CAJ | 학술논문

根据前人研究，上海地区地下水的水质变化主要受人工开采和回灌活动的影响，然而，由于地下水流动的复杂性,水质变化与采灌井的开采量及回灌量之间存在着完全非线性的关系；另一方面，人工神经网络技术具有较强的解决矛盾样本(即非线性问题)的能力，因此，作者在收集已有观测资料的基础上，利用人工神经网络技术对上海地区第III承压含水层的矿化度变化进行了预测。算例表明，应用该法进行单井预测的相对误差只有2.07%，多井预测的相对误差小于1.5%。
근거전인연구，상해지구지하수적수질변화주요수인공개채화회관활동적영향，연이，유우지하수류동적복잡성,수질변화여채관정적개채량급회관량지간존재착완전비선성적관계；령일방면，인공신경망락기술구유교강적해결모순양본(즉비선성문제)적능력，인차，작자재수집이유관측자료적기출상，이용인공신경망락기술대상해지구제III승압함수층적광화도변화진행료예측。산례표명，응용해법진행단정예측적상대오차지유2.07%，다정예측적상대오차소우1.5%。
According to previous study, the variety of grou ndwater qualityis mostly influenced by human activities of exploitation and rei njection in Shanghai. The correlation between these two factors is nonlinear, an d, artificial neural network has the ability to solve this problem. Therefore, b ased on collected data of groundwater quality, quantity of exploitation and rein jection，this article predicts the variety of salinity in III confined aquifer i n Shanghai area using artificial neural network. The results indicate that the r elative error is only 2.07% in single well prediction, and less 1.5% in multi-we ll prediction.