SPECTRUM REDUCTION FACTOR IN STEEL BUILDINGS

ABSTRACT The Spectrum Reduction Factor used in some of the current seismic design codes are intended to account for damping, energy dissipation capacity as well as for overstrength of the structures. The usual design approach is based on the classic elastic analysis and the non-linear seismic behavior is represented by lateral forces obtained from an elastic spectrum, reduced by this factor. The Chilean codes define this factor as a function of the fundamental period. The code makes no difference between steel moment resistant frames and steel braced frames. That means it uses the same spectrum reduction factor for all of them, i.e., the same seismic design loading. Carlos Aguirre is Professor of Structural Engineering at Universidad Tcnica Federico Santa Mara, Valparaiso, Chile. He has been involved in research on seismic behavior of steel structures, experimental behavior of connections and rack supporting structures He has taught different subjects, with emphasis in steel structures and seismic engineering. He has undertaken development, consulting and he had served as Head of the Structural Division and Chairman of the Civil Engineering Department. He currently participates in Code Committees of the Chilean Seismic Regulations. He is a licensed Professional Engineer and he owns and engineering firm who works in building and industrial projects. A comparative analysis of earthquake demands and building capacities for a set of different types of steel buildings, including moment and braced frames, from 6 up 32 stories is presented. All the buildings were designed according to the Chilean code and practice. The earthquake requirements of the buildings were obtained from nonlinear analyses, using some representative earthquake records. The building capacities were obtained by two methods: push over analysis and dynamic non linear analysis under scaled records. It was found that the demands of the Chilean earthquake on the buildings sample are smaller than the demands of Northridge or Kobe earthquakes on the same sample. Several differences in the behavior of the structural systems, as well as differences between the code and the calculated reduction factors were also found.

INTRODUCTION Lateral design forces specified in the earthquake resistant design provisions are normally lower and in some cases much lower than the lateral forces required to maintain a given structure in the elastic range, in the event of a severe earthquake occurrence. The usual design code approach is to stipulate the lateral design forces in terms of a Reduced Elastic Design Spectrum by a factor called the Spectrum Reduction Factor (sometimes called Strength Reduction Factor). Under this approach the design is based on an elastic analysis and the non-linear seismic behavior is represented by the reduction of the elastic spectrum. It is assumed that the factor takes into account different aspects related o the seismic behavior, such as the energy dissipation capacity of the structure; redundancy; overstrength, changes in damping and period and observed performance of different structural types. The level of reduction is primarily based on the observed performance of different structural systems during previous destructive earthquakes and it is normally associated to the ductility ratio of the structure. Several researchers have expressed their concern about the lack of rationality in the reduction factors currently specified in building codes (Bertero, V.V., 1986; Fischinger, F. and Fajfar, 1990; Uang, C. M., 1991). There are a number of studies (Newmark and Hall, 1973, Lai and Biggs, 1980, Ridell and Newmark, 1979, Elgadhamsi and Mohraz, 1987, Ridell et al, 1989, Arias and Hidalgo, 1990, Nassar and Krawinkler, 1991, Vidic et al., 1992, Miranda E., 1993) performed from the standpoint of the seismic demand on a single degree of freedom system (SDFS). They recommend to use a period dependent reduction factor, the function increases from 1 (on T=0) up to the ductility ratio (when T is large). Some codes use to apply a constant reduction factor, the Chilean approach, (NCh 433-96 Code), includes this factor as a function of the fundamental period. The shape of this function for different kinds of soils is shown on Fig. 1. It is based on the non linear analysis of SDFS subjected to a set of several ground motion records (Riddell, Hidalgo and Cruz 1989). The code makes no difference between moment resistant frames concentrically braced and eccentrically braced frames, which means the same spectrum reduction factor for all of them, i.e., the same seismic design loading.

