Abstract
Civil engineering structures are commonly monitored to assess their structural behaviour, using alarm thresholds to indicate when contingency actions are needed to improve safety. However, there is a need for guidelines on how to establish thresholds that ensure sufficient safety. This paper therefore proposes a general computational algorithm for establishment of reliability-based alarm thresholds for civil engineering structures. The algorithm is based on Subset simulation with independent-component Markov chain Monte Carlo simulation and applicable with both analytical structural models and finite element models. The reliability-based alarm thresholds can straightforwardly be used in the monitoring plans that are developed in the design phase of a construction project, in particular for sequentially loaded structures such as staged construction of embankments. With the reliability-based alarm thresholds, contingency actions will only be implemented when they are needed to satisfy the target probability of failure.
1. Introduction
Observation of structural behaviour is standard practice in civil engineering, in particular for structures of high importance or high risk. As the cost for sensors and other equipment reduces, more and more structures are being monitored. Examples include large dams, bridges, nuclear power facilities, and geotechnical structures such as tunnels and excavations [1–7]. The purpose can be, for example, validation of design assumptions and evaluation of need for design alterations or remedial measures to ensure structural safety or satisfactory serviceability. Observations of structural behaviour can also be used to gain information about engineering properties of existing structures in assessments of their structural safety. Additional information generally implies that uncertainties are reduced and that the calculated structural reliability is improved; thereby, costly replacement or strengthening interventions may be avoided. This principle is widely applied in reliability-based design and reliability-based safety assessments of civil infrastructure; see e.g. [8–17].
