Improving Machinery Reliability 3 Episode 7 pdf

244 Improving Machinery Reliability RELATIONSHIP MATRICES Goals and metrics vs. critical business issues I ', CBI Relationships mu Direct Indirect None Critical business issues vs. obstacles Figure 4-1. Relationship matrices. tactical level. Tactical level goals and metrics also have associated CBIs. These CBIs become the operational level goals with their associated metrics. This top-down approach ensures that operational level goals and metrics relate to tactical and strate- gic goals and metrics, Figure 4-2. This goal focus sets enterprise direction. Metric data are used to determine the impact of change on an operation. These data are a valuable enterprise asset that must be maximized. Metrics have three important aspects: Determining what metric data to gather and use Gathering the metric data Using the metric data. To be effective, metrics must be carefully selected to measure what is important. What is measured will improve. Selecting what is most important to the operation, and measuring it, will focus improvement efforts. After what is to be measured is determined, appropriate data must be collected, stored, and made accessible. Metrics are like other tools: If they are not used, they have little value. The value of metric data lies in the skills of those using the tools or metrics. To provide value, the metric asset must be used within the operation. For the value of the metric asset to be maximized, the data must be shared with other organizations. If common metrics are established for each enterprise operation, the data can be shared among many operations and results can be compared. Comparison of metric data benefits all organizations sharing the information. Value is realized by adapting (not adopting) what is learned from the information. The competitive advantage is not in the metric data, but in how the information is used, just as it is with any other tool. Maintenance and Benchmarking Reliability 245 TOP-DOWN GOALS DEVELOPMENT ENTERPRISE Critical business issues MAINTENANCE OPERATIONS Operational level Goals/metric Critical business issues Figure 4-2. A top-down approach ensures that operational level goals and metrics relate to tactical and strategic goals and metrics. Measurement Attributes To be efficient and effective, metrics must have the following attributes. Measure- ments must be: Appropriate. They must accurately measure the specific operational aspect that requires measurement. 0 Acceptable. They must be considered, by all concerned, to measure the desired operational aspect. Simple. They must be easy to understand, easy to gather, and easy to interpret. Unambiguous. Metric results must communicate a clear message relating to the 0 Comparable. Metric data must be capable of being analyzed in relationship to pre- operation being measured. viously gathered internal and external data. Strategic Level Maintenance Measures The following strategic level metrics are suggested as a common set of mainte- nance measurements: Maintenance costs as a percent of cost of goods sold (or total production costs) * Maintenance costs as a percent of machinery and equipment replacement value Number of service maintenance employees as a percent of direct labor employees Spare parts inventory as a percent of machinery and equipment replacement value a Spare parts inventory turns a Percent of routine scheduled maintenance hours 246 Improving Machinery Reliability Certifiable training costs per employee Maintenance-related downtime To assure accurate comparisons, it is extremely important that measurements be calculated in the same manner from consistent data. Calculations also must be clear- ly understood for meaningful comparisons. For example, the exact elements that generate “cost of goods sold” or “total production cost” vary from business to busi- ness and must be calculated consistently for an accurate comparison. Another example is in the use of the number of service maintenance employees. The appropriateness of this measure depends on the level of facility automation. More maintenance workers may be required to support the automated operation, although fewer direct labor employees may be needed. The formulas for calculating the suggested maintenance measurements are shown in Table 4-1. No one measurement can paint the entire picture of an operation. These suggested strategic level maintenance measures attempt to cover all the major maintenance ele- ments: human resources; materials; machinery, facilities, and production equipment; and the maintenance processes. Table 4-1 Formulas for Calculating Strategic Level Measurements Measurement Formula Maintenance costs as percent of cost of Maintenance costs Cost of goods sold x 100 goods sold x 100 Maintenance costs as a percent of machinery Service maintenance costs Machinery and equipment replacement value and equipment replacement value X 100 Number of service