智能网络化水资源的实时预测分析  

WATER RESOURCE QUANTITY FORECAST IN REAL-TIME BY NEURAL NETWORKS

作　　者：倪红珍[1] 贾传义 王浩[3] 张庆华[1] 王仲发[1] 

机构地区：[1]山东农业大学水利土木工程学院,山东泰安271018 [2]东平县水资源办公室,山东东平271500 [3]中国水利水电科学研究院,北京100044

出　　处：《山东农业大学学报（自然科学版）》2004年第2期231-237,共7页Journal of Shandong Agricultural University：Natural Science Edition

摘　　要：在水资源严重短缺地区 ,对水资源较准确可靠的实时预测预报是充分利用当地有限的水资源、实施地表地下水资源的联合调度和最优配置极为重要的前提和条件。本文利用BP神经网络的纠错性、智能性 ,对复杂的水资源产汇流黑箱模型进行预测计算 ,并应用于太子河流域枯季的水资源实时监测监控系统 ,较传统的统计回归计算方法 ,其计算速度与拟合优度都有较大的提高与改进。In the area where water resources is very short, it is very important to forecast water resource quantities in real time accurately for utilizing water resources efficiently and for the optimum ground-underground water united-plan and allocation. Making use of the ability of BP neural networks-correct errors and its aptitude, this paper simulates the complicated runoff-yield and confluence process which is called black-box model. Later, the model was tested in the water resource forecasting and plan of Taizi drainage area of Liaoning province in dry season. Comparing with the traditional regression analysis, the method has much more superiority in calculating speed and precision.

关 键 词：智能网络化 水资源 实时预测模型 入库径流 BP网络算法 

分 类 号：TV213.4[水利工程—水文学及水资源]