Sunteți pe pagina 1din 5

Live Load Distribution for Assessment of Highway Bridges in American and European Codes

Janusz Hoowaty, West Pomeranian University of Technology in Szczecin, Poland. Contact: Janusz.Holowaty@zut.edu.pl
DOI: 10.2749/101686612X13363929518018

Abstract
Bridge capacities are assessed by different methods which differ in accuracy. The level of sophistication depends on the type of existing structure and the criteria of assessment required. However, in most cases, simplified models and the load distribution methods are sufficient and safe. In this paper, the live load distribution methods adopted for the analysis of existing bridges are presented. British and American distribution factors (DF) for the assessment of the structural adequacy of bridge structures in service are discussed. Also, in Poland a simple method of live load distribution has been adopted for load rating of existing road bridges. Examples of DF methods for analysing highway bridge decks are presented. The results are compared with the simplistic distributions and the accurate computer method. Keywords: existing bridges; codes for assessment; load distribution; distribution factors; bridge rating.

(a)

(b)

Fig. 1: Five-girder concrete bridge. (a) Side view and (b) bottom view

Introduction
The live load transverse distribution methods, based on characterising parameters, were predominantly used in Europe1,2 and America for the design and assessment of bridge decks prior to the era of the modern computer methods. In the USA and Canada, the methods were developed according to the DF technique and codified. In these countries, they are still the basic procedures for a global analysis of simply supported square bridges. In Europe, the DF methods were not introduced to designing as they seemed to be too simple and more conservative than computer-based accurate analyses. The computer methods became very popular and accessible with the quick development of microcomputers and software.2,3 However, the simplified methods of bridge analysis remain very useful and essential for preliminary design and checking calculation as they make bridge analysis simple and quick.
574 Technical Report

With the introduction of the road bridge replacement, rehabilitation and strengthening programs in a number of countries, bridge assessment codes need to be introduced or modified. It was recognised that the use of design standards for assessment would result in bridges failing the assessment unnecessarily. In the UK, the procedures for bridge assessment were re-evaluated and new rules were established, with the departmental standards and additional guidance issued within the design manual for roads and bridges.4,5 In the USA, the AASHTO LRFD Specifications were introduced for the design of all new bridges from 2008. The AASHTO Standard Specifications have been applied to the maintenance and rehabilitation of existing structures since then.6 In Poland, new set of instructions for the assessment of existing bridge capacities was published.7 The simplified load distribution methods have been adopted in structural analysis for the first level of assessment and checking of bridge decks.

t trucks at the end of the century. In some European countries, gross limits are even higher. In bridge codes it was reflected by heavier traffic models. Old bridges require to be strengthened or replaced not only because of heavier vehicles but also because of aging and damages. Many old bridges were assessed to be substandard with lower weight limits. However, the methods used in the past to calculate bridges were necessarily conservative to provide margins of safety required in service life. Now the margin of safety can be reduced, using modern numerical methods and the real capacities can be assessed properly.

Levels of Assessment
The origins of modern bridge revalidation programmes were the new maximum weight limits for lorries introduced by new construction and vehicle use regulations, increased traffic loading and the deterioration of older bridges. The maximum weight limit for lorries was raised several times and is now 44 t under the EU directives for international transport road vehicles. The levels of bridge assessment in stages of increasing complexity were established.5 The bridge capacity is to be determined with minimum effort and cost, with a more refined level of assessment accepted when a lower level has failed. Generally, the suggested levels of assessment are: Level 1: Assessment, using simple analysis which gives a conservative

Needs for Assessment and Load Rating


Highway infrastructure protection and safety is usually considered while determining the weight limits for trucks and the costs for construction and maintenance to serve the present and future traffic. Increased axle weights and permitted gross mass of vehicles were significant during the last century, from 12 t vehicles at the beginning of the 20th century to 44

Structural Engineering International 4/2012

estimate of load capacity. All the partial safety factors from the codes are used. Level 2: The level incorporates more rened methods of analysis. They include grillage analogy and nite element computer analysis. Levels 3, 4 and 5: More sophisticated and use bridge-specic live loading, nominal values of loads or a reliability analysis of specic bridge structures. The preferred methods of analysis for the assessment of different types of bridge decks are refined. The British guidance4 uses a simplified method of DF at level 1 of assessment. It is advisable to assess a bridge deck by the methods used in the design, introducing the same code. When there are no design drawings or calculations, the first level of assessment should be undertaken by checking the structural identity of the structure. Simplified methods of analysis should be conservative to provide a sufficient safety factor. If the structure is adequate at the first stage, then no further analysis would be required. Using codified simple procedures, all the rules and requirements in the codes should be followed.

