Optimal R&R Planning Using Advanced Quantified Risk Approach for Water Distribution Pipe

In 2009, Anchorage Water and Wastewater Utility (AWWU) began using a qualitative risk matrix to evaluate the relative importance of projects and better prioritize capital projects. Assets were evaluated and systematically assigned consequence and likelihood of failure, and, as a result, a risk score. The two limitations of this initial approach were: 1) subjective rather than objective measurements, and 2) difficulty in identifying the optimal timing of capital work. In order to address these two limitations, AWWU contracted with InfraPlan to assist in developing a process to identify the optimal timing for replacement for water pipe assets using objective measureable factors. This process was a collaborative effort involving two tools provided by InfraPlan, Casses and Economic Pipe Replacement Model (EPRM), along with scripting developed by AWWU to quantify consequence of failure. Casses is an advanced statistical tool that helps quantify the likelihood of failure of each pipe each year in the future using historic causal factors. InfraPlan assisted AWWU in developing the original datasets and trained AWWU staff in operating the software. In parallel, AWWU staff developed an in-depth triple bottom line consequence analysis quantified in dollar terms at the pipe segment level using advanced Geographical Information System (GIS) analytical techniques and scripting processes (FME Workbench). This resulted in being able to quickly (push of a button) run updated system-wide consequence analysis, and to update the scripting with additional consequence factors as the information becomes available. These two steps have removed the subjective nature of risk categorization and also allowed for evaluation of future risk. A final tool (EPRM) is then used to combine forecasted inflation and cost of capital rates, the quantified likelihood and consequence, and replacement costs to determine the optimal timing for R&R work. This analysis can be continuously updated when new data becomes available (or its quality improves).