Integrated Infrastructure Management Decision-Making Using ArcGIS™

Lynne Higgins, P.Eng
Infrastructure Services Department
Saskatoon, Canada

Jan-Mark Gustafson, P.Eng., Ph.D.
Infrastructure Services Department
Saskatoon, Canada

Dale V. Clancy, P.Eng.,
Infrastructure Services Department
Saskatoon, Canada

Abstract
A holistic approach to managing infrastructure in a public right-of-way is achievable through ArcGIS™. The City of Saskatoon has developed an initial framework for considering spatial and temporal relationships between multiple assets for use in developing interdependent rehabilitation and funding strategies. ArcGIS™ accounts for the influence of one asset on another through spatial relationships, leading to targeted work plans and more appropriate overall spending.

INTRODUCTION

In a perfect world, utility companies and municipal governments would work closely to develop short- and long-term strategies for rehabilitating assets within the public right-of-way at minimum long-term cost. The reality is that assets are managed separately by municipal governments and/or publicly-held and private corporations, all operating under separate boards or councils. Ultimately, these agencies are responsible to their communities for the levels of service delivered and for the cost of providing the service. With increasing community expectations for value in tax and utility payments and with legislative requirements to develop long-term plans for the sustainability of assets, an overall asset management strategy provides opportunities to minimize the costs of providing services.

The economic life of assets within an urban environment is highly variable. Some, such as sewer mains, can be still in use after hundreds of years. Water mains are expected to last 10 to 300 years; sidewalks 50 to 150 years. Pavements, however, generally have a shorter economic life, typically only 10 to 50 years. This difference in economic life affects how these assets are managed separately and how integrated management of these assets can be handled.

Predictive and economic models have been developed to consider the long-term economics of individual assets or asset groups. In most cases, these models have been developed using asset-specific software. ArcGIS allows for integration of the modeling output using a common platform and a consistent spatial reference. This facilitates the development of rehabilitation plans beyond what will be done next year. With multi-year plans, other assets within the right-of-way can be considered well in advance of the planned rehabilitation date. Where information or analysis has not been completed, asset stakeholders can plan and budget for inspection, analysis, and rehabilitation that will complement the rehabilitation programs of all assets.

To ensure the sustainability of an asset, a long-term planning horizon is needed. Asset stakeholders need to know what work has been done, what work needs to be done, and to understand this information in a framework that considers all associated assets within the right-of-way. With ArcGIS, asset stakeholders can look beyond the confines of their specific responsibilities and can now consider information about other assets that may influence treatment selection and timing for their own assets. This leads to integrated action plans to improve inspection scheduling and rehabilitation decisions. While making optimum decisions is the goal in this work, making better decisions that result in better economic outcomes has considerable benefit.

BACKGROUND

Centrally located on the Canadian Prairies, Saskatoon was incorporated in 1906 to facilitate the construction of water and sewer systems. The City currently has a population of 208,000 people serviced by 980 km of road, 850 km of water mains, 850 km of sanitary sewers, 510 km of storm sewer mains, and 58,000 sewer services.

After decades of system expansion combined with limited maintenance, the need for long-term preservation of existing assets was identified. Rehabilitation programs for assets such as cast iron water mains and Arterial roads were introduced in 1982, with additional programs for other assets added over time.

The need for better management tools was acknowledged with the development of Geographic Information Systems to map the City's infrastructure. As technology developed, the City moved from a paper-based information system to a database system for the storage of asset data, including inventory, condition, and maintenance and rehabilitation treatments. Although this work was done in the same department, these databases were developed separately and in different formats.

Comparable initiatives were occurring throughout the organization and the need for a common corporate platform was identified. In 2001, ArcGIS was chosen. This created an opportunity to bring information about different assets together into a single platform. The first key GIS components to be identified corporately were the parcel and transportation models. Training on Esri software began in the summer of 2001.

Organizationally, changes were also happening that would make integration more feasible. In 1997, as part of the corporate strategic plan, an Asset Preservation Group was formed within the Public Works Branch with a mandate to provide:

"Strategic level analysis to sustain the Public Works infrastructure
at minimum long-term costs."

