A RLC parameter identification method for synchronous buck converters based on simulated annealing coupled cuckoo algorithm

Fault detection technology based on parameter identification is advancing rapidly. For power electronic circuits, parameter identification is essential for accurately evaluating performance. The commonly used identification algorithms in engineering include least squares method and particle swarm optimization algorithm, etc. The least squares method has the disadvantage of large identification error and low identification accuracy, while the convergence instability and susceptibility to local optima of particle swarm optimization algorithm also need to be improved. To address the shortcomings of the least squares method and leverage the advantages of the cuckoo algorithm, such as strong global search capability and fast convergence, this paper focuses on the synchronous buck circuit. A local linear model of the synchronous buck circuit is established to mitigate the influence of nonlinear signals, and the signals collected include the terminal voltage V_S2, inductor current i_L and output voltage V₀. The cuckoo algorithm is employed to identify the parameters. To overcome the local optimum problem associated with the cuckoo algorithm, an improved cuckoo algorithm is proposed that integrates simulated annealing to identify the parameters L, C, R_L and R_C in the circuit. To demonstrate the robustness and convergence of the proposed algorithm, we use the MATLAB/Simulink platform for comparative analysis against the conventional least squares method. The results show that the improved cuckoo algorithm achieves faster convergence and greater robustness. Finally, validation on a physical experimental platform indicates that the data extracted from the actual circuit maintains high identification accuracy and convergence speed, with identification errors below 5%, further verifying the effectiveness of the algorithm.