基于灰色自记忆模型的城市月需水量预测  

Forecasting monthly city water demand with gray self-memory model

作　　者：段海妮[1] 莫淑红[1] 沈冰[1] 韩海军[1] 张高锋 王积科 

机构地区：[1]西安理工大学西北水资源与环境生态教育部重点实验室,陕西西安710048 [2]西安市水务局,陕西西安710002 [3]宝鸡市水利局水资源处,陕西宝鸡721000

出　　处：《干旱区地理》2011年第4期591-595,共5页Arid Land Geography

基　　金：国家自然科学基金项目(50779052);陕西省教育厅重点实验室研究计划(08JZ55);陕西省自然科学基础研究计划(2010JM5001)

摘　　要：建立了以灰色微分方程为动力核的自记忆模型,并首次应用于城市月需水量预测。先将具有季节性特征的实测历史数据通过季节指数法获得相对平滑的序列,再用所建灰色自记忆模型进行模拟、预测。研究表明,灰色自记忆模型建模方法简单且模拟和预测精度较为满意,平均相对误差为5.74%,灰色GM(1,1)模型预测平均相对误差是8.77%,其精度比GM(1,1)模型有较大的提高。With a gray differential equation as its core,the gray-self memory model is widely used in hydro-meteorologic prediction,when the phenomena,such as river runoff,precipitation,evaporation etc.have long-term observed data series.In this study it is applied to simulate and predict the urban water demand for the first time.Baoji is the second large city of Shaanxi Province facing the water shortage blocking its social and economic development.In order to realize the rational and efficient use of limited water resources,it is necessary to predict the water demand based on the characteristics of water supply and demand of the city.The monthly water use data from 2000 to 2009 are used to establish model,of which the data from 2000 to 2008,totally 108 samples as the found of the model,and the data from the year 2009,totally 12 samples as prediction validation data.The monthly water use data should be smoothed first,because they have the feature of seasonal fluctuation.And after prediction is made,the results should be restored to the original mode.It is shown by case study that the presented method is simple and practical,and the simulation and predict results are rather satisfactory.The relative error of gray self-memory model is blow 10%,with the average relative error of 5.74%,the average relative error of gray GM(1,1) model is 8.77%,and the average error of gray GM(1,1) model with data untreated by seasonal exponent method is 10.47%.This can be seen that the accuracy of the gray self-memory model is better than that of gray GM(1,1) model.So the gray self-memory model used to predict urban water demand is successful and effectual,and it can give some valuable reference to water supply planning.The gray self-memory model used in the field of urban water demand prediction may enhance the technology in balancing the water supply and demand.

关 键 词：月需水量 灰色理论 自记忆方程 

分 类 号：P333.1[天文地球—水文科学]