Study On Live Load Model For Highway Bridges Based On The Characters Of Wheel Axles

Live load model plays an important role in many aspects of bridges, such as bridgedesign, analysis of bridges in-service, reliability and decision assessment of bridges. Moreand more overloaded vehicles drive on bridges, which imposes a severe situation for bridgemanagement. Less effort is put in the research on the live load model for bridges than otheraspects such as the material feature and the structural performance of bridges. Generally,gross weight of vehicles on bridges is treated as a stochastic process during bridge lifetime,and the distribution of the maximum load of it is deduced while researching into bridgelive load model, like the load model in the code for highway bridge design in China. Sucha treatment neglects the contribution of the characters of wheel axles, such as the amountof vehicles wheels, the distance of wheel axles and the apportionment of the gross weighton each wheel. As a result, such a model will not describe the real loads of vehicles.In this paper, the general method to research the live load models for bridges isdiscarded. Vehicles are separated into individual wheel axles on bridge. With theinformation of vehicles tested on bridges, chracters of wheel axles, like the axle distanceand the axle weight, are analysed, based on which the live load models of motorcade trafficand of renewal random field are proposed. The main wok and conclusion are stated asfollows.1) Preponderance of weigh-in-motion (WIM) by means of bending plate based on themeasurement of strains is demonstrated. A fiber-optic WIM system based on fiber Bragggratting technology is developed. Such a WIM system is employed for in-site testingaccompanied by a commercial one. Datum from them is compared by nonparametric test.As a result, the population distribution of such two systems is equal, which demonstratedthe reliability and accuracy of the developed WIM system.2) Based on a dreat deal of vehicle information tested in-site, distributions of theweight of the wheel axles on each lane are analysed and described by a3-peak normaldistribution by means of nonline least square fitting. Distributions of the weight of thefront wheel axle of all kinds of vehicle are analysed. Distributions of the ratio of the weightof back wheel axles to the front wheel axle are also analysed. Parameters of them are alsoobtained. 3) Based on the vehicles’ information, distributions of axle distance of each vehicleare analysed and alse described by a3-peak normal distribution. Distributions of the headdistance and the following distance of vehicles in every running conditons are analysed. Asa result, the distribution of distances sumit to exponential distribution while vehicles run inordinary condition and composite condition, and sumit to normal distribution in crowdedcondition. Further more, distribution of the distance of the wheel axles separated fromvehicles are analysed, which sumit to multi-peak normal-exponential distribution ormulti-peak normal distribution.4) Based on the characters of wheel axles and the reguralities of the distributionparameters on each lane, combined with real traffic on bridge, a live load model ofmotorcade traffic is proposed. Such a model consists of cars, trucks and heavy lorries.Parameters and regulations of their association are also analysed. Bridges of various spansand different patterns are used as example to calculate the effect of such a live load model,which is greater than the effet of the live load model specified by the code for bridgedesign.5) Based on the characters of wheel axles, methodology of describing the axle distanceby means of random field is proposed. Theory of random field is also applied tointerpretate why the distribution of vehicles sumit to exponential one in ordinary andcomposed running condition. Methodologies of multi-peak normal-Poisson renewalrandom field and multi-peak normal renewal random field are used to describe the distanceof wheel axles on bridges separated from vehicles. Live load model of renewal randomfield is proposed and parameters of the model are acquired simultaneously. Examples ofbridges for the effect of such a model show that, to the bridges of large span, the effect ofthe renewal model is greater than that from the code for bridge design, and it will beopposite to bridges of short span. Such a model eyes on the average characters of wheelaxles and kicks away the extreme characters of them.