考虑经济发展趋势的中长期电力负荷SALSSVM预测

Medium and Long Term Power Load Forecasting Based on SALSSVM Model Considering Economic Development Trend

准确的中长期电力负荷预测对电力系统的合理规划具有重要作用。考虑了经济发展趋势,提出了一种应用模拟退火算法(SA)优化最小二乘支持向量机(LSSVM)的中长期负荷预测新方法(SALSSVM)。首先将反映经济发展趋势的指标和历史负荷数据作为输入变量;其次运用SA优化选择用于负荷预测的LSSVM模型最优参数值;最后将该方法与未考虑经济发展趋势或未经SA优化的LSSVM预测方法进行对比。实例验证结果表明,考虑经济发展趋势并经SA优化的LSSVM模型具有更高的预测精度。该方法是有效可行的。

Accurate medium and long term power load forecasting plays an important role in the rational planning of power system. Considering economic development trend, a new load forecasting method （SALSSVM） is proposed that simulated annealing algorithm （SA） is applied to optimize the least squares support vector machine （LSSVM） model. Firstly, the indicators that reflect the economic development trend and historical load data are as the input variables; secondly, SA is used to select the appropriate parameters values of LSSVM model; finally, this method is compared with the LSSVM model that doesn’t consider the economic development trend or isn’t optimized by SA. The calculation result shows that the LSSVM model that is optimized by SA and considered the economic development trend has the highest forecasting accuracy. This proposed method is feasible and effective.