Multiparameter Identification for Buck Converter Based on Time-Domain Mapping

This paper introduces a multi-parameter identification method for a Buck converter based on time-domain mapping. First, an analytical solution model of the converter is developed, and waveforms over several periods are sampled. The one-to-one time-domain relationship between parameters and waveforms is then explored by iteratively updating the model parameters using Particle Swarm Optimization (PSO) to approximate the sampled waveforms. When the difference between the generated and sampled waveforms becomes sufficiently small, the corresponding parameter set is considered equivalent to that of the physical converter. Simulation results indicate that the identification errors for all parameters remain within 1%. Experimental validation, including tests under simulated component aging, demonstrates that the proposed method accurately tracks parameter degradation at low sampling rates. Furthermore, PSO maintains good robustness across various operating conditions, with most parameter errors remaining below 5%. These findings verify that the method maintains high estimation accuracy even when sampling rates are low, thereby reducing hardware requirements and implementation costs. This advantage makes it particularly suitable for long-term monitoring of component aging and degradation in power converters, enabling timely maintenance and improved system reliability in practical engineering applications.