Cost-benefit analysis for SHM systems based on minimum detectable parameter change

Abstract

Value of information (VoI) analysis can be used to assess the benefit of monitoring the health of infrastructure such as bridges and waterway locks. Such information supports decision-making processes, guiding operators towards the best actions for the infrastructure. In recent studies, VoI analysis is integrated with the modeling of the temporal evolution of structural conditions, so that the value of Structural Health Monitoring (VoSHM) can be assessed over the life cycle of the infrastructure. However, none of these studies considers the ability of the measurement system to detect small changes and its impact on the assessment of the VoI. This paper presents an approach to VoI assessment based on the concept of minimum change detectability and Bayesian linear filters. By exploiting a feature of the Kalman filter whereby posterior covariance is independent of observations, it is possible to predict the minimum shift in the mean value of a structural parameter for which its posterior density function would exceed a user-defined safety threshold. By constructing the decision tree and considering the cost of monitoring and maintenance, the value of detecting smaller changes can be estimated, which is the value of early change detection and performing preventive maintenance. As a proof of concept, this approach is applied to assess the VoSHM for a navigation lock built with precast reinforced concrete elements, where the steel bars in the chamber are damaged by fatigue and corrosion. The VoSHM is compared with the VoI associated to other Non Destructive Tests. The paper then discusses the implications of applying Bayesian linear filters in SHM for parameter change detection, the main one being the development of refined scales for structural condition assessment before any data from the changed state is available. Ultimately, the VoSHM is the benefit brought by adopting these refined scales in maintenance management.