基于遗传算法的BP神经网络模型在地下水水位预测中的研究——以泰安市城区旧县岩溶水水文地质单元为例  

Research on BP neural network model based on genetic algorithm in groundwater level prediction -- Taking the hydro-geological unit of karst water in Tai′an-Jiuxian County as an example

作　　者：王琪 WANG Qi(The Third Exploration Team Of Shandong Coalfield Geologic Bureau,Shandong Taian,271000)

机构地区：[1]山东省煤田地质局第三勘探队,山东泰安271000

出　　处：《地下水》2024年第6期11-13,17,共4页Ground water

基　　金：泰安市城区-旧县岩溶水水文地质单元基于BP神经网络的地下水水位动态预测模型(鲁煤地科字(2022)31号);泰安市地下水超采区评价报告编制服务采购项目(ZFZB2021-06-03)。

摘　　要：地下水水位变化是一个受多种因素影响高度复杂的系统,为合理预测地下水水位,本次以泰安市城区-旧县岩溶水水文地质单元为例建立遗传算法的BP神经网络模型,对地下水水位进行预测。为保证模型对地下水水位变化趋势精确预测,选择2015年1月-2017年12月研究区的月地下水开采量、月降水量、河流入渗量作为输入层,地下水水位作为输出数据,将2017年1月-2020年12月数据作为检验样本,与实际数据进行对比。对比结果表明:遗传算法优化BP神经网络训练阶段和测试阶段所有所有监测井的QR值均大于90%,且MSPE值小于0.05,拟合效果为“好”,说明模型对于所有监测井均表现出很好的拟合效果。因此该模型可为地下水研究提供一种有效的地下水水位预测方法,具备一定应用前景。The change of groundwater level is a highly complex system affected by many factors. In order to reasonably predict the groundwater level, this paper takes the hydrogeological unit of karst water in Tai 'an City-Jiuxian County as an example to establish a BP neural network model of genetic algorithm to predict the groundwater level. In order to ensure the accurate prediction of the change trend of groundwater level by the model, Monthly groundwater exploitation, Monthly precipitation and river infiltration in the study area from January 2015 to December 2017 were selected as the input layer, and the groundwater level was used as the output data. The data from January 2017 to December 2020 were used as test samples and compared with the actual data. The comparison results show that the QR values of all monitoring wells in the training stage and test stage of BP neural network optimized by genetic algorithm are greater than 90 %, and the MSPE value is less than 0.05, and the fitting effect is ' good ', indicating that the model shows good fitting effect for all monitoring wells. Therefore, the model can provide an effective groundwater level prediction method for groundwater research, and has certain application prospects.

关 键 词：GA-BP神经网络 地下水水位 预测模型 泰安市 

分 类 号：P641.74[天文地球—地质矿产勘探]