Analyze the Crosstalk of Multi-core Twisted Wires and the Effect of Non-matched Impedance Based on BSAS-BPNN Algorithm

Abstract

In this paper, a multi-core twisted wire model with random non-uniform twists is established. The random combination of complete and non-complete pitch sections is used to accurately simulate the randomness of actual multi-core twisted wires. On the basis of the model, the cross section of the cascaded MTL is obtained, and a neural network algorithm is used to describe the complex relationship between the arbitrary position of the multi-core twisted wires and the unit length parameters. The unit length parameters at any position are obtained by cross section rotational transformation and random transposition transformation between conductors. Finally, the crosstalk in electromagnetic compatibility performance is calculated, and different termination impedances are analyzed. The results show that the crosstalk of multi-core twisted wires is susceptible to the effects of twisting and termination impedance at high frequencies, and the reliability of the proposed method is verified by comparison with full-wave simulation.