Based On The Neural Network Model To Predict Water Quality

In recent years, the aquaculture industry is growing at an alarming rate, it has become the world within the scope of another leading industry. The backward management style and backward mode of aquaculture water quality lead to losses from aquaculture diseases as high as14billion yuan a year in China, the deterioration of water quality is a major cause of outbreaks of aquatic products. Due to the nonlinear and complexity of aquaculture ponds system, as well as the interaction between water quality factors, using traditional methods can’t satisfying water quality prediction model is established.Artificial neural network prediction model based on computational intelligence as the core of modern forecasting method. Although the artificial neural network is widely used in various fields, but in terms of aquaculture water quality prediction is the primary stage. Based on MATLAB2011platform, using L-M algorithm improved BP network and RBF network for water quality prediction is studied.In this paper, I used one of the aquaculture farms in haikou for example, and through actual testing for its aquaculture ponds for eight days I won6kinds of water quality factor of the data. Established contains a hidden layer of three layer BP neural network and RBF neural network, the output of the two network target for dissolved oxygen solubility, take another5kinds of water quality factors as input vector of the network, neural network prediction model is established with MATLAB software.Based on the breeding of dissolved oxygen solubility prediction results show that the BP neural network and RBF neural network can effectively apply in the aspect of water quality prediction, and it has high precision, has the very high practical value.