论文部分内容阅读
利用随机过程及时间序列分析手段，根据用水量序列季节性、趋势性及随机扰动性的特点，建立了用水量预测的实用组合动态模型。解决了标准日与周末用水量预测的衔接问题，利用加权递推最小二乘法（ＲＬＳ）进行动态参数估计，因此可很好地满足实时控制的需要。该方法经实例验证，预测误差较小，适用性强，可直接应用于供水系统的调度控制中。
By using stochastic process and time series analysis methods, a practical combined dynamic model of water consumption forecasting was established based on the characteristics of seasonal, trending and random disturbances of water consumption series. It solves the convergence problem between standard daily and weekend water consumption forecasts, and uses weighted recursive least squares (RLS) to perform dynamic parameter estimation. Therefore, it can well meet the needs of real-time control. The method is verified by an example, the prediction error is small, and its applicability is strong. It can be directly applied to the dispatching control of water supply system.