Distribution of Wheel loads in Longitudinal Beams of the code.6 For external beams, the traffic load is determined by applying to the beam the reaction of wheel loads from a simply supported slab (lever rule). If the beam spacing S is higher than in Table 3.23.1,6 the wheel load distribution is obtained by assuming the slab to act as a simple span between beams. In concrete beams, two ways of calculating wheel load distribution are possible: according to the table and according to the lever rule procedure for external girders. British assessment code (GB) The assessment code BD 21/01 and the associated advice note BA 16/974 are included in the design manual for roads and bridges as separate sheets. They were amended several times, before the commencement of a 15 year bridge renovation programme for trunk road bridges in the United Kingdom and during the implementation of the program. The live loading assessment is 40 (44) t and the restricted traffic levels cover a range of vehicles from 26 t reduced to 3 t and include fire engines. The code uses the reduction factors K for HA loading in estimating the loading categories. Factors K are given in graphs according to traffic flow and the road surface condition. The traffic flow is described as high (H), medium (M) and low (L). Road surface categories are good (g) and poor (p). This gives a total of six categories of bridges which are referred to as Hg, Mg, Lg, Hp, Mp and Lp. Graphs of load distributions for the assessment of girder bridges are included in advice note BA 16/97.4 They are used for estimating the loads carried by internal and external girders. The proportion factors KL are only intended for use with the loading specified for assessment, but can be
(a)

used for determining both moments and shears. The nominal live load bending moment in a girder is obtained by multiplying the gross moment with the appropriate proportion factor KL. The gross moment in a girder under the traffic lane is the effect of the live load from one notional lane of width aL = 2,5 m.

Simplistic Live Load Distribution Methods


Elementary load distribution methods like the lever rule or that of eccentric compression do not take into account either real rigidities of bridge deck members or types of live loads. The methods can be developed to analyse selected types of bridge decks to assure that the calculated effects of load distribution on bridge girders are approximately equal to the real behaviour of the bridge deck. The lever rule uses the simple static for wheel load distribution on girders or stringers. It is the simplest method but usually overestimates load effects and sometimes gives incorrect results. A more detailed analysis which takes into account the rigidities of bridge deck members shows that the lever rule overestimates load distribution even in non-composite or wood bridge decks. The second elementary load distribution method for comparison is eccentric compression, also known as the rigidcrossbeam method. This method, in its basic form, also ignores the rigidities of the decks but assumes that the cross section of the deck is braced sufficiently and has the cross beam at midspan, which is of infinitely high rigidity. The live load distribution is then calculated according to the eccentric compression rule (simple static). The eccentric compression procedure was adapted in the Polish instructions for service load assessment and load rating of existing bridges.7
(b)

Code-Based Methods for Rating of Existing Bridges


The live load DF methods introduced in the American and British codes for assessment are simple techniques which are based on many years of application or computer-based calculations and load tests.4 The traditional S-over formulas in Standard AASHTO Specifications have been in use since the 1930s. They are very simple, yet conservative but well accepted by bridge engineers. Standard AASHTO Specifications (US) The DF method is used for analysing typical square bridges. The DFs are described for different types of simply supported decks. The factors and procedures of analysis are valid only for standard traffic loads given in the specifications, e.g. H and HS loadings.6 In calculating bending moments in internal longitudinal beams and stringers, the transverse distribution is determined by applying to the beam the fraction of wheel load (both front and rear) according to Table 3.23.1

Fig. 2: Six-girder concrete bridge. (a) Side view and (b) bottom view Technical Report 575

