A conceptual design for a power quality meter based on an intelligent system is presented in this paper. The general performance requirements of the meter, the major parts of one type of intelligent system, and the information required to "train" the complete package are discussed in this paper. The proposed meter will ultimately detect power quality problems such as sags, swells, spikes, neutral voltage rise, harmonic contamination, etc. Individual artificial neural networks will be used to identify the exact type(s) of problems present in a given voltage or current waveform. Based on the identifications provided by the artificial neural networks, an expert system will recommend an appropriate control action.
展开▼
