Abstract:In aquaculture, dissolved oxygen is a key water quality factor to ensure the survival of aquaculture organisms. In order to ensure that there is sufficient dissolved oxygen in the water body, aquaculture plants generally adopt a regular oxygen production method. Although this ensures sufficient dissolved oxygen, it causes a large energy consumption. In response to this problem, a dissolved oxygen regulation method was proposed based on modeling prediction and relational rule database, which mainly included three parts. Firstly, an adaptive enhanced particle swarm optimization-extreme learning machine model (AdaBoost-PSO-ELM) was constructed to achieve accurate prediction of dissolved oxygen. Then, the curved surface fitting method was used to quantify the relationship between the initial concentration of dissolved oxygen, the aeration flow rate and the opening time of the aerator, and a relation rule database was built to provide a basis for controlling the aerator. Finally, based on the predicted value of dissolved oxygen and combined with current dissolved oxygen content, the computer monitoring platform called the relation rule database to reasonably control the opening time of the aerator. The dissolved oxygen prediction results showed that the MSE, MAE and RMSE of the AdaBoost-PSO-ELM model reached 0.0055mg2/L2, 0.0531mg/L and 0.0745mg/L, respectively. Compared with particle swarm optimization extreme learning machine (PSO-ELM), extreme learning machine (ELM), BP neural network (BPNN) and wavelet neural network (WNN), the prediction performance of AdaBoost-PSO-ELM was significantly improved. The results of aeration experiments showed that the priori equation based on cubic polynomial can accurately quantify the nonlinear relationship between the initial concentration of dissolved oxygen, the aeration flow rate and the opening time of the aerator, and the R2 of fitting was above 0.99. At the same time, the rule database constructed based on the quantitative results can reasonably control the opening time of the aerator, which was of great significance for saving energy and promoting sustainable aquaculture, and it had great application prospects in the future.