Optimal sensor placement method for structural parameter identification considering nonlinear correlations under dynamic loadings

Heng Ouyang

doi:10.1016/j.ymssp.2024.112049

2025-02-01

PRE

DOI：10.1016/j.ymssp.2024.112049

原文链接

原文求助

摘要

En 中文

Structural parameter identification technology, using the structural responses measured by sensors and an accurate structural model to achieve the structural parameters to be determined (TBD), is widely applied in the structural damage detection, dynamic model updating and health monitoring. In practical engineering problems, the uncertainties and nonlinear correlations are commonly present in measurement data by sensors when a structure is subjected to dynamic loadings, which can significantly affect the stability and accuracy of identified results of the TBD parameter. To address the aforementioned issue, an optimal sensor placement (OSP) method for structural parameter identification considering nonlinear correlations under dynamic loadings is proposed in this paper. To start with, the OSP problem for structural parameter identification considering nonlinear correlations under dynamic loadings is discussed in detail and transformed into a forward uncertainty propagation problem. Then, a maximum independent mean variance criterion is proposed to determine the position and number of sensors for the OSP problem. In the framework of forward uncertainty propagation, Copula functions are then adopted to quantify the uncertainties and nonlinear correlations of TBD parameters, and the feature information of dynamic responses is extracted by the multi-layer discrete wavelet transform (MLDWT) method to convert the dynamic problem into a time-invariant problem. Subsequently, an improved principal component analysis and optimal polynomial chaos expansion method (IPCA-OPCE) is proposed to determine the OSP approach accurately and efficiently. Eventually, the structural parameters are accurately and stably identified by using the Markov chain Monte Carlo method. The accuracy and engineering applicability of the proposed method are demonstrated by a numerical example of a truss structure with 25 bars and an experimental example of a 4.55 MW compound planetary gearbox.

Keyword:

Optimal sensor placement

Inverse parameter identification

Dynamic loading

Nonlinear correlation

Compound planetary gearbox