maintenance employees as Number of service maintenance employees Number of direct labor employees a percent of direct labor employees X 100 Spare parts inventory as a percent of machin- Spare parts inventory value Machinery and equipment replacement value ery and equipment replacement value Spare parts inventory turns Value of spare parts issued in past 12 months Average spare parts inventory value for past 12 months x 100 Routine scheduled maintenance hours as a Routine scheduled maintenance hours Total maintenance hours percent of total maintenance hours Certifiable training costs per employee Certified training costs Total number of service maintenance employees x 100 Maintenance-related downtime as a percent Maintenance - related downtime of total downtime Total downtime Maintenance and Benchmarking Reliability 241 Tactical Level Maintenance Measures Succeeding with strategic level maintenance measurement comparison will gener- ate significant improvements. Tactical level maintenance measures should be devel- oped to support the enterprise’s associated tactical goals. Although these measure- ments are important, no specific measurements are suggested here. These measurements can be developed in relationship to specific enterprise and mainte- nance operation situations. Operational Level Maintenance Measures The operational level is the lowest level of maintenance measurement. Compar- isons at this level are not as important as those at the strategic level. Operational level comparisons are most beneficial in the investigation of specific improvement situations. Operational level metrics are used primarily by a maintenance department to better manage its operation. Some suggested operational level metrics folldw: Total minimum maintenance cost Life-cycle cost of asset ownership Mean time between failures Mean time to restore Overall equipment effectiveness (a total productive maintenance, or TPM mea- surement) Average response time to unscheduled machine failure Percent of time machine is available to run versus scheduled run time Periodic customer satisfaction surveys @ Periodic skilled trades work constraint analysis These measurements, like the strategic level measurements, must be checked consis- tently and used to be effective. It is recommended that a maintenance engineering func- tion be established to manage and use maintenance data, among other responsibilities. Benchmarking A benchmarking industry has emerged in recent years, along with a broad range of measurements (including maintenance measurements) that theoretically define “world class.” For the most part, these are bottom-line-type measurements that have value when internal performance is compared. Table 4-2 lists some world-class value examples for strategic level measurements.* Various sources contacted for these values indicated that they would not validate the numbers. There are too many associated variables and there was generally too much risk associated with publishing the values. Commercial sources felt that their values were proprietary and that publishing the values divulges information from which they derive revenue. These are just some of the problems with benchmarking. - *Numerical values have been adjusted to reflect H. P. BlochlF. K. Geitner’s experience. 248 Improving Machinery Reliability Table 4-2 Sample Values for Strategic Level Measurements Measurement Value Best Average Maintenance costs as a percent of cost of goods sold 1.9 Maintenance costs as a percent of machinery and equipment replacement value 2.0 Number of service maintenance employees as a percent of direct labor employees 3.0 Spare parts inventory as a percent of machinery and equipment replacement value 1.0 Spare parts inventory turns 1.4 Routine scheduled maintenance hours as percent of total maintenance hours 80.0 $1,200 Maintenance-related downtime as a percent of total downtime 4.0 Certified training costs per employee 4.0 5.0 7.0 3.0 0.5 35.0 $400 10.0 As a result, some words of caution are necessary regarding benchmarking and the information on the world-class characteristics just presented. The information is pro- vided because everyone seems interested. The use of this information is, however, somewhat questionable. Questions such as the following arise: What is the world-class environment (for example, industry and manufacturing type)? Where does the measured operation fit? If, in a specific area, an operation’s measurement exceeds world class, is there no How are the measurements calculated? Is our operation calculating the compo- Is world class too expensive for our operation? need for improvement? nents consistently? These questions reflect just some of the issues that render world-class information questionable. Benchmarking and data comparison are good tools if understood and used correct- ly. A lot of good can come from comparing an operation with others that are similar and dissimilar. New ideas