Artificial Intelligence Methodology for Smart and Sustainable Manufacturing Industry

Abstract Manufacturing industry is seeing an explosion in data availability, with potential for desired optimisation of productive systems. For instance, additive manufacturing is a novel technology that is currently advancing fast and constantly. As the physics of additive manufacturing are complex there are cases where it is of interest to monitor the process, and further to this, a potential objective is to optimize process parameters. Analytics and decision support technologies can support optimisation from data and expert knowledge. We present current status of analytics methodologies, which have not been updated in years. Also, we introduce an additive manufacturing process, laser cladding. From this context, this paper presents a methodology proposal, and further to this, applicability in a laser cladding scenario. A set of experiments is performed and quality of results, both in terms of porosity and dimensionality, assessed. A study of correlation is centred on data analysis for extracting relationships among the process data and process parameters with respect to the quality of the test outcome. As a result, in our study porosity showed up as the hardest to generate and to model of the two quality indicators studied, with the model attaining modest results, while results for dimensionality were better, and various variables showed importance for correlations.