The effective management of municipal water distribution pipes is no exception and as such, many municipalities and utilities across the country and internationally have significantly changed their philosophical and pragmatic approach to managing their distribution systems.

Essential infrastructure

There is a growing need for cities and water utilities to find better ways to prioritise their infrastructure asset maintenance, rehabilitation and replacement projects and to intefrate infrastructure asset management techniques into their decision making. As infrastructure ages, it becomes increasingly more challenging to assign limited capital expenditures to the repair, rehabilitation or replacement of the assets. The prudent management of water mains requires a reasonable assessment of its current condition coupled with a reliable methodology to forecast future condition under different management scenarios.

It can be argued that water mains are one of the most difficult infrastructure assets to properly manage in a total asset management framework. This can be partly attributed to the difficulties associated with ascertaining a reliable measure for their physical condition. Faced with this challenge, the City of Hamilton embarked on an ambitious series of studies in 2006 geared towards the creation of a comprehensive tool set for managing water mains. The studies encompassed four main components:

1. Water Main Criticality Model; As the starting point for risk management, the criticality model classifies assets based on their ramifications of failure. The model encompasses economic, environmental, social and operational consequences of pipe failure. The model defines the subsequent management practices that will be used for high, low, and medium criticality pipe. For low criticality pipe, failure can be tolerated and the goal is to develop sound management policies that balance life-cycle costs with acceptable levels of service. For high criticality pipe, failure is not acceptable and hence more proactive policies driven by actual pipe condition and deterioration factors are sought. 2. Water Main Performance Model; The challenge associated with developing a performance model for water mains stems from the lack of reliable data on which to base the notion of ‘pipe condition’. The performance model is developed for low criticality pipe and uses the number of breaks as a proxy for condition. By mining the break history of various pipe vintages (ductile iron, cast iron pit cast, and cast iron spun cast), a life regression model is calibrated for times between subsequent breaks. The model is subsequently used as a forward looking predictive model to forecast the expected failure times. The model was used to assist in the development of economic intervention strategies for replacement and/or rehabilitation, long-term budget forecasts of repair and rehabilitation needs in addition to aiding at the tactical level in rationalising the coordination of capital works with sewers and road. 3. Critical Water Main Management; A unique approach for critical water mains is developed based on the proactive collection of condition information. The framework is composed of two main components: a condition assessment rationalisation framework and a condition rating consolidation framework. 4. Information Management Framework; The aforementioned components all require sound and reliable information. This framework aims to standardise the way information is used to make decisions pertaining to water main assets. The framework is currently developing standard information policies and practices that include all stakeholders who interact with water main information throughout its life cycle. In addition, the framework is investigating the most optimum use of HANSEN (CMMS) to support existing and evolving business processes within the City.

Water Main Criticality Model

The premise behind this model stems from the risk-based prioritisation and decision making concepts that are entrenched in the city’s overall approach to asset management. The intent of this model is to answer questions such as ‘Which water mains will have the greatest impact to the city, should a break occur?’ in order to focus resources and effort on these assets before they fail. Prior to the establishment of the criticality model, there was no standard method for the creation of a Water Main Criticality Model. It remains a subjective process in which the municipality must be heavily involved with the selection and ranking of parameters that they feel affect water main rehabilitation and replacement costs for the local area. Because of this, the parameters affecting cost of rehabilitation and replacement of water main infrastructure were selected and ranked within a joint effort between the consultant and the City of Hamilton.

Examples of criticality parameters include pipe characteristics such as material and underground depth, type of land use in which the pipe is situated, whether the pipe is connected to major water users or to important public health facilities such as hospitals and dialysis centres, whether the pipe is located in steep slopes or environmentally sensitive areas.

The application of GIS is particularly well suited to the development of Criticality Models because of its ability to apply many of the criticality parameters to the segments representing the water main assets through spatial analysis. Risk elements were organised into four main categories:

1. Economic – influence of the asset’s failure on monetary resources 2. Operational – influence of the asset’s failure on operational ability 3. Social – influence of the asset’s failure on society 4. Environmental – influence of the asset’s failure on the environment.

The result of the criticality model was the categorisation of the citys inventory of water mains into three distinct groups; high, medium and low criticality water mains. The model categorised approximately10 per cent of the network as having a high criticality, approximately 20 per cent of the network as medium criticality and the remainder as low criticality.

