in pressure
profiles is visible because of absence of influence of
bar for AMg6 alloy during indentation and direct impact
speed of deformation on a dynamic yield point of an
experiments is presented in Fig. 3.
investigated material. It is visible that relative deviations
Fig. 3
Verification of AMg6 models by using dynamic indentation of (a) cone indenter and (b) direct impact
BRAGOV A M, et al: Identification and Verification of Some Plasticity Models for Structural Materials by Using Pressure Bar Technique
Fig. 4
73
Simulated and experimental residual profiles of C101 copper in a condition as-received (a) and
after an annealing (b) after impacts at different velocities
between numerical and experimental profiles of samples
References:
after impact do not exceed 2.5 .
[1] Bragov A M, Lomunov A K. Methodological aspects of

studying dynamic material properties using the Kolsky
method[J]. Int Journal of Impact Engineering, 1995, 16(2):
321-330.
[2] Nicholas T. Tensile testing of materials at high rates of
strain[J]. Experimental Mechanics, 1981, 21(5):177-195.
[3] Dharan C K H, Hauser F E. Determination of stress-strain
characteristics at very high strain rates[J]. Experimental
Mechanics, 1970, 10:370-376.
[4] Abramov A V, Bragov A M, Konstantinov A Yu, et al.
Sergeich experimental design analysis of high-speed
deformation of aluminum alloy AMg6[J]. Problems of
strength and plasticity, 2004, 66:93-100.
[5] Bragov A M, Lomunov A K, Medvedev A A. A modified
Kolsky method for the investigation of the strain-rate
history dependence of mechanical properties of materials
[J]. Journal de Physique IV, 1991(1):471-475.
Conclusion
The methodology of a complex experimental-
numerical research of high-speed deformation of

materials is offered. Advantage of the offered system of
verification experiments is that the stress-strain mode of
tested specimens during these experiments differs from
those in base experiments, and rates and degrees of
deformation also differ from modes in which parameters
of models have been obtained. The numerical results
were in good accordance with experimental ones that
confirmed appropriate determination of parameters of
the model and itself choice of the models for description
of material behavior.
BRAGOV A M1, BALANDIN V V1, KONSTANTINOV A YU1,

KONSTANTINOVA YU V2, LOMUNOV A K1,3, FILIPPOV A R1
1.
603950; 2.
603004; 3.
603950
:
,
,
.
25
, 2012 1
Vol.25 No.1, Jan. 2012
Stamp Collapse under Compression

WU Jian1, HUANG Ke-zhi1*, HUANG Yong-gang2
( 1.AML, Department of Engineering Mechanics, Tsinghua University, Beijing 100084, China; 2.Department of Civil and
Environmental Engineering and Department of Mechanical Engineering, Northwestern University, Evanston 60208, USA)
Abstract: Reversible adhesion is an important feature of stamp for transferring printing, and it is realized in
recent experiments by controlling the contact area. Compression leads to stamp collapse, and controls the
contact area to change the adhesion strength by 3 orders of magnitude. Mechanics models are developed to
determine the stamp collapse and contact area under compression. The contact area for the external load and
aspect ratio are obtained, and the finding is very useful for stamp design.
Key words: roof collapse; adhesion; transfer printing
CLC number: O341
Document code: A
Introduction
Article ID: 1001-5132 2012 01-0074-04
collapse without external load. The purpose of this paper

is to determine analytically the collapse and contact area
Geckos can climb on almost any surfaces because

their feet have fibrillar structures to adjust adhesion with
of stamp under external load, which is important to the

reversible adhesion.
a surface. This has motivated extensive studies on

biomimic adhesion materials
[1-9]
. However, fibrillar
Stamp collapse process
structures are rather difficult to fabricate for synthetic

materials. Inspired by aphids, which use sagging and
The process of stamp collapse consists of four
retraction of foot pads to enlarge or diminish contact
stages as the external load p increase: (i) uncollaspsed
[10]
state; (ii) collapsed state, which corresponds to a sudden
developed a robust approach of reversible control of
increase of contact area between the stamp and silicon
adhesion by adjusting the contact area with a surface.
platelet; (iii) stamp contact, during which the contact
They fabricated polydimethysiloxane (PDMS) stamp
area remains the same as the external load increase; and
with microtips to pick up silicon platelets from their
(iv) zipping of interface, which corresponds to the
growth substrate, and print them onto a receiving
increase of contact area with the external load.
substrate. The stamp collapsed onto a silicon platelet to
2.1 Uncollapsed and collapsed state
area with a surface to modulate adhesion. Kim et al
reach maximum contact area when the PDMS stamp
The schematic diagram in Fig. 1 shows the roof
[11]
collapse due to the external load and the adhesion
developed the theory model for the mechanics of
between the stamp and substrate. The punch height h is
was pressed firmly against the platelet. Wu et al

reversible adhesion. Huang et al
[12]
studied the stamp
much smaller than the punch width and the stamp height,

Foundation items: Supported by the National Natural Science Foundation of China (11032001, 11172146); Russian Foundation for Basic Research (11-01-91217).
The first author: WU Jian (1978-), male, Huaian Jiangsu, associate researcher, research domain: dynamical mechanics.
E-mail: wujian@mail.tsinghua.edu.cn
*
Corresponding author: HUANG Ke-zhi (1927-), male, Nanchang Jiangxi, professor , research domain: dynamical mechanics.
E-mail: huangkz@mail.tsinghua.edu.cn
WU Jian, et al: Stamp Collapse under Compression
and the punch space 2b is also smaller than the punch

width and the stamp height. The elastic modulus of
PDMS is smaller 5 orders of magnitude than that for
silicon platelet. We assume the roof collapse does not
half that in the homogeneous PDMS (Fig. 2(b))
[12]
The half of the potential energy difference between
by
Ehb
(1)
y
2 K ( 1 c / b ) (b 2
c
x 2 )(c 2
x2 )
c,
(3)
given by
d
/2
K (k )
given by
where 2c is the collapse length, and the strain energy
(2)
y
2c ,
Ucollapse due to roof collapse is

1 c
U collapse
(2)
y h dx ,
2 c
where h he p / E is the deformed punch height
1 k 2 sin 2
(4)
2
y
is the stress at the contact
interface, which is the sum of
(1)
y
(2)
y
can be
[13]
p[ x (b c) / 2]
x
(b c) x bc
p[ x (b c) / 2]
x2
(b c) x bc
c.
(5)
Then, the potential energy difference,
U total ,
equals to
related to the external load p, and the plane strain elastic

y
where K is the complete elliptic integral of first kind
(1)
modulus E , and
The second part of interface stress,
U total can be given as
U collapse
can be given
[12]
collapsed (Fig. 2(b)) and uncollapsed states (Fig. 2(a))

U total
(1)
y
The first part of interface stress,
affect the boundary. The deformation energy with roof

collapse in the stamp and silicon platelet (Fig. 1(b)) is
75
E h2 e
U total
as shown in
p
E
c
b
c2
1 2
b
2K
Fig. 3.
phe
p
E
2c(b c) 2c .
(6)
Collapse length 2c is determined by minimizing
0 , and the
the above energy difference, d U total / dc

following equation is given
Fig. 1
Schematic diagram of (a) original stamp,

(b) roof collapse between stamp and
silicon platelet due to external load p
d
d (c / b)
c
b
K
K
4 h
Eb b
p h
E b
c2
1 2
b
2
p
E
p
E
1 2
2
c
b
c
c
1
b
b
0.
(7)
Fig. 4 shows the normalized energy difference

2 U total / ( E b 2 ) versus the normalized collapse length
Fig. 2
Schematic diagram of roof collapse in Homogeneous PDMS due to external load p
c / b for the normalized external load p / E

0.0168, 0.0264, 0.035 when 4 / ( E b)
0,
0.002 . There
are several states of roof collapse that are controlled by

p / E . For p / E
0.168 , the curves increase monoto-
nically with c / b such that 2 U total / ( E b 2 ) does not
have minimum. Therefore, roof collapse will not occur

if the external load is small. For p / E
Fig. 3
0.0168 , each
Schematic diagram to illustrate the
curve in Fig. 4 has a minimum. Roof collapse may occur
superposition of stress field
in order to reach this state of minimal energy difference.
76
2012
For 0.0168
p/E
0.0264 , 2 U total / ( E b 2 ) is po-
(b c ) / 2 .
(8)
sitive, which gives the minimal energy difference
The stress intensity factor decreases and the
(marked by * in Fi. 4) above the horizontal dashed line
contact length remains the same as the external load
of 2 U total / ( E b )
0 . The energy of the collapsed
increases after roof collapse, and it equals to zero when
0.026 4 is higher than that
the external load reaches pcontact. Fig. 6 shows the
for the uncollapsed state. Therefore, even though roof
normalized external load pcontact/E is approximately
state for 0.016 8
p/E
collapse may occur, it is unstable. For p / E
0.026 4 ,
proportional to the normalized height for 4 / ( E b)
the minimal energy difference is negative, and the
0.001, 0.002, 0.003 , and the effect of 4 / ( E b) can be
collapsed state is stable. Equ. (7) and 2 U total /
neglected.
0 determined the stable roof collapse, ccollapse.
(E b )
Fig. 5 shows the normalized collapse length ccollapse / b

versus normalized height h / b for 4 / ( E b) 0.001,
0.002, 0.003 . The normalized collapse length decreases
when normalized height increases, the normalized

height has small effect on the normalized collapse
length when the normalized height is larger than 0.3.
Fig. 6
The normalized external load p/E versus

the normalized punch height h/b for
4 /(Eb)= 0.001, 0.002, 0.003
The stress intensity factor still keeps zero, K I
0,
and the contact length increases as the external load is

larger than pcontact. Fig. 7 shows the normalized contact
length increases with the normalized external load for
Fig. 4
The normalized energy difference versus the
normalized collapse length, c/b, for 4 /(Eb)=0.002
Fig. 5
different normalized punch height, and the normalized

height has large effect on the contact length.
The normalized collapse length versus the norma-
Fig. 7
lized height h/b, for 4 /(Eb)=0.001, 0.002, 0.003
The normalized contact length c/b versus the
normalized external load for h/b=0.1, 0.2, 04
2.2 Stamp contact and zipping of interface
during the zipping of interface
For the collapsed state, the stress field shows in Fig.

3, and the stress intensity factor can be given as[13]
KI
E he
Conclusion
p/ E
2 1 c 2 / b2 K ( 1 c2 / b2 ) c
We studied the roof collapse process in stamp due
WU Jian, et al: Stamp Collapse under Compression
77
to the external load and the adhesion between the PDSM
mushroom-shaped fibrillar adhesive microstructure[J].
and silicon platelet. The effects of punch width, space
Journal of the Royal Society Interface, 2007(4):271-275.
and the external load on the contact length are accounted
[6] Reddy E Arzt, A Del Campo, Bioinspired surfaces with

switchable adhesion[J]. Advanced Materials, 2007, 19:
for. The normalized collapse length ccollapse/b decreases
3833-3837.
when normalized height increases, the effect of
[7] Lee J, Fearing R S. Contact self-cleaning of synthetic
normalized height on the normalized collapse length can
gecko adhesive from polymer microfibers[J]. Langmuir,
be neglected when the normalized height is larger than
2008, 24:10587-10591.
0.3. The normalized contact length quickly increases
[8] Murphy M P, Aksak B, Sitti M. Gecko-inspired direc-
with the normalized external load during the beginning
tional and controllable adhesion[J]. Small, 2009(5):170-
of interface zipping, and the normalized height has large

effect on the normalized contact length.
175.
[9] Kim S, Sitti M, Xie T, et al. Reversible dry micro-fibrillar
adhesives with thermally controllable adhesion[J]. Soft
Matter, 2009(5):3689-3693.
References:
[1] Geim A K, Dubonos S V, Grigorieva I V, et al.
[10] Kim S, Wu J, Carlson A, et al. Microstructured elastomeric surfaces with reversible adhesion and an example
Microfabricated adhesive mimicking geck foot-hair[J].
of their use in deterministic assembly by transfer printing
Nature Materials, 2003(2):461-463.
[J]. Proceedings of the National Academy of Sciences of
[2] Sitti M, Fearing R S. Synthetic gecko foot-hair micro/

nano structures as dry adhesives[J]. Journal of Adhesion
Science and Technology, 2003, 17:1055-1073.
[3] Gao Huajian, Yao Haimin. Shape insensitive optimal
adhesion of nanoscale fibrillar structures[J]. Proceedings
the United States of America, 2010, 107: 17095-17100.

[11] Wu Jian, Kim S, Chen Weiqin, et al. Mechanics of
reversible adhesion[J]. Soft Matter, 2011, 7:8657-8662.
[12] Huang Yonggang, Zhou Weixing, Hsia K J, et al. Stamp
collapse in soft lithography[J]. Langmuir, 2005, 21:8058-
of the National Academy of Sciences of the United States

of America, 2004, 101:7851-7856.
[4] Kim S, Sitti M. Biologically inspired polymer microfibers
8068.
[13] Tada H, Paris P C, Irwin G R. The Stress analysis of
cracks handbook[M]. 3rd ed. New York: Professional
with spatulate tips as repeatable fiblillar adhesives[J].
Engineering Publishing, 2000.
Applied Physics Letters, 2006, 89:261911-261913.

