Study On Research And Application Of Water Quality Assessment And Prediction Method In The Typical Villages And Towns

With the rapid development of national economy and people’s living standards improve, environmental pollution problem is getting more and more attention,but in the past environmental pollution control was mainly concentrated in large cities and large water bodies,and had little attention on villages and towns. Base on The National Key Technology R&D Program during The Twelfth Five-year Guideline Period “Rural environmental monitoring technology integration and application demonstration”, in this paper,with aquaculture water and receiving water as object of the study, deep research methods of evaluation and prediction of water quality, the main results are as follows:(1) In view of defects of the traditional principal component analysis in the evaluation of water quality,from two aspects of data standardization and the consistency of feature vectors to improve the traditional principal component analysis, examples showed that the improved principal component analysis method was more practical and more objective.(2) Through the analysis of fuzzy comprehensive evaluation’factors, then the combination weighting method and weighted average principle was put forword. the combination weighting method,which effectively combining the analytic hierarchy process and entropy weight method,not only consider the influence of the maximum pollution index on water quality, but the interactions between different pollution indexes, made the weight results more objective and reasonable. By using the multiplicative additive operator and weighted average principle, the loss of data information was avoided in the process of fuzzy evaluation, and the evaluation results was more visual and more reasonable.(3) The water quality prediction model based on SVR was established, in the process of traditional SVR parameter selection, the workload is calculated, and the optimal parameters waste long time and difficult to be determined,so a SVR parameter optimization method based on genetic algorithm is proposed.Taking DO、COD and TN of Qiao Kou Town aquaculture water and sewage water as samples in a time series, tectonic SVR prediction model, and with BP prediction model are compared from the forecasting error and the trend curve to analyze their prediction performance.(4) According to the sudden, real.time and complexity of the characteristics of pollution, the single factor and multi factor state early warning model of the aquaculture water and sewage water was established, and completed the determination of the police intelligence category,at last in view of the problem that the water environment is easy to occur, the paper puts forward measures and countermeasures.