Load forecast calibration method for large-scale electricity-dependent Corporation

The load of large-scale electricity-dependent corporation consists of the power loads of each production units; it has quite different characteristics from the load pattern of a large region. When a major maintenance happens, the load is much lower than normal level, so the result of the basic load forecast model looks much higher, and need to be calibration. In this paper, a load forecast calibration algorithm based on maintenance schedule is proposed. The algorithm aims at improving the precision of load forecasting when production line maintenance happens. The relationship between the load abnormality and maintenance is obtained by the way of power line load analysis, and calibration coefficients of each maintenance items are fixed using constrained least squares algorithm. The proposed algorithm is tested on real data, and prospective results are obtained.