A Study On Vehicle Load Effect Probability Distribution Model Base On Bayesian Principle

With the development of transportation industry in our country, the number of highwayBridges and vehicles are increase. While the assessments of load capacity for bridgecomponents become increasingly prominent. Vehicle load effect, which is the main variableload effect of highway bridges, reflects the vehicle load conditions of highway bridgesdirectly, therefore, it is reasonable that accurate determination of the vehicle load effect is ofgreat significance in the assessments of load capacity for bridge components.The vehicle load effect of highway Bridges can be described by probability models. Thevehicle load effect probability distribution models of current code is determined on the basisof vehicle platoon measured by four representatively national WIM measuring point in1990s.Thus they can not meet requirements of the assessments of load capacity when the timesections, regions and structure are being considered. Consequently, vehicle load effectprobability distribution model of existing bridges is studied in dissertation.the vehicle load effect probability distribution models of existing bridge can be dividedinto the Vehicle load effect section distribution and the probability distribution of maximumvehicle load effect during object serviceable life. The section distribution reflect loadconditions in at one point, hence, it needs to be updated as time goes on. However, theprobability distribution of maximum vehicle load effect during object serviceable life is theforecasting base on the section distribution.The main work of this dissertation is as followed.(1) The vehicle load effect samples has been obtained by influence surface moving loadmethod, which adopting vehicle platoon data measured by WIM measuring point inXingguang Bridge site area.(2) The maximum likelihood estimation(MLE) and Pearson chi-square fit test methodhas been used to the analysis of the Vehicle load effect section distribution.(3) The Vehicle load effect section distribution update strategy of Xingguang Bridge,including update method, update route, update cycle, etc. has been described. Then, abayesian approach of the section distribution updating, that base on a small vehicle load effectsample has been introduced. (4) Introduce the concept of object serviceable life. Then derive the probabilitydistribution of maximum vehicle load effect during object serviceable life, as well as thevehicle load effect characteristic values, from the section distribution on the basis ofstationary binomial process.The main conclusions of this dissertation is as followed.(1) The analysis results shows that, under both traffic states, the section distribution ofvehicle load effect of Xingguang Bridge obey lognormal distribution.(2) The Vehicle load effect section distribution update strategy of Xingguang Bridge canwell meeting the needs of the assessments of load capacity for Xingguang Bridge HealthMonitoring System. Meanwhile, it can also offers reference for vehicle load effect sectiondistribution updating of other existing bridges.(3) The result of bayesian approach of the section distribution updating shows that theposterior section distribution is more closely related to the prior section distribution, whichdetermined by more vehicle load effect samples.(4) The probability distribution of maximum vehicle load effect during object serviceablelife as well as the characteristic values of Vehicle load effect were derived from the sectiondistribution on the basis of stationary binomial process. The result shows that, since themarked difference of degree of effect samples dispersion, some components’ vehicle loadeffect characteristic values under normal traffic state is slightly bigger than that of densetraffic state.