This group combined expertise in design, construction, and operations of both surface and underground infrastructure with expertise in economic decision analysis and statistics. Economics is the group's primary consideration.

In keeping with its mandate, the group develops decision models that focus on minimizing the long-term costs, i.e. current plus discounted future costs. These models take into consideration minimum acceptable levels of service and budget/resources as constraints on the possible choices. In implementing the models, the starting point for minimum acceptable levels of service is determined by legislative requirements, e.g. safety, current practice and precedence, or that which is achievable using current budget/resources. This usually results in a backlog. City Council is initially asked to approve a financial and resource plan to eliminate the backlog as soon as possible. After the backlog has been eliminated, City Council is asked to consider improvements in the minimum acceptable levels of service in relation to the increase in budgets/resources required to achieve them.

The group has implemented models for cast iron water mains (Gustafson, Clancy1) and roads. Work is also underway on development of models for the evaluation of sewer mains (Clancy, Gustafson, Higgins2) and sidewalks. Future models are expected to consider valves, manholes, service connections, and hydrants.

SPATIAL RELATIONSHIPS IN A RIGHT-OF-WAY

This paper focuses on the infrastructure assets located within an urban right-of-way. These assets include roads and abutting sidewalks or boulevards, plus any underground and overhead utilities. Consideration is also given to abutting land parcels with associated service connections and property drainage.

Utilities such as water and sewer mains are typically installed beneath the road. In some cities, these are under the boulevard to reduce utility-driven damage to the roads. The spatial influence of all these assets is illustrated in Figure 1, a typical cross section through a right-of-way in Saskatoon. Work on any of the underground assets not only can drastically affect the road structure above, but the proximity of underground assets to one another can affect how the treatment of one impacts another.

In Figure 1, the inverted triangles indicate the trench that would result if the water, sewer, or storm sewer main were rehabilitated by open cut excavation. Where the trenches overlap, some backfill material would be re-excavated if the mains were replaced separately. The diagram also illustrates the extent of road that would be removed for an open cut excavation. There are trench overlaps in Figure 1 representing the extent of road that would be disturbed by asset rehabilitations at different times.

The City of Saskatoon owns and operates its own water distribution system and the sanitary and storm sewer collection systems. This may make asset integration in a public right-of-way more feasible than when multiple agencies are involved. The analysis presented will focus on the relationship between the water and sewer utilities and roads, but the process is anticipated to eventually include all assets within the right-of-way.


Figure 1: Assets in a Typical Urban Right-of-Way.

ASSET-SPECIFIC PRESERVATION STRATEGIES

The underlying processes for developing asset preservation strategies differ for each asset. Each strategy is a function of current levels of knowledge regarding condition and inventory, treatment choices, timing, and economics, and analysis software. The output of each strategy is the best choice in inspection timing, next treatment choice, and treatment timing for each asset. These models are at differing stages of development and at various levels of economic and network analysis.

The integration framework proposes to take the outputs of these independent strategies and collectively consider the best overall choices for all assets.

Minimum Expected Economic Loss Model For Cast Iron Water Mains

The Minimum Expected Economic Loss (MEEL) model was developed using a semi-Markov process to model the performance of cast iron and steel water mains from data collected since 1958 (Gustafson, Clancy3). The results of the performance model are utilized in a predictive model using a Monte Carlo simulation. These results are combined with cost models for repairs and replacement in a model that predicts the economic loss of replacing a water main after each of the first through 15th breaks. That break number after which replacement results in the minimum expected economic loss is the MEEL Break for a given length of pipe and category of cast iron water main (which uses year of installation as a proxy for wall thickness). By subtracting the number of breaks from the MEEL Break, the number of breaks before consideration of replacement, termed the BBR, can be determined. Where the replacement demand exceeds the annual funding allocated for replacement, a backlog of required replacements is generated, and the MEEL model determines a future date for replacements for water mains within this backlog. Where sufficient funds are allocated to replace water mains that are at their MEEL Break, the BBR is used to consider future replacement needs. Figure 2 shows the status of four blocks of water mains based on BBR. All of these blocks are far from the current standard MEEL Break of 15 breaks. This makes any road rehabilitation decision here an easy one - planned roadwork can proceed.

