The project addresses the artificial intelligence-based process design of microfinishing with the aim of reducing the consumption of human and material resources. The core of a suitable process design for microfinishing currently lies in the empirical knowledge of machine and tool manufacturers as well as experienced machine operators. However, the number of experienced specialists is gradually declining due to various factors, such as demographic change, resulting in a loss of experiential knowledge. In addition, microfinishing usually involves a multi-stage (often three-stage) machining process, the individual stages of which must be coordinated and require testing. Microfinishing is a precision machining process whose complexity limits the application of simulation techniques and analytical descriptions. For this reason, the use of AI, in particular machine learning (ML), is a suitable method for extracting process knowledge and describing high-dimensional, non-linear dependencies.

To this end, the aim is to train a machine learning model based on experimental test data to be generated. The extent to which repetitions of the same process parameters lead to comparable end results in terms of surface topography must be assessed in order to take this into account for database generation. Starting from a basic model of the first microfinishing stage, the extent to which the ML models for the two remaining tool stages can be created by transfer learning is examined in order to reduce the amount of test data required and thus reduce the amount of testing required. At the same time, this ensures transferability to tool combinations not considered in the project with little additional effort. The representative process models (surrogate models) are ultimately used as optimization objects to achieve an intelligent, adaptive, and optimized process design. Optimization is preferably carried out using reinforcement learning or genetic algorithms to predict optimal setting parameters for each machining stage across the entire microfinishing process and achieve an optimal end result. In addition to process design, the use of suitable measurement technology in the experiments will determine the extent to which the interaction between the material removal rate and tool wear can be described and how this can be fed into the ML. Another goal is to establish a quantifiable causal relationship between the initial and the final topographies, taking into account the process control variables. After successful optimization with validation, a final recommendation model for dealing with versatile adjustment and influencing variables is to be developed.