A power load forecast approach based on spatial-temporal clustering of load data

摘要

Load forecast is very important for power system operation. Pursuing higher accuracy of loadforecast is always the major target of this field. The existence of bad historical load data couldbadly affect the forecast accuracies of time series–based load forecast techniques. Within themulti-node load data, the outliers of them are merely appeared instantaneously; the bad impactof the outliers could be decreased by making use of the spatial relativity of multi-node load data.Anovel load forecasting approach based on spatial-temporal feature clustering is proposed in thispaper. The temporal regular load pattern is extracted from the total load for an individual node.The spatial distribution characteristics of the individual incremental load have been categorizedby the k-medoids clustering algorithm. The MapReduce computing mode is used to manipulatemulti-node load data to raise computation efficiency. This new forecasting approach could reducethe influence of outliers and provide reliable and efficient load forecast with high accuracy. Testingin a real power system data set with up to 6.6% of outliers, the results of this approach showa lower forecasting error, about 6.2%, compared with three other time series–based methods(between 9.8% and 10.6%).