Structural Engineering International 4/2012

Bridges and Numerical Method Used in Comparisons


Bridge Deck Considered Two reinforced concrete bridge decks are used for comparison of live load distributions for assessment. The design loading was between 20 and 30 t. They carried two lanes of traffic. The bridges were substandard and were assessed at the end of the 1990s for a road renovation program. The first bridge is a five girder bridge with a 12,6 m clear span. A view of the bridge is shown in Fig. 1. The deck was constructed as a skew and simply supported, with the girder spacing S = 2,2 m (Fig. 3a). The carriageway width was 8,10 m. The bridge was due for replacement, as it had low capacity. The second bridge is a six girder bridge with a 7,5 m clear span (Fig. 2). The deck was constructed square and simply supported, with girder spacing S = 1,6 m (Fig. 3b). The carriageway width was 7,50 m. The bridge was assessed following its low capacity and for determining the reserves of strength required for another few years service. The time of service had to be prolonged owing to changes in design alignments of the road and delays in financing. Numerical Models of Bridge Decks The grillage analogy was applied for modelling the bridge decks as the most common method used in bridge analysis. The grillage method is a well-established and refined technique2,3 and has a long history in technical literature. The cross sections, equivalent grillage members and adopted grillage meshes of the bridge decks are presented in Fig. 3. The models are rigid grillages (without hinges). The longitudinal grillage members are placed coincident with the centre lines of the girders as are the cross beams. The transverse members model cross-beams at supports and at midspan, and also slab at quarter lines (L/4) (Fig. 3b). In numerical analyses, the classical transverse load distribution approach was applied.1,8 The distribution coefficients k are computed by numerical programs for each bridge deck (Fig. 3). The values of k are given under each girder and they form influence lines (numerical model in Figs. 4 and 5). The sum of k under truck wheels (Fig. 4 green lines) gives DF. The same is for HA loading (Fig. 5) but area w is used because there is no wheels. Adopting
576 Technical Report

(a)
10,90 m s Grillage members

(b)
9,50 m s

47 8,50 m

Fig. 3: Cross-sections and grillage meshes. (a) Five-girder bridge and (b) six-girder bridge (Units in m)
10,90 m

s = 2,20 m

13,50 m

(a)

US

0,5 1,0

DF DF

= 0,846 = 1,163

DF

= 0,564

DFUS = 1,202

(b)

US

0,5 1,0

DF DF

= 1,051 = 0,982

DF

= 1,345

DFUS = 1,202

(c)

P18

P18 = 71,2 kN

1,83 m

Influence line according to: Lever rule Eccentric compression Numerical model American code (AASHTO) US

Fig. 4: Load distribution in five-girder concrete bridge. (a) External girder, (b) internal girder and (c) AASHTO assessment vehicle (Units in m, DF values are dimensionless).

Structural Engineering International 4/2012

the classical deflection approach for numerical load distribution makes the computed distribution coefficients independent of the type of loading.

9,50 m

Comparisons of Distribution Methods for Assessment


The simple DFs in codes for assessment are suitable for the initial assessment of bridges at level 1.4,5 They are compared with the results of calculations made by very simple procedures. All the results are also compared with a more detailed analysis made with computer-aided grillage numerical method. The results are presented as influence lines of transverse load distribution for external and internal girders in two concrete bridges. Example 1 (US code)Five-girder bridge In this example, the AASHTO standard vehicle loading is used for comparison purposes. The external girder DF for the grillage model is DF = 0,052 + 0,125 + 0,219 + 0,450 = 0,846 (Fig. 4a). The simplified procedure gives the values: DF = 0,564 for the lever rule and DFn = 1,163 for eccentric compression. According to the Standard AASHTO Specifications6 (Table 3.23.1) the distribution factor is DFUS = S/1830 = 1,202 and it is the safe value. The lever DF is decidedly too small. For the internal girder, the DFs are respectively equal to (Fig. 4b): for the numerical model DF = 0,159 + 0,259 + 0,328 + 0,305 = 1,051; for the lever rule DF = 1,345; for eccentric compression DFn = 0,982 and for AASHTO DFUS = 1,202 again. The AASHTO factor is also safe. Example 2 (GB code)Six-girder bridge Type HA loading is used to compare nominal areas for the distribution of load. Nominal area w is the area of appropriate influence line of transverse load distribution under the HA loading. The load distribution area for the external girder by the grillage model is w = 1,034 (Fig. 5a). The simplified procedure gives the values: w = 0,632 for the lever rule and w n = 1,413 for eccentric compression. According to the British code,4 the DF is w GB = KL aL = 0,495 2,5 = 1,237. The value of load distribution according to the British code is the safe value. The

s = 1,60 m UDL (KEL) GB

(a)

0,5

= 1,034 = 1,413
GB

= 0,632 = 1,237

1,0

UDL (KEL)

GB

(b)

0,5

= 1,332 = 1,290
GB

= 1,590 = 1,337

1,0

UDL (KEL)

(c)
aL = 2,50 m

Influence line according to: Lever rule Eccentric compression Numerical model GB British code

Fig. 5: Load distribution in six-girder concrete bridge. (a) External girder, (b) internal girder and (c) HA loading for assessment (Units in m, w values in 1 m).