[5] Gorb S, Varenberg M, Peressadko A, et al. Biomimetic
1*
1.
2.
100084;
60208
.
3
,
:
.
.
25
, 2012 1
Vol.25 No.1, Jan. 2012
Incubation-time Based Approach for Dynamic Yielding of Metals

GRUZDKOV A1,2, PETROV Y1,3*
( 1.Department of Mathematics and Mechanics, Saint Petersburg State University, Saint Petersburg 198504, Russia;
2.Department of Information Technology, Saint Petersburg Technological Institute, Saint Petersburg 190013, Russia;
3.Institute of Mechanical Engineering of Russian Academy of Science, Saint Petersburg 199178, Russia )
Abstract: Transition of the materials from elastic to plastic state is modeled by incubation-time based
criterion. This approach is similar to the structural-temporal theory for brittle solids fracture developed in
St.-Petersburg State University. This approach allow us to model anomalous behavior of yield stress in case
of elevated temperature, brittle-to-ductile failure mode transition and some other effects.
Key words: dynamic yielding; incubation-time approach; failure criteria
CLC number: O347.2
Document code: A
Article ID: 1001-5132 2012 01-0078-05
The structural-temporal theory of quasi-brittle
rate effect. Although experimental investigations show
fracture of solids had been proposed in 1980s. Later it
that this approach cannot be applied for explanation of
was developed to the incubation-time approach that
some phenomena such as yield delay. This phenome-
proved to be an effective tool of analysis for various
non is described in paper [2] and in many other works.
phenomena: ductile or brittle failure of defect-free
The applied stress is supported on the constant level.
specimen, fracture of specimen with macro-defect
Despite the fact that the values of applied stress is
(crack), plastic yielding, cavitation in liquids, electrical
significantly exceed static yield limit the yielding occurs
breakdown in solids, phase transition etc. Some
not immediately but after a definite time (yield delay).
principal results obtained during two last decades are
This delay tends to grow with decreasing of temperature
[1]
presented in recent monograph . New results concer-
(Fig. 1) and for low temperature it may reach the values
ning dynamic yielding of metals are to be discussed
of several seconds. Dependence of yield delay ( t* ) on
below.
applied stress (
) is in a good agreement with
following scaling law:
Dynamic yield criterion for metals

and alloys
log t*
log
Const .
(1)
Eq. (1) corresponds to straight lines in logarithmic

scales; the slopes of these lines are defined by different
Dependence of yield stress on loading time had
values of the shape parameter
that significantly
been observed in numerous experiments during last
decreases with growth of temperature (Fig. 1). The Eq.
decades. This dependence is usually treated as a strain-
(1) proved to be valid for numerous experimental data

The first author: GRUZDKOV A (1965- ), male, Doctor, research domain: material behavior under dynamic loading. E-mail: gruzdkov@mail.ru
*Corresponding author: PETROV Y (1957- ), male, Correspondence Academician of the Russian Academy of Sciences, professor,
research domain: dynamic fracture. E-mail: yp@yp1004.spb.edu.ru
GRUZDKOV A, et al: Incubation-time Based Approach for Dynamic Yielding of Metals
79
Analyzing experimental data for various materials we

propose approximations for these dependencies.
Y
where TP ,
T
,
TP
exp
exp
U
,
kT
(5)
W
kT
1 exp
0
(4)
(6)
, U ,
, W
are material
[6-7]
constants. Further investigation have shown

Fig. 1
that
Dependence of yield delay on applied stress for
dependencies Eqs (4-6) are enough good for many
mild steel (C 0.22, Mn 0.36). Experimental data from [2]
materials (different kinds of steel, molybdenum,
obtained using different loading history.
niobium, titanium alloys, aluminum etc.) in wide range
The necessity to take into consideration a real

loading history was emphasized by many famous
[3]
researchers
of temperatures and strain rate.

It necessary to note that at present time thermal
. Various applications to engineering
effect on yielding is often considered in very simplified
problems, analysis and comparison of experimental data
manner. Many existing models consider only thermal
obtained by different techniques require yield criterion
softening, i.e. decreasing of yield with increasing of
applicable for arbitrary loading history. Such criterion
temperature (Eq. (4)). The other approach is based on
was proposed in our previous works [4-5]:
the concept of rheologically simple material
1
Y
where
t
Y
limit and
(s)
ds 1 ,
(2)
incubation time of yielding,
[3]
. For
such materials the change of temperature is equivalent

to change of time scale (Eq. (5)). The example of such
static yield
approach is well-known hypothesis of equivalency
shape parameter. They are considered as
between increasing of strain-rate and decreasing of
the material parameters. The transition to the plastic

state corresponds to earliest violation of the condition
temperature.
Our results demonstrate that such simplified
(2). It is easy to show that for loading pulse of short
approaches
duration the criterion (2) is in a good correspondence
investigated materials. The Eq. (6) demonstrates that the
with Eq. (1) while in case of slow loading the condition
most of metals and alloys couldnt be considered as
(2) is in close agreement with quasistatic yield criterion
rheologically simple material. The change of tempe-
(t )
(3)
Therefore criterion (2) is valid in a very wide range

of loading rates
[4-7]
are
unsatisfactory
rature not only shift diagram
for
the
most
of
vertically (thermal
softening) and horizontally (time-temperature correspondence), but also radically changes the shape of these
diagrams.
Time-temperature correspondence
The change of shape of the diagram
can
lead to reverse temperature dependence of yield stress in

The duration of loading (loading rate) is not the
the range of very high strain-rate: the dynamic yield
only parameter that influence the processes discussed.
stress for elevated temperatures may be higher than for
One of the most important parameters is the temperature
low temperatures. This phenomenon contradicts to the
during testing. The material characteristics (
concept of thermal softening but it is predicted by our
proved to be strongly dependent on temperature.
model. The experimental confirmation of such ano-
80
2012
malous behavior of yield stress was obtained for pure

titanium and single-crystal aluminum
[7]
surfaces gives us a curve in plane T
that separate
domains of brittle and ductile fracture. This curve can be
One of the possible applications of yield criterion
treated as a strain-rate dependence of transitional
is the prediction of the conditions of brittle-to-ductile
temperature. Preliminary calculations for some metals
transition in fracture of solids. Brittle-to-ductile transi-
show that this dependence might have been non-mono-
tion is usually considered as the result of competition
tonous. Fig. 2 presents the calculated dependence for
between two mechanisms of deformation and fracture.
chromium-nickel-molybdenum steel. Material parame-
The assumption that these processes can be analyzed
ters have been determined on the basis of experimental
separately is known as Ludwik-Davidenko-Orowan
data
hypothesis. Using simultaneously criterion of plastic
ponds to brittle fracture.
yielding (2) and similar criterion of brittle fracture
[1]
[8]
. Domain below to the calculated curve corres-
we can estimate the time necessary for each of the
Generalization of yield criterion
processes (fracture delay and yield delay). Hence

we are able to predict which of them (cleavage or plastic
yielding) should take place earlier.
Temperature and strain-rate are usually considered
Dynamic yield criterion (2) can be considered in

frame of more general approach based on the notion of
fading memory
[3]
. According to this approach the
as the most important parameters influencing the failure
current value of applied stress in quasi-static criterion (3)
mode. It was as early as 1930s when Davidenko N N
should be replaced by relaxed stress, i.e. by functional
and his collaborators (Vitmann V F, Stepanov V I, et al.)

had shown that increasing of strain-rate shifts the
I (t )
( s ) K (t s ) ds
transition temperature to the domain of higher tempera-
where K (t ) is the function of fading memory. K (t )
tures. But recently obtained results on the behavior of
is nonnegative non-increasing function satisfying the
materials in case of extremely high strain-rate (anoma-
condition:
lous behavior of yield limit etc.) require renewing some

traditional views.
Eqs. (2,4-6) allow us to calculate dependence of
yield stress on both-strain-rate and temperature, i.e.
*
( , T ) . Similar dependence can be obtained for
resistance to brittle fracture. Intersection of these
K (t ) dt 1 .
0
Criterion (2) is evidently obtained if

1/ , 0 t
,
K (t )
0, otherwise.
So incubation time is the time of full fading which
supposed to be finite. If we deal with simultaneous
processes with different characteristic time it may be
useful to replace single incubation time by relaxation
spectrum. We define
t
I (t )
( s ) H ( s ) ds dp ,
( p)
0
1
t
p
where p is frequency,
Fig. 2
Dependence of transition temperature on
strain-rate for chromium-nickel-molybdenum steel [8]
( p ) is spectral function sa-
tisfying condition
( p)
dp 1 .
p
0
Incubation-time model corresponds to
GRUZDKOV A, et al: Incubation-time Based Approach for Dynamic Yielding of Metals
(p
( p)
where
1)
also been analyzed.
(t ) is Dirac delta function. Consider for
example two-level model. Each level is defined by

characteristic time
and threshold level of stress
Then we obtain following criterion

1
( s ) H ( s )ds 1 .
(7)
The calculated curve for following values of parameters

2
4.5 10 6 s,
2.4 10 8 s,
Incubation-time approach is applied to the
should be experimentally determined. Criterion is valid
Conclusion
temperature only two additional parameters of material
( s) H ( s )ds
t
modeling of yielding of metals and alloys. For given
81
103 MPa,
135 MPa compared on the Fig. 3 with experimen-
tal data for mild steel [9].
for wide range of loading rate; it allows using this

criterion for the analysis of thermal effect and failure
mode transition. Analysis shows that simplified models
fail to explain behavior of real materials. Proposed
temperature dependencies allow to include both
normal thermal softening and anomalous increasing
of yield stress. Similar approach has been used for the
analysis of martensitic inelasticity in shape memory
NiTi alloy [10].
References:
[1] Bratov V, Morozov N, Petrov Y. Dynamic Strength of
Continuum[M]. Saint Petersburg: Saint Petersburg University Press, 2009.
[2] Kraft J M, Sullivan A M. Effect of grain size and carbon
content on yield delay time of mild steel[J]. Trans of Am
Soc Metals, 1959, 51:643-648.
(a) semi-logarithmic scale
[3] Rabotnov Yu N. Elements of hereditary solid mechanics

[M]. Moscow: Mir Publishers, 1980.
[4] Gruzdkov A A, Petrov Y V. On temperature-time corresponddence in high-rate deformation of metals[J]. Doklady Physics, 1999, 44(2):114-116.
[5] Gruzdkov A A, Petrov Y V, Smirnov V I. An invariant
form of the dynamic criterion for yield of metals[J].
Physics of the Solid State, 2002, 44(11):2080-2082.
[6] Gruzdkov A A, Sitnikova E V, Petrov Y V, et al. Thermal
effect in dynamic yielding and fracture of metals and
alloys[J]. Mathematics and Mechanics of Solids, 2009, 14
(1/2):72-87.
[7] Petrov Yu V, Gruzdkov A A, Sitnikova E V. Anomalous
(b) linear scale
Fig. 3
Dependence of yield limit of mild-steel on tempe-
rature and strain rate model (Eq. 7) and experiment [9]
It should be noted that in case of multilevel model
behavior of yield stress upon an increase in temperature

under high strain rate conditions[J]. Doklady Physics.
2007, 52(12):691-694.
[8] Vashchenko A P, Leonov V P, Tokarev V M, et al. Effect
the shape parameter is not introduced (
1 ). Some
of rapid deformation and temperature on the strength and
other variants of choice of spectral function
( p ) have
plasticity characteristics of chromium-nickel-molybde-
82
2012
[10] Gruzdkov A, Krivosheev S, Petrov Yu, et al. Martensitic
num steel[J]. Strength of Materials, 1992, 23(9):951-954.

[9] Campbell J D, Ferguson W G. The temperature and
inelasticity of Ni-Ti shape-memory alloy under pulsed
strain-rate dependence of shear strength of mild steel[J].
loading[J]. Materials Science and Engineering 2008, 481-
The Philosophical Magazine, 1970, 21:63-68.
482:105-108.
GRUZDKOV A1,2, PETROV Y1,3*

1.
2.
198504;
3.
190013;
199178
:
University
St.-Petersburg State
,
.
25
, 2012 1
Vol.25 No.1, Jan. 2012
An Interrupted Dynamic Tensile Testing for Pure Copper Bars

MA Dong-fang, CHEN Da-nian*, WU Shan-xing, WANG Huan-ran, CAI Can-yuan
( Mechanics and Materials Science Research Center, Ningbo University, Ningbo 315211, China )
Abstract A high-speed tensile fa ility (HSTF) invented by us was applied to interrupt tests for pure copper
bars controlled locally at different levels of elongation. It was realized to isolate and identify the various
stages of the dynamic fracture process of pure copper bars under impact tension. The results of scanning
electron microscopical (SEM) investigation of recovered pure copper specimens were shown. The
axisymmetric necking rod model with a central void under static tension presented by Ragab was extended
to predicting localized necking and fracture of pure copper bars under impact tension.
Key words: interrupted tension; high strain rate; constitutive model; necking; void evolution; pure copper bar
CLC number: O347.3
Document code: A
Article ID: 1001-5132 2012 01-0083-06
As dynamic diffuse necking of metal bars develops
involved in analyzing localized necking and fracture of
into localized necking, voids within the necking zone
bars in interrupted tests. The axisymmetric necking rod
nucleate, grow and coalesce. Extensive research has been
model with a central void under static tension presented
directed toward modeling dynamic void growth and
by Ragab[8] was extended to predicting localized necking
coalescence in order to obtain a qualitative understanding
and fracture of pure copper bars under impact tension.
of the influence of void distribution, material properties

and loading on dynamic response of materials including
fracture (see, e.g. Curran et al[1], Thomason[2], Tonks et
al[3], Bai et al[4], Chen et al[5]). In order to properly
characterize dynamic necking and fracture process of
[6]
bars, a technique has been developed by Barton
to
carry out interrupted tensile testing at high strain rates.
An interrupted impact tension

testing to characterize necking
and fracture of pure copper bars
A technique to carry out interrupted tensile testing
at high strain rates on the flying wedge[9] was developed

by Barton[6]. In order to properly characterize necking
A high-speed tensile facility (HSTF) invented by
and fracture process of pure copper bars, we introduced
was applied to interrupt tests for pure copper bars at
Bartons technique[6] on a high-speed tensile facility
different levels of elongation. It was realized to isolate
(HSTF) designed by us[7]. The HSTF consists of two
and identify the various stages of the dynamic fracture
assemblies: a gas gun system to propel a projectile and a
process of pure copper bars under impact tension. The
tension mechanism to grip and strain bars. The projectile
results of scanning electron microscopical (SEM)
impacts the target plate connected to the bars, pulling out
investigation of recovered pure copper specimens show
the bars. The end blocks of the bars attached to the target
nucleation, growth and coalescence of voids in the
plate are monitored by electronic pins, and the impact
necking zone. When diffuse necking became localized
velocities in tension of the bars can be measured with
necking, damage evolution in the necking zone was
attainable precision. Multiple-bars can be conducted in a
[7]
us
Received date: 20111030.