Figure 2: Expected Future Water Main Replacement based on BBR.

Road Rehabilitation Model

The road model utilizes a Markov process in determining strategic times to select a treatment given funding limitations (VEMAX4). The software utilizes existing condition and expected road performance to evaluate the benefits of all treatment options and selects the mix of treatments that will give the best overall benefit. These treatment programs represent the ideal; in reality, it is not always going to be possible to carry out the exact program generated for the year. Factors such as coordination with planned underground rehabilitation, planned traffic safety and capacity improvements, resource limitations, detour issues, and special events routinely lead to adjustments to the original program. This adjustment in treatment timing is reasonable from a performance perspective; the least cost to maintain a road in reasonable condition lies at the bottom of a "bathtub" curve, as shown in Figure 3 . Given that the curve will be relatively flat for a portion of the life of an asset, this means that doing a treatment in any one of several years can still be considered the "best timing".

Figure 3: Typical "Bathtub" Curve for Least Cost Maintenance and Rehabilitation.

The City of Saskatoon has developed a 3-year road rehabilitation program for longer-term logistics planning and to facilitate multi-asset integration. Figure 4 illustrates part of the output from the Road Rehabilitation Model. Planned 2002 work is shown in green, 2003 work in blue, and 2004 work in red.

Figure 4: Three-Year Road Rehabilitation Program (Reconstruction and Overlays only).

Sewer Rehabilitation Model

The sewer rehabilitation model development is based on video camera inspection of those locations with the highest probability of requiring rehabilitation or where the consequences of failure would result in significant property damage. Sewers in these categories are inspected by video camera and then evaluated using the North American Association of Pipeline Inspectors (NAAPI) score system with some modification for local conditions.

Total and peak scores are considered in the development of a rehabilitation program. For concrete sewers where deterioration is due to corrosion and tends to be more or less consistent along the entire length of the pipe, the total score is considered and full linings are normally selected. Lining alternatives that utilize particular treatments are considered first in rehabilitation of clay tile sewer mains where defects are localized along pipe lengths.

Figure 5 illustrates the output of the modified NAAPI-based evaluation process. The higher NAAPI score indicates where the risk of failure is the highest and lining would be done soon.

Figure 5: Sewer Overall Score Using the NAAPI Scoring Process.

Other Assets

The above models independently determine the candidates for rehabilitation of roads, sewers, and water mains. This asset rehabilitation work represents a significant portion of annual preservation budgets. Asset-specific programs are also carried out on an annual basis on sidewalks and appurtenances associated with underground mains. Although not currently modeled for optimized timing of replacement, valves, hydrants, water and sewer service connections, and manholes are also replaced or rehabilitated as part of an ongoing operating and rehabilitation program.

INITIAL INTEGRATION FRAMEWORK

ArcGIS brings together the disparate outputs of the water main, sewer main and road preservation processes. Every integration decision considers inter-related factors such as proximity, treatment choices and timing, treatment impact or influence, and economics.

Interrelation of Factors

The spatial relationship between assets is key to determining preservation strategies across multiple assets. For example, it would not make economic sense to replace a water main that has reached the end of its economical life and leave behind a deteriorating sewer main that would require the removal of the water main in the future. It would also make little sense to reconstruct a road and two years later excavate and replace the sewer main. The issue is to consider the economics and expected life of existing assets in determining a treatment strategy by looking at either a top-to-bottom or bottom-to-top strategy in defining the Asset Preservation Plan for all assets in a given right-of-way.

In the top-to-bottom strategy, the road agency considers the remaining life of underground utilities in its strategy planning. Where the surface asset will be impacted by an underground utility within the expected life of the rehabilitation treatment, consideration should be given to selecting a less costly, intermediate road treatment that still achieves safe driving conditions.

Traditionally, road agencies would replace a road's riding surface and sometimes the underlying structure when the road could no longer be economically maintained. Now there are intermediate treatments that can be utilized to maintain the investment in the road. Each has different best intervention timing, cost, life expectancy, and economic value. It may be economically appropriate to use one of these treatments to extend the life of the surface until the underground utility approaches the end of its economic life. In the best outcome, the surface asset would be maintained in safe condition at the least cost until optimal timing for renewal of the underground asset is reached. Alternatively, it could be a better economic decision to undertake the road rehabilitation early, foregoing the remaining economic life of the underground asset, where the cost of maintaining the road in a safe condition exceeds the value of the remaining life of the underground asset.

