基于BP神经网络的岩溶水库渗漏评估——以贵州林歹迎燕水库为例  

Leakage Evaluation for Karst Reservoirs through Application of BP Neural Network——A Case Study of Yingyan Revervoir of Lindai in Guizhou Province

作　　者：彭三曦[1,2] 李义连 单慧媚[3] PENG Sanxi;LI Yilian;SHAN Huimei(School of Environmental Studies,China University of Geosciences,Wuhan 430074,China;College of Earth Sciences,Guilin University of Technology,Guilin 541004,China;3.College of Environmental Science and Engineering,Guilin University of Technology,Guilin 541004,China)

机构地区：[1]中国地质大学(武汉)环境学院,湖北武汉430074 [2]桂林理工大学地球科学学院,广西桂林541004 [3]桂林理工大学环境科学与工程学院,广西桂林541004

出　　处：《安全与环境工程》2018年第2期1-6,共6页Safety and Environmental Engineering

基　　金：国家自然科学青年基金项目(41502232)

摘　　要：岩溶地区修建水库的渗漏问题是水库能否正常运行的瓶颈。岩溶区水库渗漏的影响因素复杂繁多,难以用具体的数学模型建立各参数之间的非线性关系,以定量评估水库的渗漏量。以贵州岩溶地区林歹迎燕水库为例,通过收集57个同类水库渗漏实例,在详细分析库区水文地质条件的基础上,构建了BP神经网络模型,利用其反向传递并自我修正误差的功能优势,定量预测评估迎燕水库的渗漏量。结果表明:迎燕水库的渗漏量约为0.003m3/s,与库区实际监测结果基本一致,表明BP神经网络模型可以有效地应用于岩溶地区水库渗漏量的评估,可为岩溶水库渗漏评价提供一种新思路,对我国岩溶地区水库修建的选址以及对邻近矿区地下开采的影响评估具有重要的参考价值。Many complex factors influence karst reservoir leakage,and it is difficult to establish the nonlinear relationship of these factors and to evaluate leakage amount by using a certain mathematical model.Therefore,the leakage is the key bottleneck for the normal operation of the reservoir.Taking Yingyan reservoir of Lindai in Guizhou province as an example,this paper collects the leakage data from 57 similar reservoirs and analyzes the hydrogeological condition of Yingyan reservoir.Then the paper constructs a BP neural network model to predict Yingyan reservoir leakage by taking advantage of its function of back propagation and self-correcting.The results show that the leakage of Yingyan reservoir is 0.003 m 3/s which is basically consistent with the actual monitoring results,indicating that the BP neural network model can be applied to predicting the leakage quantity of karst reservoir effectively and providing a new idea for leakage evaluation.The study gives a reference for the site selection of the reservoir construction in karst areas and the evaluation of underground mining in nearby mining areas.

关 键 词：岩溶地区 水库渗漏 BP神经网络 定量评估 迎燕水库 

分 类 号：X141[环境科学与工程—环境科学] TV697.32[水利工程—水利水电工程]