Mark your Calendar!
MARCON-2007 May-8-11, 2007
Knoxville TN

MARCON-2007

Schedule

Registration Form (PDF)

Expo

Hotel

Workshops

Abstracts

Contacts

Reliability Engineering Principles (REP) is a basic training course for engineers. Reliability Engineering Principles covers the fundamentals of reliability. The course is long on practical problem solving and short on statistical calculations (we use software for the statistics). The course shows how improving reliability boosts business performance by reducing the high cost of unreliability. The reliability tools are used for finding cost issues rather than simply functioning as technical marvels.

Subjects discussed in the course are generally connected to money issues for finding affordable levels of reliability by use of reliability tools.

Manufactures that incorporate root cause analysis into their reliability efforts have found most root cause analysis results uncover human error as a contributor to the event being analyzed. Equipment start–up is always a concern because maintenance and operations spend log hours working to restore equipment availability as soon as possible. Many times during or shortly after start-up equipment fails. When investigations are completed a bearing was installed incorrectly, an impeller was installed backwards, the wrong lapping compound was used, and the list goes on.

You can manage human errors like these much like you manage the safety of your employees. In fact, proper understanding of human error traps can keep employees safer. When employees are aware of potential human error traps they can take actions to avoid results that may injure the employee and or damage company assets. This process can best be introduced using first line supervision as the driver to uncover and manage potential human error traps. When the concepts of this class are fully utilized safety incidents are decreased and reliability is improved through longer runs without incident. More than 7000 supervisors have benefited from taking this course.

PART 1:  The Application of Six Sigma to Maintenance and Reliability - led by Paul Casto, Eastman Chemical

Value creation is the language of management and Six Sigma is a tool used to create value which is understood by company management. Six Sigma is traditionally used to

improve the manufacturing process and is not often applied to Maintenance and Reliability (M&R).  Integrating Six Sigma into the M&R process will link company management to M&R, as well as, produce significant bottom-line results. 

This workshop will begin with a discussion on building the business case for maintenance and reliability.  This will include an introduction to the fundamentals of finance applied to maintenance.  The workshop will discuss how the Maintenance and Reliability function can create value that shows up on the company income statement, where to find value, how to measure value, the magnitude of the potential value and how to capture this value.  The lecture material will encompass proven approaches and techniques to properly apply both Six Sigma and reliability tools to maintenance and reliability.  The use of these tools will be illustrated through case studies and a review of actual project results. 

PART 2:  Value Stream Mapping for Maintenance - led by Dr. Rupy Sawhney, University of Tennessee College of Engineering

This workshop presents a Maintenance Value Stream Map (VSM) to the participants. This unique VSM has been developed for maintenance organizations and illustrates improvement opportunities within maintenance organizations. The VSM specifically allows one to identify non-value added activities as well as problems with coordination and scheduling activities. A Lego exercise is used to illustrate the concepts and attendees will participate in the Lego exercise and design an optimal maintenance operation. The  participants will then be guided in applying the learned concepts to their operations. 

1.    The participants will have the following learning experience:

       A.  Maintenance based Value Stream Map technique

       B.  How to develop the VSM

       C.  Simulating the VSM with Arena software 

2.    Analyzing the relevant metrics from the Simulation Model

       A.  CPM/ PERT evaluation of processes

       B.  Developing a PERT network

       C.  Analyzing maintenance operations through PERT networks 

3. Application of Lean and Six sigma concepts to improve processes

      A.  Identifying concepts of lean to stream line maintenance operations

      B.  Implementing six sigma tools (statistics) to reduce variation in maintenance

Eastman Chemical is a producer of intermediate chemical products and has been a leader in the application of reliability based technologies and practices for many years. The focus on reliability at Eastman is a key element of the company's overall strategy to be a dependable supplier to customers, as well as, a method to improve availability and increase production in constrained product streams. Inherent availability is a function of reliability and maintainability, which says that availability is increased by improving reliability and maintainability. Reliability is measured based on failures and maintainability is measured based on repair times. By definition, to improve availability, failures must be reduced and repair times shortened. This is done by focusing on the maintenance function. Maintenance can be defined as “activities aimed at retaining an item in or returning it to an acceptable condition”. This definition encompasses both reliability (retaining an item) and maintainability (returning it to an acceptable condition). As a result, a maintenance program encompassing reliability, standardized work and maintainability is a critical tool in improving availability.

