饮用水源水质评价的模糊聚类模型的研究

Establishment of Fuzzy Clustering Analysis Model in Drinking Water Source Quality Assessment

考虑到饮用水源样本并没有严格的属性,以及它们在形态和类属方面存在着中介性,确定了对人类生理健康影响较大的高锰酸盐指数、生化需氧量、总氮、氨氮、氟化物、总磷、溶解氧等9项评价因子;将特征空间中一组没有类别标记的矢量按某种相似性准则划分到若干个子集中,实现每个子集代表整个样本集的某个或者某些特征和性质;建立起基于模糊聚类法的饮用水源水质评价模型。最后,将所建模型应用到阜新市目前饮用水源地水质评价中,应用效果表明:所建模型既能有效地判断饮用水源水质类别,又能确定水体中主要污染物和主要污染源,通过不同评价方法结果的比较分析,体现了所建模型的良好有效性和适用性。

In view of no-strict attribute and betweenness in drinking water source form and class attribute,nine assessment factors which had much more influence on human health were identified, such as permanganate index, biochemical oxygen demand, total nitrogen, ammonia nitrogen, fluoride, total phosphorus and dissolved oxygen. A vector without any identification code was set to several subset based on its similari- ty. Due to above solution, the function which every subset could stand for some character of whole sample set was carried out. Then fuzzy clus- tering analysis model in water source quality assessment was established and applied to main water source quality assessment in Fuxin city, which results showed that not only could the new model identified the source water quality level, but found out the main pollutants and main pollution sources. The final analysis between the results of different assessment methods reflected that fuzzy clustering analysis model in drinking water source quality assessment had favorable validity and applicability.