Foundation item: Supported by the National Natural Science Foundation of China (11032001, 10872100, 11072118) and by Russian Foundation for
Basic Research (11-01-91217).
The first author: MA Dong-fang (1981), male, Ningbo Zhejiang, research domain: dynamic response of materials. E-mail: enwaill@sian.com
*Corresponding author: CHEN Da-nian (1940), male, Ningbo Zhejiang, professor, research domain: dynamic response of materials. E-mail: chdnch@nbu.edu.cn
84
2012
single test to examine the scatter of experimental data,
elongation and the initial diameter of the bar in the
and to control the bars at different levels of elongation. A
elongation-controlled part, respectively,
cross-sectional diagram of the HSTF with special fixtures

is shown in Fig. 1. In order to get a better estimate of the
localized strain defined as

*
2 ln( Dc 0 / Dc ) ,
impact velocity in tension of the bars, the whole test on
where, Dc is the minimal diameter of the bar in the
the HSTF was numerically simulated by using
elongation-controlled part. The recovered pure copper
[10]
LS-DYNA
. The special fixture was clamped around a
is the
(1)
bars are shown in Fig. 3.
certain part of the bar and only a certain amount of

elongation of the bar was allowed in tension testing. The
control effects of the fixture and internal shoulders on the
deformation of the bar were also numerically simulated
by using LS-DYNA[10].
(a)
60
M14
R1
R2
17
17
R1
20
39
16
18
50
39
20
(b)
Fig. 1
A cross-sectional diagram of the HSTF
Fig. 2
with special fixtures
(a) Photograph of pure copper bar and special split
fixture for interrupted test on the HSTF, (b) Parameters
A series of impact tension tests at different

velocities for pure copper bars with special split fixtures
of pure copper bar and special split fixture for

interrupted test on the HSTF
were carried out on the HSTF. When the pure copper bar
The results of scanning electron microscopical
with the fixture shown in Fig. 2 was subjected to impact
(SEM) investigation of pure copper specimens shown in
tension, the elongation of the bar was locally controlled
Fig. 3 (b) are shown in Fig. 4 to Fig. 6 to clarify damage
by the fixture. The typical experimental results are listed
evolution within the elongation-controlled parts of the
in Table 1, including diffuse necking, localized necking
specimens. As shown in Fig. 3 and Figs. 4 to 6, the
and fracture of pure copper bars in interrupted tests. In
different stages of the dynamic fracture process of pure
Table 1, Lc0, L and Dc0 are the initial length, the axial
copper bars under impact tension were isolated and
Table 1
Lc0/mm
17.78
17.76
(A)17.70
17.70
17.72
(B)17.82
18.02
18.00
(C)17.82
17.96
18.12
(D)18.04
Typical experimental and computed results of interrupted tests on the HSTF
L/Lc0/mm Dc0/mm
0.44
0.54
0.56
0.33
0.43
0.56
0.23
0.38
0.60
0.23
0.44
0.51
6.10
6.18
6.10
6.12
6.06
5.98
5.98
6.08
6.10
5.84
5.94
5.88
experimental
localized
strain
0.87
1.39
1.53
0.41
0.81
1.43
0.23
0.46
1.54
0.32
0.54
1.07
computed localized strain *

containing void
no void
J C model Z A model
J C model
Z
0.75
0.75
0.58
1.34
1.36
0.97
1.48
1.43
1.14
0.40
0.38
0.39
0.66
0.64
0.53
1.40
1.34
1.16
0.25
0.26
0.25
0.47
0.45
0.44
1.54
1.50
1.23
0.24
0.24
0.25
0.72
0.64
0.58
0.98
0.94
0.75
A model
0.56
0.94
1.07
0.36
0.51
1.12
0.25
0.42
1.19
0.25
0.56
0.73
MA Dong-fang, et al:
85
identified by controlling the bars at different levels of
flow stress) in the necking zone is higher than that in the
elongation and making examination of the bars before
other region of the bars. Furthermore, the stress
final and complete fracture. As the controlled elongation
triaxiality is highest in the middle of the minimum
increased, diffuse necking of the bars developed into
cross-section of the necking zone of the bars. In general,
localized necking, and voids within the necking zone
void nucleation, growth and coalescence are directly
nucleated, grew and coalesced.
dependent on the stress triaxiality as indicated by Figs. 4

to 6. The computed final (at the controlled elongation)
L/Lc0=0.44
b1
L/Lc0=0.43
L/Lc0=0.54
b2
L/Lc0=0.56
L/Lc0=0.56
b3
L/Lc0=0.33
stress triaxialities and strain rates at the positions where

voids were detected are also indicated in Figs. 4 to 6,
respectively. Different mechanisms of void coalescence
in the pure copper could be observed from Figs. 4 to 6:
(a) V=14.0 ms
-1
stable coalescence through impingement of primary
-1
voids; unstable coalescence through necking and
(b) V=21.3 ms
shearing of the primary inter-void ligament.

L/Lc0=0.23
L/Lc0=0.23
However, Figs. 4 to 6 clearly show that void

evolution at a position near the surface of the minimum
L/Lc0=0.44
L/Lc0=0.60
cross-section of the necking zone is more severe than that

in the middle of the necking zone, which may be
L/Lc0=0.38
connected with the findings discussed by Alves et al[11]:
L/Lc0=0.51
(c) V=23.4 ms-1

Fig. 3
(d) V=34.4 ms-1
certainly not from the middle where the triaxiality is a
Recovered pure copper bars
The stress triaxiality ST =p /
(a) ST 0.36,
Fig. 4
a crack started to run probably from the notch root;
(p is pressure,
-1
=3 780 s
is
maximum.
(b) ST 0.33,
-1
=4 012 s
(c) ST 0.31
-1
=450 s
SEM micrographs of the pure copper specimen: (a) shown in Fig. 3(b1) in the middle of the minimum cross-
section of the necking zone, (b) shown in Fig. 3(b1) at a position near the surface of the minimum cross-section of the
necking zone, (c) shown in Fig. 3(b1) at a position of the maximum cross-section of the necking zone
(a) ST
Fig. 5
0.41,
=12 050
s-1
(b) ST
0.36,
-1
=12 400 s
(c) ST
0.30,
=20
s-1
section of the necking zone, (b) shown in Fig. 3(b2) at a position near the surface of the minimum cross-section
of the necking zone, (c) shown in Fig. 3(b2) at a position of the maximum cross-section of the necking zone
86
2012
(a) ST 0.34,
Fig. 6
=2 110 s-1
(b) ST 0.33,
=2 200 s-1
(c) ST 0.32,
=1 280 s-1
section of the necking zone, (b) shown in Fig. 3(b3) at a position near the surface of the minimum cross-section of the
necking zone, (c) shown in Fig. 3(b3) at a position of the maximum cross-section of the necking zone
A combined theoretical and

numerical approach to analyze
the interrupted impact tension
testing for pure copper bars
The interrupted tensile tests at high strain rates for
pure copper bars were numerically simulated by using

LS-DYNA[10]. It should be noted that the stress
tensor
ij
is expressed as follows in LS-DYNA[10]:

( p q ) Sij ,
ij
(2)
where p , q and Sij are the pressure, the artificial viscosity

and the stress deviator tensor, respectively. Up to the
Hugoniot elastic limit, p is given by K ln(
where K ,
/ ) ,
are the bulk modulus, the initial density
and the instantaneous density, respectively. For stress

above the Hugoniot elastic limit, the Gruneisen equation
of state is adopted, and p is given by
p
where
parameter
C0
(1 (1
1 ( S1 1)
/
and
1,
the
C0 S1 are
material
C0
km/s,
3.94
S1
) )
(3)
U,
0
and U are the Gruneisen
specific
constants.
1.49
energy,
For
0
respectively.
pure
2.0.
copper,
Deviatoric
stresses Sij are computed by Hookes law and dynamic

constitutive models. There are a number of constitutive
n1
)(1 (T * ) m ),
)(1 C ln
(4)
where
T*
0
Tm Tr ,
T Tr
, B, n1 , C and
(5)
m are experimentally determined
material parameters, Tr is the reference temperature at

which
is measured, Tm is the melting temperature.
Zerilli and Armstrong (Z
A)[13] proposed two
microstructurally based constitutive models. For FCC

metal, Z-A model is
C0
C2
1/ 2
*
3
(6)
exp( C3 T C4 T ln ),
*
where C0 , C2 , C , C4 are material parameters.

The original and adjusted parameters of J C and
Z A constitutive model for the pure copper are listed in
Table 2 and Table 3, resperively.
Table 2
Original and adjusted parameters of J C
constitutive model for the pure copper

Constitutive
relation
J-C
Parameters
0
/MPa
B/MPa
n1
Original
85
348
0.62
0.030
1.20
Adjusted
85
308
0.54
0.025
1.09
Table 3
Original and adjusted parameters of Z A
constitutive model for the pure copper

Constitutive
model
Z A
Parameters
C0 /MPa C2 /MPa C3* / K -1

*
C4* / K -1
models of materials under impact loading obtained from
Original
84
800
0.003 2
0.000 13
TSHB tests to describe the flow stress
Adjusted
85
770
0.003 1
0.000 113
of metals as a
function of the strain , strain rate and temperature T.

[12]
Johnson and Cook (J-C)

model:
proposed the following
When diffuse necking developed into localized

necking, evolution of voids within the necking zone
should be involved. It was concluded by Niordson[14] that
MA Dong-fang, et al:
87
for cracked specimens not only the void volume fraction,
containing a central void and no void are compared with
but also the typical void size is of importance to the
experimental results in Table 1.
fracture strength of ductile materials. A model for static

ductile fracture based on internal necking of spheroidal
voids was presented by Ragab[8]. The predictions of
strains and void fractions at fracture cover axisymmetric
and plane strain loading conditions for unnotched and
notched bars made of various alloys for which the
micro-mechanical
data
were
mostly
identified
experimentally.
(a)
Fig 8
(b)
(c)
(d)
(e)
Comparison between experimental and numerically
A novel numerical simulation of the dynamic tensile
simulated deformation for the recovered bar shown in Fig. 3
copper bars are performed as shown in Fig. 7, which was
(b1): (a) computed with adjusted J C model, (b) computed
presented by Ragab[8] for static tension testing. For the
with adjusted Z A model, (c) experimental, (d) computed

with original J C model, (e) computed with original
recovered bars shown in Figs. 3(b1) and 3(b2), the
Z A model
computed deformations of bars with a central void using

the original and adjusted constitutive parameters are
compared with the experiments in Fig. 8 and Fig. 9,
respectively. The micro-mechanical data for the pure
copper are[8]: fi =0.002 5,
void fraction,
1i
1i
=1.6, where fi is the initial
is the initial void aspect ratio. In fact,
the original constitutive model determined in a certain

range of strains was adjusted in numerically simulating
large deformations of pure copper bars including
localized necking.
(a)
(b)
(c)
(d)
(e)
Fig. 9 Comparison between experimental and numerically

simulated deformation for the recovered bar shown in Fig. 3(b2):
(a) computed with adjusted J C model, (b) computed with
For typical tests, the localized strains
of bars in
1a
adjusted Z A model, (c) experimental, (d) computed with

original J C model, (e) computed with original Z A model
Conclusion
(1) A high-speed tensile facility (HSTF) invented by

[12]
us
was applied to interrupt tests for pure copper bars
controlled locally at different levels of elongation.

(2) It was realized to isolate and identify the various
stages of the dynamic fracture process of pure copper
bars under impact tension.
1a
(a)
(b)
Fig. 7 Definition of a representative element: (a) array

of regular parallel-piped unit cells each cont aining an
isolated ellipsoidal void; (b) axisymmetric necked rod
with a central void
the elongation-controlled parts computed with the

adjusted constitutive parameters in the cases of
(3) SEM micrographs of recovered pure copper

specimens show that void evolution at a position near the
surface of the minimum cross-section of the necking
zone is more severe than that in the middle of the necking
zone, which may be connected with the findings
discussed by Alves et al[11].
(4) The axisymmetric necking rod model with a
88
2012
central void under static tension presented by Ragab[8]
critical impact velocity of copper in tension[J]. Chine
was extended to predicting localized necking and fracture
Phys Lett, 2008, 25:1049-1051.

[8] Ragab A R. Application of an extended void growth
of pure copper bars under impact tension.
model with strain hardening and void shape evolution to

ductile fracture under axisymmetric tension[J]. Eng Frac
References:
Mech, 2004, 71:1515-1534.
[1] Curran D R, Seaman L, Shockey D A. Dynamic failure of
[9] Sturges J L, Cole B N. The flying wedge: A method for
solids[J]. Phys Reports, 1987, 147:253-388.
high strain rate tensile testing. Part 1. Reasons for its
[2] Thomason P F. A three-dimensional model for ductile
development and general description[J]. Int J Impact Eng,
fracture by the growth and coalescence of microvoids[J].
2001, 25:251-264.
Acta Metallurgica, 1985, 33:1087-1095.
[10] Hallquist J O. LS-DYNA Keywords Uses Manual
[3] Tonks D L, Zurek A K, Thissel W R. Void coalescence
(Version 970)[M]. USA: LSTC, 2003.
model for ductile damage[C]//Furnish M D, Thadhani N
[11] Alves M, Jones N. Influence of hydrostatic stress on
N, Horie Y. Shock Compression of Condensed Matter.
failure of axisymmetric notched specimens[J]. J Mech
American Institute of Physics, 2001:611-614.
Phys Solids, 1999, 47:643-667.
[4] Bai Y L, Xia M F, Ke F J, et al. Statistical microdamage
[12] Johnson G R, Cook W H. A constitutive model and data
mechanics and damage field evolution[J]. Theoretical
for metals subjects to large strain, high strain rates and
Appl Frac Mech, 2001, 37:1-10.
high temperatures[C]//Proceedings of seventh inter-
[5] Chen D N, Fan C L, Xie S G, et al. Study on constitutive
national symposium on Ballistics. Netherlands, Hague,
relations and spall models for oxygen-free high-
1983:541-547.
conductivity copper under planar shock tests[J]. J Appl
[13] Zerilli F J, Armstrong R W. Dislocation-mechanics-based
Phys, 2007, 101(6):3532-3541.
constitutive relations for material dynamics calculations
[6] Barton D C. Determination of the high strain rate fracture
[J]. J Appl Phys, 1987, 61:1816-1825.
properties of ductile materials using a combined experi-
[14] Niordson C F. Void growth to coalescence in a non-local
mental/numerical approach[J]. Int J Impact Eng, 2004, 30:
material[J]. Eur J Mech A/Solids, 2008, 27:222-233.
1147-1159.
[7] Hu J W, Jin Y H, Chen D N, et al. Measurement of
,
,
315211
(HSTF)
:
,
,
.
.
Ragab
,
,
.
:
25
, 2012 1
Vol.25 No.1, Jan. 2012
Effect of High Strain Rate on TiNi Shape Memory Alloys

