Parametric Optimization of Microhardness of Electroless Ni-Zn-Cu-P Coating Using Taguchi Design and Artificial Neural Network

In this investigation, the AZ31 magnesium alloy was coated with a quaternary electroless Ni-Zn-Cu-P coating for less than 600 s, resulting in a coating thickness of 5 mu m. To optimize the controllable coating parameters (nickel sulphate, zinc sulphate, and sodium hypophosphite), the Taguchi L27 orthogonal array was employed to maximize the microhardness of the coatings. By using these coating parameters as inputs and the microhardness of the coatings as an output, the applicability of an artificial neural network (ANN) was examined. To predict the microhardness of the coatings, ANNs with feed-forward back-propagation neural networks were trained using the Levenberg-Marquardt algorithm with 1 neuron for the first ANN, 2 neurons for the second, and so on up to ten ANN. The network with seven neurons in the hidden layer (ANN 3-7-1) shows the maximum correlation coefficient (R-2), indicating that ANN 3-7-1 accurately predicts microhardness. For ANN 3-7-1, the root mean squared error and R-2 were 8.8475 and 0.982, respectively. The surface morphology, composition, and crystallinity of the coatings were investigated and determined by field emission scanning electron microscopy, energy dispersive x-ray spectroscopy, and x-ray diffraction analysis.

Automated summary

In this study, a quaternary electroless Ni-Zn-Cu-P coating with a thickness of 5 μm was successfully coated on the surface of AZ31 magnesium alloy. The controllable coating parameters were optimized using Taguchi L27 orthogonal array to maximize the microhardness of the coatings. An artificial neural network (ANN) model was developed and ANN 3-7-1 showed the highest correlation coefficient and accurately predicted the microhardness of the coatings. Surface morphology, composition, and crystallinity of the coatings were also investigated using various analytical techniques.