Study For Cable Tensions Identification Of Short Suspender

For present-day bridges.cable tensions testing is a vitally important jov in course of construction.The tensions condition of cables plays an especially important part in the course of construnction and operation.In order to ensure the safety of operation and lengthen the service life,the exact testing result is being required to obtain.Incable stayed bridges and arch bridges.the short suspenders existing and the boundary condition as well as the bending rigidity has an greater effect on the cable forces testing.This paper based on Additional Mass Method,using the method of GA-BP Neural network after Genetic Algorithm optimization to identify the cable forces of short suspenders and a preferable research result is obtained.Firstly,this paper introduces a variety of cable forces testing method at present and analyzes the serviceability as well as defect of each kind of method.The often-used vibration method at hom and abroad is described in detail and by summariziry the development course and research status of vibration method,the limitation of it in cable tesions testing of short suspenders is analyzed.Secondly,its solutions are dexcribed briefly.In addtion,the frequency equations under the three boundary conditions are presented,and the effect of bending rigidity on cable forces is analzed emphatically.Thirdly,the foundamendal thoery of AMM is introduced briefly.Based on AMM,making use of the GA-BP toolbox to forecast the cable forces of suspenders,the aim can be achived,but it calls for a quantity of data and the speed of iteration is slow.Morever,the iterating is unstability.The GA method which is adopted in this paper optimizes BP Neural network and it improves the stability of iteration sa well as the convergence speed.Therefore,the combination way of GA and BP Neural network is discussed in detail.Making use of GA-BP Neural method to forecast the cable forces,needs to get enough resonable training date firstly.This paper analyzes how to establish numerical computation model and makes use of AMM to do the numerical computating,in addition,using the resonable calculation result as training data.Then the building course of the GA-BP Neural network structure is discussed in detail,using numerical computation,in order to check calculating the stability,converical computation,in order to check calculating the stability,convergence and validity.Lastly,the realizing course of test verification is introduced in detail,In addition,it makes the comparative analysis between the results forecasted by the GA-BP Neural network structure after training and the test data,verifying the applicability of this computing method.