Measuring Reliability & Maintenance Effectiveness On A Global Basis
EP Editorial Staff | October 18, 2010
Industries everywhere are finding themselves operating in a new reality. Here’s a new standard for dealing with it.
As some markets fl ourish and others recover from economic recession, it’s clear that the onetime buzzword “globalization” is now a reality. Today, industries compete on a global scale where, given acceptable product reliability and quality, price is the ultimate currency of success. With instant global communication capabilities available at virtually all levels, it is getting harder to sustain business performance at those companies that choose to isolate themselves and operate in a self-imposed cocoon or vacuum. Successful manufacturing performance is no longer tied to a technology, process or location. Instead, it’s basically a function of who can make a quality product at the lowest cost.
The value of performance benchmarking
All for-profit businesses operate under the same simple equation: Profit = Revenue – Cost. From this equation, it is easier to earn a dollar more profit by reducing expenses (which is your choice) than to earn an additional dollar of revenue (which is someone else’s choice.) In practice, there are interactions between expense reductions and revenue growth that also need to be understood to effectively optimize this deceptively simple equation. While there is no single business model to do this, one general strategy that can provide valuable insights is performance benchmarking—or, using Solomon’s terminology, Comparative Performance Analysis™ (CPA).
Benchmarking enables a company to measure and compare its performance against peer companies in a constructive and confi dential manner. The quantitative differences computed between a plant and other similar plants using a detailed data taxonomy can provide invaluable information regarding improvement opportunities. This is a way of effectively extending a “lessons learned” exercise across multiple companies.
A critical attribute of effective reliability and maintenance benchmarking is the ability to compare disparate assets. But even small differences for similar plants can alter the value of the comparison. Each asset-performance value (e.g., maintenance cost) is divided by a computed standard to normalize the results. Our company computes these standards using patented methodologies [Ref. 1 and 2] that are accepted around the world as integral parts of CPA studies in refining, petrochemicals, power generation, pipelines and terminals (see Sidebar below).
Historically, maintenance cost performance has been measured by dividing expense data by the plant replacement value (PRV). The generally accepted assumption is that the amount of money required to maintain a plant’s physical assets varies directly with the amount of money required to replace it. Intuitively, there is a logical connection in the sense that the more equipment a specific type of plant has, the higher the replacement and maintenance cost.
Individual companies use various methods to calculate or estimate PRV for insurance and other internal purposes—and in most cases, these values are accepted, internal standards. Aside from The Society for Maintenance and Reliability Professionals (SMRP) Work Management Guideline 1.0, Determining Asset Replacement Value, published in 2009, there’s no generally accepted common method for computing and auditing PRV calculations [Ref. 3]. Consequently, using PRV divisors for intercompany comparisons can add uncertainty to performance metric comparisons. Moreover, PRV inherently contains additional uncertainty because these calculations are seldom, if ever, verified in practice. There is little actual data on the true cost of plantreplacement at a given site in a given year since companies do not actually undergo such an ordeal.
For CPA studies of reliability and maintenance (RAM), our company has developed a new maintenance normalizer called Equivalent Maintenance Complexity (EMC™). The intent is not to displace the use of PRV, but rather to provide a more accurate means of benchmarking that does not possess the fi nancial, market and computational uncertainties inherent in PRV.
Since the objective of a standard is to bisect the actual performance data, roughly 50% of the actual performance is above and below the standard values. It’s not intended to be a predictor of future or past performance. Prediction techniques are intended as precise estimates that aren’t relevant to performance assessment. For example, if a prediction model were completely accurate, all asset ratios (Actual/Divisor) would be one—there would be no assets with above or below-par performance. The prediction model would include this modeling as part of its calculation, resulting in little comparative analysis value.
How EMC works
The EMC standard is computed as the nominal number of routine maintenance labor hours required to maintain a plant with its specific characteristics. For RAM benchmarking analyses, Solomon has chosen routine maintenance labor hours as the standard variable (instead of labor or other costs) because of uncertainties in exchange rates and regional labor cost differences. Labor costs required to maintain similar assets can vary widely across regions or countries, but labor hours should be the same.
Our method first determines, from data and expert opinion, the primary, direct drivers of routine maintenance labor hours. These variables are called “first principle characteristics.” In general, these characteristics fall into the following categories: location, process unit size, process type, process severity and the equipment counts for specific types like centrifugal pumps, electric motors, etc. After these variables are identified, a database of unit information is compiled and a non-linear optimization model is used to calculate the EMC coefficients as a function of first-principle characteristics.
Value limitations or constraints are applied to these coefficients to ensure that the results reflect realistic relationships. For example, all coefficients may be required to be positive, because increasing values can imply more (not fewer) labor hours.
Although PRV usage will continue to be a valid normalizing methodology, intercompany variability associated with the determination of PRV can and does influence this normalizing methodology—as illustrated in Fig. 1, which plots the actual rotating-equipment routine maintenance labor hours for refining process units as a function of the stated PRV.
While the general correlation is apparent in Fig. 1, the large amount of variation illustrates the limitations of using PRV as a performance-normalization variable. Using the same data set, EMC provides a more robust standard for comparison, as seen in Fig. 2.
The EMC standard is computed using non-turnaround maintenance labor hours and first-principle-characteristic data from study participants. One advantage of this methodology is that any plant with the same first principle characteristics values will have the same EMC. Variations in performance compared with the standard using the quotient [Actual/Standard] can point to weaknesses in specific types of maintenance practices using the data-collection taxonomy.
The measurement aspect of using EMC is just the beginning. The detailed data collected enables opportunity areas and the amount of savings to be identified. It is from these types of activities that CPA studies can be applied on a periodic basis to assess competitive positions.
