基于回归分析和灰度模型GM(1，1)的民勤县需水量预测

Water Demand Forecast Based on Regression Analysis and Gray Model GM(1,1) in Minqin County

解决水资源供需矛盾的关键在于准确进行需水预测。为解决民勤县的水资源供需矛盾,对民勤县2008—2014年的用水量结构及特征进行分析,并分别使用回归分析法和灰度模型GM(1,1)对2015年的分项需水量和总需水量进行预测,与2015年实际用水量进行比较。结果表明:回归分析法和灰度模型GM(1,1)预测精度良好,与实际值的相对误差均小于1%,各分项用水量的预测均通过差异性检验。与灰度模型GM(1,1)相比,回归分析法具有较高的精度和便利的操作性。2种模型预测2018年民勤县的需水量分别为33 926万m3和34 236万m3,分项用水量预测结果差异性小,可为民勤县水资源合理配置及供水系统优化调度提供参考。

The key to solution of conflict between supply and demand of water resources is to forecast accurately water demand. To elimi-nate this conflict in Minqin County,structure and characteristics of water consumption in Minqin in 2008-2014. Furthermore,both re-gression analysis method and gray model GM（1,1） are applied respectively to forecast partial water demand and total water demand in2015. The forecast result is compared with the actual water demand in 2015. The comparison demonstrates that the forecast precision ofboth regression analysis method and gray model GM（1,1） are excellent. The relative errors of the forecast values to the actual value bothare lower than 1%. The forecasts of partial water demands all pass difference testing. Compared with the gray model GM（1,1）, the re-gression analysis method is with high precision and convenient operation property. Through the forecast by these two models,the water de-mands in Minqin county in 2018 are 339. 26 million m3 and 342. 36 million m3 respectively. The forecast result of the partial water de-mand is small in deference. This can provides the rational arrangement of water resources and the optimum dispatch of water supply sys-tem in Minqin county with reference.