基于ARIMA模型对东江水源需水预测研究

Research on water demand prediction of Dongjiang water source based on AMIRA model

在取水系统的科学决策与现代化管理中,水量预测是关键环节,其准确性直接关系到调度方案的可靠性。深圳市作为全国严重缺水城市之一,以提高水资源的抗风险能力和科学调度水平为首要任务。引入西江引水工程,旨在缓解东江水源工程的供水压力;通过ARIMA时间序列模型对水厂端用水需求进行预测,以期构建一个高精度的组合预测模型,为城市供水安全提供更加快速、有效的保障。结果显示,ARIMA模型的平均相对误差为3.72%,且误差超过5%的情况较少,表明模型具有较好的预测效果。

In the scientific decision-making and modern management of the raw water intake system,water quantity prediction is an indispensable part,and the accuracy of its predicted values is closely related to the reliability of the scheduling plan.There are no major rivers or reservoirs in Shenzhen,and the water resources in the area are extremely scarce.The water supply in the city is in a tight saturation state.Therefore,this article introduces the Xijiang Water Diversion Project to alleviate the water supply pressure of the Dongjiang Water Source Project.The ARIMA time series model is used to study the water demand of the water plant end.The ARIMA neural network model can handle the characteristics of time series data well,with an average relative error of 3.72%,and there are few cases where the relative error exceeds 5%,indicating good prediction results.