In the past decade, sewer rehabilitation has focused on trenchless technologies to deal with the high cost of open cut and trench box strategies. Where the main can be effectively rehabilitated using a trenchless strategy, the decision can be made to proceed with a higher cost road treatment. There is still some risk that the decision could lead to emergency and planned excavations if the sewer is not lined before the surface is rehabilitated. However, this would be expected to be a localized event that does not alter expectations for a good economic decision.

In a bottom-to-top approach, an underground asset is nearing the end of its economic life. Any rehabilitation strategy decision will impact the surface asset in key ways. In maintaining the water and sewer service, emergency and planned repairs will be required, resulting in excavations in the road. Failure of the water system may result in the slow escape of water into the road's structure. This increases the moisture content of the trench backfill, affecting not only the localized area of the failure, but potentially weakening the road structure along the length of the pipe as water seeps along on the outside of the water pipe.

If the deteriorated asset is the water main, rehabilitation choices are currently limited to open cut, trench box, or shaft-and-auger. The road surface is likely to have deteriorated due to past repairs to the water main. If the sewer asset is deteriorated and sewer lining is not expected to be an alternative in the future, replacing both the water and the sewer main at the same time may be the best decision. Conversely, it may be more economical to defer the water main and incur additional failures. This may be a good economic decision but is not without risk due to the costs of emergency repairs and the inconvenience to the customers that must also be considered.

For a deteriorating sewer main, the same excavation treatments are available as for water mains, but with the emergence of trenchless technologies, utilities have additional alternatives. If the sewer can be lined, it may still be more economical to replace the water main in one open cut or it may be more economical to wait until the road surface had deteriorated further before carrying out any underground rehabilitation.

Whether considering a top-to-bottom or bottom-to-top approach, the rehabilitation decisions are dependent on not only the asset being at the end of its economical life, but also on the existence and condition of all the other assets within the right-of-way. ArcGIS provides a platform to consider all assets within an urban right-of-way. Where the analysis has been carried out, the opportunity exists to account for the condition and expected performance of other assets on the decisions regarding one particular asset. This facilitates the development of integrated Asset Preservation Plans for a given right-of-way.

Initial Integration

Figure 6 shows the first combination of planned water main, sewer main, and road programs for 2002. Roads are in green, water mains are in blue, and sewer mains are shown in red. All of these programs were developed independently. However, this comprehensive map is the starting point for the integration of decision-making. Conflicts are immediately apparent and other factors can be brought into consideration early on.

Figure 6: Initial Program Integration - 2002 Major Treatment Programs for Water Mains, Sewer Mains, and Roads.

This spatial representation replaces the lengthy lists used in the previous non-integrated asset preservation process. Our former process was generally top-to-bottom with asset stakeholders operating independently. The list of proposed road rehabilitation locations was circulated to the other asset managers, who then independently assessed this information against their own planned programs. Considerable time was spent collecting targeted condition information and proposing revisions to treatments or their timing based upon this current knowledge. This meant that the original programs could be revised frequently as conflicts and issues were dealt with. Eventually, a comprehensive map would be created for that year's work and publicly released. This process was largely inefficient with a potential to lead to uneconomic decisions.

With ArcGIS, integration begins with the comprehensive map to identify areas of conflict for checking. Our initial framework was structured to consider all other assets around those locations where major - expensive and longest-lasting - road treatments such as overlays and reconstruction were scheduled. Any impact due to work on other assets could prove costly.

At all of these locations, a series of questions are answered, based on the level of knowledge about all assets, treatment choices, and particularly about the economic implications of options, regardless of which asset is driving the final treatment choices. From the roadwork perspective, for example, these questions would be:

"Does the asset location indicate that
it is likely to be impacted by the road work?"

"Is the asset condition such that
it should be dealt with prior to the road work?"

"Is the asset condition such that
the road work should be delayed?"

