Study On Intelligent Water Quality Monitoring System For Pearl Mussel Circulating Water Aquaculture

In China,the freshwater pearl mussel aquaculture plays an important role in agricultural production.The freshwater pearl mussel’s growth process does not harm the environment,but the traditional aquaculture model has caused problems such as serious water pollution,eutrophication and environment destruction.Facing such a situation,it’s necessary to develop a new freshwater pearl aquaculture model,so paper showed a industrialized circulating water aquaculture model(using multi-layer dimensional cages and feed pearl mussel with high quality microalgae).In order to monitor the quality of aquaculture water around the clock,this study developed a distributed monitoring system for the water quality monitoring based on wireless sensor network.In addition,this paper proposed SVM water quality classification model based on gray wolf optimization algorithm,and SVR water quality regression prediction model based on particle swarm optimization,these two models can analyze the water quality data of freshwater pearl culture timely and effectively,and provide the basis for decision-making of scientific culture.(1)In order to meet the needs of water quality monitoring in the industrial circulating water culture mode of freshwater pearl mussel,a distributed water quality monitoring system based on wireless sensor network is developed in this paper.The system used the three-layer architecture of the field sensing layer,transmission layer and application layer,and mainly includes the following 4 parts: water quality monitoring subsystem,meteorological monitoring subsystem,equipment control subsystem,and monitoring data center.This system implemented multi-parameter water quality and meteorological multi-parameter monitoring functions,start-stop control of circulating water treatment equipment,automatic control of algae feeding,and on-site video monitoring.In addition,a SQL Server database is set up on the host computer in the monitoring center to store real-time water quality monitoring data,providing a basis for subsequent data processing and analysis.(2)This paper proposed a water quality classification evaluation model based on the combination of improved gray wolf optimization algorithm and support vector machine.In order to improve the uniformity and ergodicity of the initial distribution of wolves,the method of Tent chaotic mapping is introduced to optimize the initial population of the original GWO algorithm.Then,the convergence factor a is reduced in a non-linear manner,so that the search range of the algorithm is enlarged at the beginning of iteration,and the convergence speed is faster in the later stage.Through the optimization experiments of the four benchmark functions,the improved GWO algorithm has faster convergence speed and higher accuracy.Finally,the measured water quality data were used for testing.The result of water quality classification of this model is similar to the actual data,with an accuracy of 98.64%,which could be used as the basis for classification and evaluation of pearl mussel aquaculture water quality.(3)A water quality prediction model based on the combination of improved particle swarm algorithm and support vector regression machine is established in this paper.In order to improve the optimization accuracy of traditional PSO algorithm and avoid premature and stagnation problems,the evaluation mechanism of chasing particles and mean square difference of fitness is introduced to find and adjust the partial aggregation of particles in time.Then,by introducing the mutation factor and memory factor,the particle position information in the local optimal range can adaptively mutate,so it can escape from the constraints of the local optimal region,and search the global space.The IPSO algorithm can effectively avoid local optimization and has faster convergence speed when benchmark function is used for optimization test.Finally,the measured water quality data is used as a sample for experiments.The results show that the IPSO-SVR model has a higher prediction accuracy,with a mean square error value of only 0.0028,it’s a very small difference from actual water quality grade,which can meet the needs of practical use.