采煤和岩溶水开采对晋祠泉流量影响的随机模型分析（摘要）

冉曦，顾盼，连雷雷，左建，邱瑞明

（长江水利委员会水文局汉江水文水资源勘测局，湖北　襄阳　441022）

【摘 要】 以晋祠泉域为例，在泉域驱动因素的基础上，将泉域岩溶水系统概化为一个多输入-双输出的系统，输入项为当年及前七年的降水、岩溶水开采量和采煤排水量；输出项为晋祠泉出流量和侧排量。然后以多元线性回归和BP神经网络两种随机模型，建立泉域总排泄量（包含出流量和侧排量）与各输入项的关系，再假定无岩溶水开采和无采煤排水两种情景，在这两种情景下模拟泉域的排泄量，对比分析采煤和岩溶水开采对晋祠泉的影响。研究结果表明：多元线性回归模型，在1956—1994多年平均条件下，岩溶水开采减少晋祠泉排泄0.42m3/s，采煤排水减少出流0.23m3/s，人类开采总减少0.65m3/s；BP神经网络模型，1956—1994多年平均岩溶水开采减少泉域排泄0.30m3/s，采煤排水减少排泄0.27m3/s，人类开采总减少0.65m3/s。20世纪80年代后采煤活动加剧，两种随机模型均反映出采煤排水对泉域有严重的影响。

【关键词】 晋祠泉；多元线性回归；BP神经网络；采煤排水；岩溶水开采

作者简介：冉曦（1972— ），男，四川南充人，高级工程师，主要从事水文泥沙分析研究。

E-mail：hjzhanghx@cjh.com.cn(www.zuozong.com)

The Impact of Coal Mining and Karst Water Exploitation Against the Flows of Jinci Spring

RAN Xi，GU Pan，LIAN Leilei，ZUO Jian，QIU Ruiming

（Bureau of Hydrology and Water Resources of Hanjiang River，Hy drology Bureau of Changjiang Water Resources Commission，Xiangyang Hubei Province 441022）

Abstract：We took the Jinci spring area as a multi-input-dual output system，based on the analysis of the area' s various driving factors，and using multiple linear regression and BP neural network model，established these two models between the area's flow and various entries，then assumed two scenarios that non-coal mining and non-karst water drainage，next simulated the flows of Jinci spring area under these two scenarios，last analysis of the impact of coal mining and karst water exploitation.The results showed that：the multiple linear regression model，under the 1956—1994 annual average conditions，karst water exploitation reduced the flows of Jinci spring area 0.42m3/s，coal mining reduced the flow 0.23m3/s，the total influence with 0.65m3/s；BP neural network model，under the 1956—1994 annual average conditions，karst water exploitation reduced the flows of Jinci spring area 0.30m3/s，coal mining reduced the flow 0.27m3/s，the total influence with 0.65m3/s.Two stochastic models reflected a serious impact on Jinci spring area，when coal mining and karst water exploitation after 1980s.

Key words：Jinci Spring；Multiple Linear Regression；BP Artificial Neural Network；Drainage of Coal Mining；Karst Water Exploitation