R 16
12 8 4 0 0.00 Rock Sand Soft Soil

1.00

2.00

3.00

T 4.00 [sec]

Fig. 1 - Spectrum Reduction Factor according to Chilean Code Steel structural systems exhibit important differences in strength, energy dissipation properties of the components and failure mode. Those differences cannot be explained only from a standpoint of a SDFS approach, the hysteretic properties of the different components and the higher mode effects require to study the seismic demand on the whole structure. This paper presents the non linear seismic analyses of a set of different types of steel buildings, designed according to the Chilean code and practice. There were included moment frames and braced frames from 6 up 32 stories. The earthquake demand was obtained from non-linear analyses, using some representative earthquake records, such as Chile 1985, Mexico 1985, Northridge 1994, and Kobe 1995. The building capacities were obtained by two methods: push over analysis and dynamic non linear analysis under scaled records.

1. Earthquake Requirements. A set of buildings from 6 to 32 stories with the same plan were designed according to the Chilean NCh 433 Seismic Code. A typical plan and elevation for eccentrically braced frames is shown on figure 1, similar geometries were considered in the case of moment and concentrically braced frames. The nonlinear response under ground motions records occurred since 1985 in Chile, Mexico, United States, and Japan were obtained for all the buildings structures.

9.0 m

4.5 m

3
4 a 32=29.28

Elevation D

3 a 7.32 = 21.96 m

3 Plan

Fig. 2 -Typical Plan and Elevation of the Buildings Selected The records selected represent some important earthquake events of the last two decades. Observed damages as well as other information have been extensively reported. Table 1 shows some of the records properties.

Table 1 Properties of the Records EARTHQUAKE Chile 3/3/85 Chile 3/3/85 Mxico 19/9/85 Northridge 17/1/94 Japn 17/1/95 RECORD Via del Mar (S20W) Llo-lleo (N10E) SCT (E00W) Sylmar (N00E) Kobe (N00E) SOIL Alluvium Sand Soft Clay Stiff Soil Alluvium EP. DIST. (KM) 84 45 400 15 40 DURAT. (SEC) 116 116 62 60 40 RICHTER MAGNIT. 7.8 7.8 8.1 6.8 6.9 PEAK (g) 0.36 0.67 0.17 0.84 0.84

2. Seismic Building Capacity. Every structural typology is more or less sensitive to different lateral forces, so, there is not a unique approach to determine the building capacity. Two methods were applied in this study: Pushover Analysis: by applying a lateral incremental loading. There were used a uniform lateral load distribution and a triangular lateral load. It is accepted that the first one represents a lower bound to the dynamic forces and the second one represents an upper bound [Bertero, 1984]. Non Linear Dynamic Analysis: by applying Scaled Ground Motions. This second method intends to include the dynamic nature of the earthquake load in the estimation of the building capacity. Even though scaled records are not the real ones, the properties of the probable ground motion are not known, so this alternative approach seems to be reasonable.

OUTLINE OF THE BUILDING ANALYSIS

All the buildings were designed according to the Chilean practice, based in the Chilean code NCh 427, Structural Design of Steel Buildings and NCh 433, Earthquake Resistant Design of Buildings, the buildings sample includes from 4 up to 32 story buildings. The non linear analysis of the buildings frame was performed by Ruaumoko program [Carr, 1996] in the case of braced frames, and Drain-2D Program [Kanahn and Powell, 1973] for moment resistant frames. The Remennikovs model [Remennikov and Walpole, 1995] included in Ruaumoko program was used for the bracing elements and the Ricles model [Ricles and Popov, 1989] for the links. The ultimate limit rotation adopted, for moment resistant frames and links was 0.01 rad and 0.07 rad respectively. The maximum axial deformation in the case of braces obtained from different authors is shown on Figure 3.

6,0%
Axial Deformation

5,0% 4,0% 3,0% 2,0% 1,0% 0,0% 0,0 50,0 100,0 150,0

Slenderness

Fig 3.-Ultimate Axial Deformation [Remennikov, 1995; Balendra, 1995 and Martinelli 1998].

1.

Moment Frame Buildings.

A typical pushover curve is shown on figure 4, the red curve was obtained by a uniform incremental loading distribution and the blue curve was obtained by a triangular incremental loading distribution, on the same curves it is indicated the position of the beam yielding and column hinges. Pushover analysis was stopped when a collapse mechanism occurred.
24

18

Uniform Load Hinges at Columns

100Q/W

Triangular Load

12 Collapse Mechanism

6 First Yielding at Beams

0 0 1 2 3 4 5 6

Drift

Fig. 4 Pushover curves for a six story moment frame A summary of the base shear capacity determined from push-over analyses is shown on Fig. 5, the analysis was performed until a collapse mechanism occurred or the rotation exceeded 0.01rad. The same figure shows the earthquake demands obtained from non linear dynamic analyses under different ground motions, as well as the design base shear according to the NCh 433 Chilean Code.