and concepts from other operations can be identified and adapted. The most productive use of benchmarking and value comparison is to iden- tify where improvements are needed and determine the impact of change. The productive way to use external benchmarking is to determine what is internal- ly important (goals and CBIs) and use benchmarking to measure those operational aspects. In the true benchmarking concept, the internal information can then be com- pared with information from other organizations. The comparison must be controlled and focused on the operational aspects being studied. This approach ensures that apples are compared with apples. Sharing internal benchmark information serves only to optimize the asset. Again, the benchmark data are the tool, and how they are used is what makes the data productive. The biggest problem with measurements and benchmarks is high development and utilization costs. Everyone in the organization must be educated concerning the costs of measurement collection and utilization and their associated values. Value is not received unless data are collected and used. The tools and time used to input, store, Maintenance and Benchmarking Reliability 249 and retrieve the measurement data are expensive. Everyone involved must be moti- vated to accurately and consistently collect data, and then to use the information. Using the information can involve the cost of establishing a data utilization func- tion with dedicated people. Measurement values and costs are high, and management must accept the associated values and costs. Everyone is looking for the quick and inexpensive answer. Therefore world-class measurements are attractive. The process is simple and inexpensive, but the results are questionable. True benchmarking is even more involved and costly. Once an operation knows what it wants to measure, it needs to collect internal data and compare like data with its benchmarking partners. There are collection costs and costs to locate benchmark partners, as well as costs of travel to the benchmark partners’ operations to collect the required data. Data must then be analyzed and new approaches determined. All these procedures take time and money-more than management usually understands and wants to expend. In spite of the perceived high measurement costs, the resulting benefits are high and the payback period is short. Wise enterprises seek out benchmarking partners and share data. Without the measurement road map, the business direction is not known until it is too late to adjust. Organize to Manage Reliability* An analysis of maintenance costs in hydrocarbon processing industry (HPI) plants has revealed that attitudes and practices of personnel are the major single “bottom line” factor. In reaching this conclusion a world-renowned benchmarking organization examined comparative analyses of plant records over the past decade. They learned that there was a wide range of performance independent of refinery age, capacity, pro- cessing complexity, and location. Facilities of all extremes in these attributes are included in both high-cost and low-cost categories. Those in the lowest quartile of per- formance posted twice the resource consumption as the best quartile. Furthermore, there was almost no similarity between refineries within a single company. The Record. Figure 4-3 illustrates rising profitability during 1986-1988 because of market conditions. As the militaries mobilized into the Arabian Gulf in 1990, profits climbed sharply. By 1992, margins relaxed somewhat to pre-war levels. But if we look at trend data of a constant trend group of 68 refineries, the picture is different. From the top curve in Figure 4-3, we note that the difference in profitability between the highest and lowest quartiles was about 5% in 1986. But by 1992, the gap had widened to 12%. This divergence is not merely an industry average phenomenon. It is clearly a difference in performance of two groups, which is not related to the industry average performance. Similar performance differences can be found in industry maintenance data. Fig- ure 4-4 portrays a six-year trend of maintenance cost for the same trend group of refineries. The data represent total annual refinery maintenance costs per unit of *Adapted by permission of R. Ricketts, Solomon Associates, Inc. Dallas, Texas. From a paper pre- sented at the 1994 NPRA Maintenance Conference, May 24-27, 1994, New Orleans, LA. 250 Improving Machinery Reliability 40 35 30 25 20 15 10 Figure 4-3. Cash basis, ROI, in percent. 