ABSG Consulting has developed a process and associated software for risk-informed performance-based asset management (RIPBAM) of power plant facilities over recent years. The RIPBAM process applies a tiered set of models and supporting performance measures (metrics) that can ultimately be applied to support decisions affecting the allocation and management of plant resources (such as funding, staffing and scheduling). In general, the goal of the RIPBAM process is to continually support decision making to maximise the facility’s net present value (NPV) and long-term profitability for its owners. The current RIPBAM beta software is programmed using an integrated combination of off-the-shelf software including Microsoft Excel, Decisioneering Crystal Ball, and Microsoft Visual Basic for Applications.

In practical applications of RIPBAM at the South Texas Project Nuclear Operating Company (STPNOC), it has become apparent that applying predicted NPV and/or payback period alone in developing, evaluating, and implementing any changes is insufficient to support rigorous, prudent, comprehensive decision-making. Other plant metrics and measures such as reliability, availability, heat rate, safety levels (such as reactor core damage frequency or large early release frequency), frequency of undesired events or states, input parameter uncertainty, and so on, have been determined to play a key role in a prudent comprehensive decision support system. Also, experience has shown that decision support metrics must be applicable to ‘packages’ or sets of multiple improvement options in order to support effective and efficient resource optimisation.

The RIPBAM approach complements and integrates existing activities like probabilistic risk assessment (PRA), preventive maintenance optimisation (PMO), and lifecycle management (LCM) methodologies. RIPBAM involves the integrated assessment of many characteristics and performance measures associated with a power generating station.

Some issues providing motivation to electric utility companies for development and implementation of RIPBAM programmes are as follows:

  • Electric generating industry transition to a deregulated environment.
  • Increasing pressure to decrease incremental generation costs and increase long-term profitability at the generating station level and at the ‘fleet of plants’ level of the energy business.
  • Increasing electricity pricing competition among competing generation sources.
  • General corporate pressures to reduce costs and increase productivity.

Clearly, one of the strongest motivations for electric utility companies to implement a RIPBAM programme is to prudently and efficiently increase power generation revenues and reduce operations and maintenance costs that ultimately impact corporate profitability and competitiveness.

METHODOLOGY OF RIPBAM

The general concept of RIPBAM is to develop a rigorous systematic risk-informed approach to assessing, analysing, predicting, and monitoring power plant economic performance while maintaining high confidence that pre-established safety limits are not and will not be breached. In general, the RIPBAM process involves the modelling and probabilistic quantification of decision support performance indicators to aid plant decision makers in determining not only which plant improvement investment options should be implemented, but also how to prioritise plant resources for their implementation based on their predicted levels of profitability.

The performance indicators applied in RIPBAM analyses are defined and described in detail in the following documents:

  • The Standard Nuclear Performance Model – A Process Management Approach, EUCG Task Force Report, Revision 1 (December 2000), Nuclear Energy Institute.
  • Nuclear Power Financial Indicators for a Competitive Market, TR-1003050 (2001), Electric Power Research Institute, Palo Alto, California, USA.
  • Regulatory Assessment Performance Indicator Guideline, NEI 99-02, Revision 1 (April 2001), Nuclear Energy Institute.

In RIPBAM, future projections of these performance indicators are determined using a risk-informed approach. Unlike most conventional asset management approaches, RIPBAM includes the incorporation of low frequency, high-consequence events that can occur over the long term, as well as shorter-term ‘expected’ events, activities, and conditions, into its performance indicator prediction process. Thus, using the RIPBAM approach, these performance indicators incorporate predicted cost aversion issues as well as the more conventional direct operations and maintenance (O&M) and capital cost issues. In this way, the RIPBAM performance indicators are effectively ‘risk-informed’ and can be applied in evaluating cost aversion and revenue-impacting issues as well as conventional cost savings issues.