Q/W
0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0.5 1.0 1.5 2.0
NCh VdM Llo Nor Kob Mex Rec Tri

T [sec]

Fig. 5 - Seismic Demand and Building Capacity Base Shear Ratio

The earthquake demand is much larger than the code design forces and the building capacities are most of the times larger than the earthquake demands. The worse condition to determine the capacity from a push over approach is the analysis under triangular incremental load, especially for the smaller buildings. Northridge ground motion produces a collapse mechanism to the 8 story building and Kobe produces the collapse for the 4 and 8 story buildings. In the case of the Mexico earthquake, the demand exceeded the building capacity only for the twenty story building, when a triangular lateral load was applied during the push over analysis. Even though for the smaller buildings the demand of the Chilean earthquake became larger than the pushover capacity determined with an inverted triangle lateral load, those buildings did not form a collapse mechanism. It would seem that the Chilean Code provisions are safe for the seismic design of steel buildings. Fig. 6 shows the Base Shear Capacity divided into the Base Shear Demand. On Fig. 6a was determined by applying a rectangular lateral load distribution during the pushover capacity analysis, fig. 6b shows the same ratio when an inverted triangle lateral load was applied. Those ratios near 1.0 indicate a near failure behavior; this condition is more likely to occur when the pushover analysis is performed by applying a triangular load. Kobe and Northridge earthquakes are the worst demand condition for the steel building, the largest plastic rotation obtained with the Northridge earthquake record was 0.01 rad (for the 8 story building) and the largest obtained with Kobe earthquake was 0.013 (for the 4 story building). For the Chilean ground motion records, the largest plastic rotation obtained was 0.005 rad (for the 20 story building with Via del Mar ground motion and for the 8 story building with Llo-lleo).
3.00 2.50
VdM Llo Nor Kob Mex

Qc/Qd

2.00 1.50 1.00 0.50 0.0

T [sec]
0.5 1.0 1.5 2.0 2.5 Fig. 6a Base Shear Capacity/Base Shear Demand (Rect. load)
VdM Llo Nor Kob Mex

2,50 2,00

Qc/Qd

1,50 1,00 0,50 0,00 0,0 0,5 1,0 1,5 2,0

T [sec]
2,5
Fig. 6b Base Shear Capacity/Base Shear Demand (Triang. load)

Even though different earthquakes can demand the same base shear to a building, the damage of the structural and non structural components is essentially related to the lateral displacements and the earthquake plastic hinge requirements. Fig. 7 shows the ratio of the largest demanded roof displacements respect to the design roof

displacement, it can be seen that the actual displacements can be from 4 up to 8 times the lateral displacement obtained from an elastic analysis. For the Chilean earthquakes, the relationship is around 4, which is larger than the amplification displacement factor specified by the Chilean Code. However, the building displacement capacities (brown and green lines) are larger than the Displacement Demands, except in the case of some of the smaller buildings where a collapse mechanism was formed. Such a larger displacement amplification factor could become important when it is necessary to define a building separation, because the standard elastic analysis is not able to predict the real displacement. The Chilean Code requires the designer to amplify the elastic displacements by R/3, which in some cases is less than half of the actual lateral displacement. Wrong displacement estimation can produce hammering between adjacent buildings and sometimes the collapse of any or both of them. It can be observed that the demanded displacements of Northridge and Kobe earthquakes are larger than the demanded displacement of the Chilean earthquakes. Displacement Capacities obtained from push over analysis are the largest.

