多项式回归模型在径流式水电站动态出力系数预测中的应用

Application of Polynomial Regression Model in Predicting Dynamic Output Coefficients of Run-of-River Hydropower Stations

文章针对目前水电站出力系数采用设计运转特性曲线插值偏差较大以及多因子推导模型建模难度大的问题,通过分析径流式水电站需要维持出入库平衡的运行特性,将出力系数的影响因子简化至入库流量单一因子,并以卡洛特水电站为例,进行多项式建模分析,并根据运行机组数量将模型分段拟合。通过模型计算的出力系数相对曲线插值得到的出力系数平均偏差为-1.31%,符合预期,表明该模型性能可靠,精度较高。该方法进一步实现了发电计划制作的自动化处理,提升了卡洛特水库运行管理效率,在径流式水电站具有一定的推广应用价值。

In response to the significant deviation observed when using interpolation of design operational characteristic curves for hydropower station output coefficients,as well as the complexity of modeling with multiple factor-derived models,this paper simplifies the factors influencing output coefficients to a single factor:inflow discharge,considering the need for run-of-river hydropower stations to maintain a balance between inflow and outflow.Taking the example of the Karot Hydropower Station,a polynomial modeling analysis is conducted,segmented to fit the model according to the number of operating units.The average deviation of the output coefficients calculated by the model relative to those obtained from curve interpolation is-1.31%,meeting expectations and indicating reliable performance and high accuracy of the model.This method further automates the generation of power generation plans,enhancing the operational management efficiency of the Karot Reservoir,and holds potential for broader application in run-of-river hydropower stations.