Neural network based swarm concept for prediction asphaltene precipitation due to natural depletion

Asphaltene is the heaviest component in crude oil. Initially, asphaltene is dissolved in crude oil. Once asphaltene is separated from the crude oil due to pressure loss or composition change in addition of solvents, asphaltene may deposit over surface rock and plug some pore throats that produce more flow resistance for oil in porous medium. In this work, the model based on a feed-forward artificial neural network (ANN) optimized by hybrid genetic algorithm and particle swarm optimization (HGAPSO) as an intelligent approach to forecast asphaltene precipitation due natural depletion is proposed. Hybrid genetic algorithm and Particle swarm optimization (HGAPSO) is carried out to decide the initial weights of the neural network. The HGAPSO-ANN model is implemented to the experimental data from one of northern Persian Gulf oil field. The forecasted outputs from the HGAPSO-ANN model and BP-ANN were compared to the experimental precipitation data. Low deviation between forecasted results of proposed model and experimental data validate good precision and accuracy of the model. The good performance of the proposed HGAPSO-ANN model ascertain by comparison between the prediction of this model and corresponding experimental data. ► Integrating artificial neural network and HGAPSO has been illustrated. ► HGAPSO-ANN model combines local and global searching ability of the ANN and HGA PSO, respectively. ► Predictive model of asphaltene precipitation has been introduced. ► To prevent improper convergence parameters of algorithm are carefully assigned.