An artificial neural network-particle swarm optimization (ANN-PSO) approach to predict the aeration efficiency of venturi aeration system

In the present study, the artificial neural network-particle swarm optimization (ANN-PSO) approach was adopted for optimizing the standard aeration efficiency (SAE) of the venturi aeration system. Standard aeration efficiency (SAE) is the main parameter which is used to evaluate the effectiveness of an aerator. Aeration test were carried out in a 220 litre capacity tank to find out the impacts of operating variables on SAE using venturi aeration system. The operating variable includes throat length (L), hole diameter (d), and flow rate (Q), however, the output variable was selected as SAE to determine the model response. A 3-6-1 ANN model was developed for predicting the SAE of venturi aeration system, and the PSO algorithm was applied to obtain the optimum values of the operating variables of venturi aeration system. The effectiveness of the ANN modelling was evaluated using the coefficient of determination (R2), root mean square error (RMSE), and mean absolute error (MAE) from the test data and forecasted data. The most appropriate optimal values of the operating variables i.e., L, d, and Q were found to be 1.82 mm, 2.00 mm, and 0.31 l/s, respectively. The optimal SAE under ideal working conditions was optimized to be 0.8333 kg O2/kWh. The applied ANN-PSO approaches can predict the optimal values satisfactorily with a maximum of 7.73% difference with the experimental values. Therefore, the suitability of the ANN-PSO approach can be useful to predict the aeration efficiency of venturi aeration system.