RAZOV A
( Saint-Petersburg State University, Saint-Petersburg 198504, Russia )
Abstract: In recent years, an increasing interest in shape memory effects and TiNi alloys can be observed
from researchers, engineers and designers due to improving technologies of manufacturing and processing
of such alloys. There is also some progress in the understanding of shape memory alloys behavior both
during and after high strain rate loading. This paper presents the state-of-the-art in the investigations of
TiNi-shape memory alloys behavior at high strain rate loading. Chronology of investigations is also
presented. Thermo-mechanical response at various temperatures and strain rates is observed. Effect of high
strain rate on functional properties of TiNi-shape memory alloys is presented. In order to describe dynamic
mechanical behavior of TiNi shape memory alloy in martensitic state an attempt to apply the principles of
yielding based on the concept of the so called incubation time was suggested. For this purpose magnetic
pulse loading was used. This method allows controlling the amplitude of electric impulse and its duration.
To determine the parameters of incubation time concept the critical amplitudes of force leading to
inelastic strain of TiNi in martensitic state was determined and the characteristic time of this process was
calculated.
Key words: shape memory effect; TiNi alloys; high rate strain; functional properties; incubation time
CLC number: O341
Document code: A
Use of materials with shape memory effect (SME)

in the technique currently receives a new impetus by the
Article ID: 1001-5132 2012 01-0089-05
technologies and the manufacture of materials with the

SME.
discovery of new alloys, a considerable progress in the
The first studies of dynamic properties of TiNi-
production of semi-finished products and technologies
alloys, conducted mainly by analogy with studies of
in their processing, overcoming conservative thinking
conventional materials, were directed on determination
engineers, designers and managers. However, many
of the deformation diagrams during dynamic loading at
areas of modern application development aim to
various temperatures. For example, in [1] the alloy
miniaturize and reduce the response time of devices, the
Ti-51at.
use of fast processes. In this regard, a great importance
investigated at high strain rate 2102 7102 s-1 in
is the study of thermo-mechanical and functional
compression mode. The split Hopkinson pressure bars
properties of the alloys with SME in dynamic conditions,
(SHPB) technique was used. The temperatures were
when the deformation occurs at speeds of about 10

3
-1
Ni annealed in vacuum at 673 K for 1 h was
varied from 201 K to 366 K, while the total strain in
10 s and above, or for a quick response of the working
each experiment was about 4 . The resulting
elements they are heated at high speed. The results of
dependence of the stresses corresponding to 2
such studies may also be useful in developing
on the temperature in the case of dynamic loading has
strain

Author biography: RAZOV A (1956 ), male, professor, research domain: shape memory alloys. E-mail: razov@smel.math.spbu.ru
90
2012
qualitatively the same shape as for the quasi-static
increasing strain rate up to 103 s-1 the austenite to
loading, which was performed at 10-4 s-1. This de-
martensite transformation initiating stress increases
pendence has a minimum in the temperature range
linearly, increasing by about 500 MPa. A further increase
between Ms and As. The stresses at dynamic loading
in strain rate leads to a sharp increase in these stresses.
were always higher in the 50 100 MPa than at quasi-
Investigation of the structure of the alloy showed that at
static loading. The second sharp turn at the dependence
this point there is a change of deformation mechanism
was caused by an intermediate rhombohedral R-phase. It
(from the formation of stress induced martensite to
was noted that the recovering of the strain after dynamic
dislocation deformation of austenite phase). The effect
loading was the same as after the quasi-static
of annealing for 30 min at 296 K, 473 K, 573 K, 723 K,
deformation. In [2] the TiNi alloy with characteristic
823 K, 873 K, 923 K of TiNi alloy with the characteristic
temperatures Mf = 39
temperature Af = -10
, Ms = 59
, As = 74
, Af = 93
was investigated at high rate compression up to

-1
on the superelastic properties
under high rate compression was investigated in [5].
total strain.
Tests were carried out at temperatures range from 77 K
It was noted that the quasi-static and dynamic diagrams
to 400 K; the strain rate was varied from 103 s-1 to
in compression up to 5
strain is almost the same. Then
4 200 s-1. Transformation stress at which the direct
the dynamic diagram exceeds the quasi-static one for a
martensitic transformation start under load decreases
250 300 MPa. It was mentioned that the stabilization
with increasing annealing temperature from about
of martensite was observed regardless of the strain rate,
800 MPa after annealing at 300 K to 500 MPa after
which led to the conclusion that the mechanisms of
annealing at 873 K. It was established that the energy
deformation of TiNi alloy in the martensitic state does
dissipation in dynamic and quasi-static cases were about
not depend on the rate of loading. TiNi alloy in the
the same. However, with repeated loadings super-elastic
super-elastic state at high-rate loading was first
changes in the dynamic case is less than in the
investigated in [3]. The Ti-55, 8
quasi-static one. In addition, it was noted that the strain
3 000 s at room temperature to about 20
Ni alloy (SE508,
NDC) with characteristic temperatures Af = 5 18

Md 150
was investigated. The SHPB technique
with pulse shaper was used. Strain rate was varied from
-1
130 s
-1
to 750 s . Comparison of the quasi-static and
rate during unloading was poorly controlled and in all

tests was about 102 s-1. In [6] Ti-55, 6
Ni (NDC SE508)
alloy with characteristic temperatures Mf = -36

-8
, As = -23
, Af = +2
, Ms =
was investigated by
high rate loading showed that the super-elastic
tensile and compressive strain at a rate of about 1 200
properties depend on the loading rate. Under quasi-static
s-1 using the split Hopkinson pressure bars with and
loading at rate of 10-3 s-1 up to 3
without a pulse shaper. The loading was carried out at
strain the recovering
was complete, while after dynamic loading the
different temperatures: -196
super-elastic loop was not closed, but a complete
CT, 100
recovering of the strain took place in some time (from
the behavior of the alloy in tension and compression at
several seconds to several hours). In [4] the Ti-50.4at.
high rate loading was discussed.
, 200
, 300
, -100
, 400
, -50
, 0
. The difference in
Ni alloy with characteristic temperature Af = 296 K was
All the above-listed works did not pay attention to
investigated at room temperature using asplit Hopkinson
the influence of high rate loading on the functional
-1
pressure bars with strain rate from 5 800 s to 17 600

-1
properties of TiNi alloys, the one-way and two-way
s . The influence of additional pulse shaper was
shape memory effects, recovery stresses. For one of the
investigated in details. It was shown that the use of
first papers in which the question of whether the shape
pulse shaper is necessary to obtain relatively constant
memory effect after the dynamic loading was studied,
strain rate of TiNi alloy. It was established that with
should be attributed to [7]. In the experiments described
RAZOV A: Effect of High Strain Rate on TiNi Shape Memory Alloys
91
in this paper, the discs 40 mm in diameter and 10 mm
compression at a speed of 10-3 s-1 was performed on the
thick made of equiatomic TiNi alloy were subjected to
test
machine
INSTRON.
Dynamic
loading
was
-1
performed using the split Hopkinson pressure bars
from a gas gun at different temperatures. On the
technique. The pulse duration in all experiments was
opposite side of the disc a suspended steel ball was
about 120 s, and the strain rate was varied from 3
positioned. This ball deformed the disc during impact
102 s-1 to 1103 s-1, 5103 s-1. After straining the
and created an imprint on it surface. At room
specimens were thermo cycled at the rate near 1 Kmin-1.
temperature TiNi alloy was in martensitic state and
At the first heating the strain recovery was observed due
subsequent heating led to the initiation of the shape
to shape memory effect, while at the subsequent cycles
memory effect and reduce the depth of indentation.
of cooling and heating, the variation of length was
Comparison
dynamic
resulted from the repeated two-way shape memory. Fig.
experiments showed that the strain recovering due to
1 shows the dependence of shape memory effect on
SME in the latter case was approximately 10
residual strain and Fig. 2 represent the dependence of
impact by flat projectile with a velocity of 100 ms
of
the
quasi-static
and
higher.
The first studies of the two-way shape memory initiated
two-way shape memory on irreversible plastic strain.
by dynamic deformation, and its comparison with the

two-way shape memory obtained after quasi-static
deformation were carried out in [8]. The 1 mm thick
target made of equiatomic TiNi alloy was subjected to
impact by flat projectile with a velocity of 100 ms-1
from a gas gun at room temperature. Then the samples
with dimensions of 50 mm5 mm1 mm were cut from
the target and were tested in three-point bending mode,
conducting thermal cycling through the temperature
range of martensitic transformations. In the quasi-static
case, the plate in the martensitic state was deformed in
Fig. 1
Dependence of the shape memory effect on
the same mode. It has been found that the two-way
the preliminary residual strain after quasi-static 1
shape memory after shock loading higher than after
and impact 2 loading
quasi-static loading. In addition, it has been established

that the two-way shape memory formed during dynamic
deformation, more resistant to external counteracting
stresses than after quasi-static effects.
Unfortunately, these works did not give precise
quantitative result on change in the functional properties
of TiNi alloys under dynamic loading.
The first systematic study of the influence of high
rate loading on SME and two-way shape memory was
carried out in [9]. Cylindrical samples of 5 mm in length
and diameter made of Ti-50.8at.
Ni alloy were used.
Fig. 2
Dependence of the two-way shape memory effecton
After annealing at 773 K for 1 h, samples were com-
the irreversible plastic deformation strain after
pressed in the martensitic state at room temperature
he quasi-static 1 and impact 2 loading
at different strain rates. Quasi-static deformation by
It should be noted that the maximum SME strain
92
2012
under dynamic loading is shifted to smaller residual
zero deflection. Fig. 4 presents the results of the one of
strains in comparison with the quasi-static loading.
the series of testing. Threshold stresses in each case
When the residual deformation is less than 5
were calculated based on elastic beam theory and their
the shape
memory effect after the dynamic compression is more
values were obtained in the range 30
complete than after quasi-static loading, and when
pulse durations between 400 ms and 800 ms (Fig. 4).
residual deformation more than 6
900 MPa for
the situation is
reversed. The situation with two-way shape memory is

very similar. When the plastic strain is less than 4
the
two-way shape memory after the dynamic compression

is more than after quasi-static loading.
During the deformation of alloys with SME two
processes can simultaneously develop-deformation due
to martensitic inelasticity (twinning, reorientation of
martensite, the formation of stress-induced martensite)
and conventional dislocation plasticity. In order to
determine the characteristic parameters of these
Fig. 4
Determination of the critical force amplitude
processes the magneto-pulse technique and the theory,

based on the concept of incubation time[10] were used.
The dynamic three-point bending tests were carried out
using magneto pulse installation with energy content of
up to 15 kJ[11]. This installation allowed generating
impulses of pressure with durations of microsecond and
amplitudes up to 2 GPa. The pressure pulse is passed
from the copper buss to the specimen through triangle
steel striker (Fig. 3). Diameter of wire made of
equiatomic TiNi alloy, annealed at 500 C during 1 h
(Mf = 36.5 C) was 2 mm, distance between supports
(2L) has been varied in the range of 26-49 mm.
Fig. 5
Dependence of threshold stress on pulse duration
After heating of the samples it was found that the

critical force FCR, initiating martensitic inelasticity
(reversible channel of the deformation) and a normal
dislocation plasticity are different from each other. For
example, for the experiment with 2L=30 mm and striker
mass m=6,7 g the force that initiates the shape memory
Pm: pressure initiated by magnetic field, i: current, C: capacity, Q:
effect was equal to 17,1 kN, and the force that initiates
switchboard, R: formative vylite resistor, L own inductance of pulse current
dislocation plasticity was equal 18,8 kN. The incubation
generator; 1: Output bus, 2: striker, 3: investigated sample, 4: supports
time of transient process was calculated on the base of
Fig. 3
Scheme of magneto-pulse installation and

scheme of loading
experimental data. Its value about 2 ms characterizes the

range of appearance of dynamic properties of TiN alloy[11].
In each test the residual deflection and force
The currently available data shows strain rate
amplitude have been measured. In each series we
sensitivity of the thermo-mechanical and functional
determined the critical force FCR by extrapolation to the
properties of the TiNi shape memory alloys.
RAZOV A: Effect of High Strain Rate on TiNi Shape Memory Alloys
93
temperatures[J]. Mechanics of Materials, 2006, 38(5-6):
High strain rate loading of TiNi alloys can lead
463-474.
both to an increase in functional properties and to their
[6] Adharapurapu R R, Jiang F, Vecchio K S, et al. Response
suppression.
of NiTi shape memory alloy at high strain rate: A
There is still no complete answer to the question
systematic investigation of temperature effects on
about the quantitative relationship between shape
tension-compression asymmetry[J]. Acta Materialia, 2006,
memory effects and strain rate, about their dependencies

on value and mode of straining for different temperatures.
54(17):4609-4620.
[7] Likhachev V A, Patrikeev Yu I. Shape memory effet in
titanium nickelide after quasi-static and shock loading[J].
References:
Rubezhnoe, 1990:128-129.
[1] Ogawa K. Characteristics of shape memory alloy at high
[8] Likhachev V A, Shimanskii S R. Effect of deformation

rate on the reversible shape memory of titanium
strain rate[J]. J Phys IV, 1988, 49(C3):115-120.
nickelide[J]. Problemy Prochnosti, 1988(2):65-68.
[2] Liu Y, Li Y, Ramesh K T, et al. High strain rate

deformation of martensitic NiTi shape memory alloy[J].
[9] Belyaev S P, Morozov N F, Razov A I, et al. Shape

memory effect in Titanium-Nickel after preliminary
Scripta Materialia, 1999, 41(1):89-95.
dynamic deformation[J]. Materials Science Forum, 2002,
[3] Chen W W, Wu Q, Kang J H, et al. Compressive
394-395:337-340.
super-elastic behavior of a NiTi shape memory alloy at

strain rates of 0.001-750 s-1[J]. International Journal of
[10] Morozov N, Petrov Yu. Dynamics of Fracture[M]. BerlinLondon-New York: Springer Verlag, 2000.
Solids and Structures, 2001, 38(50-51):8989-8998.

