Vehicle Load And Temperature Effect Analysis Of Bridge Based On Extreme Strain Estimation

As the crucial factors affecting bridge safe operation,vehicle load and temperature are main loads acting on bridges in service.It is essential to estimate the maximum load effect of a bridge during the lifetime based on long-term monitoring data of vehicle load effect and temperature-induced response.A large volume of research work has been conducted to investigate the vehicle load effect of bridges and its probability distribution presently.Study on temperature mainly focuses on temperature field of bridge,while little attention has been paid to temperature-induced response of bridge.Moreover,the research on the combination effect of vehicle load and temperature on bridge is also insufficient.Based on the long-term monitoring data from December 2014 to July 2016 of Taiping Lake Bridge in Anhui Province,the strains due to vehicle load or temperature alone can be obtained by signal decomposition method.According to the extreme value statistical theory,the probability models of strain peaks are built.The extreme strains due to vehicle load,temperature and the combination effect of them during the remaining service period of the bridge are estimated.The estimates can be utilized for further bridge safety assessment and reliability analysis,and also possess reference significance for design load criterion of newly-built bridge in the same district.The main research contents of this dissertation are described as follows:Firstly,the parameters of General Pareto distribution are estimated by the simulated annealing algorithm and maximum likelihood estimation algorithm,respectively.The results show that the estimates by simulated annealing algorithm are more accurate than that by maximum likelihood estimation algorithm,while computing time of the former is hundreds or thousands of times longer than the latter.Therefore,when the sample size is small,it is suggested to use the simulated annealing algorithm to estimate parameters of General Pareto distribution.Contrarily,when the sample size is large,the maximum likelihood estimation algorithm is more suitable to achieve high efficiency.Secondly,both the analytical mode decomposition method and l1 trend filter are used to decompose the measured strain-time history into two parts,i.e.,strain due to vehicle load and temperature strain.Then a comparison is made of the two methods in terms of the decomposition effect and time cost.It is shown that each method can be utilized for strain signal during a short sampling period,while the analytical mode decomposition method is suggested for strain data collected over a long period of time.Thirdly,strain peaks due to vehicle load do not follow the five common distributions which are Normal distribution,Log-normal distribution,Weibull distribution,Gamma distribution and Gumbel distribution.It is demonstrated that both General Pareto distribution and the mixed distribution of one Weibull distribution and two Normal distributions can be considered as the probabilistic model for the tail data of strain peaks.Since the parameters of General Pareto distribution are fewer than the mixed distribution,General Pareto distribution is more suitable as the parent distribution of these tail data.Fourthly,according to the characteristics of long-term strain data of bridge,a new threshold selection method based on the relationship between threshold and expectation of 100-yearly extreme value is proposed.Then the method is applied to estimate the weekly extreme strain due to vehicle load on the Taiping Lake Bridge.Results show that the estimates on the basis of the threshold obtained by the proposed method are closer to the measured results than the commonly used methods.Furthermore,the extreme strains of the measuring points during the remaining service period of the bridge are estimated by the peak-over-threshold method.Finally,the relationship between the ambient temperature and the thermal strain of each measuring point is studied.The extreme strains due to temperature and the combination effect of vehicle load and temperature during the remaining service period of the bridge are estimated by the average conditional exceedance rate method.It is demonstrated that the expectations of the extreme stress distributions due to temperature are about 2-3 times larger than that due to vehicle load.The expectations of the extreme stresses due to the combination effect are close to that due to temperature alone.