Surface Water Quality Evaluation And Monitoring Management System Based On Hyper-sphere Support Vector Machine

With the rapid development of China economy,human demand for water is constantly increasing.Large quantities of domestic sewage and industrial waste water are randomly discharged,which has caused serious pollution of surface water and the deteriorating trend of the quality of the environment.In addition,the surface water is one of the important sources of human water,so the water environment monitoring has become an important problem for the country to concerned about.However,the monitoring of water quality in most parts of the country is still backward and the evaluation of water quality is not accurate and comprehensive.This has brought certain difficulties to relevant environmental protection departments to monitor and treat the water environment timely and effectively.In view of the current serious water environment pollution in our country,water quality monitoring consumes time and labor,and the data collection is not timely and not effectively.What's more,the warning information is not generated,and the water quality evaluation is not comprehensive.Faced with those situations,this paper proposes an effective and intelligent monitoring system and solution,which can manage and analyse the monitoring data.In addition,we also design a new water quality evaluation model.This paper takes a certain river station in Huangshan District,Huangshan City,Anhui Province as an example and collects monitoring data of surface water through the developed surface water monitoring management system,and then uses the designed water quality evaluation model to evaluate the water quality which can reflect the water quality more accurately.It also can help the relevant departments of the environment monitor the quality of surface water in real time,efficiently and accurately so as to take timely control measures.The research contents of this paper are as follows:(1)The IPOS-HSSVM Model is designed for multi-classification water quality assessment.Firstly,we use PCA to extract the principal components of the sample as the input variables of the model,which can solve the problem that the original sample variables overlap with each other and the large number of input parameters led to model training for a long time with a bad result.By using the improved Particle Swarm Optimization(IPOS),the fixed learning factor and inertia weight of the traditional particle swarm optimization algorithm are changed,and adaptive inertial weight adjustment method is used in combination with the dynamic change learning factor to optimize the hyper-sphere support vector machine(HSSVM)model parameters.This model can be used to evaluate the water quality level.Finally,we compared with the traditional single factor evaluation method,the results found that the model evaluation results more comprehensive and accurate.(2)We design a surface water monitoring management system platform.The system platform includes the system requirements analysis,system architecture design,system integration framework design,database design,functional modules detailed design.In addition,it contains a detailed monitoring data management and review process design.We used the automatic audit,manual review,data addendum,in order to ensure the validity of the monitoring data collected.(3)System implementation and function testing.Firstly,we use code to implement the system function module and deployed to meet the requirements of the test environment.Finally,we use the written test case to test and analyse the system strictly in order to ensure the reliability and integrity of the system.