Probabilistic performance prediction modeling for bridges considering maintenance effects within a combined computation, visualization and programming environment

Various models have been developed in recent past for lifetime performance prediction of bridges. Studies for developing new theoretical models and improvement of the existing ones are part of ongoing research in bridge deterioration and management. However, improvement of such theoretical models so that they can be practically implemented in bridge management systems is a challenging task. In order to achieve such task, newly developed models must be tested and further improved to achieve practicality. Lifetime performance prediction for bridges can be performed using either a safety or condition criteria. In a well designed bridge management system, it is well established that both of these criteria should be implemented and monitored at the same time. In this study, a bridge performance prediction model developed by researchers in this field has been implemented in Matlab environment for further improvement. Computation, visualization and programming tasks are performed as part of this application specific solution using the toolboxes. For the simulation of the random variables, Latin hypercube sampling technique is used which is available in the computational environment. Random variables may be assigned normal distributions. A function is also developed to generate samplings from other distributions such as triangular distribution. Instead of using superposition of the effects of maintenance actions on no-maintenance profiles, the newly developed algorithm consists of a simulation loop which contains a time loop embedded in it. Such an algorithm enables the generation of the performance profile over the whole lifetime at once for a single simulation of all random variables involved. This approach may prove to be a faster algorithm than the one previously developed. Mean value and standard deviation profiles for the condition and safety indices as well as the probability density distributions of these indices over the bridge lifetime are graphically generated using the integrated modeling and simulation environment. Using the developed environment, effect of various maintenance-repair-replacement scenarios on condition and safety profiles of bridges is investigated