EECL lab is applying machine learning techniques in various ways to the power electronics.
First, machine learning is applied for high-performance motor control. We develop technology to effectively reduce torque ripple using artificial neural networks. Second, we are also conducting research in sensorless control of motors using machine learning to replace the existing sensorless methods.
Furthermore, machine learning is used to estimate the temperature of the drive system. It improves the reliability, and reduces cost in measuring temperature by removing the temperature sensors.
Finally, machine learning is applied for the fault diagnosis on the drive system. In the fault diagnosis of drive system, we are studying fault classification and abnormality detection. The application targets are industrial 6-axis robots and drive systems of electric vehicles, which can improve system reliability by diagnosing failures in advance.