Research On Model Predictive Control For Bidirectional DC-DC Converter Of New Energy Vehicles

Energy issues have always attracted much attention.New energy vehicles have been developed due to their clean and renewable characteristics.Hybrid systems consisting of super-capacitors and lithium batteries are favored because they can both effectively extend the life of lithium batteries in electric vehicles and increase their range.As the core of the bidirectional flow of energy,the bidirectional DC/DC converter largely determine the performance of the hybrid power system.This thesis aims to improve the dynamic response speed and immunity of bidirectional DC/DC converter.The research and design of bidirectional DC/DC converter are as follows:The overall scheme of bidirectional DC/DC converter is designed.Based on the requirements analysis,the typical topology of bidirectional DC/DC converter is analyzed.A bidirectional interleaved parallel topology is designed according to the technical requirements.The topology principle is analyzed,and the power circuit parameters are calculated and selected.The bidirectional DC/DC converter heat sink is designed and verified by simulation.A bidirectional DC/DC converter prediction model is established.The key parameter is selected as the state variable.Based on the state-space time average method,a state-space model of the step-up and step-down operation modes of the bidirectional DC/DC converter is established and discretized.Then a discrete prediction equation of the inductor current is established.A model predictive controller based on cuckoo optimization algorithm for bidirectional DC/DC converter is designed.First,the objective function is defined according to the performance requirements,the constraint conditions are given,and the prediction model is brought into the objective function for simplification.Then,the meta-heuristic algorithm of cuckoo optimization is used to solve the problem,and the specific algorithm flowchart is drawn.Finally,the sampling delay is compensated and simulated to verify the effectiveness of the algorithm in this thesis.A bidirectional DC/DC converter experimental platform based on DSP + FPGA control architecture is designed and verified by experiments.Aiming at the problem of large calculation amount caused by the cuckoo search algorithm in the rolling optimization process,the control structure of DSP + FPGA is adopted.The DSP collects real-time data and sends it to the FPGA for control amount calculation.The calculation result is then transmitted to the DSP for real-time control.The experimental results show that the bidirectional DC/DC converter designed based on cuckoo optimized model predictive control algorithm has fast dynamic response speed and good immunity.