[4] Nemat-Nasser S, Choi J Y, Guo W G, et al. Very high
[11] Gruzdkov A, Krivosheev S, Petrov Yu, et al. Martensitic
strain-rate response of a NiTi shape-memory alloy[J].
inelasticity of TiNi-shape memory alloy under pulsed
Mechanics of Materials, 2005, 37:287-298.
loading[J]. Materials Science and Engineering: A, 2008,

481-482:105-108.
[5] Nemat-Nasser S, Guo W G. Super-elastic and cyclic

response of NiTi SMA at various strain rates and
TiNi
RAZOV A
,
198504
TiNi
.
,
TiNi
,
,
TiNi
; TiNi
TiNi
.
25
, 2012 1
Vol.25 No.1, Jan. 2012
Hypervelocity Impact of the Icy Droplet on Al Shell at

Nanoscale: A Molecular Dynamics Probe
YUAN Quan-zi, ZHAO Ya-pu*
( State Key Laboratory of Nonlinear Mechanics, Institute of Mechanics, Chinese Academy of Sciences, Beijing 100190, China )
Abstract: Large-scale molecular dynamics (MD) simulations are conducted to investigate the hypervelocity
impact of the icy droplet on spacecraft Al shell at the atomic level. The velocity of the droplet is set to move
at a typical cosmic speed. Part of the kinetic energy of the droplet is converted to thermal energy, resulting
in a sudden increase of the temperature. With the increase of the droplet speed, the aluminum plate
experiences elastic, plastic deformation, melting and finally be penetrated by the droplet. The formation and
dynamic behavior of a debris cloud is observed and investigated in the simulations.
Key words: hypervelocity impact; molecular dynamics; debris cloud; icy droplet
CLC number: O341
Document code: A
Introduction
Article ID: 1001-5132 2012 01-0094-04

simulations can be complementary powerful tools to
explore the extreme conditions as in the hypervelocity
With the fast development of space technology and
impact. The structural dynamic plastic response is con-
the increase of humans space activities, the quantity of
trolled by the response number[3]: Rn ~ ( v 2 /
space debris increases significantly in the last decade.
where
[1]
, v and
)( L/ l ) 2 ,
represent density of the droplet,
According to the statistics from NASA , the appro-
velocity of the droplet and yield strength of the metal
ximately estimated population of particles is 19 000
plate, respectively. Here, L and l represent diameter of
larger than 10 cm; 500 000 between 1 cm and 10 cm;
the droplet and the thickness of the metal shell,
and exceeds 10 000 000 smaller than 1 cm. Owing to
respectively. L~l, which makes the problem independent
their hypervelocity, these space debris would make great
of the length scale. When space debris impacts on the
threat to the orbit spacecraft. Even the collision of a tiny
metal plate, the debris would be compressed adiabatically,
icy droplet at cosmic speed would result in serious
and liquefied or even vaporized instantly[4], while the
damage on the huge spacecraft. In order to ensure the
metal plate experiences elastic, plastic deformation and
safety of the spacecraft from the hypervelocity impact,
rupture finally[5].
the detailed investigation of the impact and optimization

of the structure are necessary[2].
Results and discussions
The hypervelocity impact experiments for icy

droplet with spacecraft are still extremely difficult to
In this article, we investigated from the atomic
realize till now even in laboratory. So numerical
level the hypervelocity impact of the icy droplet with

Foundation items: Supported by the National Natural Science Foundation of China (11032001, 11011120245) and by Russian Foundation for Basic
Research (11-01-91217).
The first author: YUAN Quan-zi (1983 ), male, Beijing, doctor, research domain: solid mechanics. E-mail: yuanquanzi@lnm.imech.ac.cn
Corresponding author: ZHAO Ya-pu (1963 ), male, Beijing, doctor/professor, research domain: impact dynamics, physical mechanics.
E-mail: yzhao@imech.ac.cn
YUAN Quan-zi et al: Hypervelocity Impact of the Icy Droplet with Al Shell at Nanoscale: A Molecular Dynamics Probe
95
1. The thickness of the Al shell is 2.03 nm. Since the

shock-induced yield strength of Al is about 125 150
GPa[7], the corresponding perforation velocity v ~
M
is at the order of 10 kms-1. Hence, the
velocity of the icy droplet (~3.5 nm in diameter) was

selected to be typical cosmic speed, i.e., 7.9 kms-1, 11.2
kms-1 and 16.7 kms-1, which is a usual speed of space
debris.
Fig. 2 illustrate the process of the icy droplet
The gray and white balls represent the water molecules and
impacting the Al shell at the 1st cosmic speed. Since the
the Al atoms, respectively
Fig. 1
Illustration of the MD simulation domain
kinetic energy of the droplet is less than the yield
spacecraft Aluminum (Al) shell at nanoscale using
strength of the shell, the droplet did not perforate the
molecular dynamics (MD) simulations accomplished in
shell. The plastic region accumulated and then released
[6]
LAMMPS . The simulation domain is illustrated in Fig.
(a)
into the Al surface due to the dislocation starvation of
(d) side view of the impact; gray and white balls represent water molecules and Al atoms, respectively; (e) (h) the development of the
plastic region; (i) (l) top view of the wave propagation in the Al shell
Fig. 2
The icy droplet impacts the Al shell at the 1st cosmic speed
1st cosmic velocity
Fig. 3
The evolution of the pressure with respect to time
96
2012
the nano shell (Fig. 2 (e)-(h)). The impact energy
kms-1. The impact made the shell vibrate with a period
spreads in the form of wave. The wave velocity in our
of 2 ns (Fig. 3). Because the impact velocity is much
-1
MD simulations is about 3.15 kms , which is in
higher than the speed of sound in the Al shell, the shock
accordant with the average theoretical value ~3.20
wave energy accumulated and exerted a tensile force on
Fig. 4
The icy droplet impacts the Al shell (a) (d) at the 2nd cosmic speed;
and (e)
Fig. 5
(h) at the 3rd cosmic speed
Total energy with respect to time (a) 1st cosmic speed; (b) 2nd cosmic speed
Fig. 6
The development of the plastic region (2nd cosmic speed)
YUAN Quan-zi et al: Hypervelocity Impact of the Icy Droplet with Al Shell at Nanoscale: A Molecular Dynamics Probe
97
the below surface of the Al shell.
mechanisms in the hypervelocity impact for spacecraft.
References:
Conclusions
[1] NASA.
Hypervelocity
impact
technology[EB/OL].
[2011-11-23 ]. www. nasa. gov.
In conclusion, we performed the MD simulations
[2] Yuan J, Qu G, Sun Z, et al. Optimization methodology for
to investigate the hypervelocity impact of the icy droplet
shield structure against space debris[J]. Journal of
with spacecraft Al shell at the atomic level. The kinetic
Astronautics 2, 2007, 28(2):243-248.
energy of the droplet and the yield strength of the shell
[3] Zhao Y P. Suggestion of a new dimensionless number for
competed in the dynamic process of hypervelocity
dynamic plastic response of beams and plates[J]. Archive
impact. The droplet did not perforate the shell at 1st
of Applied Mechanics, 1998, 68(7):524-538.

[4] Engel O G. Crater depth in fluid impacts[J]. Journal of
cosmic speed. Most of the plastic deformation was
Applied Physics, 1966, 37(4):1798-1808.
released from the solid surface. The Al shell vibrated
[5] Yu T X, Zhang L. Plastic bending: Theory and
almost elastically with a period of about 2 ns. The
applications[M]. Singapore: World Scientific Pub Co Inc,
droplet perforated the shell at 2nd and 3rd cosmic speed
1966.
(Fig. 4). Part of the deformation would be released in
[6] Plimpton S. Fast parallel algorithms for short-range
about 0.3 ns (Fig. 5). The plastic region propagated and
molecular-dynamics[J]. Journal of Computational Physics,
finally the plate was ruptured (Fig. 6). Accompany with
1995, 117(1):1-19.
[7] McQueen R, Fritz J, Morris C. Shock waves in
the perforation, a debris cloud would form with its
condensed matter-1983[M]. Amsterdam: North Holland
diameter dependant on the impact speed.
Physics Publishing, 1984.
We hope our large-scale MD simulation results

may help in further understanding the dynamics and
100190
.
.
25
, 2012 1
Vol.25 No.1, Jan. 2012
Threshold Characteristics of Pulse Loads Causing Fracture on an

Example of Concrete and Rocks
( Mathematics and Mechanics Faculty, Saint Petersburg State University, Saint Petersburg 198504, Russia )
Abstract: The technique of definition of threshold parameters of an impact is shown. The experimental
results on crack extension in concrete and rocks under the short pulse (2 s) of loading are presented. The
loading was carried out on the crack surface by means of an installation of conductor explosion. The
dependence of the crack length on the stored energy allows to define the threshold amplitude of the impact
for the given duration of the pulse. The possibility in principle of definition of effective surface energy of
fracture within the limits of the developed approach is shown.
Key words: threshold pulse; crack extension; electrical explosion of conductor; fracture energy; concrete;
rocks
CLC number: O347.2
Document code: A
Introduction
Article ID: 1001-5132 2012 01-0098-05
parameters should be entered into the fracture theory:

duration of the action, fracture time, the time between a
The principle of limit loads is used to estimate the
maximum of the loading in the point of fracture and the
material strength under slow loading. For example, the
moment of fracture, etc. Nikiforovsky and Shemjakin
criterion of maximum static strength of a material is
have suggested the criterion of a power pulse [2-3]:
introduced into tests for rupture [1]:
Jc .
(2)
(1)
This means that failure occurs at the value of the
This principle has well proved in the case of
pulse greater than the critical pulse. However, the
quasi-equilibrium processes in a material. However,
momentum depends on the duration T, amplitude P and
there is still no common point of view on the fracture
velocity of the pulse front P , hence, the threshold
criterion under the action which leads to failure in the
amplitude and duration determine the threshold pulse
moment of transient processes of energy distribution on
(minimal fracture pulse):
the object. The given fact arises from the circumstance
that researchers are trying to extrapolate classical
Jc
J ( P, T , P ),
J c ( P* , T * ).
(3)
principles of fracture theories from quasi-statics in the
Here it is necessary to notice that the threshold
domain of nonequilibrium processes. Nevertheless, the
values of amplitude and duration can be any for the
simple cases of quasi-statics should follow as a special
same value of the threshold pulse. If we increase the
case of the dynamic process
[2]
pulse duration, then we obtain the criterion for the
Since we speak about dynamic processes, time
quasi-static case, and the threshold pulse amplitude

The first author: SMIRNOV I (1985-), male, post of graduate student, research domain: fracture mechanics. E-mail: ivansmirnov@math.spbu.ru
SMIRNOV I, et al: Threshold Characteristics of Pulse Loads Causing Fracture on an Example of Concrete and Rocks
99
aspire to the so-called static strength of the material (Fig.

1). If the duration of the action is comparable with the
duration of fracture, then various dynamic effects of
failure become apparent. The matter is that any material
has the finite time of failure (for a given scale of
fracture), and the duration of stress pulse can be less
than the duration of the fracture process.
1: charging device, 2: discharger, 3: exploding wire,
4: laser interferometer, W: half-width of the sample
Fig. 2
The experimental scheme for testing of

samples with a notch
with a thickness of 15 mm. The notch with a length of

70 mm and width of 3 mm was made in the middle of
the sample for crack imitation.
The tests were carried out on the installation of
electrical explosion of conductors with parameters:
capacitance C
The threshold amplitude P* vs. the duration T
Fig. 1
t
max
energy
50 kV, stored
7.5 kJ. A copper wire (diameter of 0.6 mm)
and
was placed between the crack faces and perpendicular to
represent minimum and maximum time
the plane of the plate at a distance of 20 mm from the
(for the pulse of triangular shape). t

*
6 mF, voltage Uc
min
required for a given fracture.
In works [4-5], it is shown that strength parameters

of materials, such as spall strength, fracture toughness,
notch tip. The wire was passed through cambric of PVC

with a diameter of 0.7/3.5 mm for better acoustic
contact.
etc. depend on threshold loads. Overload of the
This method of loading raises the question: how to
threshold pulse does not give an objective picture about
determine the amount and shape of the impact? It is
strength, so it is important to study strength charac-
necessary to associate the energy of the explosion with
teristics at threshold parameters of the action. The ques-
the parameters of the stress pulse. A series of separate
tion arises about the definition of these thresholds.
experiments was conducted for this purpose.

In these experiments, the half of the sample width
Experimental procedure and

calculations
W was varied from 10 to 70 mm, and the velocity of the

lateral surface was measured using a laser interferometer
(Fig. 2). Actually, explosion of a wire leads to formation
As an example of definition of threshold loads, we
of a cylindrical wave, but the sample dimensions allow
will consider one of experiments which were carried out
to use the formula for calculation of the pulse amplitude
in St. Petersburg State University. In the experiments the
P in a case of a plane wave [6]:
crack length was measured as a function of energy
0.5 cw,
(4)
stored in the capacitor. The experimental setup is shown
where
in Fig. 2.
waves, w is velocity of the free surface. The duration of
Samples of sandstone, dolomite, gabbro-diabase

and concrete were in the form of plates 150/150 mm
is material density, c is velocity of longitudinal
the bell-shaped pressure pulse was T = (20.1) s.