1986 1988 1990 1992 Figure 4-4. Maintenance index, U.S. $/EDC. refinery capacity and complexity (“EDC,” or equivalent daily capacity). The mid- range curve is the industry average, revealing an increasing expenditure of about 6% per year over the six-year period. This increase will surprise no one, It is characteris- tic of inflationary pressures and increasing emphasis on control of refinery emis- sions. But when the data are viewed in terms of the performance spectrum, a very different relationship unfolds. Those refineries represented by the lower curve are the lowest-cost quartile. They posted increases of less than 1% annually. On the other hand, the highest-cost quartile’s spending (upper curve) doubled during the same period. Maintenance and Benchmarking Reliability 251 The message is that management practices are widely divergent. They represent some very different approaches to managing resource consumption in a competitive environment. Those in the high-cost group may find it hard to remain viable. Cost vs. Availability. A concern closely related to maintenance cost is the impact of maintenance activities on availability of processing facilities. Figure 4-5 charts the performance of eleven U.S. refineries that recorded the greatest increase in costs dur- ing 1986-1992. Note that accompanying the rise in costs was a major decline in mechanical reliability. Nine refineries with the greatest improvement in mechanical reliability are represented in Figure 4-6. They climbed dramatically from average into the best quartile of mechanical reliability. So it seems that, at least in refining, improved mechanical reliability isn’t related to the amount of maintenance effort. Maintenance Spending. About 35% of the average U.S. refinery maintenance budget is for higher-profit processes as shown in Figure 4-7. All other types of processes combined account for 35%. Utilities, marine, and offsites consume the remainder of the budget at 30%. These statistics raise a question of the primary focal point of the maintenance budget. Figure 4-8 provides the answer. Fifty-four percent off the average refiner’s budget is earmarked for equipment repair and programmed replacement, while energy conservation, environment and safety, and other require- ments consume 38% of the budget. Reliability improvement programs account for a mere 8% of the expenditure. Analysis of high-cost and low-cost segments of the industry reveals a wide variation in performance and trend data. The performance gap is getting wider. These differ- ences are not related to age, size, or location as it would be tempting to believe. Fur- thermore, our benchmarking organization, Solomon Associates, has not observed dif- ferences in craft competence between the high- and low-cost performers. The reason 98 97 96 95 94 93 92 1986 1988 1990 1992 Figure 4-5. US. maintenance results, eleven refineries that had largest cost increases, 1986-1 992. 252 Improving Machinery Reliability 35 98 25 E i2 0 3 95 p 96 - 10 1986 1988 1990 1992 Figure 4-6. U.S. maintenance results, encompassing nine refineries that had the great- est improvement in mechanical reliability, 1986-1 992. 20% offsites and marines Figure 4-7. Refinery maintenance expense. for the differences is simply that the lowest-quartile cost group has less demand for repair maintenance and thus does less work in this area. Table 4-3 is taken from a recent worldwide maintenance management study by the same analysts. The lowest quartile’s craftsmen have four times more pieces of rotat- ing equipment per person to maintain than the highest-cost quartile. Those in the highest-cost quartile are kept busy repairing failures. They have no opportunity to examine the causes of these failures. They thus can’t formulate actions to make per- manent repairs or to devise preventive and predictive remedies. Organization There are two types of organizational approaches: repair focused and reliability focused. Maintenance and Benchmarking Reliability 253 15% safetv. 34% environment, equipment all others % replacement programs improvement Figure 4-8. Refinery maintenance benefits. Table 4-3 Equipment density (Per million units of capacity and complexity) Lowest cost per Highest cost Per Items Items quartile craftsmen quartile craftsmen Punips 630 4.8 700 1.1 Compressors 35 0.3 40 0.1 Pressure vessels 490 3.1 500 0.8 Heat exchangers 390 2.9 525 0.8 Safety valves 840 6.3 1,150 I .8 Repair-Focused Organization. This organization style embraces the philosophy that equipment will fail and that the mission of the maintenance force is to respond quickly to equipment in distress. Failures are expected because they are the norm. Management and craftsmen stay occupied in repair activity and have no opportunity to examine failure cause. Staffing is designed to accommodate rapid repair, often including sizeable maintenance crews on non-day shifts. When failures do not fully occupy the workforce, the organization focuses on lower priority (frequently unnec- essary) minor projects to “stay busy.” Reliability-Focused Organization. Maintenance repairs in this style are viewed differently. They are not expected to happen. They are viewed as an exceptional event and a result of a flawed aspect of maintenance policy and management focus. The specter of a recurring failure and its incumbent cost is unacceptable. The organi- zation is sized to manage a condition-based monitoring system and assigns high pri- ority to the elimination of failure. Unnecessary work is not performed regardless of the current work load. [...]