残差修正法在电力负荷预测中的应用

APPLICATION OF RESIDUAL ERROR AMENDMENT METHOD IN POWER LOAD FORECASTING

影响电力负荷的因素很多 ,而且这些因素都具有不确定性 ,即这些信息具有模糊性。因此 ,为了准确进行负荷预测 ,最好采用模糊预测来研究和处理电力负荷预测。一般对于电力系统中期负荷预测采用回归分析模型 ,但其预测结果往往有很大误差。为了提高电力负荷预测技术的水平 ,作者以基于实数输出值的模糊回归分析及用三角函数拟合残差的方法 ,提出了预测电力负荷的模糊线性回归——残差修正预测模型。该模型是在模糊线性回归模型的基础上推导出来的 ,它可以寻找最合适的线性函数使理想线性回归中的线差和达到最小。

There are lots of factors influencing power load and all of them possess indetermination. Therefore, it would be the best to research and process the load forecast by fuzzy forecasting. Generally, comparing with actual data of load the middle term load forecasted by regression analysis method often contains notable error. To improve the accuracy of load forecasting, by use of residual error amendment forecasting model based on real number output and fitting residual error by trigonometric function, a fuzzy linear regression— residual error amendment model is put forward. The residual error amendment model is derived from fuzzy linear regression model, it can find the most suitable linear function to make the line difference sum in ideal linear regression minimum. The correctness and feasibility of the presented model are validated by the simulation calculation of fuzzy forecasting of electricity sale.