Аңдатпа
Given the increasing dependence of different industries on power electronic systems, particularly isolated power DC-DC converters, efficient predictive maintenance strategies have been developed to guarantee the reliable and efficient operation of these systems. Because of their complexity and dynamic nature, these converters are difficult to maintain using conventional methods like time-based or condition-based processes. Artificial neural networks (ANNs) are a useful tool for increasing the accuracy and efficacy of predictive maintenance for isolated power DC-DC converters. The model predictive, which serves as an expert in this work, provides the data used to train the suggested ANNs. Once tuned, the ANNs are used directly to predict faults in isolated DC-DC power converters. LTspice and MATLAB simulations are used to validate the performance of the proposed artificial neural networks. The results of the simulation demonstrate that the performance of the ANNs-based isolated power DC-DC converter is improved when a trained ANNs model is used. Consequently, the F1 score of the artificial neural network is 99.34%, the accuracy score is 89.83%, the precision score is 94.64%, and the recall score is 95.83%. Based on the scores results, we suggest that artificial neural networks (ANNs) could be useful for predicting isolated power DC-DC converters.