Updating numerical models towards time domain alignment of structural dynamic responses with a limited number of sensors

Yuhao Fu,Ying Wang

DOI: https://doi.org/10.1016/j.ymssp.2023.110759

IF: 8.4

2023-12-01

Mechanical Systems and Signal Processing

Abstract:Due to the inherent uncertainty and complexity of civil infrastructure, it is challenging to simulate their structural dynamic responses accurately and efficiently. Existing model updating methods normally need distributed sensors and use modal parameters to construct objective functions, which can achieve frequency domain agreement between the numerical model and the real structure. Nevertheless, time domain alignment is still difficult to achieve, especially when the sensor number is limited. To enhance the simulation accuracy, this paper proposes a two-step model updating approach through updating in frequency and time domains sequentially. A virtual force is introduced in the second step to represent structural complexity and input error, and is derived by using three representative algorithms. The approach was verified on a scaled wind turbine model in the laboratory. Parametric studies were performed on different arrangements of sensors and virtual force. The results show that the virtual force should be placed at a sensor point and close to the acting point of external loads. The approach can achieve time domain alignment between numerical model and real structure with a normalized mean square error averaging 3.25% in the optimal setting. Based on the results, the proposed framework may find vast application in the establishment of numerical models towards structural dynamic twins.

engineering, mechanical