Eastman's reliability heritage and the company's culture of continuous improvement have resulted in the developed of a business driven Reliability Based Operations and Maintenance (RBOM) strategy. This strategy has been internally developed over the past two years and is focused on increasing reliability, improving maintenance work processes and capturing the knowledge and expertise of an aging workforce. This strategy goes beyond the technical implementation of failure based maintenance and involves all members of operations and maintenance in the prevention, control and/or mitigation of equipment and process failures. This paper will discuss the genesis of the RBOM strategy, the development of implementation tools and work processes, and the information management infrastructure required to support the program.

The implementation of the RBOM program and the results of the program will be reviewed in a second paper on this subject.

A recently study for the National Nuclear Security Administration found that service lives of mechanical equipment can be substantially longer—twice as long, for some equipment—than suggested in the technical literature. Replacing major mechanical equipment based on the longer life estimates could reduce overall facility M&R costs by 5 percent or more. However, additional study is needed before these findings can be generalized to other facilities.

Done in partnership with Lawrence Livermore National Laboratory (LLNL), this study derived survivor curves for selected equipment from an unusually rich (24,000 records) historical database. Median age at retirement—a typical service life measure—is reported for six equipment types in the following table. As shown, the median values derived from the LLNL data were substantially longer than those published in a popular reference.

The implications of the findings are intriguing. Service life is a key assumption in facility life cycle planning and long term budgeting. Simple experiments using the MARS Facility Cost Forecast System demonstrated that doubling the assumed life of major mechanical equipment could reduce total M&R costs by 5 percent or more for some facilities. However, a number of qualifications must be recognized.

Lets take a look at a “day in the life” of a maintenance craftsman… We’ll call him Jim. The day shift is just arriving, the coffeepots are brewing, computers are booting up and the night shift is turning over the plant. After a little chitchat with his co-workers, Jim picks up the day’s assignments from his supervisor. This morning, he has to go to the job site to checkout the job before making a trip to the storeroom. When he gets to the job, he finds it’s a pretty straightforward job to replace a worn seal in a pump. He needs to find out the part number of the seal, so he locks out the pump and proceeds to take it apart to get the seal number. It’s not really a problem to have the pump down for a couple of hours. Jim would have gone to the manual but the last time he looked for the manual, it was nowhere to be found. He tries searching the CMMS for the right part with no luck. Jim goes through some old notes from the last time he worked on the pump and finds the seal number. A quick trip to the storeroom to get the seal and he’ll be good to go. Bad news, the seal isn’t in stock…well here we go again, thinks Jim, same old story.

In many maintenance organizations, 50% of the annual maintenance budget is spent on material and spare parts … not to mention that a significant amount of time is “spent” acquiring these spare parts. This expenditure is often a significant percentage of a plant’s total operating costs. Strangely enough, many of these organizations have almost no control of their materials. One part of the organization feels there is too much inventory and the other part thinks there isn’t enough. Why do these problems exist? What can be done about them? How do you match the demand for parts with the supply?

This presentation will take a humorous but serious look at the issues (good and bad) associated with material management for a maintenance organization. We will uncover the hidden inefficiencies of poor material management that occur thousands of time a day and their costs. These inefficiencies cumulatively add up to incredible losses to our businesses. We will also present ideas and solutions to improve these poor practices. We will make the business case for implementing enhanced materials management processes, practices, tools and techniques that are fully integrated into the work management processes. The presentation and paper will be peppered with real life first hand examples and situations that our clients and we have experienced.

