A GIS-Based Decision Support System for Optimal Renewal Planning of Sewers

1
A GIS-Based Decision Support System for Optimal
Renewal Planning of Sewers
Mahmoud Halfawy, Leila Dridi, Samar
Baker NRC-Centre for Sustainable Infrastructure
Research
INFRA 2007 5-7 November 2007 Québec
2
Background

• Fragmentation of sewer management data and
  processes.
• Need for proactive and optimized renewal
  planning.
• State-of-practice in sewer management software.
• Challenges for integrating sewer management data,
  processes, and software systems.

3
NRC-CSIR Integrated Asset Management Project

• Objective
• Develop consistent/generalized models and
  protocols for asset management process
  systematization and data integration Bridging
  the vertical (departmental) and horizontal
  (cross-disciplinary) gaps.
• Develop algorithms and a set of interoperable
  GIS-based decision support tools for optimizing
  and coordinating asset management plans for
  water, sewer, and road networks.
• Project Partners City of Regina.

4
Integration of Municipal Asset Mgt Processes
Vertical Integration
Horizontal Integration
Optimized Coordinated Plans
5
The Renewal Planning Step-Wise Algorithm
6
Data and Process Integration Using Centralized
Shared Repositories
7
Part of the UML class diagram for the integrated
sewer data model
8
Renewal Planning Algorithm Implementation

• Define an integrated data model and build asset
  data repository (inventory, hydraulic data,
  condition data, repair/incidence records, risk
  parameters, cost data, service levels, etc.).
• Define an integrated condition rating index.
• Define /calibrating deterioration curves.
• Define risk assessment model.
• Define asset prioritization criteria.
• Define renewal technologies database, and
  algorithm for selecting feasible options.
• Define a multi-objective optimization (MOO)
  algorithm (maximize condition, minimize risk, and
  minimize budget).

9
(No Transcript)
10
(No Transcript)
11
(No Transcript)
12
(No Transcript)
13
Sanitary sewers and Vitrified Clay (2881 records)
Sanitary sewers and pipe condition 3 (823
records)
14
Deterioration Modeling

• Largely depends on the quantity and quality of
  available condition data.
• May use deterministic or probabilistic models
• Employs many different techniques regression
  analysis, Markov processes, ANN, fuzzy models,
  etc.
• Our approach
• Store a library of known or previously calibrated
  models
• If a sewer or a group has sufficient data to do
  regression analysis, define a new model or
  calibrate an existing model.
• If data is not enough, assist user to select a
  suitable model based on his/her
  intuition/experience with the system and the data
  available.
• As more data become available, the models can be
  re-defined or re-calibrated.

15
(No Transcript)
16
Condition, Risk Prioritization Analysis Tool
17
Risk Assessment Prioritization Models

• Risk Consequence of Failure Probability of
  Failure
• Criticality factors affecting consequence of
  failure
• Sewer type, diameter, depth, embedment soil, land
  use, road classification, traffic volume,
  proximity to critical assets, socio-economic
  impact, site seismicity, etc.
• Procedure
• Calculate a Risk Factor (1-5 scale) that reflects
  the consequence of failure using a weighted
  average equation.
• Calculate the likelihood of failure index (LFI)
  for the sewer, based on its current age and
  expected service life.
• Risk Index RF LFI
• Prioritization ranks sewers based on their
  priority index (1-5 scale) to select candidates
  for renewal.
• Priority index is derived from the condition
  index and risk indices according to a set of
  user-defined rules.

18
Renewal Methods Selection Tool

• Applicability criteria for method selection
• Sewer characteristics (diameter, material, depth,
  type gravity or pressure, structural condition
  state, hydraulic capacity, etc.)
• Method characteristics (renewal type (NS/SS/FS),
  limitations, site and installation requirements)
• Site characteristics (soil type, traffic, water
  table, etc.)
• Cost vs. Condition Improvement
• Expected condition improvement
• Technology construction cost
• Technology overall cost (socio-economic cost)
• Expected operational cost after improvement

19
Renewal Methods Selection Tool (Cont.)
20
(No Transcript)
21
GA-Based Multi-Objective Optimization Tool

• Trade-offs of the renewal costs vs. network
  condition and risk improvements.
• Three objective functions for cost, condition,
  and risk.
• Solution using the NSGA II algorithm and the Open
  Beagle class library.
• Calculate 2 Pareto fronts for condition-cost and
  risk-cost criteria.
• Evaluate feasible solutions and select or
  synthesize a solution.

22
(No Transcript)
23
Conclusions Future Work

• Integrated approaches and DSSs are critical for
  supporting proactive asset management strategies.
• The proposed approach and software prototype
  provided promising results.
• Work is ongoing to refine/improve several models
  employed in the prototype, and to validate the
  software with more data sets. These activities
  are conducted in collaboration with the City of
  Regina as well as other industry partners.
• The software modular architecture facilitated
  incremental development and testing, and would
  also facilitate future extensions, refinement,
  and interoperability with other (e.g. legacy)
  software systems.
• Need to define an industry-wide agenda for
  developing and adopting open/standard integrated
  data process models, and software architectures.

24
(No Transcript)