Turbot Intensive Aquaculture Water Quality Parameter Modeling Method Research

There are many different kinds of Marine resources,it is an important research direction,the fishery resources is a Marine resources development and utilization of the first,but as a result of facing sea fishing capacity surplus and the reality of the offshore fishery resources recession,development of mariculture technology is becoming a hot spot of research.It is one of the main development directions of aquaculture industry in China,which has the advantages of environmental friendliness,controllable production,simple operation and convenient management.Good aquaculture environment is the basis of the aquaculture,therefore,for real-time online monitoring of water for aquaculture environment can effectively improve the benefit of breeding,better control of the breeding process,reduce the pollution of breeding.However,due to the limitation of measurement technology and measurement cost,it is difficult to realize real-time monitoring of water quality parameters such as ammonia nitrogen in water.According to the above problem,this research established independent research and development based on Programmable Logic Controller(PLC)and Windows Control Center(Win CC)laboratory intensive mariculture in the circulating water system,in the case of turbot,based on the soft measurement technology,this paper puts forward a kind of based on Genetic Algorithm(GA)to optimize the method of Least Squares Support Vector Machine(LSSVM)methods,the use of water for aquaculture environment p H value,conductivity,temperature,dissolved oxygen content as auxiliary variables,combined with the experimental measurement of water ammonia nitrogen value,for aquaculture water ammonia nitrogen content in the soft measurement modeling.In order to realize the on-line real-time monitoring of ammonia nitrogen content in aquaculture water environment during intensive farming.Through Matlab simulation,better simulation results are obtained.Compared with BP neural network model,the method selected in this paper is more effective.In this paper,the research background is the aquaculture water environment of Turbot,the main method is GA-LSSVM,the main research contents are as follows:1.Introduce the development of mariculture,breeding for aquaculture water environment parameters in the process of the measurement method used,as well as the soft measurement technology,in the process of mariculture application direction,with emphasis,in view of the two common modeling method,the BP neural network and least squares support vector machine algorithm is presented,and analysis the advantages and disadvantages of each.2.Because in the process of mariculture,nonlinear correlation between water quality parameters,although the BP neural network can approximate nonlinear function relationship between indefinitely,but in the actual operation process,tend to fall into local optimal solution of the dilemma,the least squares support vector machine algorithm for small sample,high latitudes,the practical problems of nonlinear fitting has a better effect,at the same time introducing the genetic optimization algorithm,optimize the model,so as to get the optimal training effect.3.By analyzing the process of mariculture easy measured parameters and water quality in the process of the generation and nitrification of ammonia nitrogen in water,based on the laboratory research independent establishment of the measured data of the mariculture of turbot intensification system,choose water p H,water temperature,water conductivity,the content of dissolved oxygen in water as the input variables,for difficult to realize on-line monitoring variables of ammonia nitrogen in water content,using different modeling method to establish the corresponding soft sensor model,comparing the differences between different modeling methods,choose the most ideal model.