Abstract Industry 4.0 and 5.0 approaches in thermal spraying seek to use AI modeling to tackle research issues. One common multi-criteria decision-making task is the optimization of deposition conditions to achieve desired coating characteristics and, based thereon, performance targets of specific interest. With most thermal spray technologies, this is a complex task due to the large number of influencing factors. Innovative, human-centered decision support needs to address two key challenges: (1) use experimentally collected small and smart data sets of coating characteristics and corresponding performance; and (2) integrate expert knowledge into modeling in order to obtain user-friendly, transparent decision recommendations. The present contribution illustrates how a model-based description of relationships between process conditions, coating characteristics (e.g., hardness), performance measures (e.g., sliding and abrasive wear, cavitation), and resource efficiency criteria (e.g., deposition rate) for smart data from HVAF- and HVOF-sprayed WC-CoCr coatings can be used to establish decision support. With the fuzzy pattern classification models used here, it is possible to describe patterns in a multidimensional feature space taking uncertainties and small data into account. The models are capable of formally representing complex relationships between input, process and output variables and enable predictions for the adjustment of input parameters such as nozzle type and particle size in order to optimize the coating performance flexibly, according to the defined application-related goal criteria.
Bocklisch et al. (2026) studied this question.