基于主影响因素的城市时用水量预测  

Urban hourly water demand prediction based on principal factors analysis

作　　者：麦志彦[1] 何中杰[2] 汪雄海[1] 

机构地区：[1]浙江大学电气工程学院,浙江杭州310027 [2]杭州电子科技大学自动化学院,浙江杭州310018

出　　处：《浙江大学学报（工学版）》2012年第11期1968-1974,2006,共8页Journal of Zhejiang University：Engineering Science

基　　金：国家"973"重点基础研究发展规划资助项目(2009CB320602)

摘　　要：针对城市用水量影响因素的多样性、随地域时间多变的特点及与时用水预测时间尺度难匹配的问题,提出基于主影响因素的城市时用水量预测方法。该方法先以改进的灰色关联法对用水量影响因素进行分析讨论,把得到的主影响因素作为算法输入量对预测日用水量做预测;而后针对预测日的时用水特征,在动态模糊聚类中引入聚类中心距离评价函数及非线性约束条件,寻求预测日的各时段水量分配模式,并依据该模式做时用水量的最终预测及合理调度。用萧山南阳镇历史用水量为例,做预测仿真验证,结果表明,该方法能够很好跟踪预测日的时用水量变化,稳定跟踪能力及跟踪精度上远优于传统时间序列预测法,易满足供水系统的调度精度需求。It was difficult to find the principal influencing factors on urban water demand because of their diversity,geographical time variability and difficulty in matching the time-scale with hourly water demand,so the hourly water demand forecasting method based on principal influencing factors was proposed.First,the improved grey relation analysis method was used to analyze the influencing factors of water demand,and the principal influencing factors obtained were used as the algorithm inputs to predict the daily water demand.Then,according to the characteristics of the hourly water demand in prediction days,the dynamic fuzzy clustering method with distance evaluation function of cluster centers and nonlinear constraints was put forward to recognize different hourly water demand modes,in order to make the final prediction and reasonable scheduling of hourly water demand.Case Simulation was presented with historical water demand of Xiaoshan Nangyang town.The results indicate that the method has good behaviors of hourly water demand forecasting,far better than the traditional time series forecasting method in tracking stability and precision,and is easy to meet the accuracy needs of water supply system scheduling.

关 键 词：主影响因素 时用水量预测 灰色关联分析 动态模糊聚类 

分 类 号：TP273.1[自动化与计算机技术—检测技术与自动化装置]