9 - A robust multiple open-switch fault diagnosis approach for converter in wind energy system

Abstract

In wind power generation system, when a converter fault occurs, it will directly damage expensive equipment such as the generator, impede the operation of the wind turbine, or even pose a great threat to the entire power grid. Since converter faults occur independently and randomly, fault diagnosis can be considered as a feasible methodology for quick detection and isolation of faults. In this chapter, a novel data-driven fault diagnosis strategy of wind power converters is proposed based on ensemble empirical mode decomposition (EEMD), intrinsic mode function (IMF), and norm entropy (NE). The output three-phase voltage signals of the converter are processed and used to detect and identify the converter open-circuit faults. In the proposed method, the voltage is first processed by EEMD to obtain a series of IMFs. Then, NE is calculated based on statistical characteristics of the IMFs, and the extracted IMF-NE information is used to describe the diagnostic features. Finally, a support vector machine (SVM) is used to classify and identify converter faults. The effectiveness and reliability of the proposed method are validated in a simulated 1.5MW doubly fed wind power system, and the robustness of the proposed strategy to wind speed variation and sensor noise is also tested. From the simulated results, the proposed method shows outstanding performance in terms of robustness, high accuracy, and simple implementation without complex parameter tuning.