Thus, the initial amplitude was determined by
100
2012
fitting the experimental values of the stress pulse

amplitude at aspiration of the sample width to enough
loading in the form of the Heaviside step function
f (t )
H (t )
[7]
. We used numerical modelling to avoid
small size (Fig. 3). It was found that the value of the
bulky calculations and to consider the solution taking
pulse amplitude corresponds to one fifth of the value of
into account the real sizes of the sample.
the stored energy.
Change of the stress at the crack tip without

account of crack propagation is shown in Fig. 5. The
arrival of longitudinal wave leads to compression in the
crack tip. The rapid increase of stress begins after the
arrival of the Rayleigh wave.
experiment: square points; extrapolation: dashed line
Fig. 3
The attenuation of the pulse amplitude in the
gabbro-diabase sample (U= 25 kV). W is half-width

of the sample (see Fig. 2)..
The stress-wave pattern in the vicinity of the crack
Fig. 5
Simulation of the stress-wave pattern
tip in the conditions of this experiment can be obtained
at the crack tip
from a solution of the plane problem. An elastic

isotropic plane (x, y) with a semi-infinite crack (y = 0,
Results
x<0) is considered. At time t = 0 a concentrated pulse of

pressure act on the crack face in the point with
coordinates (y = 0, x = -L) (Fig. 4). Thereby,
2
The dependences were obtained by extrapolation the
experimental points to zero length (or minimum visible
y
t
(
,
0,
)
x
t
P ( x L) f (t ), x
y
xy ( x, 0, t )
v( x, 0, t )
The experimental results are presented in Figure 6.
0,
0, x
length) of the crack. The amplitude corresponding to

0
(5)
x
0
zero length of the crack is threshold amplitude. For

example, the threshold for sandstone is 0.32 GPa. Thus,
fracture will not occur when a pulse with duration T < 2
s and amplitude of P < 0.32 GPa.
It is important to note that we must always bear in
mind the scale of observation. This means that we must
agree on what is fracture? at the stage of the problem
formulation. For example, in our case, it is a crack
visible with the naked eye. However, we could look for
an initial crack with a microscope or to accept its
minimum allowable length. All these are different scales
Fig. 4
Statement of the analytical problem
This analytical problem has been solved for
of failure, which require different contributions of

energy.
SMIRNOV I, et al: Threshold Characteristics of Pulse Loads Causing Fracture on an Example of Concrete and Rocks 101
The angle between the Y-axis and the curve is the
in a case of composites requires further research, since
fracture energy d / dl . In our example, the estimation
threshold loads can differently affect on the elements of
of quantitative values of the energy consumption can
the composite material.
contain large errors due to the indirect determination of

the input energy. However in [8], it is shown that the
Conclusions
energy required for fracture depends on the pulse

duration and can greatly exceed the Griffith surface
energy.
Strength characteristics of a material depend on

threshold parameters of the impact. This paper presents
an example of defining the threshold pulse amplitude
and duration in simple experiments. It is shown that
failure of the considered rocks and concrete has
threshold character.
The analysis of the results shows the possibility of
application of electric explosion of conductors to study
dynamic strength parameters of materials, including
rocks and concrete.
The crack length was measured as a function of
explosion energy. The threshold value of the pulse
amplitude for a given duration was defined by
extrapolation of the experimental points to zero length
of the crack. This approach makes it possible to
determine not only the threshold characteristics of an
impact, but also effective fracture surface energy.
Acknowledgments
The authors would like to thank the Russian
Foundation for Basis Research for the support of these
studies; and Mr. Ezio Cadoni from DynaMat Laboratory
of SUPSI for the samples of concrete.
References:
[1] Beer F P, Johnston E R, DeWolf J T and Mazurek D F.
points: the result of the experiments; dashes: the linear extrapolation
Fig. 6
The crack length vs. the stored

energy of the capacitor
The analysis of the fracture surface of concrete has

shown that there are several ways of crack extension:
through the matrix, between the inclusion and the matrix,
Mechanics of Materials[M]. 5th ed. New York: McGrawHill Companies, 2008.

[2] Bratov V A, Morozov N F, Petrov Yu V. Dynamic
strength of continuum[M]. St. Petersburg: St. Petersburg
university press, 2009.
[3] Nikiforovskii V S, Shemyakin E I. Dynamic Fracture of
Solids[M]. Novosibirsk: Nauka, 1979.
through inclusion. Each way requires the various
[4] Petrov Yu V. Incubation time criterion and the pulsed
contribution of energy. Determination of fracture energy
strength of continua: racture, cavitation, and electrical
102
2012
Academic Press, 1967.
breakdown[J]. Doklady Physics, 2004, 49(4):246-249.
[7] Freund L B. Dynamic Fracture Mechanics[M]. Cam-
[5] Petrov Yu V, Smirnov I V, Utkin A A. Effects of strain-
bridge: Cambridge University Press, 1998.
rate strength dependence in nanosecond load duration
[8] Gruzdkov A A, Krivosheev S I, Petrov Yu V. The fracture
range[J]. Mechanics of Solids, 2010, 45(3):476-484.
energy of materials under pulse microsecond-scale load-
[6] Zeldovich Ya B, Raiser Yu P. Physics of shock waves
ing[J]. Physics of the Solid State, 2003, 45(5):886-889.
and high-temperature hydrodynamic phenomena Volume II

[M]//Hayes W D, Probstein
R F,. New York and London:

,
198504
2 s
.
:
25
, 2012 1
Vol.25 No.1, Jan. 2012
Radon-based ULM
,
,
315016
Radon
,
Radon
: Radon
: TP391.41
:A
: 1001-5132 2012 01-0103-05
,
.
Radon
.
R-L
,
.
,
1.1
,
,
,
.
.
[1-2]
,
[3-8]
,
,
,
.
[1-8]
[3]
[4]
1)
[5]
Cannon
,
,
.
[6]
Radon
(Uniform Linear Motion, ULM).
[7-8]
,
Radon
[7]
,
,
[8]
,
,
Radon
2011 10 09.
1977
:
. E-mail: hwen713@gmail.com
104
2012
(a)
(b) 3
(c) 4
G (u , v)
sin( uL)
.
lena
o
60
Matlab
,
,
lena
(u, v)
sin( uL)
1.2
(
(
)
Radon
:
g ( x, y )
f ( x, y )
h ( x , y ) n ( x, y ) ,
, g ( x, y )
; f ( x, y )
h ( x, y )
(1)
(PSF); n( x, y )
h ( x, y )
1 / l, x2
y2
l, x / y
tan ,
(2)
0, else.
Radon
2.1
(2)
2.1.1
PSF
h ( x, y )
:
1 / l ,| x | l , y
0, else.
(3)
:
0,
(3)
bit
(1) ,
Fourier
(
G (u , v)
;
)
F (u, v) H (u , v)
F (u , v)
h ( x, y )e
l
F (u , v)
0
F (u, v)
1
e
L
j 2 ux
j 2 ( ux vy )
dxdy
.
Matlab
dx
sin( uL)
e
uL
j uL
,
lena
3
.
2.1.2
e
j 2 ( ux vy )
. (4)
105
0.89.
4
2.2 Radon
2.2.1 Radon
.
Radon
(
4(b)), Radon
,
,
1
,
.
(a) lena
(
),
Radon
.
g( , )
.
( , ),
g( , )
Radon
( , )
:
g( , )
g ( x, y ) (
x cos
y sin )dxdy ,
x cos( )
(b) FFT
lena
(5)
y sin( )
()
.
FFT
2.2.2 Radon
,
.
(
,
4(c)),
.
Radon
5
2
Otsu
(a)
Radon
(b)
lena
Radon
,
.
delta
,
,
5(
60
Radon
(c)
1)
90
106
2012
1.97,
3, 4.9,
80, 170.
,
.
6
,
1
,
.
Radon
2.3
5
Radon
Matlab
,
.
(5)
(a)
3
10
0.1),
1.
1
10
0.1
11
10
9.4
20
23
20
21.5
30
33
30
32.1
40
43
40
42.2
50
51
50
50
60
60
60
59.8
80
85
80
83
170
174
170
174.9
4,
,
.
2
;
,
,
,
,
.
1
,
0.1
,
;
(b)
.
2.38,
80, 170.
0.1
Radon
,
,
,
107
512 512
,
Radon
.
3(a)
7
,
:
.
[1]
[J].
.
, 2004, 24(9):4-6.
[2] Yitzhaky Y, Milberg R, Yohaev S, eta1. Comparison of

direct blind deconvolution methods for motion blured
images[J]. Applied Optics.1999, 38(20):4325-4332.
[3]
[J].
[4]
, 2004, 25(4):297-299.
,
PSF
[J].
, 2001, 21(9):40-41.
[5] Cannon M. Blind Deconvolution of spatially invariant
image blurs with phase[J]. IEEE Trans Acous. Speech
Signal Process, 1976, ASSP-24(1):58-63.
7
Radon-based ULM
[6]
[J].
.
, 2008, 28(6):220-222.
[7] Krahmer F, Lin Y, McAdoo B, et a1. Blind image Deconvolution motion blur estimation [EB/OL]. [2006-05-12].
,
.
http://www.ima.umn.edu/2005-2006/MM8.9-18.06/activit
ies/Wohlberg-Brendt/team5_pres.pdf.
:
;
[8]
,
[J].
Radon
, 2008, 30(9):51-57.
Research on Radon-based ULM Image Blur Direction Detection Algorithm

HUANG Wen, ZHANG Xue-yan
( Department of Information, Ningbo Radio & TV University, Ningbo 315016, China )
Abstract: Address a Radon-based uniform linear motion(ULM) image blur direction detection algorithm by
analyzing the advantages and disadvantages of the previous algorithms. First assume ULMs rationality and
reason characteristics of the image spectrum. Second apply Radon transform to the pretreatment image, and
compute its motion direction. Further experiments show that to get more accurate direction reduces the excessive
step, which greatly larges operational costs while the effect is small.
Key words: radon transform; point spread function; Otsu; fast fourier transform
25
, 2012 1
Vol.25 No.1, Jan. 2012
1.
315211; 2.
315211
.
;
: C939; TP301
:A
: 1001-5132 2012 01-0108-05
,
.
,
[2]
,
,
.
,
,
[1]
,
.
(
)
,
.
Uschold&King
Gruningner&Fox
[3]
TOVE
, Cyc
.
,
,
1959
,
.
,
2011 09 03.
1989
1971
,
,
,
,
,
,
,
.
,
,
:
. E-mail: 598919619@qq.com
:
. E-mail: lbs@nbu.edu.cn
109
2.2
,
;
Protg
KAON
,
,
Open
WebOnto
,
,
HTML Widgets
Loom
,
. Protg
.
,
Ontosaurus
,
Protg
.
[4]
2.3
,
gua
2.1
: Ontolin-
OKBC (Open Knowledge Base Connectivity)
OCML (Operational Conceptual Modeling Language)

Frame Logic
[5]
, web
,
,
web
, OWL
,
2002
, W3C
OWL
.
DAM+OIL
,
.
,
,
.
.
,
,
2002
OWL
2.4
,
,
,
,
20252
4614
15638
.
.
(
110
2012
)
(
1).
2.4.1
2.4.4
,
[6-7]
Mcleod
. Hammar
,
<canonical-
term, term, ontology>;
<term1, ontology1, relationship, term2, ontology>.
Guarino
2.4.2
Lexical Conceptual Graph

,
,
.
Protg
,
,
, OWL
.
2.4.5
2.4.3
protg
.
(1)
.
.
Ontoclean
2.5
.
(2)
Uschold & King
(
.
4
)
Bernaras
Gruninger & Fox
,
(
Ontometric
2.5.1
,
,
,
.
(3)
.
.
: part-of
kind-of attribute-of
instance-of. part-of
, kind-of
2
, attribute-of
2.5.2
, instance-of
.
(4)
111
,
.
,
,
,
,
.
,
.
,
2.5.3
.
:
.
(1)
,
.
,
.
.
(2)
.
,
,
<
>
(3)
2.5.4
(4)
,
.
.
,
(5)
(6)
2.6
(1).
,
,
.
,
,
,
.
,
,
,
,
,
,
,
.
112
2012
3
[4]
:
[1]
[J].
, 2007, 33(4):41-44.
.
[J].
, 2010(2):33-
[5]
[2]
[J].
37.
,
240.
[J].
[6]
, 2004, 41(7):1041-1050.
. OWL
[J].
, 2006(12):30-45.
[3] Uschold M, Gruninger M. Ontologies: Principles, methods,

and applications[J]. Knowledge Engineering Review,
, 2007, 30(2):236-
[7]
.
:
1996, 11(2):93-155.
[J].
, 2010, 23(2):23-28.
Thesaurus-based Ontological Construction Using Ocean Domain Ontology

ZHOU Jun-gen1, LIU Bai-song2
( 1.Faculty of Business, Ningbo University, Ningbo 315211, China; 2.The Library, Ningbo University, Ningbo 315211, China )
Abstract: With a view of better conduct ocean management, this paper applies some concepts in ontology to the
ocean management area, and explores the methods for ontological establishment. From the ocean status in quo
and based on the existing Ocean Thesaurus, the ontology construction method is proposed. Under the guidance
of the introduced method the ocean domain ontology is constructed for ocean management. This ontology solves
effectively the coordination and information sharing for different administrative departments and advances the
ocean management techniques.
Key words: ocean management; domain ontology; thesaurus
25
, 2012 1
Vol.25 No.1, Jan. 2012
315211
:
,
;
J2SE
ARM
Linux
:
,
: U447; TP273
:A
: 1001-5132
2012
01-0113-06
,
[1]
,
;
[5-6]
.
,
,
,
GPRS
[2-3]
.
,
1
1.1
,
.
,
RISC(
32
ARM
[4]
,
,
,
.
ARM
.
,
,
2011 09 03.
1986
1960
,
,
Y1110161 ;
,
,
,
,
LCD
2011A610181 ;
,
:
:
,
PC
2010R405064, 2010R405066 .
. E-mail: zhouhq1212@sohu.com
. E-mail: zhouyu@nbu.edu.cn
114
2012
1.3
.
1
(1)
,
,
,
,
,
A/D
,
ARM
Linux)
Jam VM
GNU
jikes
(2)
LC0136M
IC
.
,
(3)
,
,
EM9130N.
,
.
(4)
1.2
(
,
,
2).
ARM
. ARM/Strong ARM
,
,
.
,
MMU
ARM
,
2
,
,
Nor Flash
(Nand Flash
SDRAM
(5)
,
) LCD
,
S3C2410
classpath
,
J2SE
,
.
ARM920T
Linux 2.6
Java
JAL (Java
ARM920T
.
CRC
115
:
,
Dn1/ 2 P T H rs (1)VDn1/ 2 ,
1/ 2
n
T
p
(5)
V Em ,
1/ 2
n
E PD
(6)
(7)
G
[7-8]
2.1
,
,
[9].
2.2
,
.
.
ERA
.
,
/ERA
U (k ) , p
Y (k ) ,
X (k 1) GX (k ) BU (k ),
Y (k ) CX (k ),
, X (k )
, G , B, C
(1)
3
, B [ X 0 | X1 |
0,1,
,
(2)
Y (k ) CG B.
(i
, m 1)
| X m 1 |] , X i
ERA
,
. ERA
,
(
Hankel
Y (k ) Y (k 1
Y (k 1 Y (k 2
H rs (k )
Y (k r 1 Y (k r
H rs (0)
,
.
.(3)
Y (k r s 2
, P, V
, D
H rs (0)
), ERA
Y (k s 1
Y (k s
PDV ,
, D diag (d1 ,
(4)
, dn , dn 1 ,
, dl ), l
min(rp, ms ) .
:
T
p
T
m
16
[ I p Op
O p ],
[ I m Om
Om ],
diag (d1 ,
, d n ),
Dp
n
1.
3.1
,
,
1
/m
16
154.7
/m
0.15
/MPa
2.11010
/(kgm-1)
84.1
/kN
4.4103
/
27.1
/Hz
0.789
116
2012
,
.
3000
.
4~
.
,
80 Hz,
6
,
(
).
(a)
2.
(b)
(a)
(b)
(a)
(b)
117
3
2
/
/m
5
5
/Hz
Hz
Hz
Hz
0.789
0.813
3.04
AR
0.789
0.758
3.93
0.789
0.804
1.90
Burg
0.789
0.750
4.94
0.789
0.758
3.93
0.789
0.781
1.01
Mcov
0.789
0.788
0.13
Music
0.789
0.766
2.92
ERA
0.789
0.797
1.01
SSI
0.789
0.781
1.01
AR
0.789
0.820
3.92
0.789
0.813
3.04
,
,
,
,
ERA
5 Burg
3.2
ERA
ERA
, ERA
,
3.3
4.
,
,
.
4
0.789
0.820
3.93
Music
0.789
0.801
1.52
ERA
0.789
0.787
0.25
SSI
0.789
0.783
0.76
AR
0.789
0.781
1.01
0.789
0.781
1.01
Mcov
0.789
0.781
1.01
Music
0.789
0.766
2.92
ERA
0.789
0.779
1.68
SSI
0.789
0.793
0.51
5 Burg
Mcov
AR
0.789
0.813
3.04
Burg
0.789
0.961
21.80
Mcov
0.789
0.961
21.80
Music
0.789
0.742
5.96
ERA
0.789
0.761
3.55
SSI
0.789
0.838
6.21
,
4
/Hz
/kN
/Hz
/kN
C15
0.86
3 922
0.85
3 831
C17
0.80
3 765
0.78
3 579
C18
0.78
3 821
0.76
3 627
C19
0.76
4 317
0.75
4 204
C20
0.72
4 281
0.71
4 163
C16
0.89
4 342
0.88
4 245
C17
0.84
4 172
0.83
4 073
C18
0.79
4 167
0.77
3 958
C19
0.76
4 186
0.75
4 076
C20
0.71
3 951
0.69
3 731
118
2012
[J].
[5]
[6]
[J].
.
, 2008(1 :75-76.
[J].
[7]
, 2008, 12:1-4.
[J].
, 2005, 7(7):23-24.
.
[8]
,
[J].
:
,
[J].
.
, 2007, 24(2):110-114.
[9] Ni Y Q, Ko J M, Zheng G. Dynamic analysis of large-
, 2002(5):79-82.
diameter sagged cables taking in account flexural rigidity
[J]. Journal of Sound and Vibration, 2002, 8(5):620-627.
,
[J].
[3]
ARM
[J].
WSN
[2]
, 2008, 6(7):230-234.
[1]
, 2007, 25(2):9-12.
[4]
.
, 2008, 5(8):57-71.
The Embedded Remote Detection System for

