Neural computation to correlate HVOF thermal spraying parameters with the magnetic properties of FeNb alloy deposits

The effect of the HVOF process parameters on coating porosity and magnetic properties of FeNb deposits was studied using artificial neural network methodology. A first artificial neural network was used and optimized to relate process parameters to the magnetic properties of the coating. The effect of process parameters on magnetic properties was quantified by a second network. Predicted magnetic properties correlated with coating porosity were obtained using these optimized network structures. It was then possible to identify the role of porosity with regards to improvement of coercivity and saturation magnetization. (c) 2005 Elsevier B.V. All rights reserved.