Sensor Placement for Qualitative Fault Diagnosability in Large-Scale Adaptive Structures

The integration of actuators and sensors into buildings, allowing them to dynamically adapt to different load situations and external forces, is a promising approach towards a reduced resource consumption in the construction sector. The capability to detect and identify malfunctions in the sensors and actuators of these adaptive buildings is essential for ensuring their functionality, and must therefore be guaranteed prior to their construction. In this study, a method for sensor placement in adaptive buildings was developed, which ensures the desired capability for diagnosing faults in sensors and actuators with as few sensors as possible. The proposed method employs a qualitative measure of fault detectability and isolability based on the Chow-Willsky scheme. Two different approaches for finding optimal sensor configurations with respect to this measure were evaluated and compared to random configurations. The results demonstrated that the developed method finds suitable sensor placements fulfilling the objective of a maximal qualitative fault diagnosability.