This paper proposes a novel approach based on the chaos eye method (CEM) and extension neural network (ENN) for fault diagnosis of wind power systems. First, we used sensors to capture the vibration signals of the wind power system to detect subtle changes. Subsequently, the chaotic synchronization detection method was used to form a chaos error distribution diagram. The distribution diagram centroid, called chaos eye in this paper, was used as the fault diagnosis feature to reduce the number of extracted features. This reduction in diagnostic features enables considerably reducing the computation time and cost of hardware implementation. The ENN-based method was then used to design a fault diagnosis system for the tested wind power generation. The feasibility and practicability of the proposed method were validated using a simulation system. The patent for the proposed method is currently pending, and this method contributes to the key technologies of large-scale wind power generation systems in Taiwan.

Included in

Engineering Commons