Role of Sensors in Error Propagation with the Dynamic Constrained Observability Method

Area of Science:

• Structural Engineering
• System Identification
• Uncertainty Quantification

Background:

• Structural system identification is challenged by ill-conditioning, affecting parameter uniqueness and stability.
• Error propagation from epistemic and aleatory uncertainties, particularly sensor accuracy, complicates analysis.
• Uncertainty quantification (UQ) is crucial for assessing the impact of uncertainties on identified parameters.

Purpose of the Study:

• To review existing UQ methods for parameter identification in structural systems.
• To introduce and validate a novel dynamic constrained observability method (COM).
• To analyze the influence of sensor placement and uncertainty types on system identification.

Main Methods:

• A comprehensive literature review of UQ approaches in structural parameter identification.
• Development and application of the dynamic constrained observability method (COM).
• Experimental validation using a reinforced concrete beam and a real bridge system.

Main Results:

• The dynamic COM addresses shortcomings of existing UQ methods in structural identification.
• COM demonstrated superior performance and applicability in identifying parameters of a real bridge.
• Optimal sensor placement depends on the presence of epistemic uncertainty and structural knowledge.

Conclusions:

• The dynamic COM offers an efficient approach for uncertainty quantification in structural system identification.
• Sensor placement strategies must account for both sensor accuracy and epistemic uncertainty.
• As structural knowledge grows, epistemic uncertainty diminishes, allowing for refined sensor placement optimization.