Machine Learning based fault diagnosis method for 3-phase passive rectifiers in a Wind Energy Conversion System

There are a handful of ways to find faults in a converter. Traditional methods of fault diagnosis rely on expert knowledge and manual inspection, which can be time-consuming and costly. Data-driven is one of the methods which is in fashion these days as Artificial Intelligence is setting its root day by day. In this paper, a novel method was discussed to classify the fault on a residential wind turbine converter that uses a diode rectifier to convert AC to DC. By using fault data in a MATLAB model, data is authenticated by using a wind emulator and by the hardware of an uncontrolled 3-phase rectifier and boost stage in Normal mode operation. A correlation between input parameters and output parameters was made by using Python. As wind turbines are becoming an increasingly important source of renewable energy, it is important to ensure their proper functioning and minimize downtime due to faults. Wind turbine gets more than 30% of fault alone from the converter and semiconductor devices, so it is crucial to find a low-cost and less complex method to prevent such things. After collecting data, we used various machine learning algorithms and techniques that have been used for classification problems such as Support Vector Machines (SVM), Decision Trees, Random Forest, etc. Overall, the use of machine learning for fault diagnosis in wind turbines shows promising results in improving the reliability and efficiency of wind energy systems.