16.0
VdM Llo Nor Kob Mex Rec Tri

12.0

Du/Dd

8.0

4.0

0.0 0.5 1.0 1.5 2.0

T [sec]
Fig. 7 Roof Displacement Demand/Design Roof Displacement According to the results, the largest plastic rotation demanded by the Chilean earthquakes was 0.005 rad. As a consequence of the weak beam - strong column design criteria, the non linear behavior was concentrated at the end of the beams. In spite of the number of connections who reached the plastic range with Llo-lleo record (more than 80% for 4 to 12 story buildings and 50% for the taller buildings); none of the buildings reached a collapse mechanism. The reliability of the moment frame buildings can be improved by using special connections, but the results seem to indicate that there is no need to improve the plastic rotation capacity for the Chilean ground motions conditions. The Northridge earthquake maximum rotation demand for the 8 story building was 0.01 rad, and it finally collapsed forming an unstable mechanism. This plastic rotation is practically twice larger than the Llo-lleo record, which is the most severe of the Chilean 1985 records. (Both records have almost the same peak ground acceleration). The buildings sample under Kobe earthquake exhibit a similar behavior than the Northridge earthquake, with the larger damage concentrated on the smaller buildings. Mexico earthquake record, due to the characteristics of the soil, has lower frequency content, so it affected mostly the taller buildings. The maximum plastic rotation demanded was 0.009 rad for the 16 story building.

2. Braced Frames. The roof displacement for two different eccentric frames is shown on Fig.8 and Fig. 9. It can be observed, just like the moment frame analysis, that the effect of the Mexican Earthquake on the short to medium period buildings is not significant. Displacement demands are low and the structure remains in the elastic range. Llo-lleo and Via del Mar ground motion records demand non linear roof displacements around 10 cm, the non linear behavior concentrates on the braces.

The Northridge ground motion record produces a 15 cm displacement at the roof and the nonlinear behavior took place in the first story members, including braces, some beams, and columns. In this case, after the earthquake, the structure ends with a permanent 2.4 cm roof displacement. Kobe ground motion produces the largest roof displacement, twenty seven cm after 5.4 seconds. Although the inelastic seismic activity was important under Kobe record, it was not enough to reach some limit condition of the frame, either in the ultimate member displacements or the collapse mechanism of the frame.

0 ,3

1 K obe 2 L L o lleo 3 M xico 1 4 N o rth rid g e 5 V i a

0 ,2

0 ,1 Displacement [m]

0 4 -0 ,1 2 5

-0 ,2

-0 ,3 0 10 20 30 T im e [s] 40 50 60

Fig. 8 - Roof Displacement for the eight stories eccentric building Figure 9 shows the influence of the long period waves of the Mexican Earthquake on the tallest building of the sample. The largest displacement was 68 cm, it occurred after forty seconds from the record beginning. A summary of the largest roof displacement demands for the eccentric buildings is shown on Table 2.
1 ,2 1 Kobe 2 L L o lleo 0 ,8 3 Displacement [m] 0 ,4 4 2 0 1 -0 ,4 3 M xico 4 N o rth rid ge 5 V i a

-0 ,8

-1 ,2 0 10 20 30 T im e [s] 40 50 60

Fig. 9 - Roof Displacement for the thirty two stories eccentric building

The Northridge record produces the sudden collapse of the four story building after 4.8 seconds shaking; the collapse occurrence is near to an important peak of the ground motion record. This earthquake produced the buckling of the first story braces and plastic rotation in some joints. The relationship between base shear and roof displacement as well as other parameters were also obtained. The curves were not included in the paper. A summary of the largest base shear obtained are shown in table 3. More detailed results can be found in the original report [Uribe, 1999]. Table 2 - Roof Displacement Demands. Chilean Code Displacements [cm] Max. Separat. Kobe Llolleo Mxico Northridge [cm] 0.91 13.3 8.0 0.4 25.1 3.31 27.4 9.8 2.1 14.9 9.95 26.6 13.9 10.5 30.3 23.30 40.1 18.3 44.6 42.1 58.58 18.5 23.4 67.7 52.7

Via 0.7 12.2 12.3 15.5 12.3

4 Stories 8 Stories 16 Stories 24 Stories 32 Stories

Figure 10 shows the relationship between roof displacement and base shear for a thirty two story building. It was obtained from a pushover analysis. Triangular and rectangular static lateral loading results are shown, as well as bilinear equivalent curves. First yielding and ultimate lateral deflection are shown for both cases.
700 1 2 3 4 1 T r i a n g u la r la t e r a l l o a d i n g R e c ta n g u l a r l a t e r a l lo a d in g B i l in e a r a p r o x im a t io n t o 1 B i l in e a r a p r o x im a t io n t o 2

