LAUSR

ABSTRACT PERFORMANCE-BASED BUDGET ALLOCATION MODEL FOR WATER NETWORKS Mostafa Ismaeel, MASc Concordia University, 2016 The assessment of water network is a challenge that concerns municipalities worldwide. Most of the water distribution systems around the world are deteriorating and, thus, their rehabilitation become urgent while it costs billions of dollars. According to the Canadian Infrastructure Report Card (CIRC, 2016), the Canadian water distribution systems are graded as “good: adequate for now” with 35% graded from “fair” to “very poor” and the estimated replacement cost is almost CAD 60 billion. The American Society of Civil Engineers Report Card has evaluated the condition of drinking water networks in the United States as “poor” with a grade “D”, stating that the United States water networks need USD 126 billion in order to reach a grade “B” by 2020 (ASCE, 2013). Thus, it is obvious that the necessity of providing continuous potable water under tight budgets plunks extra pressure on municipalities and triggers the need for a proper performance assessment. Accordingly, this research aims at developing a Water Networks Performance-Based Budget Allocation (WNPBA) model, composed of two sub-models: (1) Water Networks Performance Assessment (WNPA) model to precisely assess the performance of the water network components and (2) Budget Allocation (BA) model to optimally allocate budget according to the performance assessment. The WNPA model encompasses two key indices: (1) Pipes Performance Index (PPI) and (2) Accessories Performance Index (API). These indices reflect the status of network components and their deterioration levels and they propose consequent, preventative actions. The WNPA utilize the Fuzzy Analytical Network Process (FANP) to identify and evaluate the weight of functional performance criteria (i.e. physical, operational, quality of service and environmental) of pipes and accessories. It also exploits both the Preference Ranking Organization Method of Enrichment Evaluation (PROMETHEE) and the simple Multi Attribute Utility Theory (MAUT) to compute the functional and global performance indices of the network components. Moreover, the BA model utilize genetic algorithms (GA) and Greedy Heuristics (GH) to optimally allocate the available funds. The required data for this research is collected from experts and two water municipalities (Montreal, QC and Moncton, NB). The developed models are applied to the two water networks. The results show that most of City of Moncton sub-network 2 components are in a good or medium state, except for pipes 4 & 10 and accessory 7; those are in a poor state, while sub-network 1 is graded excellent for 1 accessory, good for 9 accessories and 7 pipes, medium for 23 accessories and 14 pipes and poor for 2 accessories and 14 pipes. The pipelines in city of Montreal sub-network are graded excellent for 16 pipes, good for 32 pipes and medium for 5 pipes while the accessories are graded excellent for 49 accessories, medium for 8 accessories and poor for 21 accessories. All the sub-networks are generally in a medium state (4