In addition to primary economic performance indicators, the RIPBAM methodology provides safety performance indicators, like core damage frequency and large early release frequency, which give the decision makers high confidence that investment implementation will not breach any pre-established safety criteria for the plant. In this way, decisions can be made to implement investments with positive impact on long-term profitability and reliability, while also providing for associated beneficial or insignificant impact on plant safety.

RIPBAM can also supply interim performance indicators, like projected plant trip frequency or projected generation loss, to characterise the predicted impact of recommended investments on plant reliability. Experience has shown that a structured evaluation of these quantitative decision support performance indicators not only provides valuable relative prioritisation information to the decision makers, but also ‘injects’ a more rigorous, systematic approach into the overall decision-making process than might otherwise be applied without their consideration. This process has been applied to ‘real’ decisions evaluated by STPNOC.

While most power stations have established processes that address portions of the profitability prediction process, most do not have a truly comprehensive top-down integrated focused approach for resource management like RIPBAM. The RIPBAM methodology incorporates that ability to predict important decision support performance indicators at the station or corporate level, while maintaining the capability of displaying, in an integrated fashion, useful breakdowns of contributors to the major decision support performance indicators. This ability to perform ‘risk decomposition’ and ‘risk roll-up’ comparative analyses within the RIPBAM methodology has proven to be very valuable in the decision-making and decision-implementation processes at power stations.

RIPBAM encompasses the consideration of all historical and potential future cost-impacting and revenue-impacting events, activities and conditions at the power plant (or plants). The scope of a fully developed RIPBAM process includes assessment, prediction, and monitoring of all significant factors impacting station costs and revenues, including spot market prices for electricity and associated electricity sales contracts. However, the primary focus of a practical RIPBAM process is generally on the factors that can be controlled or significantly influenced by station or corporate management and support staff. These factors include, but are not limited to, the following:

  • Planned outage frequency and duration.
  • Nuclear safety (for example, reactor fuel damage frequency, large early release of radioactivity frequency, frequency and magnitude of unplanned radiation exposure to the general public).
  • Reactor trip frequency.
  • Station reliability and availability performance (control of generation losses).
  • Unit thermal efficiency (heat rate) performance.
  • Frequency and magnitude of lost time industrial safety incidents.
  • Frequency and magnitude of liability lawsuits.
  • Direct O&M and capital costs (including costs for availability model, PRA model, heat rate model, and other associated RIPBAM elemental model implementation and maintenance).
  • Short- and long-term electricity sales contracting arrangements.

All of this information is generally available at power stations. The RIPBAM approach uses a top-down logical approach to identify and analyse cost and revenue impacting factors that make the most significant quantitative contributions to station or company long-term profitability and value, and therefore, likely present the most significant opportunities for plant improvement investment option development and implementation. Thus, RIPBAM can serve as an effective strategic decision support tool. Figure 1 presents a general overview of the asset management model concept and supporting generating station model and information source relationships to RIPBAM.

RIPBAM generates predictions probabilistically so that performance indicator information can be supplied to managers in terms of probability distributions as well as point estimates. Using the information provided by projected performance indicator probability distributions, managers and other decision makers can apply the concept of ‘confidence levels’ in their critical decision-making processes.

The ability to incorporate confidence intervals in decision criteria is a significant advantage of RIPBAM over conventional ‘point estimate only’ supported decision making. RIPBAM incorporates uncertainty in input data, analysis assumptions, and model success criteria. Figure 2 presents an example showing probability density distributions of projected investment NPV for two different investment options, both with the same predicted mean (or point estimate) NPV performance indicators ($300,000 in this case). However, to most plant managers, investment option 1 would be preferable, and would be recommended over option 2 in the RIPBAM process, because the outcome is more certain, and because there is much lower chance of a negative payback result with option 1.

RIPBAM includes both monitoring and trending of historical performance as well as prediction of future performance. After the baseline RIPBAM process is developed for a specific power station or company, it can be applied in the station asset management process. There can be many differing sources of motivation for plant investment. Investment can be motivated internally at the station by monitoring, trending, and analysis of selected plant performance indicators. In this process, the analysis of performance indicators includes peer group comparative analyses and benchmarking analyses to aid in the identification of potential investment or improvement options and strategies. Investment can also be motivated from outside sources such as regulatory agencies or owner/stockholder groups.

