6 Considerations For Navigating FERC’s Draft RIDM Guidelines
March 16, 2021
DEAN B. DURKEE, PH.D., PE
This blog was written prior to FERC finalizing its dam safety regulations on December 16, 2021. To learn about the finalized regulations, refer to FERC news.
Dam failures in recent years have prompted the Federal Energy Regulatory Commission (FERC) to scrutinize dam safety practices. While hydropower owners in the U.S. have invested considerable resources in performing Potential Failure Mode Analyses (PFMA) for their dams to meet FERC requirements for dam surveillance and monitoring, PFMAs stop short of providing an accurate portrayal of the true portfolio risk profile.
Risk, by definition, includes both likelihood and consequences of failure. The current FERC PFMA process does not allow for consideration of the consequences of failure or provide an adequate methodology for assessing likelihood. In addition, the use of the current FERC categories does not provide an appropriate portrayal of the overall portfolio risk. Because of this, FERC has released Draft Risk-Informed Decision Making (RIDM) Guidelines to supplement its Dam Safety Program.
Implementation of the Draft RIDM Guidelines has been slow due to uncertainty of the benefits and costs on the part of licensees and lack of experience with risk analysis among consultants. As a result, FERC has endorsed licensees to conduct pilot project risk analyses in specific cases, utilizing the methodologies presented in the Draft RIDM Guidelines. Several licensees have opted to perform pilot projects utilizing the Draft RIDM Guidelines. These recent efforts have not been without challenges and lessons learned, which can be categorized into six general areas.
1. Use of Previously Developed Potential Failure Modes
Since the early 2000s, licensees, through five-year periodic inspections and associated PFMAs, have developed potential failure modes (PFM) that could form the basis for the implementation of RIDM. If formulated correctly, these PFMs would be initially evaluated utilizing the Draft RIDM Guidance for evaluating the likelihood and consequences of failure. However, in many cases, the previously developed PFMs are of insufficient detail to accurately portray the initiation, continuation, and progression to failure. Because of this, most licensees have been required to spend excessive amounts of time and money brainstorming, screening, and developing failure modes that, had they been properly developed previously, would have formed the basis for evaluating risk.
In addition, over the last 20 years, many licensees have been reluctant to thoroughly brainstorm, screen, and develop PFMs; and many credible PFMs were overlooked, dismissed as non-credible, or ruled out. This has also added to the need for licensees to spend much more time during the implementation of the Draft RIDM Guidelines, brainstorming, screening, and developing PFMs which requires significant investment to move from the PFMA process to a RIDM process. This could be addressed by performing a thorough portfolio screening.
2. Limited to No Flood Frequency or Probabilistic Seismic Hazard Information
Moving from qualitative risk analysis (PFMA) to semi-quantitative (SQRA) and quantitative risk analysis (QRA) requires a better understanding of probabilistic relationships for flood and seismic events. The FERC PFMA process utilizes deterministic parameters, such as the probable maximum flood (PMF) and maximum credible earthquake (MCE) to evaluate PFMs. This clearly provides very limited information for estimating the likelihood of events of lesser magnitude, and, as a result, the critical load cases (loads of higher probability that result in similar consequences) would be missed. To properly assess likelihood moving into semi-quantitative and quantitative analysis requires that licensees perform these analyses. Consider performing probabilistic seismic hazard analysis (PSHA) and stochastic event flood modeling (SEFM) early in the process for use during the semi-quantitative and quantitative analyses.
3. Limited Understanding of Consequences of Failure
Because the PFMA process does not specifically require a rigorous assessment of consequences, very little information is typically available beyond inundation mapping and an estimate of the population at risk (PAR). While this provides an upper bound for SQRA, it is likely to grossly overestimate the potential life loss (PLL), which should include assessment of breach location and duration, warning time, depth and velocity, and proximity to safe locations during a failure event. As a result, most licensees will need to perform some form of consequence analysis prior to performing the SQRA. Consider using the Reclamation Consequence Evaluation Method (RCEM) to estimate PLL for various failure scenarios.
4. Availability of Subject Matter Experts
Since the implementation of risk-informed decision making has been slow to develop outside of the federal dam owners in the U.S., the broader dam engineering community is not well versed in the use of risk analysis methods. As a result, there are limited numbers of experienced engineers to serve as subject matter experts, outside of those with previous U.S. Army Corps of Engineers and U.S. Bureau of Reclamation experience, where risk analysis has become commonplace over the last 10 to 20 years.
To help mitigate this problem, consider using subject matter experts (SME) who may lack an extensive risk background but are recognized for their strong technical credentials and structure the project such that there is sufficient time for the SMEs to seek training in risk analysis. Industry is responding with more and more opportunities for training through our professional organizations as well as USACE and FEMA.
5. Coordination and Schedule
The high demand for the relatively small pool of professionals who have the requisite experience may pose a challenge in finding individuals who have availability to perform the work. Plan ahead for your SQRA to allow flexibility and seek input from FERC regarding the important issues at your dam.
6. Simulation and Calculation Results
Don’t rush to lock in estimates that potentially lead to overconfidence, particularly relating to erodibility and consequence estimation. Estimate a low and high around a reasonable confidence interval, then perform adequate sensitivity analyses and properly evaluate “what-if” scenarios.