IMPROVING DECISION MAKING FOR INFRASTRUCTURE MANAGEMENT

点击翻译按钮获取中文摘要

Challenges in managing our infrastructure are associated with the growing costs in maintaining its level of performance in a changing climate, the associated uncertainties and growing budget constraints. For example, the MTQ manages over 5500 major structures on its network with an additional 4300 structures from the municipal network. The annual budget for the maintenance of these structures is over 640M$ and is predicted to increase with ageing of the infrastructure. The proposed research program is an evolution of the subjects studied in the previous Discovery Grant and is focused on improving the effectiveness and efficiency of infrastructure asset management by reducing uncertainties related to the evaluation of the current condition and residual life of infrastructure components (e.g. bridges) and improving decision analysis for the scheduling of maintenance, repair and replacement of components through network analysis. The industry is increasingly relying on structural health monitoring, non-destructive techniques, and numerical models to reduce uncertainties on the current and future condition of structural components and to supplement information obtained from visual inspections. The ability to detect structure deterioration at earlier stages can facilitate implementation of appropriate remedial measures and prevent this damage from becoming critical. The research will focus on evaluating the potential of the impulse response (IR) test and vibration analysis for the condition assessment of concrete plates and to develop guidelines as a function of the type, location and severity of defects for field investigations and quality control. The ability to accurately predict the residual life of a component given its current condition is the next important element in preventing damage to reach critical levels and avoid the spiralling high repair costs that we are observing. Current Bridge Management Systems rely mainly on historical records of condition states derived from visuals inspections to predict the residual life of components using Markov Chains models. This approach does not favor preventive maintenance since visual signs of deterioration occur only once it has reached critical levels and cannot be used for planning future interventions in the context of climate change. An approach based on physical and numerical models of deterioration is proposed for predicting condition states for different climate change scenarios. Finally, recent experience with major repairs in Montreal highlight the lack of planning and coordination of major rehabilitation projects and their negative impact on traffic and the economy in general. The last element of the research proposal is to address the issue of planning interventions in space and time over a network by considering the uncertainties on the current state of bridges, their residual life, climate change scenarios, financial constraints and impacts on users.