Vollmer Werke Maschinenfabrik GmbH is a SME mechanical engineering company and one of the leading manufacturers of machines for processing rotary tools, circular saws and metal-cutting band saws. The company's portfolio includes around 60 types of grinding machines, which generate a varying amount of measurement data per grinding process, depending on the sensor equipment. Currently, this data is evaluated manually and, if necessary, by process experts.
In the future, the analysis of process data, e.g. for process monitoring and optimization, will become more and more significant for both manufacturers and operators of grinding machines. The task of this QuickCheck was to investigate what information can currently be obtained from process data of a plant that reflects the current state of the art.
The solution approach in this QuickCheck primarily involved an explorative data analysis. Historical process data was first converted into a uniform format and then evaluated using classical statistical methods. Correlation within the measurement data was searched for, and a look was taken at the probability distributions. Finally, various concepts for extended data evaluation, e.g. using ML algorithms, were presented for the present use case with regard to the IT architecture.
The results of the QuickCheck showed that although a statistical evaluation can already be performed with the current process data, the data basis must still be increased somewhat for the use of AI/ML algorithms. This can be achieved through additional sensors. Based on the QuickCheck, it can be concluded that plant monitoring, the classic area of application of machine learning in production, would be well suited for grinding machines.