600 5 4 9 ,6 500 4 6 8 ,7 Base Shear [T] 400 3 3 2 ,5 300 2 9 6 ,1 200 3 4

F i r s t s i g n if i c a n t y ie l d in g

100 0 ,4 0 6 0 0 ,3 1 1 0 ,5 1 1 ,5 2 2 ,5 2 ,8 3 3 ,1 3 ,5

R o o f D i s p la c e m e n t [ m ]

Fig. 10 - Roof Displacement vs. Base Shear for the thirty two stories eccentric building The largest base shears obtained for the entire building sample and loadings are shown on Table 3. Similar results were obtained by using scale ground motions and dynamic non linear by performing non linear analysis; the records were scaled until the collapse mechanism occurrence. Lateral displacement capacities obtained are a little smaller than those obtained by using pushover analysis, except the case of the four stories building, where the largest value was obtained by applying the non linear dynamic analysis and scaled ground motion records. On the other side, base shear obtained with this second approach is a little higher than the first one.

Table 3 - Base Shears Requirements

Building Loading Kobe Llolleo Mxico Northridge Via D. Uniforme D. Triangular 4 Stories Qmx [T] 159.0 147.4 18.00 149.0 36.40 157.8 150.1 %P 40.1 37.2 4.53 37.5 9.17 39.7 37.8 8 Stories Qmx [T] 241.8 221.1 86.3 236.9 224.8 264.6 238.6 %P 32.6 29.8 11.6 31.9 30.3 35.7 32.2 16 Stories Qmx [T] 258.0 137.0 133.0 267.9 130.0 291.8 244.0 %P 15.9 8.43 8.18 16.5 8.00 17.9 15.0 24 Stories Qmx [T] 234.1 162.0 223.5 269.2 96.6 302.0 211.1 %P 11.2 7.50 10.3 12.5 4.47 13.9 9.77 32 Stories Qmx [T] 139.5 147.4 399.2 354.4 87.8 549.6 468.7 %P 4.07 4.30 11.7 10.3 2.57 16.1 13.7

STRENGTH REDUCTION FACTORS The Strength Reduction Factor depends basically of the structural system and the nature of the earthquake motion. It allows us to perform a linear analysis of the structures with reduced design forces; this approach assumes that the structure is able to behave in a stable non linear way when a destructive earthquake occurs. The comparison of the earthquake requirements and the building capacities allows us to estimate the response modification factors. The approach given by equation (1) is shown on Fig. 11 (Uang, 1991). According to the figure, the reduction factor has two components a first factor associated to the ductility ratio of the structure and a second factor dealing with the over-strength of the structure.

R=

Qe Q y * = R * Q y Qs

(1)

Where: Qe: Base Shear from Elastic Structural Behavior Qy: Base Shear at Yield Frame Level Qs: Base Shear at First Significant Yielding Occurrence R: Reduction Factor due to Ductility : Reduction Factor due to Overstrength

Base Shear
Qe Elastic Behavior

Actual Behavior Qy Qs Bilinear Model max

Displacement

Fig. 11 Strength Reduction Factor Determination

The following curves show some comparisons intending to look for a better understanding of the relationships between earthquake demands and building capacities. Figures 12 and 13 show the ratio of the base shear according to the elastic dynamic analysis and the base shear capacity obtained from a pushover analysis. In the case of the most severe records the ratio is around 3, for less severe earthquakes the ratio can become smaller than one, the black curve is the R factor according to the NCh 433 Chilean code.

