Application of GA and SVM to estimate ammonia nitrogen in aquaculture

The ammonia nitrogen content in aquaculture water is an important factor affecting the health of the aquaculture objects. It is an important indicator of the most common pollution emission and monitoring in intensive aquaculture. At present, it is difficult to realize on-line real-time monitoring of ammonia nitrogen in aquaculture water, and the laboratory method has the disadvantages of high cost, long monitoring time and easy to cause pollution. To solve this problem, a circulating water culture system of turbot was established in the laboratory firstly, then the auxiliary variables such as temperature, dissolved oxygen, conductivity and pH value were chosen, and a soft measurement model was proposed based on genetic algorithm (GA) and support vector machine (SVM) finally. It realized the estimation of ammonia nitrogen content in the industrialized aquatic environment. The results show that the estimation error of this method is small and the precision is better than BP neural networks and partial least square (PLS) regression, which confirm the practical value of the model in the intensive aquaculture environment.