A Framework for the Development of Soft Sensors for Process Plants

Abstract

By monitoring sensors, problematic system states in process engineering plants and production lines can be detected and corrected before significant damage occurs. Soft sensors as a combination of data analysis and digitalization are tailored to a specific application. During their development, large amounts of time series data from sensors are therefore analyzed, interpreted in consultation with the plant operators, and applied as the basis for the design of data processing methods using digital models, which form the core of soft sensors. The framework presented here is employed in development processes to make the research, development, and integration of soft sensors into operational and technical processes more effective and efficient.