Application Of Non-constant Speed Algorithm In Bridge Weigh-in-motion System

Bridge weigh in motion(BWIM)system is based on the Bridge's dynamic response to identify the axle load of vehicles.The algorithm proposed by Moses is a commonly used method in BWIM which is based on the bridge influence line to quickly identify the axle load of the vehicle crossing the bridge,and it is also the algorithm adopted by the current commercial BWIM system.Therefore,it is important to obtain the influence line consistent with the actual bridge structure to improve the accuracy of the axle load identified.However,this method is greatly affected by the dynamic effect of the bridge,and it will lead to the o ccurrence of ill-posed problems.In addition,the current influence line algorithm only considers the situation that the vehicle travels at a constant speed,so its application range is limited.In this paper,taking a simple beam bridge in I-78,Alabama,the United States,as an example,Moses algorithm is studied in depth in many aspects.The main work and result are described as follow:(1)Taking vehicles with any number of axles as an example,the calculation formula of Moses algorithm is derived,and its effectiveness is verified in the calibration test and random traffic flow test.The results show that the error of Moses algorithm is small and within the acceptable range when calculating the total axle load and group of axles load,while the error of single axle load is large and recognition accuracy needs to be improved.(2)A quasi-static influence line algorithm was proposed based on Moses influence line algorithm,and its effectiveness was verified in the calibration test and random traffic flow test.The quasi-static influence line is obtained by filtering out the dynamic part of the calibrated influence line,and its theory is based on the spectrum analysis of influence line and strain.The calculation results of the two tests show that the quasi-static influence line algorithm is at the same accuracy level as Moses algorithm.It can be concluded from the spectrum analysis of strain and influence line that the shape difference of the calculated influence line mainly depends on vehicle speed and the natural frequency of the bridge.(3)The ill-posed equation theory is introduced,and the generation mechanism of the ill-posed problem in Moses algorithm is analyzed.An improvement method is proposed for the generation mechanism,which is verified in the numerical simulation method.In addition,this paper proves that the Gaussian noise in strain data has little influence on the accuracy of axle load identification,so Gaussian noise should not be used to simulate the noise in real measurement.(4)Considering that the vehicle is traveling at a non-constant speed,a new improved algorithm,namely the non-constant speed algorithm,is proposed and the calculation formula is derived.The precision of the algorithm is analyzed in real bridge tests and numerical simulation.In the calibration test and random vehicle flow test,because the speed is strictly controlled,the advantage of non-constant speed algorithm can't be reflected.In the numerical simulation calculation,considering the vehicle driving at different accelerations,the final calculation results show that the calculation accuracy of the non-constant speed algorithm is generally at a high level,and especially its calculation accuracy of single axle weight is obviously higher than that of Moses algorithm.