"Are there factors that justify altering
the selected road treatment?"

"What is the risk of a poor decision regarding the other assets
if the road work proceeds as planned?"

For many of the locations, the assets do not impact on each other and rehabilitation work can proceed as planned. These are the "No Brainer" decisions. It is, however, for the other locations that the use of ArcGIS is vital to making an informed economic decision.

Repair work, including any needed replacement, needs to be done in advance of the major roadwork. The time required to repair associated appurtenances in advance of the roadwork is a potential delay that can be reduced by using existing records to define what assets needed further review. For example, instead of checking all valves in the field, their condition and most recent inspection date can be reviewed through ArcGIS to easily determine which ones still need inspection or repair and which can be safely ignored. This focused, targeted work plan for one asset could then be better coordinated with the planned roadwork.

Other appurtenances can also be identified and evaluated with this process. The water and sewer repair database is accessed to determine all appurtenances recently rehabilitated or repaired and to show these for each road segment under consideration.

EXAMPLES

The following examples illustrate some of the decisions now possible using ArcGIS.

Example 1 - 25th Street from Spadina to 2nd Avenue

This Arterial road begins at one of the City's bridges into the downtown area and carries 40,000 vehicles AADT. Because it has been selected for an overlay, a major road treatment, the Asset Preservation Plan for this road must account for the underground mains and associated appurtenances.

Water mains were the first consideration. A check of the break history indicated that there were no major concerns. No water mains were close to the number of breaks at which replacement was economically justified (Gustafson, Clancy3). Water mains with no breaks to date are assumed to continue without breaks. Primary water mains, regardless of the number of breaks, are not considered for replacement; however, if a primary water main does not yet have cathodic protection, this work is done before the roadwork.

Table 1 shows that those water mains which had any breaks at all still have a low probability that they would have enough breaks to justify replacement within the expected 20 year life of the overlay. This does not, however, preclude the possibility that there may still be some water main breaks sometime within the life of the overlay, but this risk is considered acceptable.

Table 1: Water Main Break Analysis - 25th Street.

IDTYPESIZEBREAKSCONDITION FINDINGS
550ci30013 % probability of replacement
551steel6001Not included in analysis
552steel7507Not included in analysis
553ci15011% probability of replacement
554ci3000Not under consideration for replacement
555ci30029 % probability of replacement
556ci1000Not under consideration for replacement
557steel7500Not included in analysis
558ci1500Not under consideration for replacement
559steel7500Not included in analysis
560ci3000Not under consideration for replacement
561ci3000Not under consideration for replacement
562ci20011% probability of replacement
563steel7500Not included in analysis
564steel7500 Not included in analysis
565ci20011 % probability of replacement
566ci3000Not under consideration for replacement
569steel7500Not included in analysis
570ci300118 % probability of replacement
571ci3000Not under consideration for replacement
572steel7500Not included in analysis

The video camera inspection database for sewers is also accessed as part of the process to form an Asset Preservation Plan. Those sewer mains that have been video inspected are colour-coded by the year of the last inspection as shown in Figure 7. Dotted lines indicate sewers that have not been inspected and, therefore, require a complete investigation. For this example, all of the sewers have been recently inspected and no work is required on any of them.

Figure 7: Sewer Video Inspection Requirements - 25th Street.

Time becomes a factor in the video inspection decision. If the video inspection had been done too long ago, it is possible for conditions to change sufficiently to justify another inspection. This reflects the comfort levels within the decision - the age of the data versus the importance of the treatment. We have identified the need to set parameters relative to this factor, but have not formally documented them.

In Figure 8, the Select by Attributes feature in ArcMap is used to first identify the manholes falling within the roadwork limits that have already been inspected. The majority of manhole inspections are carried out as part of an ongoing cleaning process for maintaining sewer mains. The inspection database is attached to the manhole layer and those manholes that have been inspected are shown in yellow and a required treatment for each manhole has been identified. Most of the manholes along 25th Street have not yet been inspected; these are indicated in blue. Manholes outside of the scope of the roadwork are smaller black dots.

Figure 8: Manhole Inspection Requirements - 25th Street.

