Hybrid Uncertainty Quantification Method For Electromagnetic Coupling Of Automotive Harness And Its Reliability Evaluation

Along with the automobile industry and electrical equipment fast development,there are more and more types and quantities of electrical equipment installed on automobiles.The substantial increase in the type and quantity of on-board electronic equipment not only enriches the practical performance of vehicles,but also brings serious challenges to electromagnetic safety and electromagnetic compatibility.As the carrier of transmission of electrical signals between various on-board electronic devices,the phenomenon that some functions or electronic devices of the vehicle are degraded or even fail due to the electromagnetic coupling of the harness is also increasingly serious.Considering the uncertainties in the manufacturing process and binding and installation process of automobile harnesses in practical engineering,the uncertainty quantification of electromagnetic coupling effect of automobile harnesses has become a hot issue to be solved urgently in the field of automobile EMC.Since the electromagnetic simulation analysis method is a common method to solve electromagnetic compatibility problems in the early stage of vehicle development,the progress evaluation method of electromagnetic simulation model also has very important research significance.In response to the above problems,this paper mainly carried out the following work:1.Based on classical electromagnetic principles,this paper deduces the multiconductor transmission lines theory,using chain parameter matrices and Thevenin’s theorem.Two methods for calculating the unit length distribution parameter matrix,namely,the image analytical method and the finite element method,were derived.The analytical method was used to calculate the unit length distribution parameters of ordinary wire harnesses and the results obtained by Q3 D were compared.To predict and analyze the electromagnetic coupling of car harnesses,this paper applies the multi conductor transmission line equation to calculate the induced current of car harness crosstalk and radiation sensitivity,and uses the calculation results of CST software as a reference to verify the accuracy of the electromagnetic coupling model of car harnesses constructed.2.Considering the mixed uncertainty problem of electromagnetic coupling of vehicle harnesses due to manufacturing processes,bundling and installation,actual operating conditions,and other reasons in engineering applications,this paper uses chaos polynomial method and Bayesian optimization algorithm to quantify the mixed uncertainty and sensitivity analysis of electromagnetic coupling effects of vehicle harnesses based on probability box theory.Firstly,the traditional hybrid uncertainty quantization framework based on Monte Carlo method and grid search method is introduced.To solve the defect of the low computational efficiency,this paper proposes a hybrid uncertainty quantization method using chaotic polynomial method and Bayesian optimization algorithm,and deduces the chaotic polynomial method and Bayesian optimization algorithm.Further,in order to perform sensitivity analysis on the input parameters of the model,the principle of Sobol sensitivity analysis method is derived in this paper.Finally,combined with an example of vehicle wiring harness,the accuracy and efficiency of the method are verified by comparing the mixed uncertainty quantification results of the traditional method and the proposed method.3.When conducting uncertainty analysis,simulation model credibility evaluation can objectively and quantitatively evaluate uncertain models.Therefore,this article conducts relevant research on the reliability evaluation methods of uncertain electromagnetic simulation models,summarizes the research significance of model reliability evaluation,and introduces the solution process and mathematical properties of reliability evaluation indicators.Considering the defect of current methods for evaluating the reliability of electromagnetic simulation models,this paper proposes to use area metric factors to quantitatively evaluate uncertain simulation models,and to evaluate the reliability of the established hybrid uncertain electromagnetic simulation models,further proving the correctness and availability of the method.Aiming at the mixed uncertainty analysis of electromagnetic coupling of car harnesses,the method based on chaos polynomial and Bayesian optimization algorithm proposed can greatly reduce the cost and function calls while ensuring the correctness of the model compared with the traditional mixed uncertainty quantification method.At the same time,it realizes the sensitivity analysis and reliability evaluation of the mixed uncertainty model of electromagnetic coupling of car harnesses.