Water Main Performance Model

The performance model is geared towards addressing the needs of low criticality water mains. The model is based on a statistical approach founded on developing mathematical models that utilise past failure history to forecast future trends and variations in breakage rates. As mentioned in the introduction to this paper, the ST-LR approach models the time between successive water main failures. The approach models each failure number (i.e. 1st failure, 2nd failure, etc…) as a separate and distinct condition state. Five distinct cohorts were modelled; three vintages of Cast Iron (spun cast), pit cast iron, and Ductile Iron. Other material types were in use (Hyprotec ductile and plastic) but did not have sufficient failure history to warrant a standalone analysis.

Sample results of the calibration process are shown in table 1. Results show a relatively long mean time to first failure for ductile iron pipes compared to cast iron pipes. However, for mean time to subsequent failures (especially after the 4th failure) cast iron pipes tend to outperform ductile iron. No significant differences were noticed in the performance of spun and pit cast iron pipes.

The performance model was used to predict the following:

* Long-term funding requirements for repair and replacement/rehabilitation under different level of service scenarios. Level of service was measured by number of failures experienced on a water main segment. * Using a Monte Carlo simulation to predict future failure patterns on various water main vintages, the optimum time to replace/rehabilitate was computed as a function of the ratio to repair to replacement cost. This minimum expected economic loss (MEEL) was found to be somewhat large for typical cost ratios (8-12 failures on a water main segment). This indicated that the governing criterion is more likely to be a level of service indicator set by the municipality rather than pure economics.

In addition to the use of the model at the strategic planning level, the performance model was used to develop two important tools for water main management at the tactical/operational level. The coordinated infrastructure renewal tool and the early replacement tool set. Critical Water Main Management

The approach for managing critical water mains differs considerably from that for noncritical mains. Some of these key differences include:

* Repair policy: With noncritical water mains, breaks can be tolerated and hence a run-to-end of service life approach can be accepted. Conversely, critical water mains with zero tolerance for failure, a proactive maintenance and rehabilitation policy should be sought. * Tolerance to uncertainty: Whereas with noncritical water mains and their run to failure management approach uncertainty in condition state can be tolerated, no such tolerance can be allowed for critical water mains. This has ramifications on the amount of information being collected and the level of detail at which this information should be stored.

The critical water main management framework consists of three main tool sets.

1. Assessment rationalisation framework: 2. Condition rating consolidation framework: This tool set attempts to standardise the way the results of assessment techniques are interpreted and subsequently used to drive decisions. The framework is developed for ductile iron and cast iron water mains as they compose the majority of the city’s critical inventory. The framework utilises Fuzzy Logic and the Analytical Hierarchy Process to combine condition rating results into an overall rating for the condition state as well as the expected deterioration rate of the pipe. 3. Planning cycle decision analysis tool: The purpose of this tool is to equip the asset manager with a consistent methodology for decision-making during each planning cycle. Within a planning cycle, the asset manger must make one of three decisions for the critical water main inventory; schedule intervention, schedule inspection and revisit at next planning cycle.

In order to make an informed decision, the asset manager must consider the following aspects:

* Condition State. This is analogous to the probability of failure. * Pipeline Risk. This corresponds to the consequence of failure. * Extent of condition/deterioration information currently available. The tool performs a trade-off between the available amount of condition/deterioration information and the risk associated with operating the pipeline. * Level of uncertainty associated with inferring the condition state. Associated with the condition state that is inferred from the consolidation tool will be a measure of uncertainty. This factor must be considered in the decision process.

This study is still ongoing and aims to re-evaluate what information is collected, and the way information is stored and handled throughout the lifecycle of the water main assets.

Conclusion

With these tools built, the city has established the foundation for the effective management of its watermain infrastructure. Furthermore, as these tools are integrated into the daily business decision process, they will be refined and improved to reflect the increasing knowledge growth within the city. They will also form the basis for clearly articulating the ramifications of decisions. This includes the need to focus resources, particularly financial, on the assessment of critical infrastructure, which in many cases has not yet failed or otherwise caused operational issues. These types of studies are often expensive and do not result in new tangible assets, but rather an improved understanding of the probability of failure. As such, these types of expenditures can often be challenging for cities and utilities to get funding approval for with out being able to demonstrate the non tangible benefits.