A New Combination Method Of Vehicle And Temperature Load Effects Based On Shm Data

With the growth of the service time, more and more long-span bridges sufferedstructural damage due to overload, so a new load effects combination method isneeded urgently during structural reliability design. Wide application of structuralhealth monitoring system which provides accurate response information of bridgesis fairly helpful to evaluate the probability characteristics of single load. However,the structure suffered a variety of loads which can be depicted as stochasticprocesses because their intensities vary in time and in space. To attain an accuratevalue of combination, load effects processes are modeled as Poisson processes orstationary binomial stochastic processes because of their goodness to fit actualphysical processes.Based on monitoring data of Nanjing3rd Yangtze River Bridge, this thesisdecouples vehicel and temperature load effects through moving average method andextracts sample of vehicel and temperature load effects. Applying generalized Paretodistribution and two weighted normal distribution to fit the empirical distributionfunction and frequency histogram, probability distribution function and probabilitydensity function of the load effects metioned above is derived. After modelingvehicle and temperature load effects as Poisson impulse process and rectangularpulse process respectively, atheoretical expression ofthe cumulative distributionfunction and probability distribution function ofthe maximum intensity of thecombination of above load processes during a reference period is deduced referingload coincidence method. For complex convolution integral is involved in thetheoretical expression, numerical integration is applied to calculate the curves of cdfin different design reference period. Then using Monte Carlo simulation method,probability density function is calculated from the curve of cdf. In the end, acomparison is investigated analytically between proposed method and two widelyused methods (e.g., non-conservative Turkstra rule and conservative Borges rule) tocertify the accuracy and applicability of this new method.The proposed method builds probability model of corresponding load effectusing monitoring data in a certain period. Through stochastic process combinationtheory, maximum intensity of load effects combination in different design referenceperiod is calculated. Furthermore, this method will make up the lack of monitoringdata and its predictability will become the basis for decision making on maintenanceand management of existing bridges.