An Improved Optimization Model to Predict the Microhardness of Ni/Al2O3 Nanocomposite Coatings Prepared by Electrodeposition: A Hybrid Artificial Neural Network-Modified Particle Swarm Optimization Approach

Abstract This study has employed a particle swarm optimization-based artificial neural network approach to predict the microhardness of Ni/Al2O3 nanocomposite coatings prepared by electrodeposition. At first, in order to collect the experimental data, the experiments were designed using a factorial D-optimal array. By considering the effective operating parameters in the electrochemical deposition as independent variables, 105 repeated tests were performed, and the microhardness of the coatings was determined as a dependent variable. Various ANN-MPSO networks were validated using the correlation coefficient, mean bias error, root mean square error, and mean percentage error as criteria. The experiments confirmed the possibility of providing coating with the microhardness of approximately 870 HV. The results demonstrated that the proposed model was an appropriate, applicable, and precise approach to predict the microhardness of Ni/Al2O3 nanocomposite coatings.