Digging in the Dark: Making Asset Management Decisions in the Age of Imperfect Information

Abstract

The City of Atlanta Department of Watershed Management has been systematically improving the knowledge of its asset base for the last few years. The Sewer System Evaluation Surveys (SSES) program involved inspection of 1,600 miles of sanitary sewer pipelines, took a decade, and was recently completed. The department has now begun to make advancements in trying to gather knowledge of its drinking water systems through a valve and hydrant location program and geographic information system (GIS) mapping of drinking water lines. The classical approach to asset management suggests that an organization work on knowing what asset it has before beginning condition assessment programs. Atlanta is finding that this approach of developing a comprehensive asset register is too expensive, time consuming, and impractical for an organization that needs to be constantly thinking about the best ways to allocate capital funding and to make decisions on asset renewal, rehabilitation, and replacement. Moreover, the ready availability of robust/current operations and maintenance data (O&M) always tends to be a challenge. In this paper, the authors show how the City of Atlanta is developing models based on risk and criticality evaluations with available data for its drinking water and sanitary sewer pipelines. Atlanta's approach is to evolve a model that is continually updated as the data quality gets better because of targeted efforts in that space. Meanwhile, the city uses heuristics (rules of thumb), targeted risk and consequence of failure scoring, and sensitivity analysis of the scoring results to create lists of assets that can then be subject to conventional condition assessment programs. Because conventional condition assessment programs tend to be invasive and expensive, it is incumbent on Atlanta, which like most utilities is fiscally constrained, to develop desktop models that can serve as cost-effective simulations for narrowing down the number of miles of pipelines (both drinking water and wastewater) that need to be targeted for traditional condition assessment. This paper will detail the evolution of this microtargeting approach, based on available data, and then detail the mechanisms used for smoothing the data. Development of a robust and sustainable long-term asset management program for pipelines requires the ability to evaluate knowledge about a system's condition using mechanisms that are data driven and yet easily accessible. This paper will look at how Atlanta's Department of Watershed Management is developing an iterative approach to system condition information management and decision making.