Font Size: a A A

Research On Construction Of Mathematical Model Of Improve DHNN Water Quality Evaluation And Practically

Posted on:2021-08-31Degree:MasterType:Thesis
Country:ChinaCandidate:G LuFull Text:PDF
GTID:2491306530958439Subject:Municipal engineering
Abstract/Summary:PDF Full Text Request
Along with the development of computer technology and algorithms,the mathematical model of water quality in the basin has gently developed into the aid of ANNs.Aiming at the problems of DHNN’s easy over-fitting,accuracy to be improved,and research on regional water pollution control,the main research content:constructing a watershed evaluation index system as an input to an improved DHNN water quality evaluation mathematical model,and the model output is coupled with a fuzzy membership function to achieve In the case of continuous changes,you can quickly determine how close the water quality category is to the upper and lower levels.According to the output results of the model,the law of regional ecological evolution is analyzed,and finally,the driving mechanism of the water environment in the basin is analyzed from different angles,and suggestions for the treatment of water pollution in the basin are put forward.The main research conclusions are as follows:(1)Construction of evaluation index system.Through SPSS20 software statistical analysis of sample data,construction of watershed evaluation index system.The results show that the factor analysis method not only makes up for its inability to realize water quality classification,but also simplifies the input of the DHNN model,reduces the degree of overfitting,accelerates the convergence to a certain extent,and improves the timeliness,indicating the coupling effect of the factor analysis method and the DHNN model strong applicability.(2)The improved DHNN water quality evaluation mathematical model was constructed.According to the characteristics of comprehensive water quality evaluation and the fuzzy characteristics of water quality information,DHNN neural network self-feedback recognition function,SVD algorithm does not need to be calculated in blocks,coupling builds a DHNN mathematical model based on SVD algorithm,and the output result is input into the fuzzy membership function.The successful construction of the improved DHNN water quality evaluation mathematical model shows that the DHNN neural network is no longer over-fitting when the water quality category changes continuously.At the same time,the fuzzy membership function is added to achieve the water quality evaluation category evaluation membership value,so that the evaluation the method is closer to objective reality.(3)The applicability and expansibility of the improved DHNN water quality evaluation mathematical model were analyzed.Through comparison and analysis with the evaluation results of the comprehensive pollution index method,the results show that the improved DHNN water quality evaluation mathematical model can quickly realize water quality classification,with high accuracy and small deviation,which is close to the actual water quality status.Therefore,the improved DHNN water quality evaluation mathematical model has strong applicability and expansibility,and has certain research and utilization value.(4)Analysis on the law of watershed temporal and spatial evolution and its driving mechanism(1)The time distribution characteristics of the water environment quality of the basin show that the intranasal water quality category of the basin changes fromⅤ0.50toⅢ0.99,and the water environment quality has improved significantly,but the Yang River Basin is more than ClassⅣand the Sanggan River Basin is more than ClassⅢ.Relative water quality is relatively good;the water quality during the high and normal water period is better than that during the low water period,indicating that the water environment quality is greatly affected by hydrological factors,but the water environment quality is the effect of multiple factors,which should be considered comprehensively and focused on governance.(2)The characteristics of the spatial distribution of water environment quality in the basin show that the over-standard factor of the monitoring section is TN>F->TP>COD>CODMn;in terms of spatial distribution,the northwest region with Zuowei as the core has more pollution than the southwest region with Xiao Dukou as the core,From north to south,from west to east,the trend is gradually aggravated,but the water environment quality of the Sanggan River Basin is better than that of the Yang River Basin.(3)The correlation between natural factors and water quality shows that:The hydrological monitoring data of the Yang River Basin is represented by the monitoring data of the Zhang Jiakou Hydrological Monitoring Station and the Xiang Shuipu Hydrological Monitoring Station,and the monitoring hydrological data of the Shi Xiali Hydrological Monitoring Station and the Qian Jiashawa Hydrological Monitoring Station are used to indicate the Sanggan River.According to the hydrological monitoring data of the Sanggan River Basin,TN in the Yang River Basin is greatly affected by flow,water level,and velocity,COD and CODMnare greatly affected by the water level,and TN in the Sanggan River Basin is greatly affected by the water level.Therefore,the upper reaches of the Yongding River Basin should ensure the water level of the basin to reduce the impact of TN,COD,and CODMnpollution loads from the source.(4)The correlation between social factors and water quality shows that the secondary industry has a strong correlation with water quality,with the highest correlation coefficient,which is one of the main factors affecting water quality;The tertiary industry has a strong correlation with forestry,animal husbandry and fishery and water quality,and agriculture and water quality have a strong correlation.The degree of correlation is moderate;therefore,the secondary industry is the main non-point source pollution source,followed by the tertiary industry and non-main non-point source pollution sources such as forestry,animal husbandry and fishery,agriculture,urban residents’lives,and population.(5)Suggestions on water pollution control in River BasinThe results of the driving mechanism show that frequent production activities,soil erosion and rainfall drive have aggravated the loss of pollutants;hydrological elements are one of the main reasons for the TN pollution load exceeding the threshold;the key source areas with high pollution load are mainly distributed in the Yang River Basins,frequent production activities,soil erosion and rainfall drive have aggravated the loss of pollutants;hydrological elements are one of the main reasons for the TN pollution load exceeding the threshold;the secondary industry is the main source of non-point source pollution,followed by the tertiary industry with forestry,animal husbandry and fishery,agricultural non-point source pollution,urban residents’lives,population and other aspects have little impact on river basin non-point source pollution and are not the main non-point source pollution source.Therefore,relevant government departments should adjust the industrial structure,especially the discharge of industrial wastewater.The root cause,cure its root.Recommendations:First,build vegetation filter belts,carry out sloping farmland management and other water and soil conservation measures;second,optimize industrial structure,comprehensive management at the source and end,and third,strengthen ecological monitoring,early warning and emergency response mechanisms,and establish effective river basin environmental public policies.
Keywords/Search Tags:Improved DHNN water quality evaluation mathematical model, SVD algorithm, fuzzy membership degree, ecological evolution law
PDF Full Text Request
Related items