| 摘要: |
| 【目的】 探求洛惠渠灌区地下水的动态变化情况，为该灌区地表生态环境的改善提供科学依据。【方法】 以陕西洛惠渠灌区实测数据为例，采用3层BP网络建模方法，对灌区自然-人工-生物条件下地下水位动态进行研究，运用缺省因子检验法分析各因子对地下水埋深影响的敏感性；并采用灰色关联方法分析各因子与地下水埋深的关系。【结果】 人工神经网络模型具有较高的精度，能够定量描述地下水位动态与各因子之间的响应关系，蒸发量是影响该灌区地下水位动态的主要因子，各因子之间相互作用，形成了复杂条件下的耦合关系；灰色关联排序结果基本符合敏感性因子分析结果，很好地验证了各因子的敏感程度。【结论】 将这2种方法结合运用到灌区地下水位动态评价中是切实可行的，是对传统地下水位动态研究方法的补充与完善。 |
| 关键词: 洛惠渠灌区 地下水位动态 人工神经网络 敏感性因子 灰色关联分析 |
| DOI： |
| 分类号: |
| 基金项目:国家科技支撑项目（2006BAD09B02）；国家自然科学基金项目（40771124）；中日合作项目“沙漠化防治规划研究”（SBS-379）；西安理工大学优秀博士学位论文基金项目（106-210911） |
|
| Simulation study on groundwater level dynamic in Luohuiqu irrigation district |
|
|
| Abstract: |
| 【Objective】 It can provide scientific basis for the improvement of the ecological environment to explore the dynamic variation of groundwater in Luohuiqu irrigation district. 【Method】 Groundwater level dynamic under natural artificial-biological conditions was studied with measured data as example based on application of BP network of three layers.The additional momentum method and self adaptive tactic for training were adopted;sensitivity analysis about groundwater depth was conducted according to each factor by using default factor method;and the grey correlation analysis method was applied to analyze the relations between factors and groundwater depth.【Result】 The results show that the artificial neural network model can express quantitatively the response relationship between groundwater level dynamic and various factors with sufficient high accuracy.The evaporation is the main factor affecting groundwater level dynamic in this irrigation district,and the interrelationship amongst various factors form coupling relationship under the complicated condition.The ranking results correspond to the analysis results about sensitive factors basically,and the grey correlation analysis method can also verify the sensitivity degree amongst various factors.【Conclusion】 It is feasible and practical to combine the two methods in the appraisal of groundwater level dynamic in irrigation district,which supplements and perfects the traditional research methods of groundwater level dynamic. |
| Key words: Luohuiqu irrigation district groundwater level dynamic artificial neural network sensitive factor grey correlation analysis |
