Machine Learning-Based Optimization of Synchronous Rectification Low-Inductance Current Secondary Boost Converter (SLIC-QBC)

For recycling waste batteries, high-gain DC-DC provides a great solution. In this article, the design and optimization of a high-gain converter–synchronous rectification low-inductance current secondary boost converter (SLIC-QBC) is studied. The optimization objective of this article is to propose an automatic design method for passive components and the switching frequency of the converter to improve efficiency and power density. A machine learning-integrated optimization method is proposed to minimize the converter mass and power loss of the converter. In this method, first, a component-based automatic design model is built with embedded SLIC-QBC simulation. Then, a series of design schemes is generated within a large parameter range, and training data for machine learning are collected. Support vector machine (SVM) and artificial neural network (ANN) are used to validate the converter design scheme, where ANN establishes the mapping relationship from design parameters to optimization objectives. After the optimization, an experimental prototype is built for experimental verification. The simulation and experimental results verify the practicability of the proposed method.