| Abstract: |
The formation of a coating is the result of several physicochemical mechanisms involving dozens of influencing factors that have nonlinear and multiply interacting dependencies. In the concept of Digital Shadow in production technologies, employing ensemble of computer-aided methods is inevitable to understand and control these correlations in the complex coating technologies. This study aims at integrating the artificial intelligence methodologies to predict and optimize the deposition efficiency (DE) in plasma spraying. In the first step, analytical models are developed to predict the DE. These models include numerical analysis of the process with Computational Fluid Dynamics (CFD)-simulations, data-driven Support Vector Machine (SVM)-algorithms and knowledge-based Fuzzy Logic (FL)-techniques. In the next step, experimental data are generated via design of experiments (DOE) to enrich the data sets of the analytical models. Thus, the prediction results in each iteration are validated by experiments and subsequently the prediction accuracy will be increased. This expert system can be used as a tool for quality prediction and monitoring in plasma spraying using data- and knowledge-based techniques and hence, the integration of coating process into production chain could be improved.
Keywords: Plasma spraying, artificial intelligence, Digital Shadow, CFD simulation, expert system
|
|
