粗糙集法解多环境因素影响的母线负荷预测问题

Solution of Multi Environmental Factor-Influenced Bus Load Forecasting by Rough Set Method

母线负荷的环境因素众多且影响关系复杂,综合考虑所有属性会引入无关随机信息,降低预测精度。为此,提出了考虑多环境因素的母线负荷预测粗糙集(rough set,RS)方法,采用快速属性约简算法(fast attribute reduction algorithm,FARA)确定对母线负荷影响较大的条件属性;基于概率规则导出决策规则集;通过距离度量法匹配规则,从而实现母线负荷预测。将所提方法应用于预测广西某地区电网220 kV母线有功负荷,结果表明该方法能从母线负荷预测的历史数据样本挖掘出有益预测规则,具有较高的预测精度。

There are numerous environmental factors influencing bus load and the relation among these factors are complex, and the forecasting accuracy of bus load will be reduced due to leading in irrelevant random information while all attributes of environmental factors are taken into account. A rough set （RS） method for bus load forecasting, in which multi environmental factors are considered, is proposed. The fast attribute reduction algorithm （FARA） is adopted to determine condition attributes influencing bus load evidently. Based on probabilistic rules the decision rule set is derived, and then bus load forecasting is implemented by matching rules based on distance measure method. Applying the proposed method in the active power forecasting of a 220kV bus in a certain regional power network in Guangxi power grid, the forecasting results show that using the proposed method the prediction rules beneficial to the forecasting can be mined from historical data samples of bus load forecasting, and a higher forecasting accuracy can be attained.