ADMO - Automatic data-driven modeling and H_2/H_∞- norm-based dimension reduction of process-oriented and cooperative systems for SHM condition analysis with methods of system identification and machine learning on exposed structures

The digital change is causing profound changes in all areas of society. In the fusion of BIM, the optimized planning, execution and management of plants, buildings and infrastructures, with Structural Health Monitoring (SHM) a digital twin functions as a central element of an efficient data organization.

The aim of this project is a method that realizes automated data-driven modeling based on the H2/H-infinite norm and methods of system identification coupled with machine learning. This enables a condition monitoring as a digital twin over the service life of the real twin, the building, which is incorporated into an SHM/BIM concept. Based on process-oriented cooperative systems, special physically interpretable indicators are able to automatically display and localize structural changes.

The numerical method works with stochastic multi-correlated output-only measurement data, with special consideration and classification of environmental and operational conditions. The automatically generated parameterized stochastic process models of the system and filter theory enables a prediction of future damage states on the examined structure. This gives the public authority a set of tools for predictive planning of maintenance measures on structures with high economic benefits.