... Governors 5.94 18 12 , 37 0 102,900 0 2,082 17 ,31 7 19 ,39 9 2 67, 024 2 67. 024 9.16 4 3, 70 0 102,900 49 ,70 0 404 11, 233 11, 6 37 160,181 209.881 13. 74 4 3, 70 0 102,900 7 1, 870 269 7, 489 7, 758 106 ,78 9 178 ,659 dig in his own files for technical papers and magazine articles that could shed light on the matter Or, the reader could simply review Appendix B of this text, which deals with common-sense reliability models... REPLACEMENT VALUE + + 3+ 89 + 3. 88 + 3. 66 + + 3* 14+ 3. 08 28 6 26 4 + + 39 3 3.60 + + 3- 00 2 9 4 + 2. 07 + 24 8 + 1 51 + + + 2.56 21 4 1. 73 0 ! I + 4. 97 !/ b i 0 l: 1: 11 2 I 13 4 : I I I I I I I I 15 16 17 18 19 20 2 22 : 1 3 Figure 4-9 Annual maintenance costs as a percentage of plant replacement value (Courtesy of HSB Reliability Technologies, as published in Maintenance Technology, July-August 79 96.) *Raymond... been reduced to 6 .7 Option 1, conversionlupgrading of 232 pumps during the next shutdown would cost $( 232 ) (34 70 ) = $805,040 This one-time expenditure would likely result in yearly savings of ( 17. 7-6 .7) (0. 232 ) (12 months) ($6 ,70 0) = $205,181 The resulting payback period would be 805,040/205,181 = 3. 9 years and savings over a 4%-year period would amount to ($205,181) (4.5) = $9 23, 3 15 Next, we’ll examine... shaded area of the diagram: [( 17. 7 - 6 .7) (54) ( 232 )/(2) (1,000)]+ L(6 .7) (54) ( 232 )/1,0001= 1 53 pumps Since all of of these repairs would incorporate upgrade components at $500 per pump, our repair expenditures in a 4%-year period would total (1 53) (7, 200) = $1,101,600 Calculating the cost of option 3, “business as usual,” is easiest Failures would continue at a rate of 17. 7 per 1,000 machines per month... anticipated pump failure rate reduction due to upgrade efforts  270 Improving Machinery Reliability 80 70 u: m 40 E 4 30 20 10 Figure 5-4 Improved mean-time-between-failure(MTBF) at three British oil refineries were attributed to seal upgrade and selection ~trategy.~ MTBF increases of (80- 57) / 57 = 40% in 4 years (refinery B) and (50 -33 ) 33 = 51% in 2 years (refinery C) It makes good sense to see more... one or more years Life Cycle Cost Studies 277 Table 5-8 Present Value and Future Value =w - - "" ".",.- ni*m,,n, D.1 ~ I Vearshencel 01 11 21 31 41 51 61 71 81 91 101 111 121 131 141 151 161 171 181 191 20 Presenlvalve olUSSl.OOI 1.0010.89( 0601 0 7 1 I 0 B l ~ 0 5 7 l 0 5 1 ~ 0 4 5 ~ 0 4 0 ~ 0 3 6 ~ 0 3 2 ~ 0 2 9 ~ 0 2 6 ~ 0 2 3 ~ 0 2 0 ~ 0 0.13I 0 1 6 ~ 0 10 0.151 1 ~ ~ 0.121 F ~ l ~ r... ~ 0 4 5 ~ 0 4 0 ~ 0 3 6 ~ 0 3 2 ~ 0 2 9 ~ 0 2 6 ~ 0 2 3 ~ 0 2 0 ~ 0 0.13I 0 1 6 ~ 0 10 0.151 1 ~ ~ 0.121 F ~ l ~ r e v a l u eUS11.0011 001 1.121 1251 1.401 1. 571 1 .76 11.9112 211 2.481 2 7? 13 11 13. 46 139 014 .36 14 88l547l6.13l 6 871 7. 691 8 811 9 6 5 ol and cash begins to flow in for a series of many years The amount of cash "thrown off" by a project is an important consideration and a helpful criterion... C started with 33 months seal MFBF and that our refinery is presently at 28 months MFBF Returning to refinery A and their overall pump MTBF, which had increased from 30 months at the end of year 2 to about 71 months at the end of year 7, we calculate an MTBF increase of (71 -30 ) /30 = 36 % in 5 years If we take into account the observation that refineries starting with MTBF figures of 30 months have experienced... ANSI-PLUS retrofit instead of 51120 7 Additional maintenancecost outlay for installing ANSI-PLUS retrofit parts insteadof 534 4,960 116 (Item 3 x Item 8) I 536 2,152 271 272 Improving Machinety Reliability out” pumps were originally furnished with conventional stuffing boxes that could accommodate only relatively small diameter mechanical seals Whenever one of the plant’s 4 17 pumps undergoes a shop repair,... MTTR is 18 hours, and MTBF = (28) (7yrs) (8 ,76 0 hrs / yr) - (33 failures) (18 hrs) 33 failures = 52,011 hours or 5.94 years In this example, the cost to repair or replace a hydromechanical governor is $12 , 37 0 Production losses are primarily influenced by the need to flare huge amounts of hydrocarbon feed for approximately 1%hours per outage event This costs the plant $72 ,820, plus $5,120 in lost profits . 12 , 37 0 102,900 0 2,082 17 ,31 7 19 ,39 9 2 67, 024 2 67. 024 9.16 4 3, 70 0 102,900 49 ,70 0 404 11, 233 11, 6 37 160,18 1 209.881 13. 74 4 3, 70 0 102,900 7 1, 870 269 7, 489 7, 758 106 ,78 9. PLANT REPLACEMENT VALUE ++ 3. 88 3. 89 ++ + 3* 14+ 3. 08 2.86 + 2.64 2.1 4 + ++ 4. 97 + + + 3. 93 3. 66 3. 60 ++ + + 3- 00 2.94 2.56 2.48 + 2. 07 1. 73 + 1.12 1 I IIIIIIII. 1994 NPRA Maintenance Conference, May 24- 27, 1994, New Orleans, LA. 250 Improving Machinery Reliability 40 35 30 25 20 15 10 Figure 4 -3. Cash basis, ROI, in percent. 1986 1988