Sensitivity-Based Parameters Identification Method Of Long-Line Structures

Railway transportation,as a popular means of transport,has gradually become the backbone of modern transportation.To be the long-line structure,the railway structrure is subject to high-speed and high-frequence running vehicles,resulting in continuous deterioration of its performance.The damage of the structure will undoubtedly bring significant dangers to the railway traffic and even lead to the major traffic accidents.Therefore,real-time identification and monitoring of vehicle parameters and damages of the track is significant to ensure the normal operation of urban traffic.Based on the support by Basic Research Program of China(973 Program,2011CB013800)and National Natural Science Foundation of China(NSFC,5177081226),the parameter identification method of the long-line structures is studied in this dissertation.By adopting the theoretical derivation,finite element numerical simulation and experimental validation,the parameter identification method of long-line railway structure under vehicle load is studied,including the following parts:(1)A new fast explicit integration method is proposed to solve the dynamic response analysis of the long-line structure with high accuracy and high efficiency.The effects of the length of the track,the length of time step,the relaxation factor and the convergence conditions are investigated to verify the stability and efficiency of the method.The maximum time step,the optimal relaxation factor and convergence conditions are obtained,which provides a foundation for the parameter identification of the long-line railway system later.(2)Deriving sensitivity of the vehicle dynamic response to vehicle parameters and structure damages,an iterative optimization algorithm based on sensitivity analysis is presented to identify the damages of the long-line structure.The correctness and stability of the proposed method are verified from different track irregularity,the position of the response,the number of units with stiffness changes,and the measurement noise.The high accuracy and efficiency and of the parameters identification method proposed in this chapter is verified by comparing with the traditional method.(3)Since the parameter identification of the long-line railway system is inefficient,a substructure is proposed to the parameter identification of the long-ling railway structure.The track is divided into multiple sub-structures.The interface force of the target track substructure is expressed by the Chebyshev polynomial.The sensitivity matrix of the vehicle’s dynamic response to track damages and interface force parameters is derived.The track damages and interface force parameters are identified simultaneously based on substructure method.Since the substructure damage identification method contains fewer number of degrees of freedom and the number of parameters to be identified in the calculation process,the efficiency of the parameter identification is greatly improved.(4)Considering the uncertainty of the railway traffic,a calculation of dynamic response boundary interval based on the high dimensional model representation is proposed for the long-line structure to estimate the bounds of the dynamic response.The dynamic responses of the long-line structure under moving loads with fuzzy parameters are expressed as an orthogonal polynomial.Afterwards,according to the Lagrange interpolation principle,the orthogonal polynomial is converted into the Lagrange interpolation function.The boundary interval of the dynamic response of the long-line structure can be calculated approximately by the dynamic response of a few sample points,which greatly reduces the computation time and improves the calculation efficiency for statistical analysis.(5)Combined with parameter identification method of the long-line structure,a method based on the perturbation method for the identification of uncertainty damages in the long-line structure is proposed.The first derivative and second derivative of the vehicle dynamic response and vehicle dynamic response sensitivity to the random part of the system parameters are derived.Afterwards,the mathematical expectations and variance of the parameters to be identified are calculated.