Quantifying the benefit of SHM : can the VoI be negative?

Verzobio, Andrea and Bolognani, Denise and Zonta, Daniele and Quigley, John; Lynch, Jerome P. and Huang, Haiying and Sohn, Hoon and Wang, Kon-Well, eds. (2019) Quantifying the benefit of SHM : can the VoI be negative? In: 13th Sensors and Smart Structures Technologies for Civil, Mechanical, and Aerospace Systems 2019. SPIE, USA. ISBN 9781510625952 (https://doi.org/10.1117/12.2518878)

[thumbnail of Verzobio-etal-ICASP-2019-Quantifying-the-benefit-of-SHM-can-the-VoI-be-negative]
Text. Filename: Verzobio_etal_ICASP_2019_Quantifying_the_benefit_of_SHM_can_the_VoI_be_negative.pdf
Accepted Author Manuscript

Download (738kB)| Preview


Structural Health Monitoring (SHM) is generally presented as a powerful tool that allows bridge managers to make decisions on maintenance, reconstruction and repair of their assets. The benefit of SHM can be properly quantified using the concept of Value of Information (VoI), i.e. the difference between the utilities of operating the structure with and without the monitoring system. In calculating the VoI, two understood assumptions are that all decisions concerning system installation and operation are taken by the same rational agent and that "information never hurts", i.e. VoI is guaranteed to be non-negative. However, in the real world, the individual who decides on buying a monitoring system, the owner, is often not the same individual, the manager, who will use it once the system has been installed, and they may be-have differently because of their different risk aversion. We develop a formulation to properly evaluate the VoI from the owner perspective, when the manager is a different individual. We demonstrate that in a decision-making process where the two individuals involved share exactly the same information, but behave differently, the VoI can be negative and that we can always find a combination of prior probabilities and utility functions which ultimately yields a negative conditional VoI. Indeed, even if the two agents have an agreement a priori, due to their different behaviors, their optimal actions can diverge after the installation of the monitoring system. We apply this formulation on a real-life case study concerning the Streicker Bridge (NJ, USA).