Some of the topics discussed will be Proactive Maintenance, Storeroom Benchmarks, Conflicting Objectives, Proper Inventory Reduction, Documentation Management Systems, Imaging Systems, Bills of Material, Stock Analysis, Cost-Risk Analysis and much more.

This presentation will be valuable to a wide range of maintenance and reliability personnel such as, Maintenance Managers and Supervisors, Purchasing Agents, Storeroom Managers, Maintenance Planners, Corporate Managers and Plant Managers.

Eastman's Kingsport-TN site is implementing an innovative process to improve reliability in the plant.  Eastman Chemical Company has been known for years as a company that is effective in the application of Predictive Technologies.  However, study has shown that these technologies, overall, are just a small part of having an effective Reliability Program. 

As a result, Eastman has implemented a new Reliability Based Operations and Maintenance (RBOM) strategy in the Ketones Production process at the Kingsport site.  This process involves a dedicated team to implement Reliability Strategies that are based on Failure Modes of the Process Equipment.  The team is made up of four full-time employees:  Reliability Engineer, Reliability Technologist, an Operator and a Mechanic from Ketones. 

Failure Modes and Effects Analysis (FMEA) is the main tool used by the RBOM Team.  Our goal in this effort is to implement Proactive strategies that are based on Failure Modes of the equipment so that 80% of the work done is Proactive.  While this process is time consuming, significant results have been gained.

Another presentation, given by Paul Casto, will present the need, theory and strategy of the RBOM process.  This presentation will detail the Implementation of the project in Ketones and the results so far.  The following topics will be covered:

·         Setting the Stage for Change

·         Change Management

·         Training Requirements

·         Implementation of the RBOM Process

o       Process Mapping

o       FMEA's

o       Creating Tasks from the FMEA

·         Implementation of Handheld Technologies

·         System's Requirements/Issues

·         The need for Strong Leadership

·         Results of Project

Degradation data analysis has recently been used to assess the reliability of highly reliable and complex engineering systems. In recently published literature it has been declared that the traditional lifetime data analysis is no longer acceptable in condition-based maintenance (CBM) offerings. Many authors believe that monitoring the item's actual performance degradation facilitates estimation of the item's remaining useful life (RUL). Performing the health condition monitoring procedures would be more informative than evaluating average reliability parameters associated with the entire population of similar items.

This paper presents a linear growth model-based methodology for estimation remaining useful life of components. Monotonic degradation taking place in aging components and systems is assumed to be a linearly growing process. The degradation process is modeled as a gradual damage accumulation process having a linear trend. Observing health condition measurements at the component of interest one is able to estimate parameters of the particular linear growth model and make a prediction regarding the degradation values in future.

The estimated parameters are subject to uncertainty associated with random deviations caused by a process noise and measurement errors. Apparently the uncertainty tends to propagate into the final RUL prediction, deteriorating the prognosis in terms of certainty. This paper presents a statistical analysis regarding the usefulness of a RUL prediction, which takes into account individual health condition data rather than average reliability characteristics of the entire population of similar objects. A criterion of usefulness, upon which the practitioner comes to a decision on which approach is going to be more beneficial, is defined to be the uncertainty associated with the RUL prediction since the ultimate goal is to make the prediction as certain as possible.

A great deal of uncertainty in health condition measurements, upon which the individual RUL prognosis will be made, can be an insurmountable obstacle in obtaining an accurate time-to-failure prediction for an individual item. On the other hand, large variability in lifetime data observed over a sample of similar items impairs the RUL prediction based on the empirical time-to-failure distribution. A linear growth model of performance degradation is considered. Bayesian techniques are applied to evaluate parameters of the model. Uncertainty analysis is performed to estimate remaining useful life of the item and uncertainty associated. Simulated data examples are given to illustrate the proposed criterion.

