Modelling and control of the thermal spray process by means of two-stage statistical prediction models

Highly efficient processes to decrease resource consumption are a growing challenge in the coating industry. In addition to the optimization of coatings, the prevention of deviations in coating properties is a particular focus of the scientific community. Especially during long-term coating processes, due to environmental influences or component wear, the coating properties may deviate. These deviations can lead to an increased need for reworking up to an elimination of coated workpieces and thus to increased production costs and higher resource requirements. This paper demonstrates the development of a predictive model for the control of coating properties during the coating process. The coating properties are estimated in two stages using response surface models. The in-flight particle properties are modelled as a function of the parameter settings, and the coating properties are modelled depending on the in-flight particle properties. The created two-stage model results from linking both models. The aim of the development of these models is a monitoring of the coating process and the adjustment of the process parameters in the event of a deviation in the coating quality as a function of the in-flight particle properties.