Research On The Application Of Water Quality Prediction And Early Warning In Aquaculture Systems

China’s agricultural industrial structure has long been showing a diversified development trend,covering a variety of agricultural structures such as planting,forestry,animal husbandry and aquaculture.Especially in the aquaculture industry,there is a lack of professional technical personnel.However,the current aquaculture industry is also facing problems such as resource shortage and environmental degradation.Due to the improvement of the refinement degree of aquaculture and the increase of aquaculture species,the water quality environment of aquaculture is also deteriorating,which is easy to cause the quality and safety of aquatic products.Therefore,the prediction and early warning of the water quality of aquaculture has become an urgent problem to be solved in China’s aquaculture industry.The prediction and early warning of water quality change in aquaculture system enables decision-makers to know the change of water quality in advance and take measures to improve the water quality.To predict and early warning of water quality factors,it is necessary to first understand the standard of water quality parameters and analyze historical data to give the predicted value.In recent years,with the establishment of more and more water quality evaluation systems,the water quality prediction and early warning technology has become increasingly perfect,and gradually become the research focus of aquaculture system and other disciplines.This paper selected several water quality parameters common in aquaculture systems of an aquaculture company in dalian city,and used regression analysis method,sequence analysis method and neural network method to predict these water quality parameters respectively,compared the prediction effects of these methods and analyzed the reasons for these differences.The results show that the average prediction error of regression analysis method and neural network is low,and the prediction and early warning effect is very good.Regression analysis shows the linear relationship between dependent variable and independent variable,which has high application value.BP(back propagation)neural network,as the most widely used method in neural network technology,has strong self-learning and self-adaptation ability in water quality data prediction and warning.