Research And Implementation Of Model Predictive Control For DC/DC Converter Of Fuel Cell

With the rapid development of new energy power generation technology,fuel cell technology has also received more and more attention.However,the output characteristics of the fuel cell are soft and the dynamic response is slow.It is necessary to use a special DC/DC converter to convert its output energy into usable electric energy.Therefore,the research of the corresponding DC/DC converter is also a hot spot.In order to reduce the output ripple of the fuel cell and improve the conversion efficiency and dynamic response performance of the converter,the research on predictive control of DC/DC converter model for fuel cell was carried out.The main contents are as follows:The topology of the DC/DC converter for fuel cells was designed.According to the requirements analysis and technical parameters,a two-stage,high step-up ratio,low ripple ratio DC/DC converter was designed.The front stage was an interleaved parallel Boost converter,and the latter stage was an autotransformer phase-shifted full bridge circuit.At the same time,the parameters of the power circuit were calculated and the devices were selected,the specifications of the radiator were designed,the thermal simulation model was built and the thermal simulation was completed.The predictive model of DC/DC converter for fuel cell was established.Because the converter is a two-stage structure and the front and back stages can be controlled independently,based on KCL law and KVL law,the state space equations of the front and back stages in each cycle were listed respectively,and the mathematical model in continuous time domain was obtained.By introducing predictive time domain P and control time domain M into discrete mathematical model,a complete predictive model was obtained,which lays a foundation for the design of model predictive controller of converter.A model predictive controller was designed.Based on the predictive model of the converter,by introducing the weight coefficients representing the primary and secondary control objectives,and a reasonable value function was designed.The optimal control quantity was obtained by minimizing the value function.By determining the current steady-state error as the main control objective,the reasonable weight coefficient value was selected,and the simulation platform was built based on the simulation software,and the prediction algorithm program module was compiled.The simulation of the converter based on the predictive control strategy was completed,and the simulation results were compared with those of the traditional PI control.The prototype of DC/DC converter for fuel cell was designed,and the actual effect of the model predictive control algorithm was verified by experiments.According to the complexity of model predictive control DC/DC converter,a control system based on DSP was designed,and the sub-program module was designed.According to the parameters of the power circuit,some parameters of the acquisition board and the driving board of the weak current part were designed.Finally,an experimental prototype of DC/DC converter for fuel cell was built,and the experiments based on model predictive control and traditional PI control were completed.The experimental results are compared and analyzed to verify the feasibility and validity of the proposed model predictive control algorithm and scheme based on DC/DC converter.