Qed/Qc
9.00 8.00 7.00 6.00 5.00 4.00 3.00 2.00 1.00 0.00 0.5 1.0 1.5 2.0
NCh 433 Rmo VdM Llo Nor Kob Mex

T [sec]

Fig. 12 Base Shear Elastic Demand/Base Shear Capacity for Moment Frames (Rectangular lateral load)

Qed/Qc
9.00 8.00 7.00 6.00 5.00 4.00 3.00 2.00 1.00 0.00 0.5 1.0 1.5 2.0
NCh 433 Rmo VdM Llo Nor Kob Mex

T [sec]

Fig. 13 Base Shear Elastic Demand/Base Shear Capacity for Moment Frames (Triangular lateral load)

The code also requires a minimum limit to the base design shear, which means that the lateral load has to be increased when the base shear is smaller than a minimum. It follows that this modification of the base shear is equivalent to reduce the Strength Reduction Factor; the modified factor is the blue line on figures 12 and 13. It can be seen that the trend of the base shear ratios for actual frame building does not behave in the same way than the reduction factor of the Chilean code. If we modify the R function according to the code shear bound, then the trend is slightly closer to the actual behavior. The ratio of the base shear obtained from the elastic analysis under the code elastic design spectrum (without reduction) and the capacity obtained from pushover analysis is shown on Fig. 14. The ratio is between 2 and 3.

Qed/Qc
9.00 8.00 7.00 6.00 5.00 4.00 3.00 2.00 1.00 0.00 0.5 1.0 1.5 2.0
NCh 433 Rmo Rect Tri

T [sec]

Fig. 14 Base Shear Elastic Spectrum/Base Shear Capacity (Moment Frames) Figures 15 and 16 were obtained for eccentrically braced buildings, they show the ratio of the Base Shear Elastic Earthquake Demand (dynamic analysis with records) to the Base Shear Capacity obtained from a pushover analysis under rectangular and triangular lateral load respectively. The energy dissipation capacity of the link elements improve the energy dissipation capacity of the braced structure compared to the moment resistant frame buildings. The base shear ratio becomes closer to the modified reduction factor of the Chilean code. Fig. 17 shows the ratio of the Base Shear Elastic Demand and the Base Shear Capacity obtained from the non linear analyses with scaled ground motion records. The Chilean earthquakes ratios are between one and three and Northridge and Kobe Earthquake lies between three and five. A trial to obtain the R factor from equation 1 was done. It requires an estimation of the component factors, i.e. the R ductility factor and overstrength factor. R factor is defined as the base shear reached by the elastic structure divided by the base shear developed by the same structure when its behavior is inelastic. The comparison has to be done for the same loading condition, i.e. the same earthquake record has to be applied to the structure who behaves in two different ways. In order to evaluate this factor it was necessary to perform two analyses under a selected earthquake, able to take the structure into the non linear zone.

Depending on the characteristics of the structure such an earthquake may be different for different structures, for instance, Northridge or Kobe are the worst condition for the smaller building because those records demand them a strong non linear behavior, however, for the tallest building, Mexico record is the worst condition.

Qed/Qc
10.00 9.00 8.00 7.00 6.00 5.00 4.00 3.00 2.00 1.00 0.00 0.30
NCh 433 Rmo VdM Llo Nor Kob Mex

T [sec]
0.70 1.10 1.50 1.90 2.30 2.70

Fig. 15 Base Shear Elastic Analysis/Base Shear Capacity (Rectangular Load, Eccentrically Braced Frames)

Qed/Qc
10.00 9.00 8.00 7.00 6.00 5.00 4.00 3.00 2.00 1.00 0.00 0.30
NCh 433 Rmo VdM Llo Nor Kob Mex

T [sec]
0.70 1.10 1.50 1.90 2.30 2.70

Fig. 16 Base Shear Elastic Analysis/Base Shear Capacity (Triangular Load, Eccentrically Braced Frames)

Qed/Qc
10.00 9.00 8.00 7.00 6.00 5.00 4.00 3.00 2.00 1.00 0.00 0.30
NCh 433 Rmo VdM Llo Nor Kob Mex

