摘要
地下水环境脆弱性具有模糊特性,现有的地下水环境脆弱性评价方法普遍采用加权评分法和模糊数学方法.加权评分法在评价因素权重的确定上人为性较大,并且该方法不能反映各评价因素指标值的连续变化对地下水环境脆弱性的影响;模糊数学方法在评价因素权重的确定和隶属度函数的构建上存在着不足.为此,建立了地下水环境脆弱性的改进BP神经网络模型.黄淮平原宁陵县的应用结果表明,改进BP神经网络法训练速度快、精度高,能较好地解决非线性的模式识别问题,如实地评价地下水环境的脆弱性.
The weight grade method and fuzzy mathematic method are commonly used for groundwater environmental vulnerability evaluation due to its fuzziness. However, the weight grade method is affected by some artificial factors in determining the weights of different evaluation factors, and the method cannot reflect the influences of the continuous changes of evaluation factors on the vulnerability of groundwater environment. Meanwhile, the fuzzy mathematic method is of some disadvantages in determining the weights of evaluation factors and establishment of the function of subordination degree. Thus, a modified BP neural network model was established for groundwater environmental vulnerability evaluation. The application of the model to Ningling County in Huang-Huai Plain shows that the modified BP neural network is of fast speed and high accuracy in nonlinear mode identification and can provide correct evaluation of groundwater environmental vulnerability.
出处
《河海大学学报（自然科学版）》
CAS
CSCD
北大核心
2007年第3期245-250,共6页
Journal of Hohai University(Natural Sciences)
基金
国家"973"重点基础研究发展规划项目(G1999043608)
陕西省重点实验室项目(02JS37)
关键词
地下水环境
脆弱性评价
改进BP神经网络
模式识别
groundwater environment
vulnerability evaluation
improved BP neural network
model identification
作者简介
李梅（1963-）,女,山东寿光人,高级工程师,博士研究生,主要从事水文水资源研究.