The RIPBAM model can be used to predict both baseline (‘base case’) performance and performance under an ‘investment case’, often referred to as a ‘delta case’ in the plant decision support framework.

Selection of decision criteria

In general, most experts agree that assessing the potential net benefit (in terms of impact on long-term NPV) and the benefit-to-cost ratio (BCR) of a planned change (or improvement option) is important to the decision process. However, there are many metrics that can be effectively applied in supporting power plant change management.

Firstly, when a RIPBAM process with appropriate safety and availability models is in place at a plant, some preliminary structure, system, and component (SSC) metrics, called importance measures, can be applied to support identification of potential areas for improvement. In a fully functional RIPBAM programme, profitability achievement worth (PAW) and profitability reduction worth (PRW) are applied to help identify those areas of future investment at the plant that have high potential for station NPV improvement.

The PAW is simply the ratio of the projected station annual profit given a specific cost category (such as component failure with associated corrective maintenance) is made perfect over the baseline station projected annual profit. Similarly, the associated issue PRW is the ratio of projected station annual profit given the same cost category is always in its maximum negative impact state (in this case the component failure is always in effect) over the projected baseline station annual profit. The baseline station annual profit is simply the projected base case station annual profit with no new implementation of changes in plant equipment, procedures, or policies.

While the PAW and PRW metrics are generally applied at the component failure mode level of indenture at STPNOC, other issues modelled in RIPBAM, such as human error modes, outage schedule change impacts, plant design modifications, and plant procedure changes can be evaluated with PAW and PRW metrics. It is important to note that assessment of component failure mode or other issue PAW and PRW metrics are only applied to help identify potential areas for improvement that could have high payback at the station, and that they are not used as a final determining factor in change implementation. In general, experience has shown that it is usually prudent to investigate changes or improvement options for those component failure modes or other issues that yield both a PAW and PRW value in the top 50% or less of the complete spectrum of component/issue PAW and PRW values.

After component failure modes or other issues have been determined to be candidates for improvement option development, an expert panel develops initial improvement option recommendations to address leading candidates for positive impact on long-term profitability. The next step is the three-tiered RIPBAM assessment process for these candidate improvement options. These items passing through successive levels of the RIPBAM screening process are ranked by two key metrics, absolute net benefit (NB – projected impact on long-term NPV) and BCR (expressed as the ratio of NB over the total cost of change implementation plus change effectiveness maintenance over plant life). Finally, a third ranked list of improvement options is developed based on increasing values of projected total integrated cost of implementation and maintenance over expected plant life.

Other supporting ranked lists of improvement options can be developed as desired based on supporting metrics such as payback period and return on investment (ROI), expressed in terms of equivalent annual percentage rate return (APR) on investment for the change option.

These lists with supporting analyses and technical documentation are then passed on to an expert panel charged with ‘packaging’ change options for potential implementation. Experience has shown that presenting the improvement options as points on a graph of BCR versus NB can be a valuable graphical display for decision makers.

In practice, selection of individual improvement options can be as simple as choosing those items on the list with positive NB and the highest positive BCR, continuing down the list of decreasing BCR options until the budget for changes is surpassed by the total projected cost of the set of high-BCR improvement options. However, this can sometimes be misleading and therefore the selection process should be supported by application of additional ‘constraint’ metrics, such as ROI, core damage frequency (CDF), and large early release frequency (LERF). Also, as illustrated earlier, it is important to apply principles of uncertainty analysis in the decision-making process. Therefore, confidence levels must be selected for the decision metrics applied.