T [sec]
0.70 1.10 1.50 1.90 2.30 2.70

Fig. 17 Base Shear Elastic Dynamic Analysis/Base Shear Scaled Record Non Linear Based on the aforementioned assumption there were selected the largest base shear ratios. Different ground motion records were applied to different buildings and the only requirement was to apply the same ground motion record for both analyses (linear and non linear) of the same building. This upper bound of the shear ratios should be ductility contribution to the Strength Reduction Factors. In order to complete the R factor, it was necessary to estimate the factor, due to overstrength. This factor was obtained as a result of the pushover analysis according to the definition (equation 1). The overstrength factors for both rectangular and triangular lateral incremental loading obtained from pushover capacity analyses are shown on Table 4. For the purpose of having a single value for every building, it was applied the average from both lateral loading. Table 4 Overstrength Factor Rectangular Lateral Load Triangular Lateral Load Qy [T] Qs [T] Qy [T] Qs [T] ] ] =Qy/Qs ] ] =Qy/Qs 153.9 127.8 1.20 144.5 118.1 1.22 260.7 187.2 1.39 234.1 156.5 1.50 283.9 196.7 1.44 236.4 166.9 1.42 287.8 187.8 1.53 221.2 163.2 1.35 546.6 332.5 1.64 459.7 296.1 1.55

Building Stories 4 8 16 24 32

The different factors are shown on Fig. 18, the green line is the overstrength factor from Table 4, the red line is the ductility factor and the blue line is the Strength Reduction Factor obtained as the product of both of them. It can be found that the blue line is a reasonable agreement to the Chilean NCh 433 Seismic Code in black color. The average difference is 13% and for most of the buildings the Code equation is smaller, which is on the safe side, except for the tallest building.

Qed/Qc
10.00 9.00 8.00 7.00 6.00 5.00 4.00 3.00 2.00 1.00 0.00 0.30
NCh 433 Ductility Overstrength R

T [sec]
0.70 1.10 1.50 1.90 2.30 2.70

Fig. 18 Strength Reduction Factor obtained from Dynamic Analyses The NCh 433 code equation is based on the non linear analyses of SDFS under different ground motions, so it does not include redundancy, overstrength or higher mode effects, because those aspects require a multi degree of freedom model and some additional hypotheses. The results show that, in spite of the elementary model applied by the NCh 433 Chilean Code equation, it provides a good Strength Reduction Factor for Steel Structures. CONCLUDING REMARKS 1. Capacity of Steel Buildings designed according to the Chilean Code and practice is larger than the demand of the Chilean earthquakes. The analyses by using Chilean earthquakes, in some cases show a slight non linear behavior of the structures, but not significant damage. It would seem that NCh 433 code is safe, mainly due to the nature of the Chilean earthquakes. The analyses suggest the necessity to review the minimum separation between adjacent buildings requested by the codes, it could be easily exceeded during a severe earthquake occurrence. Evidence on building collapses due to hammering between structures also recommends such a reviewing. Even though, the reduction factor approach, it is based on a very simple SDFS model, it provides a good agreement to the lateral design forces obtained from a Demand Capacity analysis of the structures. Redundancy, overstrength, structure deterioration, higher mode effects and experience requires being explicit included in the seismic codes. Some of them can be obtained from theoretical analyses; some others require laboratory tests, field measurements, and experience. The previous analysis shows only some aspects of this widespread approach of seismic forces definition by applying a reduction factor to the elastic response spectrum. It is clear this is not a simple matter, many codes put in this bag all those things that are not well known, so the R factor becomes a kind of useful number, even though it is not well understood what it is being taking into account. Northridge and Kobe earthquakes produces large displacements of the buildings, sometimes permanent, with extensive damage, the smaller is the building the highest is the damage. Northridge record produces the sudden collapse of the four stories building after 4.8 sec shaking. The NCh 433 code is not appropriate for these types of earthquakes conditions. The upper and lower boundaries to the base shear of the Chilean code are on the right way. They improve the R factor and put some control levels to the spectrum.

2. 3. 4. 5.

6.

7.

ACKNOLEDGEMENTS This work was partially completed under the sponsorship of Universidad Tecnica Federico Santa Mara research project No 96.26.21. Roberto Uribe, Francisco Arriaza, Jose Gonzalez and Alejandro Cornejo contributions are greatly appreciated. All of them participate as research assistants during the development of the project.

