A neuro-fuzzy system for recognition of power quality disturbances

The ability to locate the sources of disturbances in power systems is paramount to the maintenance of the quality of power supply. Once the source of disturbances is identified, solutions can be found to reduce or remove these disturbances from the system. Currently, the classification of a broad range of disturbances is carried out manually on collected data; it is a costly and inefficient task. This paper presents an automatic disturbance recognition system, its potential advantages and describes a method for building such a system. The system applies a variety of tools, which include Fourier transforms, wavelet transforms, artificial neural networks, and fuzzy logic.