The civil infrastructure including buildings, highways, bridges, and water/sewer systems, is crucial for economic growth and prosperity. Sustaining the safety and operability of infrastructure networks is a complex task due to the harsh operational environment and the limited repair budgets. Among the various infrastructure systems, educational and health care buildings represent a great challenge due to their diverse components that have different repair requirements. While many asset management systems have been introduced in the literature, few have focused on buildings or some of its individual components (e.g., roofs). Ideally, an asset management system would include important functions such as condition assessment, deterioration prediction, repair selection, and component prioritization for repair along a planning horizon. Existing systems, however, may not adequately cover all these functions and lack optimization features that are suitable for large scale networks.; This research introduces a comprehensive asset management framework to support the efficient planning of maintenance and repair programs for educational buildings. The proposed framework has unique focus on tracking the dynamics of defects in various building components and in optimally repairing these defects. The framework introduces the following novel developments: (1) a simple visual approach to support speedy and less subjective assessment of the current severities of defects associated with various building components during field inspection; (2) a modified Markov chain approach using optimization for component-dependent prediction of future severities along the planning horizon; (3) a procedure for determining the least-cost strategy to repair component deficiencies in each year of the planning horizon; and (4) a procedure for network-level optimization to prioritize components for repair, considering practical constraints and user preferences. The framework, as such, is designed to deal with large scale networks by applying optimization sequentially to project-level decisions (repair types) and then to network-level decisions of selecting the components to repair along the planning horizon.; The framework has been implemented in user-friendly prototype and its performance tested using data obtained from a school board in North America. Based on extensive experimentation with various optimization techniques and strategies on different problem sizes, the system proved to be practical and capable of optimizing repair funds for up to 1,200 components, simultaneously, using gradient-based mathematical optimization, and for a much larger number using the genetic algorithms technique. The system as such, will aid consultants and owner organizations administering large inventory of buildings (e.g., municipalities and governmental agencies) in making appropriate decisions that ensure the sustainable operation of the infrastructure assets with least cost.
展开▼
