Research Based On Fuzzy Comprehensive Method And BP Neural Network For Water Quality Evaluation Of Landscape River

As we all know, water is the primary natural resource that is indispensible in mankind survival and society development. But with the development of the society,water shortage and water pollution are gradually becoming great international issues. As a very important part of the study of water environment, water quality assessment has played an important role in water management and maintenance.In this paper, after long-time and high-density sampling and monitoring of the landscape river in the Urban Area of Tianjin, based on these practical monitoring dates, and uniting both fuzzy comprehensive models and the theory of BP neural network, the pollution characteristics of this river has been analyzed. And then, with an optimized fuzzy comprehensive assessment method, as well as the water quality evaluation method based on BP neural network, the present water quality of the primary and secondary river channel and drained channels has been evaluated. Meanwhile, a brief analysis has been made to the reasons for present water quality. It showed that:the principle of superposition membership should be used in the fuzzy comprehensive assessment method; Considering the TN index exceeding seriously, the situation that contains factor TN and the situation without it has been separately evaluated; The indexes of the primary, secondary and drained river channels were poor from July to September, and the results of the two evaluation methods were both Class â…¤.While in other months, except index TN, the indexes were preferable, and it showed that the results of evaluation on the situation without factor TN were better than Class â…£; The percentages of the Class â…¤ results of evaluation on the primary and secondary river channels were 84.06% and 71.74%, the percentages of results superior to Class â…¤ were 15.94% and 28.26%; Because of the function of drained channels, the indexes of water quality in these channels exceeded seriously.Comparing the two methods of evaluating water quality, a conclusion from this study is that the results of the fuzzy comprehensive assessment method are basically identical with that of evaluation method based on BP neural network. The two water quality evaluation methods are both comprehensive evaluation methods for they consider each water quality index. Owing to its network self-learning ability, the results of evaluation method based on BP neural network can be more objective and accurate, which makes this method be the first choice for water quality evaluation.Accordingly, the paper can provide theoretical basis for water quality analysis of the landscape river in the Urban Area of Tianjin and offer data support for water harnessing and quality research, and at the same time, we hope it may offer some useful reference function to water quality analysis in other districts.