Effectively managing aging assets: using a simulation model to improve O&M and capital planning processes.

Electric utilities must manage a variety of aging distribution assets that are critical to system reliability. They are also faced with potentially huge costs when they need to replace these assets to maintain reliability. Making intelligent decisions about asset maintenance and replacement first requires that utilities have accurate information about the failure patterns of these assets over time. However, most data elements that could shed light on such patterns--asset condition, joint use, maintenance patterns, or results of stratified inspection--are not widely available. Still, utilities must forecast their capital and O&M spending requirements every year, regardless of their understanding of such asset failures.

In addition to these gaps in data, a lack of effective analytical tools and processes make it even more difficult to support such budget allocation decisions. For the most part, capital funding decisions are being made using a simple but potentially inaccurate forecasting method of taking the average of asset failures over a certain period of time. This often means existing replacement or maintenance strategies are not linked to the costs and reliability that will be experienced in several years. Thus, most utilities cannot quantitatively evaluate alternative strategies in order to select the best ones to implement. A more rigorous methodology is needed.

Understanding Asset Failures

To truly understand how an asset class fails over time, it is essential that utility companies capture and store historical data about asset failures. A database with critical asset attribute elements can provide insight to the pattern of failures and establish the framework for a probabilistic model that can replicate the failure patterns over time in a simulated environment.

Engineers often try to generate survivor curves for assets. If a utility can track an asset group from the time the assets were placed into service until the last remaining member of that asset year group has been taken out of service, then analyses such as survivor curves can applied to that group of assets. This approach is unrealistic as most utilities do not typically define the data needed to enable a "cradle-to-grave" analysis of asset failure patterns. Further, by the time a survivor curve has been generated for an asset, newer technologies are usually replacing that asset. On the other hand, basic asset life cycle characteristics, such as when an asset was put in service and when it failed, are for the most part available.

At a strategic level, this basic data does allow a robust analysis of failure patterns and provides the insight to required capital to replace failed assets, as the failures will likely occur with similar frequency. The key requirement is assessing the probabilistic nature of the failure, understanding the differences in mitigation options, and any associated reliability and financial impacts. This would allow asset managers to readily determine the financial and reliability impact of various asset replacement or maintenance strategies and their associated risks, and understand the near- and long-term implication of asset replacement and maintenance strategies aimed at mitigating these risks. This would also enable asset managers to clearly communicate and support their asset replacement requirements to senior management and regulators.

Condition-based asset modeling provides a way to scrutinize the asset failure patterns and can enhance a probabilistic model. However, the benefits of this type of modeling cannot be realized using historical failure data. Condition assessment programs have largely been an O&M expense and have typically been the first ones utilities cut or reduce. As a result, asset condition data is unreliable, leaving utilities to speculate about condition-based hazard functions and make assumptions about asset conditions. What utilities are left with is a large number of data points and a desire to fit these data points into a recognizable failure pattern. Again, this leads us back to basic probability of failure and the failure frequency analysis of historical data points.

Developing Asset Failure Curves

Given the lack of cradle-to-grave data and information on asset condition, a suitable alternative is to gather asset inventory information that includes, at a minimum, installation dates and, for failed assets, the corresponding failure date (those assets still in service would have a failure date that is left blank). Indeed, a lot of the available asset data sets are structured in this way. Establishing a clear definition of failure for an asset class is the initial step in developing an understanding of when assets fail. Analyzing the data in terms of in-service and failure dates allows a utility to see how asset failures are distributed by age, and at what age most failures are occurring. A unique failure frequency probability curve can be created for each asset class.

Where possible, asset failure data should be stratified in order to create a separate curve for each subset of the failure. For example, transformers as a class have a unique pattern of failure, but one failure frequency curve typically generalizes the analysis. If the manufacturer or load of the transformer can be shown to have a correlation to failures and be statistically valid in stratifying the data set, then a set of six failure frequency curves can be created from a single failure frequency curve for transformers (assuming three manufactures and two load types that pass a statistical test). This in turn allows for a more detailed and robust analysis of failures and forecasting against those failures.

The strength of a frequency of failure analysis is that the historical data establishes failure trends based on the data itself and not on assumed parameters. The failure probabilities are captured on an annual basis. In turn, utilities can update the mean and errors associated with failures. These means and standard errors allow for a probabilistic model that is easy to use and produces results that can readily be communicated to decision makers concerning the results of different asset failure mitigation strategies. Some argue that the weakness of failure frequency analysis is that it assumes age as a proxy for condition. This is not true. Condition based failures are included in the failure frequency analysis; they are just not broken out in detail. If there is any weakness to failure frequency analysis, it is that newer asset technologies have not aged enough to provide sufficient insight into their patterns of failure.

