Fast Shielding Optimization of an Inductive Power Transfer System for Electric Vehicles

Abstract

The shielding design is one of the most difficult phases in developing an inductive power transfer system (IPT) for electric vehicles. In this aspect, the combination of metamodeling with a multiobjective optimization algorithm provides an efficient approach. Here, Polynomial Chaos Expansions (PCE) and Multigene Genetic Programming Algorithm (MGPA) methods are used and compared to describe the mutual inductance of the IPT system in the function of the design variables on the shielding. These metamodels are obtained based on a number of 3D Finite Element Method (FEM) computations. Then, a multiobjective optimization algorithm coupled with the PCE metamodeling technique is applied to determine the optimal design variables for a practical shielding design when considering the magnetic coupling as well as the cost of the shielding as objective functions. Such a multiobjective optimization algorithm based on a particle swarm algorithm coupled with a metamodel on PCE method is proposed, leading to improve around 104 % of the mutual inductance M and save 14 % of the cost C for the shielding compared to the initial design.INDEX TERMS Shielding design, polynomial chaos expansion, multigene genetic programming algorithm, particle swarm algorithm, inductive power transfer system.

Recent Results in Shielding Technologies for Wireless Electric Vehicle Charging Systems

“…Metamodeling with a multiobjective optimization algorithm: The combination of metamodeling with a multiobjective optimization algorithm, such as Polynomial Chaos Expansions (PCE) and Multigene Genetic Programming Algorithm (MGPA), provides an efficient approach for shielding optimization. This method is used to determine the optimal design variables for a practical shielding design, considering the magnetic coupling and the cost of the shielding as objective functions [91]. Rain optimization algorithm (ROA): The rain optimization algorithm is used to optimize the coupling coefficient in the design of the charging coils for inductive type wireless power transfer systems in electric vehicles [92].…”

Optimization of ferrite structures in inductive power transfer system for electric vehicles

“…A 3D computational approach gives reliable results about the electrical parameters (mutual inductance, transmission efficiency) and the magnetic parameters (magnetic flux density leakage) around the system, but it may lead to heavy computations that have to be repeated for each new configuration that is highly dependent on various parameters: the size of coils, geometrical characteristics of the system (e.g., ferrite plates, shielding plates), possible misalignment between transmitter and receiver while charging. In [2], a fast and efficient modeling methodology in order to assess the efficiency of RIPT systems was proposed to manage EMC constraints in EVs. The introduction of metamodeling techniques (surrogate models based on polynomial chaos expansions) allows to manage the variability of design parameters describing the electromagnetic problem and to quantitatively determine the contribution of each design variable to the observed output.…”

Optimal Design of a Low-Cost Sae Ja2954 Compliant Wpt System Using Nsga Ii

“…In Sallan et al (2009), Shevchenko et al (2019) it has been shown that the S-S topology requires less copper than other topologies, coinciding with the main objective of this work, which is to reduce copper volumes and In terms of which part of the topology is optimised, three main categories emerge. The first category focuses exclusively on the optimisation of the inductor system, including magnetic flux density (𝐵), primary coil inductance (𝐿 𝑝 ), secondary coil inductance (𝐿 𝑠 ), mutual inductance (𝑀), and the possible incorporation of ferrites and aluminium shielding (Luo, Wei, & Covic, 2018;Otomo & Igarashi, 2019;Pei, Pichon, Le Bihan, Bensetti, & Dessante, 2022;Yilmaz, Hasan, Zane, & Pantic, 2017), in this category the optimisation is performed independently of the selected compensator in a next step. The second category involves the optimisation of the primary (𝐶 𝑝 ) and secondary (𝐶 𝑠 ) resonant capacitors, as well as other auxiliary components if any, as proposed in Bertoluzzo, Di Barba, Forzan, Mognaschi, and Sieni (2021), Yang et al (2023), where the capacitors of an LCC-LCC circuit are optimised, and in Yao et al (2019), for an S-CLC configuration.…”

“…With respect to the optimisation techniques used, these can be classified into two large groups: iterative parameter sweeping methods (Bosshard & Kolar, 2016;Sallan et al, 2009;Yan et al, 2018) and advanced optimisation algorithms such as NSGA-II type genetic algorithms (Bertoluzzo et al, 2021;Luo et al, 2018;Pei et al, 2022;Tan et al, 2019) or particle swarm optimisation (PSO) in both mono and multi objective versions (Hasan et al, 2015;Pei et al, 2022;Yang et al, 2023;Yao et al, 2019;Yilmaz et al, 2017), among others. The following conclusions can be drawn from this preliminary study:…”