Study On Water Quality Analysis Method Of Qingyi River

Water quality evaluation is a basic work of water environment management in the basin.A simple,rapid and effective evaluation of water quality category can provide decision-making basis for water resources utilization,water environment planning and management.At present,the water quality evaluation methods commonly used in China are one factor index method and Nemero index method.But these two methods all have the disadvantages of overtaking the biggest influence factors.However the current research hotspot artificial neural network model needs a large number of sample data to train models,what's more,the Markov model and other artificial intelligence analysis methods can not be used to predict the specific concentration values of water factors.Therefore,there are some deficiencies in the current water quality evaluation and research in terms of objectivity,scientificity and efficiency.In order to embody the trend of water environment in the future for a period of time in water quality evaluation,on the basis of the study of the existing water quality evaluation method,in view of the advantages of Bayesian formula and time series method for less modeling sample data,high precision and simple operation,the Bayesian water quality evaluation method based on time series is proposed.The Bayes formula is applied to the water quality analysis and the Jeffreys principle is adopted.After obtaining the actual monitoring data of the water quality factor,the probability of the previously judged water quality is revised.Then the probability that the actual water body belongs to the five water quality categories is obtained.Finally,the principle of maximum probability is adopted to determine that the water body belongs to a certain category.The time series water quality prediction model of the Qingyi River was established by the water quality monitoring data of59 months?2011.1-2015.11?of six factors of DO,permanganate index,COD,BOD₅,ammonia nitrogen and total nitrogen.The monitor data of December 2015 are used as the validation number of the model.Bayesian water quality evaluation method based on time series was used to predict and evaluate water quality factors at two monitoring stations in Hongya section of the Qingyi River in January-June of 2016,and compared with the single factor index method and Nemero index method.In the comparison,the water quality of 6 time points of the single factor index method is V class,the rest are class III water bodies.The Nemero index method has two time points in the evaluation of the type IV water body,and the rest are class III.In the evaluation of Bayesian water quality evaluation method at six time points of the two monitoring stations,the biggest probability belongs to water class are 18.43%belongs to class?,32.84%belongs to class?,29.93%belongs to class?,30.14%belongs to class?,25.26%belongs to class?,50.34%belongs to class?in Guidufu.In Mucheng Town,the biggest probability belongs to water class are 42.05%belongs to class?,42.34%belongs to class?,41.16%belongs to class?,37.23%belongs to class?,40.04%belongs to class?,39.93%belongs to class?.The results show that Bayesian water quality evaluation method has the most similarity to the directly predicted concentration values in the prediction and evaluation of Guidufu and Mucheng Town.Both sites are Class II or Class I water bodies.It is most suitable for the situation where the river belongs to drinking water sources.