Similarly, we looked at valve maintenance and repair records to determine the extent of work needed prior to the overlay. The inset in Figure 9 shows a close-up of the east end of 25th Street where the two valves shown in red have been repaired or replaced within the past five years. All valves needing inspection are shown in blue with valves outside of the scope of the area of roadwork shown as smaller black dots.

Figure 9: Valve Inspection Requirements - 25th Street.

All of this information is combined into a composite Asset Preservation Plan for the right-of-way along 25th Street from the University Bridge to 1st Avenue. The Asset Preservation Plan indicates that the planned overlay can proceed with the following work needed beforehand:

Implicit in this Asset Preservation Plan is the belief that, within the planned 20-year life of this overlay, there is not likely to be a water main or sewer main replacement required, although spot repairs may be needed.

Knowing all this information about this location has enabled us to respond readily when we were advised of the intention by another Branch in our Department to modify the geometrics of this road. Because this work is not expected to proceed for two to ten years, we will not do the overlay as planned, but will simply do extensive patching - a less expensive, shorter-duration solution to maintain the road's quality temporarily.

Example 2 - 8th Street from Lorne Avenue to Broadway Avenue

8th Street is a major Arterial bisecting the east side of Saskatoon. Most of the properties along this road are zoned for either businesses or high density residential. The portion highlighted in this example links up to the main expressway around Saskatoon.

The planned treatment for this road is seal. Under our previous process, a seal wouldn't justify any in-depth review of the condition of underground mains. However, we have recently begun to utilize emerging technology to view the structural condition of the road and create a comprehensive picture of the overall road condition. This process is being adopted for all of our high volume, heavy load roads.

Figure 10: Plot of Road Structure using Ground Penetrating Radar- 8th Street Eastbound (Courtesy of Pavement Scientific International Inc.).

Figure 10 shows the underground structure of 8th Street. From this structural plot, we can determine if the road has any weak areas and if these areas are isolated or widespread. Because this plot is only indicative of a single wheel path, we extrapolate the information, deciding if further investigation is needed. From this information, we can plan the type and extent of treatment appropriate to the overall road condition. Then we use ArcGIS to illustrate the findings of the ground penetrating radar, pinpointing areas of suspected structural weakness. The intent is to locate and repair any weak areas before treating the surface of the road. In some cases, it is certainly possible that the structural repairs will be extensive enough to eliminate the need for the surface treatment.

With dynamic segmentation, isolated structurally weak areas can be plotted against the sewer condition results shown in Figure 11. The red flags in Figure 11 indicate spot locations needing specific repairs before the road is sealed and the red lines are suggested lining locations.

Figure 11: Sewer Condition and Defects - 8th Street.

Similarly, although not shown, water main details can also be plotted, creating a comprehensive "photo" of the right-of-way. Once the specific treatments needed for the individual assets have been determined, through economic analysis a targeted action plan for all affected assets can be determined that will achieve the best sequence of treatments and timing to achieve the greatest overall benefit for all assets.

CONCLUSION

Integrated decision-making can mitigate the impact of needed asset rehabilitation by facilitating planning and coordination.

We are making small decisions for now, but have begun the framework for more comprehensive integrated future decisions. Much of the information needed to make better decisions is already available, although not always in a readily-useable format. ArcGIS is the tool to use to make integrated decision-making in an urban right-of-way feasible.

REFERENCES

  1. Gustafson, J-M. & Clancy, D.V., Modeling the Occurrence of Breaks in Cast Iron Water Mains Using Methods of Survival Analysis. Proc. 1999 Annual Conference of the AWWA, Chicago (1999).
  2. Higgins, L, Gustafson, J-M. & Clancy, D.V., Economical Sewer Main Rehabilitation utilizing ArcGIS and Dynamic Segmentation, Proc. 2002 Esri Conference, San Diego (2002).
  3. Gustafson, J-M. & Clancy, D.V., Using Monte Carlo Simulation To Develop Economic Decision Criteria For The Replacement Of Cast Iron Water Mains, Proc. 1999 Annual Conference of the AWWA, Chicago (1999).
  4. VEMAX Management, Performance Prediction Technology ™ (Strategic and Tactical), IBOS Software Tools for Infrastructure Management