information analysis and fault diagnostics /
Business Unit Manager: Dipl.-Ing. Dr. Thomas Buchegger M.Sc
Our activities focus on the development and application of automated methods for the monitoring and the analysis of processes in machines, components and industrial plants. This results in high competency in the fields of data-based fault isolation and cause analysis for black-box systems.
Process and cause analysis
- In process analysis and optimization, our competence resides in detecting unknown interdependencies that are important for process modeling (process description) based on usually measured process data.
- Intelligent combination of methods from statistics and information theory enables identifying causal correlations between significant process parameters even in black-box systems for the purpose of optimizing control variables.
- An important benefit of such analysis tools is that even highly complex processes can be analyzed automatically while sparing resources (personnel and time). Wherever adequately precise first principle models are lacking or need to be complemented, the results of process and cause analysis can provide new, directly applicable knowledge.
Visual sensors
- In the field of virtual sensors, data-based approaches model process variables whose physical measurement is difficult or impossible. Even if the measurement precision cannot be achieved for technical reasons, the use of virtual sensors provides a solution option.
- Modeling realizes a virtual sensor that estimates online the simulated behavior over time of the actual state of a physical parameter and makes it available for further processing.
Process monitoring
- Based on first-principle models and/or via integration of results from process analysis, we deliver software tools for customized process monitoring.
Fault detection and isolation (FDI)
- In fault detection, we focus on robust monitoring and detection of critical states of the unterlying system.
- Depending on behavior over time of the processes and the involved machine components, both slowly accruing and abruptly occurring fault states are detected and automatically evaluated. That is, the software provides information about the wear or fault state of components and the possible causes of faults (assuming complex processes).