Bridge Cable Force Based on ERA
ZHOU Hong-qiong, MIN Jie, ZHOU Yu*, WANG Xiao-dong
( Faculty Information Science and Engineering, Ningbo University, Ningbo 315211, China )
Abstract: An embedded cable force detection system based on the modal analysis technology is designed in this
paper. In view of the characteristics of the embedded system, we select the ERA owing to its better performance
and computational accuracy to identify the modal harmonic frequency of cable, and obtain the free response
signal using the random decrement technique to preprocess the vibration signal. The detection system is
engineered on the basis of the ARM9 chips, Linux operating system and embedded J2SE technology
development. The test results from the field engineering measurement show that the improved random
decrement technique and ERA implemented on the designated embedded system can quickly identify base
frequency of most cable vibration signal from which the cable force values can be calculated for. The proposed
approach has demonstrated the efficiency and practicability of embedded cable force detection system based on
modal analysis techniques.
Key words: modal analysis; embedded system; ERA; cable force
25
, 2012 1
Vol.25 No.1, Jan. 2012
315211
,
,
;
.
;
: TH166
;
:A
: 1001-5132
[7]
2012
01-0119-06
,
[1-2]
. Korpela
,
[3]
. Hazra
[4]
1.1
,
.
,
[5-6]
,
1
,
.
.
1
,
1
,
.
[8]
,
,
Ln
,
; C
Lf
C4
C9 C11
C 12
. Narasimhan
,
,
2011 09 26.
1988
1958
70871062
, ,
, ,
;
,
,
973
2009CB326204 ;
Y1080429
,
:
. E-mail: cherishliulanlan@sina.cn
:
. E-mail: yefeifan@nbu.edu.cn
120
2012
(3)
( 1,
),
,
n
(1 E j ) /
(1 Ek ), j
n,
(1)
k 1
1
ln m
Ej
n.
zij ln zij , i m, j
(2)
i 1
(4)
:
( 1,
).
(5)
w ( w1 , w2 ,
wn ) ,
,
wj
(3)
j 1
Z W,
,
;
(6)
, R
(4)
(r1 , r2 ,
rm )
.
.
1.3
,
:
(1)
,
.
(2)
; xi
1.2
; ri
AHP
; li
; l
; D
; EPQi
.
(5)
:
,
(1)
li
( Li
, Li
n
xi ) / C p ,
Ln
Lf , Cp
i
A
:
.
1:
:
n
min f1 ( xi )
i 1
( zij )m n .
C8 DLT , DLT
A (aij ) m n , i [1, m], j [1, n] ,

, aij
(5)
(2)
2:
(li
l )2 .
(6)
121
z1* ,
min f1 ( xi )= max( f1 ( xi ))
n
max f 2 ( xi )
ri xi ,
z2*
(7)
max f 2 ( xi ) ,
i 1
s.t.
n
xi
D,
*
1
(z , z ) .
(2)
(8)
*
2
(1)
,
,
i 1
li
1,
xi
(9)
EPQi .
(10)
, (6)
( i 1, 2 )(
ui
; (7)
; (8)
; (9)
zi* ,
1, zi
*
i
1 (z
0, zi
g ( xi , li ),
zi ) / pi , z
xi , li
.
2
xi , li
*
i
ui ( x)
; (10)
pi
zi* ,
zi
g ( xi , li ),
zi*
pi
xi
g ( xi , li ),
(11)
xi
li
li ,
g ( xi , li ),
g ( xi , li )
xi , li |
,
,
xi
0,
i 1
li
Pareto
1,
xi
EPQi } .
pi
,
.
(12)
zi*
zi
,
, pi
.
[9-10]
, pi
(12)
(8)~(10)
pi
:
i
(z
*
i
min( zi )),0
1, i 1, 2 .
(13)
Matlab
:
(1)
,
z1
((6)
(7)
f1 ( xi ), z2
f 2 ( xi ) ,
min fit ( x)
(3)
2 ( w1[u1 ( x)] w2 [u2 ( x)]) .
(14)
(3)
1
Ln /
Lf /
C1/
C2/
C3/
C4/
C5/
C6/
C7/
C8/
C9/
C10/
C11/
C12/
EPQi /
( d-1)
S1
60
70
4000
1800
85
90
75
90
70
20
85
70
90
60
40
S2
150
100
4500
1700
60
85
80
75
75
15
70
80
85
75
30
S3
60
30
3000
1300
90
75
85
60
80
20
90
90
60
85
40
S4
120
160
4000
1600
75
80
60
85
60
30
60
75
80
80
50
S5
60
50
3500
1500
80
60
90
80
75
25
75
60
65
90
50
122
2012
60,
0.2,
0.8,
100.
2
2.
1000
,
1000
800
200
,
20
,
.
EPQi
14
15
1).
(1)~(6) :
W= [0.13, 0.052 6, 0.076 9, 0.142 1, 0.062 2,
0.066 8, 0.021 2, 0.211 0, 0.061 0, 0.105 0,
0.050 4, 0.020 8]T,
5
:
T
R=[0.89, 0.87, 0.88, 0.79, 0.82] .
f1 ( x0 )
0.254 7 .
Matlab 7.11
i5, 2.79 GHz,
4
4.1
2.93 GB.
2
1
w1
0.5
0.2
0.8
0.5
0.2
0.8
w2
0.5
0.8
0.2
0.5
0.8
0.2
a1
0.8
0.8
0.8
0.2
0.2
0.2
0.8
0.8
a2
0.8
0.8
0.8
0.2
0.2
0.2
0.8
0.8
x1
164.837 9
177.737 4
166.871 3
168.731 9
166.080 1
171.080 2
266.904 5
171.215 2
x2
30.004 4
30.000 2
30.008 5
30.004 0
30.036 6
30.029 5
49.743 4
30.010 0
x3
208.318 4
214.721 8
200.844 7
205.527 2
216.816 9
209.235 2
200.376 3
202.305 8
x4
165.518 0
126.246 5
164.077 8
160.593 7
165.048 8
168.878 3
142.065 9
167.934 5
x5
231.321 6
251.295 0
238.201 2
235.143 0
222.017 5
220.776 5
140.909 9
228.533 7
l1
0.742 1
0.774 3
0.747 2
0.751 8
0.745 2
0.757 7
0.997 3
0.758 0
l2
0.930 0
0.930 0
0.930 0
0.930 0
0.930 1
0.930 1
0.995 8
0.930 0
l3
0.750 8
0.766 8
0.732 1
0.743 8
0.772 0
0.753 1
0.730 9
0.735 8
l4
0.762 5
0.680 4
0.743 5
0.737 7
0.745 1
0.751 5
0.706 8
0.749 9
l5
0.662 6
0.722 6
0.696 4
0.690 3
0.664 0
0.661 6
0.501 8
0.677 1
f1
0.035 0
0.035 8
0.033 7
0.034 0
0.038 1
0.038 2
0.178 7
0.033 0
f2
676.572 7
679.038 4
676.310 3
676.825 0
677.184 8
676.964 5
684.931 1
676.585 1
123
1,
Matlab
100
1 2
,
,
,
.
(3)
,
,
,
,
,
2
.
,
3
,
,
.
,
.
,
,
,
,
4.2
2
f2
(0.8)
(0.2)
w2
.
,
w2
1(
,
,
4.3
(1)
w1
0.254 7
f1
w1
(0.8)
f1
1~8
.
f1 =0.178 7.
w1
f1
.
(2)
,
7
0,
,
,
.
0.178 7
1 2
,
,
0.5,
,
,
1
124
2012
[4] Hazra J, Mahadevan B, Seshadri S. Capacity allocation in

an electronic market[J]. Production and Operations Management, 2004, 13(2):161-170.
[5]
.
[J].
[6]
:
[1] Araz C, Mizrak Ozfirat P, Ozkarahan I. An integrated
multicriteria decision-making methodology for outsour-
[2] Amid A, Ghodyspour S H, O'Brien C. Erratum: Fuzzy
Economics, 2006, 104:394-407.
supply chain design[J]. International Journal of Production Economics, 2002, 78:187-195.
, 2010(5):44-46.
sourcing[J]. European Journal of Operational Research,

2004, :236-250.
[8]
.
[J].
2010(2):32-34.
[9]
[J].
[3] Korpela J, kylheiko K, Lehmusvaara A, et al. An

analytic approach to production capacity allocation and
.
[J].
multi-objective linear model for supplier selection in a

supply chain[J]. International Journal of Production
, 2011(3):175-177.
[7] Talluri S, Narasimhan R. A methodology for strategic
cing management[J]. Computers and Operations Research,

2007, 34(12):3738-3756.
, 2010(3):472-
477.
[10]
,
[J].
.
:
, 2006, 19(2):204-209.
Order Allocation Driven by Manufacturing Resources in Industrial Clusters

LIU Lan-lan, ZHANG Sheng, YE Fei-fan*, LI Guo-fu
( Faculty of Mechanical Engineering and Mechanics, Ningbo University, Ningbo 315211, China )
Abstract: To tackle the problem with the unbalanced use of manufacturing resources in clusters, an order
allocation approach is presented based on the comprehensive evaluation of suppliers, in which production load
equilibrium and optimal comprehensive performance of manufacturing resources are viewed as the
decision-making criteria. The manufacturing resource utilization in suppliers is taken into consideration while
the selected manufacturing resources performance is also taken into account in the proposed approach. A multi
objective programming order allocation model is constructed, and the genetic algorithm based on fuzzy logic is
applied to seek the solution to the model. Finally, a case study is selected to verify the feasibility of the proposed
approach, and the results indicate that using the proposed model and algorithm can lead to satisfactory solutions.
Key words: order allocation; load rate balance; comprehensive evaluation; fuzzy logic; genetic algorithm
25
, 2012 1
Vol.25 No.1, Jan. 2012
,
,
315211
,
;
: R384.9
:A
(Blatta),
: 1001-5132 2012 01-0125-03
(Cockroach),
.
1.3
3.5
,
1960
20
, Princis
5000
250
[5]
200 d;
,
,
.
,
[1]
7
,
;
,
[2-3]
,
.
,
,
1.1
[3]
1.4
(Peri planeta fuligiosa)

,
.
1.2
[4]
.
,
. 1
1.5
[7]
.
,
2010 07 13.
1979
2009F80003
, ,
.
,
. E-mail: liujianfa@nbu.edu.cn
[6]
126
2012
;
,
,
72 kD
36 kD.
2.5
[17]
.
,
SDS-PAGE
ELIB
[8-9]
2.1
[10]
SDS-PAGE
,
,
,
3
3.1
[18]
2.2
[11]
cDNA
.
PCR
15
[12-13]
231
N1
,
.
,
,
2.3
.
,
3.2
,
[14]
[19]
[15]
RT-PCR
PfDNV
VP1
,
. 4
,
pMD18-T
pET28a
SDS-PAGE
A2
pET26b,
pET28a-PLA
,
43 kDa
pET26b-PLA,
BL-21-codonplus(DE3)-RIL,
,4
, SDS-PAGE
2.4
IPTG
.
PfDNV
PLA2,
.
, WHO
[16]
PAGE
SDS30
4
[20]
PfDNV
,
127
[10]
.
[21]
, 2004, 4(2):
113-117.
8~9
,
[J].
[11]
[12]
[13]
[J].
, 2005, 14(9):8-10.
[J].
, 2004,41(3): 216-222.
,
[J].
, 2006,
22(10):913-917.
,
[14]
.
,
[J].
,
,
, 2004,
20(4): 291-294.
[15] Chapman M D, Vailes L D, Haeden M L, et al. Cockroach
allergens and their role in asthma editor allergy and aller-
:
[1]
gic disease[J]. Blackweil Science Ltd, 1996, 942-951.
[16]
[J].
[2]
[3]
[M].
[5]
, 1982.
.
:
[18]
, 2008, 29(5):
cDNA
, 1993, 4(5):343-347.
.
[J].
, 2006, 6(2):127-130.
[19]
A2
, 2002, 13(1):37-38.
,
, 2005,
104-107.
[J].
[7]
[J].
[J].
[J].
,
21:64-66.
, 1982, 25(4):416-422.
, 1993, 5(2):94-95.
[6]
[J].
[17]
[J].
[4]
. 2008, 20(3):14-16.
[J].
, 2006, 21(2):
181-185.
[20]
[J]. 2006, 22(10):913-917.
[J].
[8] Cochran D C, Grayson J M. Cockroach biology[M]. New

