Optimal Sensor Placement in Buildings: Earthquake Excitation

Area of Science:

• Structural Engineering
• Geotechnical Engineering
• Applied Mathematics

Background:

• Instrumentation is crucial for structural health monitoring, but sparse sensor placement is common due to cost constraints.
• Estimating structural responses at unmonitored locations is essential for comprehensive analysis.
• Gaussian Process Regression (GPR) combined with deterministic models offers a promising approach for response estimation.

Purpose of the Study:

• To develop a methodology for optimal seismic sensor placement in buildings.
• To minimize uncertainty in reconstructing structural responses at non-instrumented floors.
• To extend previous work on sensor placement optimization to seismic excitations.

Main Methods:

• A methodology was developed to determine optimal seismic sensor locations.
• The approach minimizes uncertainty in structural response estimation using a hybrid model (deterministic beam model + GPR).
• The method was validated using numerical simulations and applied to two real instrumented buildings.

Main Results:

• The proposed sensor placement strategy achieved an average 40% reduction in response uncertainty compared to random distribution.
• A 10% uncertainty reduction was observed in a 52-story building with near-uniform existing sensor distribution.
• An 80% uncertainty reduction was achieved in a 73-story building where sensors were concentrated on localized behavior.

Conclusions:

• Optimal sensor placement significantly enhances the accuracy of structural response estimation in instrumented buildings.
• The methodology is effective for seismic excitations and applicable to diverse building structures.
• Strategic sensor deployment is key to maximizing the benefits of structural health monitoring within budget limitations.