Method for Providing Parameters for Setting a Spray-Coating Apparatus

Method for providing parameters for setting a spray-coating apparatus, the method utilizing a control unit and a machine learning algorithm implemented by the control unit, wherein it further utilizes a collection of sample spraying patterns and corresponding sample spraying profiles, wherein

• the algorithm is trained on a training dataset where each datapoint contains at least a sample spraying profile, a corresponding sample spraying pattern, an operational spraying profile and a corresponding operational spraying pattern,
• wherein the operational spraying pattern corresponds to a label for the training,
• wherein each spraying profile contains values of multiple profile parameters,
• each sample spraying pattern represents a distribution of paint thickness provided by a spray-coating apparatus on a sample surface for a specific sample spraying profile, and
• each operational spraying pattern represents a distribution of paint thickness provided by a spray-coating apparatus on a sample surface for a specific operational spraying profile,
• wherein the sample spraying profiles in the collection belong to a multidimensional parameter sub-space defining an expected range of spray-coating applications, the sample spraying profiles in the training dataset belong to a multidimensional parameter sub-space defining a training range of spray-coating applications, wherein the training range sub-space is in at least some dimensions more finely covered by the sample spraying profiles and/or is broader than the expected range sub-space, and wherein the operational spraying patterns are provided by multiple different spray-coating apparatuses from different spray-coating facilities, wherein the method comprises at least one iteration, each iteration comprising the following steps:
• Receiving an iteration spraying profile;
• Selecting N sample spraying profiles with N corresponding sample spraying patterns from the collection, wherein N is a predefined positive integer;
• Inputting the iteration spraying profile and one of the sample spraying profiles with the corresponding sample spraying pattern into the algorithm and receiving an algorithm output describing an expected spraying pattern for each of the N selected sample spraying profiles;
• Combining the N algorithm outputs into an iteration output spraying pattern;
• Evaluating the iteration output spraying pattern according to a predetermined criterion;
• Adjusting the iteration spraying profile and providing it as an iteration spraying profile for next iteration if the iteration output spraying pattern does not comply with the criterion, or using profile parameters from the iteration spraying profile as parameters for setting the spray-coating apparatus if the iteration output spraying pattern does comply with the criterion.