York: WHO/VBC, 1975:75.
, 2003,
18(2):178-180.
[21]
[9] Brenner R J, Koehler P G, Patterson R S. Health implica-
.
[J]. 2003, 18(3):237-240.
tions of infestations[J]. Infect Med, 1987(4):349-358.
Development in Biology Research on Periplaneta fuliginosa

LIU Jian-fa, WANG Shi-ping, JIANG Wen-wen, ZHANG Jie-nan, LIN Lei, ZHOU Xi-wu, ZHOU Fei, HU Qi-feng
( Medical School, Ningbo University, Ningbo 315211, China )
Abstract: The present research has demonstrated that cockroaches have certain anti-tumor function and effect of
immune adjustment. With the in-depth study being conducted, the pharmaceutical value of Periplaneta
fuliginosa has attracted more and more research attention from both parasitologists and pharmaceutical scientists.
This paper reviews the research progress and biological traits of Periplaneta fuliginosa.
Key words: medicinal insects; Periplaneta fuliginosa; development biology
25
, 2012 1
Vol.25 No.1, Jan. 2012
KCaA
CO2
1,2
1.
1*
315211; 2.
CaA
CaA
.
KCaA
CO2
CO2
CaA
,
CO2
;
CO2
CO2
;
1 bar
; KCaA; CaA; CO2

:A
: 1001-5132 2012 01-0128-05

(2)
[1]
CO2
(IPCC)
, CO2
, CO2
,
,
.
, CO2
,
[15]
[3]
:
[2]
.
(
CO2
N-
(MDEA))
N2/CO2
CO2
[4-6]
35 MDEA
[7]
[8]
[9]
[11]
KCaA
26.7 .
: TK265; TQ028
21
:
,
CaA
K+
CaA
315211
[10]
0.4 MPa
CO2
TO4
CO2
CO2
.
N2/CO2
,
[12]
Al
[SiO4]
[AlO4]
,
,
[13]
,
[14]
: (1)
: Ax/n(SiO2)(AlO2)x
2011 11 03.
,
.
Si
1987
1967
,
,
Y00813VA08 .
,
,
,
,
,
/
:
:
. E-mail: fanhaiyan0723@163.com
. E-mail: hpxcm@nbu.edu.cn
mH2O (A
: KCaA
Zn
2+
Ni
Ag
2+
4 h.
,
C2H4/CO2
. 5A
Ca
CaA
2+
(PCT)
D8 Advance
,
.
CO2
4+
Mn
5A
CO2
3+
Cr
5 2
60,
s;
S4800
,
0.02,
KCaA
,
,
,
,
PCTproE&E
(PCT)
5A
CaA
PCT E&E
KCaA
CO2
,
+
0.2
(SEM)
CO2
(XRD)
KCaA
CO2
[19]
2+
CaA
CaA
CO2
1.2
5,
KCaA
(XRD)
(SEM-EDS)
0.75CaO0.25Na2OAl2O32SiO24.5H2O.
,
min, 6 min, 10 min, 15 min
LTA
[18]
Cu
100
IIB
: IB
Cl-
Cu
K+
Ca2+
[17]
2+
129
n),
LTA
[16]
CO2
CO2
1 bar
2 bar
99.999 wt ,
CO2
.
CO2
473 K
CO2
298 K,
4h
H2O.
CaA
CO2
1
1.1
H2O
,
.
CaA
)
:
1 g CaA
5 h,
180
30 mL
KCl (
,
.
1g
,
KCl
KCl
CaA
1 min.
40 mL
2
2.1 XRD
2
CaA
XRD
130
2012
CaA
,
5A
KCaA
CaA
XRD
5~60
2 =
,
,
KCl
.
4 m
4 m
3 CaA(
KCaA(
,B
LTA
SEM
[20]
25
-1
0.2 molL
,
: Ag > Tl > K > NH4+ > Rb+ > Li+ >
+
Cs+, Zn2+ > Sr2+ > Ba2+ > Ca2+ > CO2+ > Ni2+ > Cd2+>
Hg2+, Mg2+
[21]
K+
Ca2+
CaA
.
2 CaA
KCaA
XRD
KCaA
+
X
2
,
,
, K+
4
, EDS
Scherrer
Cl
,
+
,
,
,
KCaA
K / B cos ,
,D
, nm; K
0.891;
,
,
,
0.154 056 nm; B
().
4 m.
2.2 SEM-EDS
CaA
SEM
CaA
KCaA
SEM
,
,
5(a)
SEM
A Z+B
-
,Z
CO2
5(b)
:
+ -
KCaA
1 bar
XRD
.
+ -
2 bar
CO2
0
B Z+A ,
;A
;
KCaA
K+
KCaA
2.3 PCT
,
4 m,
,
CaA
CO2
,
: KCaA
: CaA
-1
3.10 mmolg ,
CO2
1 bar
2 bar
131
CO2
, CaA
-1
3.40 mmolg ; KCaA
,
CO2
(KA
3 10
-1
1 bar
CO2
3.93 mmolg , 2 bar
4.19 mmolg .
,
,
,
CO2
.
CO2
KCaA
KCaA
CaA
Cm )
,
CO 2
2+
10
, 1 bar
m),
CO2
.
CO2
.
,
(3.3 10-10 m) ,
CO2
(CaA
-10
26.7 ,
(0.99),
,
5
10
-40
CaA
,
Ca
KCaA
CO2( 14
SEM
; PCT
,
2
XRD
EDS
,
.
CaA
CaA
(1.38 )
KCaA
K+
CO2
CO2
CO2
m)
-1
-10
:
[1]
[J].
, 2006, 21(4):346-351.
[2] IPCC. Carbon dioxide capture and storage[M]. London:

Cambridge University, 2005:265-272
[3]
[4]
. CO2
[J].
, 2003, 3(10):30-33.
,
. CO2
[J].
, 1999, 7(6):
175-181.
[5]
[J].
, 2003, 18(1):1-3.
[6]
CO2
[J].
, 2005, 24(1):
1-4.
[7]
[8]
[J].
, 2009, 13:88-39.
[J].
MDEA)
, 2007, 7(4):8-12.
[9] Anthony J L, Sudhir N V, Maginn E J, et al. Feasibility of

using ionic liquids for carbon dioxide capture[J]. International Journal of Enviroment, 2004, 4(2):105-115.
[10] Feron P H M, Jansen A E. CO2 separation with polyolefin
membrane contactors and dedicated absorption liquids:
Performances and prospects[J]. Sep Purif Technol, 2002,
27(3):231-242.
5
KCaA
1 bar (a)
2 bar (b)
CO2
[11]
.
, 2008, 21(4):5-9.
[J].
132
2012
[12]
48
[14]
[J].
[18]
[19]
, 2000, 22(2):63-65.
[M].
[17]
,
:
[J].
5A
, 2010, 28(3):12-16.
and compost on soil properties[J]. Commun Soil Sci Plant

Anal, 2002, 33(3/4):595-607.
[21]
[J].
, 2011, 20(2):52-54.
, 2004:546.
.
[J].
[20] Filcheva E G, Tsadilas C D. Influence of clinoptilolite
[J].
, 2009, 31(6):76-78.
,
CO2/N2
CO2
N2/CO2
, 1998, 23(5):11-
, 1998, 12(2):1-3.
, 2009, 29(12):79-82.
[16]
. CO2
[J].
[15]
[J].
14.
[J].
, 2009, 38(6):469-473.
[13]
C2H4
MCM-
CO2
CO2
Synthesis of KCaA Molecular Sieve and its CO2 Adsorption Capacity

FAN Hai-yan1,2, KONG Chun-long2, XIA Hai-ping1*, CHEN Liang2
( 1.Laboratory of Photo-electronic Functional Materials, Ningbo University, Ningbo 315211, China;
2.Institute of Materials Technology and Engineering, Chinese Academy of Sciences, Ningbo 315201, China )
Abstract: The KCaA molecular sieve is prepared using the ion exchange method in which potassium chloride
solution is used with stirring process. K+ is added to the CaA molecular sieve to regulate its pore size and
enhance its CO2 adsorption sites. The ion exchange time and the sorption pressure are investigated in this work.
X-ray diffraction, scanning electron microscopy, energy dispersive spectrometer and the pressure composition
isotherms technique are used to characterize the molecular sieves. Results indicate that the structure and
morphology of the KCaA molecular sieve are the same with that of CaA molecular sieve; CO2 adsorption
capacity of KCaA molecular sieves first increases with the increase of the testing pressure, followed by slightly
decreasing as the ion exchange time increases; Moreover, CO2 adsorption capacity of the KCaA molecular sieve
is found to be apparently higher (26.7 ) than that of original CaA molecular sieve under a pressure of 1 bar.
Key words: ion exchange; molecular sieve; KCaA; CaA; CO2
1988
.
.
1
1.1
8000
.
1.2
4
3 5
200
4
.
2
GB3100-3102
.
3
3.1
6
.
7 cm
.
3.2
.
.
.
4
4.1
[J].
.
4.2
4.3
[M].
.
.
4.4
4.5
4.6
4.7
1
[C]//
.
[D].
[S].
.
.
[P].
.
.
/
].
[
4.8
[1]
[2
[1]
[2]
.
5
5.1
E-mail
5.2
1.2
.
http://3xb.nbu.edu.cn
word
.
5.3
.
.
5.4
5.5
http://3xb.nbu.edu.cn/zx-lg.html
.
.
5.6
315211
87600816
.
87600291.
.
xblgb@nbu.edu.cn
818
0574

Ningbo Univ 1 and 2

Încărcat de

Informații document

Titlu original

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

Ningbo Univ 1 and 2

Încărcat de

Drepturi de autor:

Formate disponibile

NINGBO DAXUE XUEBAO LIGONG BAN

BRATOV V, PETROV Y V, VOLKOV G

SMIRNOV I, PETROV Y, SUDENKOV Y

JOURNAL OF NINGBO UNIVERSITY

Jan. 10, 2012

DING Xue-yan, ZHOU Fan, HUANG Fu-yong, HE Feng, MENG Qing-hui ( 1 )

Mechanics Research (Column for Sino-Russia Workshop)

Hierarchic Nature of Dynamic Deformation and Fracture of Rock Mass

YU Tong-xi, GAO Zi-yang, HU Ling-ling ( 53 )

Surface Effects and Problems of Nanomechanics

A Three-dimensional Viscoelastic Constitutive Model of Finite Deformation

SHEN Li-jun, BIAN Zhong-jing ( 64 )

WU Jian, HUANG Ke-zhi, HUANG Yong-gang ( 74 )

Incubation-time Based Approach for Dynamic Yielding of Metals

An Interrupted Dynamic Tensile Testing for Pure Copper Bars

Communication Engineering & Computer Application Technology

HUANG Wen, ZHANG Xue-yan ( 103 )

Order Allocation Driven by Manufacturing Resources in Industrial Clusters

Vol.25 No.1, Jan. 2012

JOURNAL OF NINGBO UNIVERSITY ( NSEE )

(Portunus trituberculatus Miers)

(Portunus trituberculatus Miers),

(Bacillus subtilis Cohn),

(Carassais auratus gibebio)

(Homarus gammarus L.)

Penaeus monodon survival and growth[J]. Aquaculture,

[9] Boonthai T, Vuthiphandchai V, Nimrat S. Probiotic

bacteria effects on growth and bacterial composition of

black tiger shrimp (Penaeus monodon)[J]. Aquaculture

Nutrition, 2011, doi: 10.1111/j.1365-2095.2011.00865.x.

[10] Keysami M A, Saad C R, Sijam K, et al. Effect of

larvae Macrobrachium rosenbergii (de Man)[J]. Aqua-

culture Nutrition, 2007, 13:131-136.

[11] Shen W Y, Fu L L, Li W F, et al. Effect of dietary

performance, immune response and antioxidant activities

of the shrimp (Litopenaeus vannamei)[J]. Aquaculture

Research, 2010, 41:1691-1698.

of dietary Bacillus spp. and mannan oligosaccharides

(MOS) on European lobster (Homarus gammarus L.)

larvae growth performance, gut morphology and gut

[13] Moriarty D. The role of microorganisms in aquaculture

Aquaculture, 1999, 180:147-165.

[1] Ring E, Olsen R E, GifstadI T , et al. Prebiotics in

[15] Ziaei-Nejad S, Rezaei M H, Takami G A, et al. The effect

enzyme activity, survival and growth in the Indian white

characterization of a heat shock protein 70 gene in

shrimp Fenneropenaeus indicus[J]. Aquaculture, 2006,

swimming crab (Portunus trituberculatus)[J]. Fish &

Shellsh Immunology, 2010, 28:56-64.

shellsh[C]//Lesel R. Microbiology in Poecilotherms.

Amsterdam: Elsevier, 1990:171-176.

[17] Prieur G, Nicolas J L, Plusquellec A, et al. Interactions

between bivalves molluscs and bacteria in the marine

environment[J]. Oceanography and Marine Biology: An

Annual Review, 1990, 28:227-352.

the growth and digestive enzyme activity of white pacific

shrimp (Litopenaeus vannamei)[J]. Oceanic and Coastal

Improvement of the probiotic effect of micro-organisms

[19] Gomez R, Geovanny D, Ma S. Influence of probiotics on

[7] Bomba A, Nemcov R, Gancarckov S, et al.

oligosaccharides and polyunsaturated fatty acids[J].

British Journal of Nutrition, 2002, 88(S1):95-99.

rohita fingerlings effects on growth, nutrient digestibility