This paper is a follow-on to the 2005 MARCON presentation, “The Basics – Don’t Put the Cart Before the Horse.” This paper was also published by Reliability Magazine, Volume II Issue 4,
http://www.reliability-magazine.com/art06/the_basics.htm

This paper defines basic conditions in the context of the Japanese Institute of Plant Maintenance (JIPM) Total Productive Maintenance (TPM) approach for ‘prevention of deterioration.’ It establishes why basic conditions are a requirement for achieving zero failures and zero defects. Next this paper focuses on practical examples of what basic conditions are; lubrication, water chemistry, alignment, oil/water/compressed air leaks, contamination control, correct maintenance practices, correct machine operations and adjustments.

This paper references recognized JIPM publications on TPM implementation to show foundational principals of maintenance practices establishing basic conditions. Several case studies will be presented to emphasize practically how basic conditions can be established in an industrial manufacturing environment. These basic conditions practical examples will be communicated through numerous visuals such as digital pictures, graphs, etc.
The conclusion of this presentation focuses on the important cultural changes required to sustain basic condition improvements on manufacturing machinery.

We’re New, We Didn’t Know Any Better!

Sandia National Laboratories recently made several strategic decisions to improve the maintenance and reliability processes at their New Mexico site. These decisions included establishing a Maintenance Engineering Department and the development of in-house Reliability Centered Maintenance capabilities. Sandia’s approach to development and implementation of an RCM program provides a model for others interested in pursuing a more formal approach to Reliability Centered Maintenance. The presentation includes discussions on key management decisions: strategic sourcing of RCM Capabilities, Project Charters, Strategic Partners, Training, Buy-in by senior management and others, Resource Estimation, joining the Maintenance and Reliability Center, Benchmarking, Project Management, Project Scoping, Implementation of recommendations and results. In addition the presentation addresses the results of a unique application of RCM II concepts to human performance and configuration control applied to a complex building alarm system.

Sandia operates in a research, development, and testing environment and financial benefits and operational efficiency are difficult to quantify. In this environment the decision to implement an RCM II program was difficult to justify; however, the results of the first two events proved to be extremely valuable and confirmed the right decision was made. The first event, a traditional application of RCM II, identified several “hidden failures” where a failure could have significant impact to research processes; though hard to quantify, the results of the RCM were deemed a success. Perhaps because we are new to RCM and have yet to develop a bounded culture of “appropriate” RCM applications, we were willing to take on a non-traditional project. Our second event was focused on processes, human performance, and configuration control pertaining to a complex fire-alarm system. The reliability of the system operation (alarming correctly when needed, but also not having False alarms) is critical to the integrity of the system and those who the system protects. The results clearly demonstrate the RCM II Methodology can be applied to processes of this nature in addition to the more traditional equipment-based analyses. In this presentation, we will share our success in both initiating a program and applying the RCM concepts in a unique way.

Rotating equipment in production facilities offer optimal performance and reliability when properly installed, maintained, and operated. Condition monitoring devices and systems are utilized to keep tabs on the operational performance of key production equipment. All types of tools and technologies exist today in the predictive maintenance field to allow monitoring and assessment of such equipment. In most cases, a portable, walk-around program will suffice. In other cases, where equipment is hard to reach, inaccessible, or located in a dangerous or hazardous area, a permanent installation of sensors and surveillance hardware is necessary. Key objectives for such a maintenance program are to minimize failures, reduce downtime, and to cut costs.

Condition monitoring tools can improve production uptime, efficiency, and profitability. Candidates for monitoring include motors, pumps, compressors, fans, gearboxes, bearings, and other critical machine elements. As components become worn, dirty, contaminated, loose, misaligned, unbalanced, and improperly lubricated, machines may experience increased vibration levels and higher temperatures, thereby leading to failures and production outages.

The tools for predictive maintenance and condition monitoring must be chosen with consideration given to planned return on investment. Production personnel need to classify machines as “critical”, “essential”, or “balance of plant”, and then decide the right mix in terms of required maintenance expertise and tools.

