387 978-1-7281-3936-4/19/$31.00 ©2019 IEEE Model Predictive Control Based On Cuckoo Search Algorithm of Interleaved Parallel Bi- directional DC-DC Converter Wenwen Sun School of Automation Wuhan University of Technology Wuhan, China 1069327165@qq.com Qihong Chen School of Automation Wuhan University of Technology Wuhan, China chenqh@whut.edu.cn Liyan Zhang School of Automation Wuhan University of Technology Wuhan, China zly@whut.edu.cn Abstract 瀆In order to improve the response speed and reliability of the interleaved parallel bi-directional DC-DC converter, a constrained model predictive control(MPC) based on cuckoo search algorithm is proposed. Firstly, take the buck mode for example, the predictive model is established according to the equivalent circuit model of the converter under different two switch states. Then the cost function is built to evaluate the performance of the converter. Otherwise, the cuckoo search optimization algorithm is introduced and used to solve model predictive control optimization problem so that the speed solution was improved. Finally, the simulation was carried out by MATLAB/Simulink and the results of the model predictive control, PI control was analyzed and compared. The simulation result show that the CS-MPC converter has better dynamic response performance and steady state performance, and the algorithm is feasible and effective. Keywords: Bi-directional DC-DC converter; Model predictive control; PI control;Cuckoo search algorithm Iˊ INTRODUCTION In order to ensure the good performance of electric vehicle power supply system and improve the service life of lithium batteries, the composite power supply composed of lithium batteries and other energy sources has attracted the attention of many researchers[1][2][3]. In such composite power supply, bidirectional DC-DC converter is the key to its energy management[4]. Therefore, the research of bidirectional DC-DC converters with fast response, high reliability and high efficiency is of great significance to the research and development of electric vehicles. The bi-directional DC-DC converter between super capacitor and storage battery needs to ensure bi-directional energy flow and smoothly adjust voltage and current[5][6]. In order to achieve this function, various advanced non- linear control algorithms, such as fuzzy neural network[7] and sliding mode control[8], have been proposed one after another. Model predictive control (MPC) is a model-based closed-loop optimal control strategy, which was originally used in the field of process control, but now it has emerged in the field of power conversion[9][10][11]. Deb and Yang proposed cuckoo search algorithm[12].A. Pirooz and Maksym Khomenko apply the core idea of model predictive control to bidirectional DC-DC converter, and achieve the corresponding control effect by optimizing the value function and selecting the optimal switch state[13][14]. Meiyang et al. used the same control principle to control and optimize the performance of interleaved bi-directional DC-DC converters[15]. Although the control principle is simple and easy to implement, the dynamic performance of the system is poor and the control accuracy is not high. Mehdi Ebad DQG Byeong-Mun Song compares the dynamic response of output voltage and the tracking effect of reference value of bi-directional DC-DC converter for four different prediction models[16]. However, this paper is based on the simplest topology and has very small capacitance at low voltage side, which is not suitable for electric vehicles. In reference [17], an active full-bridge DC-DC converter is studied, and an output current control method based on input voltage feedforward is proposed. However, for interleaved bi-directional DC-DC converters, there are few research results on predictive control algorithms. In order to overcome the shortcomings and limitations of the above control methods, a constrained model predictive control method based on cuckoo search(CS) algorithm is proposed in this paper, using a two-phase staggered bi-directional DC-DC converter topology. On the basis of establishing its average state space model, the prediction model under Buck mode is deduced. Cuckoo search optimization algorithm is used to solve the optimization problem of constrained predictive control. The effectiveness of the method is proved by simulation analysis. II. TOPOLOGY AND PREDICTIVE CONTROL MODEL A. Topology