Conceptual Framework for manufacturing data preprocessing of diverse input sources

The manufacturing industry today is experiencing a never seen increase in available data. These data compromise a variety of different formats, semantics, and quality. It is often distributed in different data sources, e.g. sensor data from the production line, environmental data or machine tool parameters. Coming from the field of application the paper will discuss, within a conceptual framework, the possibilities of how to integrate the diverse existing data-sources and how to pre-process the data with high quality using advanced outlier detection algorithms and developing reasonable outlier treatment values by applying machine-learning methods. The result will be validated with real manufacturing data from an automotive use-case.