Automated Profiling of Energy Data in Manufacturing

Abstract

In order to offer energy flexibility in energy markets in short time slots a fast and efficient processing and analysis of data from shop floor to production planning and control is necessary. To this end and to gain more knowledge, different datasets and sources have to be integrated. This paper proposes a conceptual architecture and a method for profiling energy data of manufacturing systems. This includes datasets from information systems as well as physical sources such as sensors, actuators or machine data. Real-life data often come with quality problems like missing and invalid values, outliers or duplicates. The key concept is to automatically identify the necessary metadata for including the dataset in an environment where further analysis and integration of datasets can take place. Moreover, a web service for profiling and visualizing data is implemented.