To provide more insight in the benchmarking analysis, we divide total routine maintenance labor hours into the following four mutually exclusive categories. Other categories can be used given the accessibility of sufficient data.
- Rotating Equipment — Includes pumps, compressors, etc.
- Fixed Plant — Includes piping, process vessels, etc.
- Electrical — Includes motors, transformers, etc.
- Instrumentation and Control — Includes analyzers, control valves, etc.
EMC calculations are performed on each data set. The sum of the category level EMC standards represents the standard for the total routine maintenance labor hours. (Solomon currently has more than 8000 process units in its database that supports validation of the computed EMC values.)
The equipment-category-level EMC analysis provides more precision in identifying and quantifying areas of opportunity and areas that are performing well than by looking solely at the total number of hours. For example, a plant may perform better than standard performance overall, but show sub-par performance in one or more of the equipment categories [Ref. 4].
We selected routine maintenance labor hours as the foundation for this new normalization variable because they are within the control of the local facility management team—as opposed to other financial measures that aren’t. The objective is to provide a means for measuring reliability and maintenance performance by expressing it in terms of a metric that the maintenance manager can control. All too often, decisions are made regarding fixed costs that do not take into account the full impact of the maintenance function. Arbitrary cost cuts have been the norm, often resulting in ineffective or non-productive outcomes. This brings us to the second major component of reliability and maintenance benchmarking: reliability or, more precisely, availability.
The second piece
The second piece of RAM benchmarking focuses on lost margin that is attributable to reliability and maintenance causes. This value often far exceeds the total maintenance budget. By capturing and quantifying the value of reliability and maintenance- induced production losses, we can study the potential benefits derived by not cutting cost but by improving reliability. Empirical data suggests that corporations have finally begun to understand what their maintenance organizations have been telling them for decades. Although the motivation for improving reliability may historically have been driven by a desire to lower maintenance cost, it is the increase in profitability that warrants taking the time to understand the balance between reliability and maintenance from a strategic perspective.
In the next generation of the International Study of Plant Reliability and Maintenance Effectiveness (RAM Study), scheduled for 2011, our company will utilize published industry data to calculate the value of lost margin for each process family (i.e., a unique family of refining or process chemicals) that is attributable to reliability and maintenance causes. Examples of cost and margin quantifications have shown that EMC comparative performance analysis can identify opportunities. Specific decisions to change any work practices, however, are business decisions, not analytical decisions.
There are many reasons to benchmark reliability and maintenance performance. Among the most common reasons are:
- The knowledge or belief that the competition has developed or adopted better practices
- A desire to increase throughput without major capital investment
- The introduction of previously nonexistent factors as a result of global competition
- Ensuring the long-term viability of a business or manufacturing operation
- The need to set realistic performance improvement targets
- Prioritizing of alternative improvement opportunities. It’s been demonstrated, time and again, that by adopting reliability and maintenance best practices, asset availability goes up and maintenance costs come down. The EMC metric provides a new standard to help plants improve their performance in the new reality of the global marketplace. MT
1. System and method for determining equivalency factors for use in comparative performance analysis of industrial facilities, U.S. Patent #7,233,910.
2. Method and system for greenhouse gas emissions performance assessment and allocation, U.S. Patent #7,693,725.
3. Society for Maintenance and Reliability Professionals (www.SMRP.org), Work Management Guideline 1.0, Determining Replacement Asset Value (RAV), June 2009.
4. Poling, A., Jones, R., Hernu, M., “Increasing Profitability and Competitive Position by Comparing the Maintenance Effectiveness of Process Plants Based on all Plant Factors,” NPRA Maintenance Conference, San Antonio, May 2010.
Al Poling is project manager for the International Study of Plant Reliability and Maintenance Effectiveness for HSB Solomon Associates, LLC. Starting out as a professional educator, he later became a maintenance and reliability engineer and held various plant and leadership roles with several companies over the years. Poling served as Technical Director for the Society for Maintenance and Reliability Professionals (SMRP) from 2008 to 2010. He has presented at numerous conferences nationally and internationally, and has published several white papers and articles on maintenance and related topics. He completed his technical and educational undergraduate work at Fairmont State University and earned a clinical Master of Arts degree in Technology Education from West Virginia University. Telephone: (972) 739-1731; e-mail: Al.Poling@SolomonOnline.com.
Rick Jones has over 28 years of experience in corporate and industrial risk management, insurance and risk-based methods. As the director of Statistics and Risk Modeling with Solomon, he’s responsible for the development and application of numerical and statistical methods to measure and analyze operational performance from Solomon’s and customers’ databases. A frequent speaker on the role of risk management as a tool to improve reliability, Jones also has more than 50 publications and meeting contributions on various topics relating to reliability and risk management. The author of two books, Risk-Based Management: A Reliability-Centered Approach, and 20% Chance of Rain: Your Personal Guide to Risk, he received his B.S. in Aerospace Engineering and Ph.D. in Nuclear Science and Engineering from Virginia Tech. Telephone: (972) 739-1740; e-mail: Rick.Jones@SolomonOnline.com.
Driving world-class performance improvement for energy-intensive assets…
A subsidiary of the Hartford Steam Boiler organization, HSB Solomon is a provider of benchmarking services and performance improvement consulting for the global energy industry. Headquartered in Dallas, TX, the company focuses specifically on the petroleum refining, petrochemical, pipeline, terminal and power generation markets.
Regarding the Comparative Performance Analysis (CPA) product referenced in this article, while Solomon maintains this intellectual property and uniquely applies it in energy-related industries, the method is available for license to companies in industries or work areas that the company does not service.