Short Term Load Forecasting Based on Fuzzy Clustering

This paper presents selected similar days of short-term power load forecasting model based on fuzzy clustering, the method first meteorological factors subdivided into temperature, barometric pressure, wind speed, rain, etc., and then type the week, date, type of day together constitute similar factors, fuzzy coefficient feature mapping table through fuzzy rules, not only to achieve quantitative impact factors and facilitate real-time to add a new rule. On this basis, the use of fuzzy clustering method for classification, based on the level of clustering similar day is selected to reduce the number of samples to accelerate the selected speed. The model takes into account the full impact of weather and other factors on load forecasting, further weakening the load of randomness. Simulation results show that the method has higher prediction accuracy.