AI-based diagnostic systems

Diagnostic systems detect and evaluate the deviations between the actual and the target state or between the occurred behavior and the target behavior of a system. The basis of this evaluation is information based on recorded data or models. Using the methods of model building and system analysis developed at the IOSB, for example, characteristic values are determined which allow statements about the overall state of a process by classifying information about the behavior as normal or abnormal and by step-by-step drawing conclusions about the underlying causes and errors.

Fields of application

For the complex operation of power grids (text in German language), high-resolution sensor technology is increasingly being used. Researchers at Fraunhofer IOSB-AST have developed a compression method for efficient archiving and accelerated analysis of long-term measurements. Based on this, AI procedures have been developed to detect deviations from normal network operation and to identify and localize certain operational disturbances in real time.