Research On Electromagnetic Coupling Noise Prediction Of Manual Wiring Harness Based On FDTD Algorithm

With the rapid development of modern power electronics technology,the handassembled cable bundle used for internal power supply interconnection,control signal and data information transmission and distribution between industrial equipment have become more and more dense.However,as the operating frequency increases,electromagnetic interference(EMI)and electromagnetic coupling within the cable harness or between the equipment can no longer be ignored.These complex electromagnetic environments will affect the safety and reliability of components and equipment.In order to predict or suppress the influence of the above electromagnetic environment on the equipment,it is necessary to predict the electromagnetic coupling noise of the cable bundle through a reasonable and effective method.Most of the electromagnetic coupling noise problems of the hand-assembled cable bundle focus on the crosstalk of line-to-line coupling and the radiation sensitivity(RS)of field-to-line coupling,but most of them are uniform multi-conductor transmission lines,while the cable bundle in actual problems is mainly non-uniform multi-conductor transmission lines.Based on the theory of multi-conductor transmission lines,this thesis extracts the parameter matrix of the cable bundle,and analyzes the electromagnetic coupling noise through the finite-difference time-domain algorithm(FDTD)based on the cascade idea.Firstly,the multi-conductor transmission line model and its unit length parameter matrix characteristics under line-to-line coupling and field-to-line coupling are established.The different calculation methods of the parameter matrix are discussed separately,and a method for extracting the parameter matrix of the uniform twisted manual harness based on the Back Propagation(BP)neural network algorithm is proposed.On this basis,the Beetle Antennae Search(BAS)and Beetle Swarm Optimization(BSO)algorithm are used to extend it to application scenarios with a larger number of the hand-assembled cable bundle.In the 0.1MHz-1GHz frequency band,taking twisted-core,seven-core and nineteen-core twisted cables as examples,the results show that the new method has higher prediction accuracy and performance of the multi-core manual twisted parameter matrix.Secondly,on the basis of the multi-conductor transmission line equation and parameter matrix,the FDTD algorithm under different difference formats is studied respectively.The disturbed voltages of parallel double conductors were calculated under line-to-line coupling and field-to-line coupling,and a wiring method to reduce the delay effect and peak value of the voltage was proposed.Then,based on the evenly twisted manual wiring cables,a new and effective nonideal manual wiring cables model establishment method is proposed.In the modeling process,the random transposition of the wire cabele and the random rotation of the cross-section of the wire cable were processed,and the actual restriction requirements of its randomness were considered.Combined with the BSO-BP neural network model,the parameter matrix at any position of the non-ideal manual harness can be obtained.The FDTD algorithm is used to solve the crosstalk,differential mode,common mode current,and the electromagnetic coupling noise in the "worst case".A case simulation is used to verify the effectiveness of the method proposed in this thesis.Finally,relying on the Key Engineering Laboratory of Electrical Equipment and Electromagnetic Compatibility of Jiangsu Province,the actual model was measured through the semi-anechoic chamber and the S parameters of the vector network analyzer.The result data of the method in this thesis is compared with it to verify the effectiveness of the electromagnetic coupling noise prediction in this thesis.