| Maglev train is the key research direction of high-speed rail transit,and its electromagnetic compatibility performance has been paid much attention.Due to the stable suspension gap of maglev train,the on-board electrical network can not be well grounded,and it is easy to form ground loop backflow on the surface of the car body,thus causing electromagnetic compatibility hidden trouble.Therefore,the measurement method of electric current and electric potential distribution on maglev car body surface is studied,aiming to provide necessary experimental means for improving the electromagnetic compatibility performance of maglev train.Mainly completed the following work:Since the measurement problem of large continuous conductor is not suitable for the traditional near-field scanning method,some changes were made in the collection method to make it more suitable for restoring the overall distribution of regional current and potential.The electric current distribution data of the region to be measured will be solved first,and then the electric potential distribution data will be obtained by integrating relation.A theoretical model with several adjustable parameters was established under the assumed ideal measurement conditions,which can well adapt to the complexity of engineering practice.Under the framework of the model,the radiation coefficient of the surface current element was derived,the mathematical model in the form of ill-conditioned equation was established based on the moment method,and the more compact convolution form was derived.Based on the ill-posed equation form of mathematical model,regularized ridge estimation method and L-curve method were introduced to help solve the equations,and a feasible algorithm was sorted out.The algorithm can recover the current distribution accurately and effectively,but it has some limitations in practical application due to its long operation time and high calculation consumption.By analyzing the mathematical form of ridge estimation,a simplified regularized constraint coefficient was derived,which was further sorted into a feasible algorithm.Under the same model,the residual error of the algorithm is large,and the ripple of recovery current is obvious,but the algorithm has good adaptability to the severely ill-condition,and the computational efficiency is better than the previous algorithm.Based on the convolution form of the mathematical model,the fast Fourier transform algorithm was used to realize the efficient operation of convolution,and by analyzing the similarity of the mathematical form,the regular constraint method was introduced in the frequency domain,and a fast algorithm was designed to calculate the current distribution.The algorithm is poor in peak recovery,and the recovery effect of some models is better than algorithm 2.The time complexity of convolution fast algorithm is O(7)2))and the space complexity is O(),which effectively saves the storage cost and reduces the operation time,and has the potential to realize the real-time acquisition and display of current distribution.A large number of sub-models with different parameters were simulated and analyzed to compare the influence of model parameters on the three algorithms and the similarities and differences among the three algorithms.In a typical example,the recovery current of algorithm 1 could reach a high similarity of 93.8% and a relative residual of 10.2%.The similarity of recovery current of algorithm 2 could reach 91.5%and the relative residual is 19.2%.The recovery current similarity of the fast algorithm was 90.6% and the relative residual was 14.1%.It was found that the model parameters mainly change the well-posedness of the mathematical model,which had the most obvious adverse effect on the fast algorithm.There are 54 pictures,17 tables,and 54 references. |
PDF Full Text Request