A New System Of Water Quality Synthetic Evaluation And Its Application In Water Quality Early Warning

The Water Resource is a kind of fundamental natural resources and the material basis for the biological survival. As the controlling elements of the ecological environment, it is more the indispensable natural resources of human civilization development. So the water quality status of rational, scientific and comprehensive evaluation for the rational utilization of water resources, control of water pollution, reducing the life and property loss caused by the misuse of water is of great importance. The water quality synthetic evaluation is qualitative or quantitative assessment and classification of the quality of the water resources situation according to various applications of water and water quality evaluation standard. The water quality synthetic evaluation is significant with regard to water resources assessment, it provide important gist for development, utilization and protection of water resources.The water quality synthetic evaluation method include the single pollution index method and multiterm pollution index method. With the development of computer and mathematics, more and more mathematic model and methods are applied widely, the method of water quality synthetic evaluation are further extended, but water itself is diverse and complex, many factors that affect water quality with different rules, this caused the water quality synthetic evaluation is of uncertainty and fuzziness. So far no one kind of method is widely accepted and used, it is constantly in the status of development and improvement. Sorting algorithms is a supervised learning, which find classification rules to predict the category of the new data through the analysis of the training set which the single sample is known categories. In this paper, we tested results of different sorting algorithms and selected K-star model finally. What’s more, we reduce the complexity of the algorithm and improve the efficiency of classification by improving this model according to the characteristic of five parameters.Due to the population expansion and the rapid development of social economy, sudden water pollution incident happened occasionally. From the current situation of our country’s environmental statistics, a lot of incidents is statistical data after the event. The lack of advanced information support caused the lack of effective and timely warnings before the events. Water quality early warning is refers to analyse and evaluate the water quality condition within certain time and space. It use the water quality prediction model to forecast the water quality on the basis of long-term monitoring combined with the influence of the environment and human factors, when there is the possibility of water quality deterioration, the system can make the corresponding early warning and prevention countermeasures. In this paper, we established early warning system with the accumulation of long-term monitoring data of related rivers on the basis of new water quality synthetic evaluation method.The five parameters we selected is relatively mutual independence and can most represent water quality, it is easy to detect and can realize instrument integration.These all provide possibility of real-time on-line tracking detection toward different water quality distributed in different sites and provide timely and effective information and technology services for the prevention and treatment of sudden, severe accidents.The dissertation includes four chapters:Chapter1PrefaceThe water quality synthetic evaluation is very important in water environmental monitoring, Water quality early warning system based on water quality synthetic evaluation is hopeful to timely and effectively warn before the water quality deteriorate, reduce the loss of life and property. Due to the complex and changeful of natural water itself, water quality synthetic evaluation have certain subjective tendency, so far there is no objective and unified water quality synthetic evaluation method. This chapter described rationale and the basis of subject in detail by introduce common water quality synthetic evaluation methods.Chapter2The study of relationship between five parameters and water qualityThe ingredients in natural water is complicated,each ingredients have different functions toward water quality. In order to study the influence of different substances in water to five parameters,we measured a lot of five parameters data of standard solution samples. Through the changing trend of five parameters with pollutant concentration,we can find that the five parameters have different response toward different pollutants,also,the methematical expression that can express the relationship of the two is different. So we can’t use simple linear equation、multiple regression equation、indicial equation or logarithmic equation to characterization quantitative relation between water quality and five parameters,it’s feasible to use the method of fuzzy mathematics.Chapter3The establishment of new water quality synthetic evaluation methodThe key step of establishing water quality synthetic evaluation method is finding out suitable mathematic model.The K-star model is a kind of sorting algorithms.Fristly, synthesis of a large number of samples simulated of natural water and detected five parameters of those water samples, then we get a database of3000sets of data, we called this database training documents.There are many different kinds of sorting algorithm, this chapter tests about ten kinds of sorting algorithm based on training document, we choose the most suitable model by relevant parameters, excessive and dilution experiments, the accuracy of predict standard solutions and so on. We also improve the K-star model by reduce algorithm complexity.We use the K-star model evaluate a kind of natural water samples, results shows that accuracy is high.Chapter4The establishment of water quality early warning systemWater quality early warning system can predict the time-space range and possible damage of abnormal water quality, hereby provide all kinds of alarm information and corresponding comprehensive countermeasures. What’s more, quick draw kinds and concentrations of pollutants, pollution range and potential hazards, warning the existing possibility of contaminant through five parameters when the sudden events happens is very important job in early warning system. Water quality early warning system needs accumulate long-term monitoring data. In this paper, we establish early warning system about Yingtao river of east china normal university and Huangpu river (the section of binjiang park). From October2012to January2013, we monitored the water quality of these two rivers, then use x-control chart to draw the warning figure based on lots of data. For the determination of warning pollutants, we start from the pollutants that appear in water pollution incidents frequently, test these pollutants one by one the range of parameter values with it’s concentration change, select the most appropriate pollutants as the warning pollutants.