地区电网负荷预测的灰色Verhulst与系统动力学组合模型

To solve the risk of large forecast errors of a single model and the inability to make full use of historical data in the traditional regional power grid load forecasting, this paper proposes a linear combination model of regional power grid load forecasting, i.e. combination model of grey Verhulst and system dynamics. Taking the historical data of social electricity consumption as the original data, the grey differential equation and the whitening differential equation are successively established and solved to obtain the load forecast time series based on the grey Verhulst model, which is adopted to predict the load growth in accordance with the S-shaped curve or the load growth in saturation stage. Comprehensively considering the impacts of economy, population, energy substitution and re-electrification on social electricity consumption, the economic subsystem, population subsystem, energy substitution and re-electrification subsystem and power consumption subsystem of load forecasting are established, then a load forecasting model based on system dynamics is obtained, which is suitable for load forecasting with complex structure, rich original information and close connection among the subsystems. Without increasing complexity, single model is linearly combined following the minimum variance criterion to establish a combined model for regional power grid load forecasting. All areas in Zaozhuang city are employed to analyze the calculation examples. The results show that after making full use of the historical load data, the prediction accuracy of 71% of the areas is higher than that of a single model; after considering the minimum variance criterion to linearly combine the single model, 29% of the areas encounter lower risk of large prediction errors than a single model.