Employing Weibull Probability Distributions

To address the shortcoming of evaluating newer assets with limited failure frequency data, a Weibull probability distribution analysis can be used. While not easy to explain to those who do not understand probability distributions or statistics, the Weibull distribution has been used for many years to model the randomness of asset failures and understand the probabilities, risks and mitigation possibilities. A traditional 2-parameter Weibull distribution can be configured to capture "infant mortality," attrition (random events that can put an asset out of service, such as vehicles hitting a pole or cable dig-in), or failure due to aging. However, the Weibull distribution typically only captures one of these failure modes at a time. A unique attribute of the Weibull distribution is that the sum of Weibull distributions can be modeled as one distribution by adjusting the parameters and ensuring that a relative importance of each element is captured as well. A 4-paremeter Weibull distribution allows one to model the entire life cycle of an asset: infant mortality, attrition, and age.

The Weibull distribution can be used to smooth failure frequency curves and interpolate the probability of failure for an asset class in later years that data is absent from, as it is a newer technology. An example of this is XLP cable. As one of the newer underground cable technologies, there are no assets that have reached 80 years of service. This does not mean that assets that happen to age beyond the scope of a failure frequency curve will no longer fail. Forecasting the failures of these assets based on historical failures ignores the fact that some of these cable elements could continue to survive long after the failure frequency curve analysis indicates. The 4-parameter distribution takes this into account.

Today, utilities do have a choice in modeling aging asset failure patterns. They can use simple failure frequency curves, which are derived purely from available data through a very simply ratio of failures to total inventory--a method that is relatively easy to explain. Or they can use a fitted Weibull curve, which adds a theoretical dimension and is more complicated than a simple failure frequency curve, but is often misunderstood by decision makers.

Creating a Simulation Model

Creating an understanding of asset failure patterns provides does itself provide insight, but it still does not provide a means to analyzing the "what if" of mitigating the failures. What if a run-to-fail strategy is employed? What if a utility proactively replaces assets? What are the reliability and financial implications of these strategies? These questions require the ability to evaluate different strategies, and, just as important, compare those strategies to find the best solution to meet the goals and objectives of an asset management program. The ability to accurately evaluate competing asset management strategies is integral to a rigorous asset management and life-cycle analysis methodology.

A model that lends itself perfectly to this type of analysis is a probabilistic discrete-event simulation model. A discrete-event simulation model has a clock built into its execution. That is, at the start of a simulation run, the model understands the characteristics of an asset and can age it over any desired time horizon. By incorporating failure frequency probabilities or set Weibull probability distributions, a simulation model can track failures and mitigate these failures in the model environment with evaluated strategies and predict future spending and reliability impacts.

A simulation model differs considerably from standard forecast techniques. It can allow users to play out the aging process and parallel events such as replacement and maintenance by moving assets back and forth along the age spectrum. This allows users to predict failures far into the future, and also the age profile of the resulting inventory at different points in time and the reliability implications associated with various failure mitigation strategies. For example, a run-to-failure strategy can defer some capital costs, but also have a negative impact on system reliability statistics, such as SAIDI and SAIFI. A more proactive asset replacement strategy may result in increased capital spending in the short term but could actually improve reliability if the assets are replaced prior to failure.

Any model that compares strategies should provide enable utilities to evaluate a variety of strategies. This will provide managers with the flexibility to experiment with a broad spectrum of asset management programs that may involve replacing an asset when it is absolutely necessary or performing a mixture of strategies. For example, a manager may want to look at reactively replacing failed assets with a new asset class, while at the same time proactively replacing a certain portion of the total inventory each year and doing maintenance on yet another portion. When evaluating possible strategies, the two measures that utilities are most concerned with are cost and reliability. The dynamic interplay between failures, replacements and maintenance would merit a view of these measures over several years. A simulation model that clearly shows the projected rise and fall of costs over time, the evolution of a utility's asset age profile, and the consequent improvement or degradation of reliability parameters like SAIFI and SAIDI will allow for a more precise evaluation of possible maintenance and replacement strategies.

Ultimately, rigorous asset life cycle analysis will support utility rate strategies for cost recovery. By incorporating known asset failure frequency patterns and modeling different asset replacement and life extension strategies, utilities will be able to better understand and explain the financial and reliability risks of implementing different strategies for maintaining and replacing their aging distribution assets. In the end, the financial and reliability projections that a simulation model can produce will provide utilities with an effective foundation for both internal asset management decisions and the corresponding development of future rate cases.

The Weibull distribution is one of the most widely used lifetime distributions in reliability engineering. It provides the distribution of lifetimes of objects. It was originally proposed to quantify fatigue data, bit it is also used in analysis of systems involving a "weakest link". Named after Waloddi Weibull (a Swedish engineer, scientist and mathematician), it was introduced by R Rosin and E. Rammler in 1933. The American Society of Mechanical Engineers awarded Dr. Weibull their gold medal in 1972.

Patrick Delaney is vice president of Analytic Solutions with Davies Consulting, Inc., an international strategy and management consulting firm serving energy companies and utilities. Davies provides services in the areas of asset strategy, operations improvement, reliability performance, regulatory strategy, and emergency planning and response. Mr. Delaney has over 20 years experience in strategic and operational decision development, modeling and analysis.

Wiko Kabiling is an analytic consultant with Davies Consulting, Inc.