REFERENCES Balendra, T., Lim, L., Liaw, Y., (1995), Program to Reduce the Earthquake Hazard of Steel Moment Frame Structures, Federal Emergency Management Agency, and Seismic Safety of Buildings. Bertero, V.V. (1986), Evaluation of Response Reduction Factors Recommended by ATC and SEAOC, Proc, 3rd U.S. Nat. Conf. Earthquake Engineering, Charleston, South Carolina, pp 1663-1673. Bertero V. (1984), State of the Art and Practice in Seismic Resistant of R/C Frame-Wall Structural Systems, 8th World Conference on Earthquake Engineering Proceedings, Vol 5, pp 613, San Francisco. Carr A., (1996), Ruaumoko Inelastic Analysis Program, Department of Civil Engineering, University of Canterbury, Christchurch, New Zealand. Elghadasmi, F. E. and Mohraz, B. (1987), Inelastic Earthquake Spectra, Earthquake Engineering and Structural Dynamics, Vol. 15, pp 91-104. Fischinger, F. and Fajfar, P.(1990), On the Response Modification Factors for Reinforced concrete Buildings, Proc. 4th U.S. Nat. Conf. Earthquake Engineering, Palm Springs, California, Vol. 2, pp 249-258. Hidalgo, P. A. and Arias, A., (1990), A new Chilean Code for Earthquake Resistant Design of Buildings, Proc. 4th U. S. Nat. Conference Earthquake Engineering, Palm Springs, California, Vol. 2, pp. 927-936. Instituto Nacional de Normalizacin, (1996), Seismic Design of Buildings, Norma NCh433.Of 96, Santiago, Chile. Kanahn A. E., Powell G.H., (1973), General Purpose Program for Inelastic Dynamic Response of Plane Structures, Report No. 73-6, Department of Civil Engineering, University of California, Berkeley California (in Spanish). Lai, S. P. and Biggs, J. M., (1980) Inelastic Response Spectra for Aseismic Building Design, J. Struct. Div., ASCE, Vol 106, No St6, pp 1295-1310. Martinelli, L., Mulas, M., Perotti,F., (1998), The Seismic Response of Concentrically Braced MomentResisting Steel Frames, Earthquake Engineering and Structural Dynamics, vol 12, 749-769. Miranda, E., (1993), Site-Dependent Strength Reduction Factors, Journal of Str. Engineering, ASCE, Vol.119, No 12. Miranda, E., (1997) Strength Reduction Factors in Performance Based Design, The EERC-CUREe Symposium in Honor of Vitelmo V. Bertero, EERC, University of California, Berkeley. Nassar, A. A. and Krawinkler, H., (1991), Seismic Demands for SDOF and MDOF Systems, Report No 95, The John Blume Earthquake Engineering Center, Stanford University, Stanford, California. Newmark, N. M. and Hall, W. J., (1973), Seismic Design Criteria for Nuclear Reactor Facilities, Report No 46, Building Practices for Disaster Mitigation, National Bureau of Standards, U.S. Department of Commerce, pp 209-236. Remennikov, A. and Walpole, W. , (1995) , Incremental Model for Predicting the Inelastic Hysteretic Behavior of Steel Bracing Members, Department of Civil Engineering, University of Canterbury, Christchurch, New Zealand. Ricles,J. and Popov, E., (1989) , Inelastic Link Element for EBF seismic analysis, Journal of Structural Engineering, ASCE, 120(2), (441-463). Ridell, R. and Newmark, N. M., (1979), Statistical Analysis of the Response of Nonlinear Systems Subjected to Earthquakes, Structural Research Series No 468, Dept. of civil Engineering, University of Illinois, Urbana. Ridell, R., Hidalgo, P., and Cruz, E., (1989), Response Modification Factors for Earthquake Resistant Design of Short Period Buildings, Earthquake Spectra, Vol. 5, No 3. Seneviratna, G., D., P., K. and Krawinkler, H., (1996), Modification of Seismic Demands for MDFS, 11th World Conference on Earthquake Engineering, CD-ROM paper No 2129, Acapulco, Mexico. Uang, C., M., (1991), Establishing R and Cd factors for Building Seismic Provisions, Journal of Structural Engineering, ASCE, vol. 117 No 1, pp19-28. Uribe, R., (1999), Response Modification Factors on K-Braced Steel Frames, Memoria para optar al ttulo de Ingeniero Civil, UTFSM, Valparaso, Chile (in Spanish). Vidic, T., Fajfar, P. and Fischinger, M, (1992), A Procedure for Determining Consistent Inelastic design Spectra, Proc. Workshop on Non Linear Seismic Analysis of RC Structures, Bled. Slovenia.