基于粗糙集和支持向量机的水电站发电调度规则研究

Study on hydropower station operation rules based on rough sets and support vector machine

针对梯级水电站联合运行调度的实际需求,基于粗糙集和支持向量机理论提取水电站发电调度规则。首先在梯级水库确定性优化调度结果基础上,构建由时间因子、空间因子和能量因子构成的属性集;然后采用粗糙集理论去除冗余属性,降低模型复杂度;最后采用支持向量机回归模型拟合水库时段决策变量之间的非线性关系得到水电站发电调度函数,并模拟长系列发电调度运行过程。以金中梯级水电站群为对象进行实例研究,计算结果表明:相比多元回归、全变量支持向量机模型模拟运行结果,本文提出方法得到的调度函数指导梯级运行时更多保留了确定型优化调度的发电效益,体现了方法的合理性与优越性,可以有效指导水电站水库中长期调度运行。

Rough sets and support vector machine technique is adopted to extract operation rules for the united scheduling of cascade hydropower stations. In this technique, we construct attribute sets of time factors, space factors and energy factors using deterministic optimal operation results, and then apply the rough sets theory to remove redundant attributes. The last step is to use a support vector regress model to construct a hydropower operation function for simulation of the long series of operation scheduling process. Comparative analysis of the simulation results shows that, relative to other operation functions methods, the proposed function maintains good performance of optimal operation and it would provide a guide for the mid- and long-term effective operation scheduling ofhydropower stations.