Modeling and optimization design for switching power supply

Aiming at the incapacity of establishing an accurate mathematical model for the switch mode power supply, an efficient method of engineering-oriented design is studied. A topology of switch mode power supply has been described. A method is presented, which is based on the experiment and simulation results to establish a neural network model for the switch mode power supply in Matlab. The process of circuit parameters optimizing is shown, which has combined the nonlinear neural network model with the genetic algorithm. A prototype that is based on the theory mentioned above has been made. Simulation and experiment have been taken to analyze the optimized model and the prototype. The result shows that the dynamic and static performance of the prototype, which are designed by the new method, achieve the expected success.