lever method gives too small value and eccentric compression gives too high value. For the internal girder, the distribution areas are respectively equal to (Fig. 5b): for the numerical model w = 1,332; for the lever rule w = 1,590; for eccentric compression w n = 1,290 and for the British code w GB = 1,337 for KL = 0,535. The British code procedure is safe again and the distribution is very close to that obtained by the bridge grillage model.

procedure are presented. The results of load distribution for two concrete bridges are compared and discussed. The DF methods presented are very different. The standard AASHTO procedures were used for many years for design purposes and now they are used for assessments. It is advisable to assess bridges by the code used for their design. The AASHTO DFs do not seem to be exact but after many years of application to design they are reliable and give the required margin of safety. The British assessment code was produced for a 15 year bridge renovation programme for assessment of old bridges. While assessing bridge structures, traffic flow and the condition of pavement are taken into account. The live loading assessment is divided into categories under the road vehicle regulations. The accuracy level of load distribution can be increased using individual numerical grillage analysis for computing DFs.8,9 In the area of numerical techniques, the distribution methods can be used at hand for preliminary or checking calculations. An evaluation analysis of abnormal road loads, which happen
Technical Report 577

Conclusions
The increased traffic and the problems with substandard bridges in many countries have highlighted the fact that more practical assessment methods have to be established. The rules of assessment analysis for bridges are changing and new approaches have been introduced by some road authorities, using also simplified procedures. The paper presents some simplified methods of live load distribution developed in the analysis of bridge decks. The simplified approaches for the assessment of bridge load capacities used in US, Great Britain and the international crude

Structural Engineering International 4/2012

occasionally, can be done quickly. Modern DFs are validated using numerical methods and load tests so they are reliable but still simple. The simplified procedures of load distribution are easy to use, which is a very important aspect for economic reasons and the time used. A great deal of time can be saved if preliminary assessments or calculations give good results and no further sophisticated calculations are needed.

Methods in Mechanics of the XI Polish Conference on Computer Methods in Mechanics. Cejko: Kielce, 1993; 331336. [4] BA 16/97. The Assessment of Highway Bridges and Structures. Design Manual for Roads and Bridges. HMSO: London, 1997; www. dft.gov.uk/ha/standards/dmrb. [5] Das PC. Development of bridge-specific assessment and strengthening criteria. In Safety of Bridges. Thomas Telford: London, 1997; 5357. [6] AASHTO. Standard Specifications for Highway Bridges, 17th ed. AASHTO: Washington DC, 2002. [7] The instruction for assessment of service load capacity of road bridges (in Polish). GDDKiA: Warszawa, 2004; (www.gddkia.gov.pl/a/6608/ materialy-pomocniczedo-pobrania). [8] Hoowaty J. Numerical Method for Live Load Distribution in Road Bridges. The Fourth International Conference on Structural Engineering, Mechanics and Computation. Cape Town, 2010; 425428.

[9] Hoowaty J. Comparison of Load Distribution for Assessment of Highway Bridges in American and European Codes. In Joint IABSE - fib Conference on Codes in Structural Engineering. Developments and Needs for International Practice. Dubrovnik, 2010; 11851192.

SEI Data Block Owner: GDDKiA o/Szczecin Client: GDDKiA o/Szczecin Structural checking and consultant: PPDM dr in z. J. Hoowaty, Szczecin Bridge deck surface: Bridge 1 (m): Bridge 2 (m): Service date: Bridge 1: Bridge 2: 150 83 1959 1948

References
[1] Bare R, Massonnet C. Analysis of Beam Grids and Orthotropic Plates by the GuyonMassonnet-Bare Method. Crosby Lockwood & Son Ltd, SNTL: London, Prague, 1968. [2] Hambly EC. Bridge Deck Behaviour, 2nd edn. E & FN Spon: London, 1991. [3] Hoowaty J. Numerical modelling of bridge decks in practical examples. In Computer

IABSE Publications
Digitised Archive
from 1929-1999

Did you know?


IABSE publications between 1929 - 1999 can be downloaded for FREE from the IABSE Archive on the internet - by everyone. Do full text searches in any part of the papers, or browse by title, author or year. Explore the Archive with over 80'000 pages and find a treasure of information on research, projects and IABSE.

www.iabse.org/IABSE/Publications/Archive

PS: Conference Papers published after 1999 (over 3'500) are available via Ingenta. Full Papers may be downloaded with a subscription or on a pay per view basis. www.ingentaconnect.com/content/iabse

578 Technical Report

Structural Engineering International 4/2012

S-ar putea să vă placă și