Study On Automatic Precise Control Model And System Of Shrimp Culture Process In South American White Shrimp

At present,in order to reduce the cost of farming and reduce the losses caused by operational errors,automatic and precise aquaculture process control system is increasingly becoming the needs of aquaculture enterprises.Aquatic products due to its high protein content,high nutritional value,sales increased year by year;P.vannamei as the sales volume of the highest shrimp,also become the best object we study the precise control of aquatic products breeding process automation.However,due to the fact that the breeding process of Penaeus vannamei is extremely complex,the incidents of the death of aquatic products all over the pond often happen due to operational errors,which makes the demand of aquaculture enterprises for precise and automatic aquatic product breeding process management system greatly increased.At present,the existing aquaculture management system is inclined to monitor the aquaculture environment,and the management of the culture process itself is no one.In order to establish accurate automation management system of South American white prawn breeding process,this paper from the various whiteleg shrimp breeding process data with the enterprise,carries on the contrast to the breeding process and summary,get the standard of penaeusvannameibonne process,then to the south of deconstruction,analysis of white shrimp farming processes.According to the analysis results,we establish data model,workflow engine,control principle and control algorithm of the process control system,and finally form an automatic and precise control system for Penaeus vannamei breeding process.Under the guidance of such ideas,the work done in this article includes:(1)collecting the culture process of the Penaeus vannamei,and comparing them with the same process,we got the standard model of Penaeus vannamei breeding process.Two deconstruction and analysis of standard Penaeus vannamei breeding process were carried out,and the breeding process of Penaeus vannamei can be summarized as a multi-dimensional,partially irreversible,partially conditional and partially overlapping business process.It includes three key points,nine sub key points,3 sun key points,a conditional circulation process and a condition overlapping process.The flow of various key points in the process is not clear,according to the breeding experience of the farmers.According to the results of the analysis can be penaeusvannameibonne process as a complex workflow model,in order to enable the system to handle complex business processes like this,the key process of discretization,the establishment of key state data model,will cause the change of the discrete state,the establishment of event data model.(2)on the basis of determining the workflow business model above,this paper analyzes the three major workflow engines in the mainstream: JBPM,Osworkflow and Shark.And made a comparison between the advantages and disadvantages of the three were not used in the South whitening shrimp workflow control,therefore,this paper combines the three advantages and disadvantages,put forward one for penaeusvannameibonne process and the event as a driver,as issued by the state of the workflow engine,and the design of the system the structure of the workflow engine,detailed workflow engine processing principle.(3)the circulation of the key state in the shrimp culture process in the South beauty white shrimp is not clear and depends on the cultivation experience of the cultured personnel.In order to solve the problem that the system can automatically transfer the critical state,this paper proposes to use neural network model to complete the system's learning experience of breeding experience for the aquaculture personnel,so as to circulate the critical state intelligently.This paper analyzes two kinds of effective neural network model,convolutional neural network(Convolutional Neural Network,CNN)and recurrent neural network(Recurrent Neural Network,RNN)of the two made a comparison in the learning efficiency and the convergence efficiency,the system needs a higher convergence efficiency and in accordance with the neural network model,breeding experience of training vector data the sample is the convolution neural network model,and make some simplification of the convolutional neural network,which reduce the computational complexity of training.At the same time,a corresponding coupling function model is established for the breeding experience data samples of farmers,and based on the TensorFlow framework,the principle and process of model training as well as the evaluation model of training model are given.(4)on the basis of the above,this paper established an automated and accurate control system for the culture process of Penaeus prawns.In the training on the basis of breeding experience of aquaculture personnel,with the help of the TensorFlow framework MVC framework and improved to realize the control of the breeding process,while the introduction of the adaptive mechanism,the system can continue to improve the system understanding of breeding experience of aquaculture personnel,the weight adjustment model and thus more precise value,to to farmers farming experience fit degree.Through the above research and Realization of workflow engine is proposed in this paper,farmers farming experience learning neural network in penaeusvannameibonne process automation precise control is carried out by trying to some extent,and provide the reference for the fine breeding enterprise culture.