Careful review of needs and expectations can lead a potential user for these tools down several paths. Popular choices in this area of technology include route-based portable instruments, online continuous monitoring hardware, and predictive analysis software systems. Online systems include options for wireless or hard-wired connectivity. It is important to clearly understand the pros and cons of each offering.

Recent advances in electronics technology have allowed online predictive maintenance systems to be more affordable, reliable, flexible, and modular. These systems have proven to be quite effective in providing early warnings and pinpointing root causes for machinery faults and failures. Online systems can measure and record many process parameters, allowing the user to trend and trigger on alarms so that machine performance and health can be monitored along with vibration-related faults. There has been a natural tendency to integrate some of these data with traditional process instrumentation and control systems.


This paper will focus on the ideal application of continuous surveillance systems, the economies of scale, and the distinct advantages relative to implementing a traditional portable, walk-around program. While previous experiences for such systems in this industry have brought out some areas for concern, these will be addressed as well. Specific case histories and success stories will be cited to show the advantages of increased awareness with online surveillance.

The six sigma work process is used extensively in almost all areas in today's business and manufacturing environment. Six sigma is a performance breakthrough methodology that can be used to improve virtually any process and fundamentally change the firm's cost structure. It is a disciplined approach that uses a defined work process, cross-functional teams and statistical data analysis to identify opportunities, create solutions, capture value and implement long term control. The answers too many reliability problems are concealed in the failure data and statistical methods are often needed to analyze this failure data and develop solutions. These improvement methods are data driven and require a disciplined approach. Because of the high potential value of reliability projects and the similarities between six sigma and reliability data analysis methods, one would expect six sigma to be used extensively to solve reliability problems. However, in a recent survey of academic and practitioner literature very little work on this subject could be found. This paper will review the six sigma methodology and compare this with basic reliability methodology. Similarities between six sigma and reliability methods will be discussed, specific reliability tools, not normally included in six sigma training programs, will be identified, and a process to effectively execute reliability projects using six sigma methods will be discussed.

In predictive maintenance (PdM), deterioration of equipment is detected through the monitoring of operating parameters, such as vibrations and temperatures, against some predetermined acceptable operating limits. Traditionally, these limits have been set on advice of equipment manufacturers, or the experience of the maintainers.

In this research project, the opportunity to combine statistical process control (SPC) with predictive maintenance was investigated and verified to have potential for setting limits on PdM parameters, provided a preliminary period of normal operations can be sampled and appropriate SPC methods are applicable to calculate the limits. Most maintenance practitioners are unaware of this application of SPC, and need guidance on how to compute SPC-based control limits appropriately in the context of equipment predictive maintenance monitoring and trending. After a review of the limited literature on applications of SPC in PdM, we provide an original flowchart depicting the steps necessary to analyze and prepare PdM data for use of SPC as a method to estimate “normal operating limits” for such monitoring and trending.

A test of the proposed SPC method was completed on data sets from ten vibration indicators, provided to us by a contract maintenance organization. These data sets proved difficult to analyze—at least using standard SPC techniques. However, by use of advanced SPC tools such as autoregressive modeling and regression control charts, the authors were able to successfully derive operating limits that could be used to monitor two of the indicators. These successes, as well as the failures, will be fully explained. The conclusions drawn from this test are that for vibration data, at least, the use of SPC in PdM is challenging and requires the use of advanced control charting techniques not normally taught to technicians and operators. Also, there may be instances where even advanced SPC methods fail, and for those parameters the maintainers must fall back on the more traditional approaches to set normal operating limits.

Management’s primary responsibility is to take care of business to ensure satisfied customers and share holders. One major component of this responsibility is to manage disruptions in the business. Disruptions are characterized by uncertainties and are responsible for introducing unscheduled delays and variations into the system. More often than not such events have a significant effect on the organization’s throughput, on-time delivery and gross revenue. Disruptions occur throughout all business segments from planning to shipping and are often caused by more than a single factor which may be either internal, external or a combination of both. 