The following list of parametric limits represents an example of a robust set of collective decision criteria:

  • Mean BCR >> 1.00.
  • High confidence (> 80%) that NB (or DNPV) > 0.
  • High confidence (> 95%) that DCDF < established limit.
  • High confidence (> 95%) that DLERF < established limit.
  • High confidence (> 80%) that ROI > established limit (the prevailing prime rate or federal funds rate).
  • High confidence (> 80%) that improvement option total cost < established limit (determined by utility budget).
  • Reasonable confidence (> 40%) that the improvement will actually be required or demanded by plant events and/or conditions over the remaining expected life of the plant (applies to those improvement options designed to mitigate or preclude failure scenarios).

Specific plant (or corporate) staff can modify these criteria or operating companies based on local risk perception and risk acceptance levels. These change implementation decision criteria would be applied in the plant (or fleet) budgeting process year to year with the goal of continuously improving plant (or fleet) NPV and change option ROI. Projects meeting all the collective decision criteria established by this process could be ranked according to some or several of these key decision support parameters (by decreasing value of projected ROI, BCR, NB, total cost, and so on). These ranked lists could then be used to help prioritise improvement option implementation scheduling.

Alternative investment packages

As outlined above, quantitative decision support performance indicators and associated decision criteria can provide valuable information to decision makers in evaluating individual recommended investments in plant equipment, operation, and/or maintenance practices. These quantitative performance indicators and decision criteria can, when applied correctly, provide even greater support in prioritising two or more ‘competing’ recommended ‘investment packages’. Experience has shown that, when taken in context, multiple improvement options can collectively have a different impact on plant safety and economics than might be expected by the evaluation of individual options (in other words, the ‘whole’ does not equal the simple sum of the individual parts).

Expert panels and operational experience can provide valuable information enabling RIPBAM teams to evaluate packages of improvement options that are planned to be implemented simultaneously or closely in time. The decision criteria and evaluation process would be similar to the evaluation of individual improvement options, but would be based on consolidated investment package metrics versus individual option metrics. With proper input from an expert panel, the RIPBAM team can support calculations of ‘package’ decision metrics and confidence levels as well as individual improvement option metrics. The goal, as always, is to develop investment packages that result in greater improvement to projected plant NPV and ROI.

APPLYING THE METHOD

While RIPBAM is designed to be capable of providing the full spectrum of asset management decision support performance indicators, such as those previously applied in Electric Power Research Institute (EPRI) LCM and nuclear asset management methods, the RIPBAM methodology focuses primarily on internal station management issues and applications, summarised as follows:

  • Refuelling outage schedule and duration optimisation.
  • Specific equipment design modification case studies.
  • Capital spares procurement analysis and prioritisation.
  • Major equipment refurbishment and replacement case studies.
  • Unit efficiency (heat rate) improvement case studies.
  • Station major maintenance activity prioritisation.
  • Plant life extension and licence renewal case studies.
  • Plant power uprate case studies.
  • Component ageing case studies.
  • Component obsolescence case studies.
  • Operating procedure training prioritisation.
  • Maintenance procedure training prioritisation.
  • Online versus offline maintenance trade-off studies.
  • Plant severe accident mitigation alternative (SAMA) issue evaluation case studies.
  • Procurement quality assurance audit and spot check prioritisation.

The focus of these RIPBAM applications is to continuously support development and implementation of effective and efficient station and fleet improvement investment options – those asset management decisions that support improved long-term profitability and safety – in a prudent, cost-effective manner. To date, RIPBAM has been applied using projected long-term average annual earnings (sometimes called profitability) as the key overriding or guiding performance indicator, with nuclear safety (and other performance indicators) used as decision ‘constraint’ performance indicators. In practice, the RIPBAM software is executed by RIPBAM practitioners, but the evaluation of case study cost, failure rate, repair time, and other key input data parameters is developed through data analysis and expert panel elicitation processes. Figure 3 presents an example graphical results output from RIPBAM.


Author Info:

James K Liming, ABSG Consulting Inc (ABS Consulting), 300 Commerce Drive, Suite 200, Irvine, California 92602-1305, USA


FilesFigure 3. Example of RIPBAM NB/ROI profile curves
Figure 2. Example application of confidence levels in RIPBAM
Figure 1. RIPBAM conceptual model overview