Infrastructure construction and rehabilitation: Risk-based life cycle cost analysis

The objective of this research is to provide a prediction model of life cycle costs for civil infrastructure construction and rehabilitation alternatives, taking into consideration the uncertainty involved in determining the service life of civil infrastructure units. The ability to determine the uncertainty values that accompany the predicted life-cycle costs allows decision-makers to select construction and rehabilitation alternatives with a knowledge of the inherent risks. This research utilizes a risk-based approach to predict probabilities of occurrence of total life cycle costs associated with the construction/rehabilitation of civil infrastructure unit. The uncertainty component of the model is introduced through the statistical probability distributions that represent the times-to-failure (i.e., service life) of infrastructure units.; In the developed risk-based life cycle cost model, the service life of each construction/rehabilitation alternative is input to the model using random variables sampled from the fitted statistical distributions that represent the service life of the infrastructure units. Computer simulation is used to generate the probability distribution of the model outcome (i.e., life cycle costs of construction alternatives). The resulting life cycle cost distribution is then statistically analyzed to provide a measure of risk. As aresult, decision-makers can determine the probabilities associated with various values of total life cycle costs. This enables them to make informed decisions regarding the construction/rehabilitation alternative that can be applied and the level of risk they wish to accept.; This dissertation provides a detailed explanation of the process involved in developing the risk-based life cycle model for civil engineering infrastructure. These process include: (1) defining factors affecting infrastructure performance; (2) performing statistical stratification of infrastructure inventory and failure data; (3) developing reliability and statistical models; and (4) applying simulation and risk analysis techniques. The model concept and its applicability are demonstrated using highway pavement data in the province of Alberta, Canada.