基于改进模糊综合-指数平滑法的地下水水质评价和预测

地下水水质评价是进行地下水资源开发、管理的基础,地下水水质预测是对未来的地下水水质变化趋势进行预报,可为地下水污染防治提供科学的依据。在模糊综合评价中,分别采用聚类权法、加权平均算子和级别特征值法,对传统方法的权重确定、算子选择和结果向量分析等关键环节进行改进;在指数平滑预测中,根据拟合结果确定指数平滑的类型,采用试算法选取权系数,并应用于大沽河地下水源地水质评价和预测。评价结果表明,改进的模糊综合评价可解决传统方法的不适性问题,可根据级别特征值比较评价等级相等的水质,更具实用性和合理性;预测结果表明,李戈庄采区和仁兆-朴木采区未来十年水质将持续恶化,且通过卡方检验分析2016-2017年预测的误差,验证了预测方法的有效性。

Evaluation and Rediction of Groundwater Quality Based on Improved Fuzzy Synthesis-Exponential Smoothing

Groundwater quality evaluation is the premise of groundwater resources development and management, and groundwater quality prediction is the prediction of future groundwater quality change trend, which can provide scientific basis for groundwater pollution prevention and control. In fuzzy synthetic evaluation model, clustering weighted method, weighted average operator and rank eigenvalue method are respectively adopted to improve the key links of traditional methods, such as weight determination, operator selection and result vector analysis.In the exponential smooth prediction model, the model type is determined according to the fitting results, and the weight coefficient is selected by trial algorithm. which is applied to the groundwater quality evaluation and prediction of the groundwater source field of dagu river. The evaluation results showed that the water quality of the Yifeng-Nanshaliang mining area, the Ligezhuang mining area and the Renzhao-Pumu mining area became worse in turn, the main reason was that the nitrate exceeded the standard seriously(the highest was more than 2.52 times of 5 types of water). The evaluation results were consistent with the actual water quality, which proved the reliability of the model.The prediction results show that the water quality in the Ligezhuang mining area and the Renzhao-Pumu mining area will continue to deteriorate in the next decade. In addition, the chi-square test is adopted to analyze the error of the prediction in 2016-2017, which shows that all statistics are less than the critical value of 7.815, which verifies the validity of the prediction model.