Structural dynamics & system identification
The interdisciplinary research of our working group is concerned with the development of scientific methods to be used in intelligent buildings of the future. Our main competence is the development of highly sensitive and numerically precise methods based on systems theory and information technology approaches. For example, general mathematical models of mechanical structures can be parameterized on the basis of large measurement data. These models can be used both in the planning phase and over the entire service life of a building.
System identification
System identification is generally understood as the creation of a model that provides a relationship between the input and output variables of a system. Depending on the application and objective, various algorithms are available for modeling. The identified systems can be parametric (e.g. state space models) or non-parametric (e.g. in the form of a transfer function).
A distinction is made between deterministic and stochastic methods. In a deterministic approach, the system inputs are known. The excitation can be carried out with pulses or sweeps, for example. Stochastic methods, on the other hand, do not require precise knowledge of the system inputs. The excitation takes place via noise processes (wind, waves, traffic). When using stochastic methods, system identification is possible solely on the basis of measured system outputs (e.g. accelerations or vibration velocities). These methods are also referred to as output-only.
With output-only system identification, "black box models" are obtained. These are primarily of a purely mathematical nature and cannot be directly interpreted physically. One research focus at the Department of Mechanics and Dynamics at HTWK-Leipzig is the identification of directly interpretable physical parameters based on output-only system identification.
