Fault Prediction and Diagnosis of Power Electronic Systems in Smart Grids

With the rapid development of smart grids, the stability and reliability of power electronic systems have become key challenges. Traditional fault diagnosis methods are difficult to deal with the complex and changeable operating environment and massive data. Artificial intelligence (AI) technology, especially machine learning and deep learning methods, provides new solutions for fault prediction and diagnosis of power electronic systems. The review focuses on the application status of artificial intelligence technology in fault prediction and diagnosis for power electronic systems, analyzes the advantages and disadvantages of machine learning (such as decision trees, support vector machines) and deep learning (such as convolutional neural networks, recurrent neural networks), and discusses the future research direction of multi-modal data fusion. This paper aims to provide theoretical reference and technical support for fault prediction and diagnosis of smart grid. The purpose of reviewing the existing fault prediction and diagnosis technologies for AI-driven power electronic systems is clearly defined, such as sorting out the current research status, pointing out existing problems, and looking forward to the future direction. At present, AI-driven power electronic system fault prediction and diagnosis techniques that can be applied to smart grid include machine learning, deep learning and neural network related combination methods.