基于多种不确定因素相似度匹配的母线负荷预测算法

A Bus load Forecasting Algorithm Based on Similarity Matching of Multiple Uncertain Factors

母线负荷预测是合理安排生产调度计划和实施节能发电调度的保障。对母线负荷进行了精细化分解,通过考虑历史样本数据的类型,以及影响负荷变化的星期类型、气象类型、运行方式变化等外在因素,提出了基于多种不确定因素相似度匹配的母线负荷预测算法,在此基础上,利用逐点分配因子法实现各种母线负荷成分的剥离。通过算例验证得出:采用该方法预测的平均日母线负荷预测准确率和合格率均符合考核标准要求,工业行业负荷的预测准确率达到95%以上。由此证明了该方法具有可行性和实用性。

Bus load forecasting is the guarantee to the reasonable arrangement of production scheduling and the implementation of energy-saving power generation dispatching. The refined decomposition of bus load forecasting is accomplished, and a bus load forecasting algorithm based on similarity matching of multiple uncertain factors is proposed by considering the history sample data type, external factors including week type, weather type, operation mode change and so on which influenced the change of load. On that basis, the stripping of every bus load component is realized by using point by point distribution factor method. It is concluded by example verification that the prediction accuracy rate and qualification rate of average daily bus load predicted by using the method satisfied the requirements of assessment criteria, and the prediction accuracy rate of industrial load was above 95%, which proves the feasibility and practicality of the method.