Abstract
In the present work, mechanically alloyed Al-12Si and SiO₂ powder was deposited onto an aluminium substrate by atmospheric plasma spraying (APS) to obtain a composite coating consisting of in-situ formed alumina reinforced hypereutectic Al-18Si matrix alloy. The effects of spray parameters and in-flight particle characteristics on reaction intensity between selective powders were investigated. Obtained results are tested by artificial neural network (ANN) techniques. An ANN model is built, trained and tested. Multilayer perception model has been constructed with back propagation algorithm using the input parameters of arc current, spray distance, in-flight particle velocity and temperature. The ANN model was found able to predict the coating hardness, substrate temperature, alumina intensity and silicon intensity in the range of input parameters considered. This study demonstrates that ANN is very efficient for predicting output parameters of experimental studies.

Key words
artificial neural network, in-situ plasma spray, composite coating, in-flight particle, Al-Si

Full text (301 KB)