Abstract
Three-phase pulse width modulation converters using insulated gate bipolar transistors (IGBTs) have been widely used in industrial application. However, faults in IGBTs can severely affect the operation and safety of the power electronics equipment and loads. For ensuring system reliability, it is necessary to accurately detect IGBT faults accurately as soon as their occurrences. This paper proposes a diagnosis method based on data-driven theory. A novel randomized learning technology, namely extreme learning machine (ELM) is adopted into historical data learning. Ensemble classifier structure is used to improve diagnostic accuracy. Finally, time window is defined to illustrate the relevance between diagnostic accuracy and data sampling time. By this mean, an appropriate time window is achieved to guarantee a high accuracy with relatively short decision time. Compared to other traditional methods, ELM has a better classification performance. Simulation tests validate the proposed ELM ensemble diagnostic performance.
Highlights
1 Background Nowadays, induction motor drive systems fed by three-phase pulse width modulation (PWM) converters have been widely used in industrial applications [1]
In order to improve fault diagnosis, this paper develops a data-driven fault diagnosis method for PWM converter fed induction motor drive system
In this Insulated gate bipolar transistor (IGBT) open-circuit fault diagnosis, extreme learning machine (ELM) is converted to the multi-classification with multi-outputs mode, with m equals to 22
Summary
Induction motor drive systems fed by three-phase pulse width modulation (PWM) converters have been widely used in industrial applications [1]. The initial features are selected as a subset, which contains the relevant information from the input data By those features, diagnostic algorithm can analyze symptoms to make a fault diagnosis decision. Diagnostic algorithm can analyze symptoms to make a fault diagnosis decision Such signal-based methods require long processing time and the diagnostic performance is affected by fluctuation of loads. Due to its fast learning speed, ELM has already been used in detection of microgrid islanding events [28] and real-time dynamic security assessment of power systems [29] This method has not been adopted in fault diagnosis of power electronics devices. It shows that due to the fast learning speed of ELM, the proposed scheme is feasible to identify IGBT open-circuit faults with balanced diagnostic accuracy and speed. The simulation validates that the classification performance is independent of voltage ripple, and harmonics, speed and load fluctuations
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
