Estimation of remaining useful lifetime of power modules with kalman filter

• Main Author: Ritz Ridzuan Masari
• Other Authors: Josep Pou
• Format: Final Year Project
• Language: English
• Published: 2017
• Subjects: DRNTU::Engineering
• Online Access: http://hdl.handle.net/10356/71943
• Institution: Nanyang Technological University

This report presents the estimation of remaining useful lifetime (RUL) of power modules. In any system, it is essential to know the RUL so that it will be useful for maintenance of the power modules and it could be carried out more effectively. Therefore the lifetime estimation of power modules is important especially in applications that have a high reliability requirement. In this study, Insulated Gate Bipolar Transistors (IGBT) is used for testing. IGBTs are commonly used as switching devices in various applications. Some of the applications which IGBTs are present in are electrical appliances, electric cars, trains and etc. Since the use of IGBTs is increasingly common in various systems or devices, it is necessary to monitor the degradation of the IGBTs. Through proper monitoring of these power modules, the lifetime can be determined and at the same time being able to prolong the lifetime of the devices by tracking the changes in its parameters. In the experiments, the IGBTs undergo power cycling to observe its degradation behavior. This study aims to use a method of filtering named Kalman Filter (KF) to estimate the lifetime of the IGBTs. Kalman Filter (KF) is a recursive algorithm that is largely used to solve the problem of discrete data linear filtering. Ever since R.E Kalman introduced the Kalman Filter, this method of filtering have been extensively researched on and implemented in various applications namely, assisted or autonomous navigation, tracking objects and various computer vision applications. Through this process of Kalman Filter, the ON state resistance of the IGBTs is used as the main parameter to derive the equations in order to estimate the lifetime of the power modules.