Fault Detection and Visualization Technique for Hybrid Electric Vehicle Components

The aim of the paper is to efficiently distinguish the abnormalities in huge datasets by employing different mechanisms like visualization techniques and algorithms. These methods classify faulty data from available test data and represent it visually. In data visualization human interpretation is required to detect abnormalities while the algorithm classifies faulty and healthy data using rule-based algorithms. The parallel coordinates plot is developed in MATLAB. Fault diagnosis in Permanent Magnet Synchronous Motor (PMSM) was done by means of pattern recognition algorithm based on the Concordia transform. The algorithm is simulated in MATLAB with original vehicle test data. The proposed fault detection algorithm holds good for analysis of variable speed and variable torque drives.