| Nonlinear loads absorb current from the actual grid, cause bus voltage waveform distortion and cause bus voltage and load current users appear harmonic.Harmonics generated by nonlinear loads on power systems is getting worse.Of power quality problems caused by harmonic disturbance source location, it can help power supply companies to find the responsible for causing power quality.In the actual power system, due to the instantaneous value of voltage and current data are tremendous, a large number of instantaneous values of the data will not be uploaded to the data center to save by power quality monitoring terminal.At present the layout power quality monitoring system mainly to provide statistical indicators for voltage and current values, such as IEEE P1159.3/D9standard data types mentioned. The instantaneous voltage and current values are selected to provide only a few cycles based on the actual situation,and not as a real-time data to save.Therefore, according to the statistics data of power quality to determine the harmonic disturbance source is a hot issue of power quality research.PQDIF is power quality data exchange format which specifically designed for power quality characteristics by IEEE standards committee.PQDIF provides a good platform for the analysis of power quality problems and to achieve power quality management and analysis system. PQDIF also has good compression and strong data processing ability and so on.Firstly.This paper which based on the IEEE P1159.3/D9agreement, describes in detail of the power quality data interchange format.Secondly, from the mathematical point of view, harmonic disturbance source location which due to the use of valid alues for the power quality is a typical non-linear mapping problem. As the BP neural network has strong ability to learn and can be more easily implement non-linear mapping process, the BP neural network algorithm is selected to applied to the harmonic source location.The power system simulation model creates in PSCAD. The instantaneous data are extracted from the bus and each of the related branches to be studied. According to the requirements of the harmonic national standards, calculate harmonic indices of the bus and each of the related branches to be studied, form the input vector of the BP neural network, determine the target vector of the BP neural network based on the feature of the simulation model, establish BP neural network model for the training and testing. Finally, according to power quality data of Power Company, achieve harmonic disturbance source location by BP neural network to the PQDIF actual data. |