| Bulb flat steel-STC light composite deck use ultra-high toughness material to form the pavement layer and bear the force together with the steel deck can significantly improve the local stiffness of orthotropic steel deck and improve the mechanical performance of the steel deck.However,under the repeated action of high stress level of overloaded vehicle wheel load,the weld of steel bridge surface is prone to fatigue cracking,which brings great risk to the safe use of the bridge and greatly increases the maintenance and operation costs.In order to reduce the risk of fatigue crack of steel bridge deck weld,it is important to monitor the traffic volume in real time.In this paper,Zhaohua bridge in Xiangtan,Hunan Province(self anchored single tower suspension bridge with ball flat steel STC light composite deck)is taken as the research object to carry out the real bridge test.The Moses algorithm is applied to the real bridge and optimized with the axle load identification results.A single wheel algorithm considering the overall effect separation is proposed for vehicle axle load information identification.This paper mainly includes:(1)The corresponding test plan is made from the selection of measuring points,sensor arrangement and other aspects,and the field test is carried out according to the test plan.Two types of triaxial loading vehicles with a load of about 35t and 45t are used for the sports car test,and the real-time dynamic response on the longitudinal rib of the bulb flat steel is collected when the vehicle passes the bridge.(2)The relevant formulas of the Moses algorithm are deduced in detail.The Moses algorithm is used to identify vehicle information.The results show that:the basic vehicle information(vehicle axle number,axle space and speed)recognition accuracy grade is D+(20)～D(25),and the accuracy grade of axle load recognition for ordinary lanes of bridge section is C(15);The calibration of the actual bridges with longitudinal partition lanes in mid-span and bridge tower sections has poor coincidence and the longitudinal coordinate value of the influence line is less than that of ordinary lanes.The axle load accuracy of lanes with longitudinal partitions is D+(20).The overall effect component is not separated in the mid-span and the tower section,and the overall effect of the mid-span section is greater than the cross-section of the tower,thus the accuracy of the axle load recognition of the ordinary lane in the mid-span section is less than that of the tower section;The total signals of all sensors in the whole section are used in the calculation of the axle load of Moses algorithm.For the structure with significant local stress,using Moses algorithm will cause the error accumulation caused by the environment and other factors.(3)In order to eliminate the cumulative error caused by the Moses algorithm using the full-section response to participate in the calculation,the finite element model corrected by the static bridge static test data is used to calculate the lateral influence line under a single wheel load and the wheel load is analyzed according to the field test data.The lateral action range shows that under a single wheel load,the influence range of the transverse bridge direction under a single wheel load is 5 spherical bulb flat steel longitudinal ribs.Based on the significant local stress characteristics of the spherical bulb flat steel-STC light composite deck,on the basis of the Moses algorithm,a single-wheel algorithm considering the separation of the overall effect is proposed and a detailed derivation are given.(4)Single wheel algorithm is used for vehicle information recognition.The results show that in the case of only considering the left wheel effect P_L,the right wheel effect P_R and the left wheel and the right wheel effect P_L+P_R,the coincidence degree of single wheel calibration influence line in ordinary lane of midspan and tower section is fine,and the accuracy of axle load identification grade is B(10)which is high.In the case of P_L,P_R and P_L+P_R,the coincidence degree of influence line is good,the value of influence line in longitudinal partition lanes is smaller than that of ordinary lane of midspan section,and the accuracy grade of axle load identification is C(15).The calibration influence line in longitudinal partition lanes has a high coincidence degree,and the influence line value is basically the same as that of the ordinary lane under the cases of P_Lin tower section.The coincidence degree of calibration influence line in longitudinal partition lanes is poor,and the influence line value is less than that of the ordinary lane under the cases of P_R and P_L+P_Rin tower section.The identification grades of P_L,P_R and P_L+P_R axle load are B(10),E(55)and D+(20),respectively.The accuracy grade can be significantly improved by using single wheel algorithm of the P_L.The longitudinal partition increases the local stiffness,reduces the strain value on the longitudinal rib,reduces the value of the calibration influence line,and reduces the accuracy level of axle load identification.The accuracy level of axle load recognition is controlled by single wheel load accuracy.Compared with Moses algorithm,the accuracy of single wheel algorithm in ordinary lane and lane with longitudinal partition is significantly improved.(5)By comparing and analyzing the similarities and differences between the Moses algorithm and the single wheel algorithm,the single wheel algorithm has the advantages of fewer sensors,more convenient installation,lower cost and higher accuracy of axle load identification. |
PDF Full Text Request