Traditionally there has not been a body of knowledge to assist management in handling disruptions.  What little research effort has been done in Disruption Management was generally oriented towards machine scheduling and was done in high profile sectors such as the airlines industry.  However, this paper describes a body of knowledge that has been developed.  Further, this paper presents how the concept of disruption management can be applied to the maintenance function of an organization; identifying primary disruption types and their root causes, and offering appropriate tools and possible solutions.

Utilizing a “three wave” reliability strategy, Alcoa is achieving global excellence in manufacturing across the thirty-one Primary Metals Business Unit locations worldwide. The foundation of the process is a cooperative partnership between maintenance and operations. Operations own equipment reliability with maintenance as an equal partner. A familiar analogy, you as the driver/operator of an automobile own the reliability of your car, not the dealer/mechanic.

Wave 1 – This phase starts with business case creation and presentation to educate senior management of the reasons and benefits in improving equipment reliability. Focus is on education at the plant level to align plant management, engineering, operators and maintenance personnel.

Wave 2 – This phase assesses the current conditions and includes a financial analysis that estimates the value of closing the gaps to the cost of implementation. This phase produces a preliminary master plan that outlines the sequences of processes and methodologies to close gaps.

Wave 3 – This phase focuses on implementation of the master plan and includes workshops for educating proper techniques, coaching/mentoring to ensure correct execution, and implements metrics to measure progress and determine effectiveness.

These waves stress the importance of sponsorship through corporate and plant leadership. Early results have far exceeded expectations and plants are becoming more stable, lean, predictable, reliable and cost competitive. Sustainable ROI’s vary from 3:1 to 34:1 across different locations.

The life and reliability of greased machinery depend on proper bearing lubrication.  Unfortunately, inadequate grease lubrication is a very common failure mode.  With greased bearings it is difficult to determine lubricate state along with knowing when to grease, how to grease, and when to stop greasing. Fortunately predictive technologies and work processes are being successfully applied to monitor and optimize the lubrication condition.   

Improper lubrication creates energy that manifests itself as mechanical energy (ultrasonic stress) and thermal energy (heat).  By analyzing this energy using predictive technologies, much can be learned about the bearing lubrication condition.  This paper discusses and compares the use of ultrasonic analysis, conventional vibration analysis, and advanced vibration signal processing techniques as means toward determining lubrication state and achieving optimum lubrication.  

Specifically this paper will look at: 

·         the characteristics and progression of lubrication stress,

·         the characteristics of various lubricate states obtained in a controlled test environment,

·         the application of ultrasonic analysis, conventional vibration analysis, and advanced vibration signal processing for grease lubrication monitoring,

·         the strengths and weakness of each monitoring method,

·         equipment lubrication case studies,

·         using predictive technologies to aid greasing, and

·         practical steps toward setting up a grease monitoring program.

Traditional greasing programs use time based preventive maintenance.  This method can result in over or under greasing, depending on the recommended periodicity of greasing, operating conditions and run time of the machinery.  One plant surveyed stated that 90% of all motor failures were grease related – both over and under greasing.  This particular plant was using motor manufacturer’s recommendations as the foundation of their greasing program.

An alternative greasing program uses predictive maintenance (PdM) processes and technologies.  Using PdM, greasing activities are based on the condition of the grease. Several predictive maintenance technologies for grease monitoring are available.  Vibration tends to be the most common tool used.  However, most commonly used vibration techniques do not detect early lubrication faults.  Infrared is also an option, although not preferred since initial lubrication starvation has little impact on temperature.  Some sources maintain grease sampling and analysis is the best method.  However, in most circumstances this is not practical.  Many bearings do not have an access port to the grease and those that do present their own problems.  The grease samples taken most likely do not represent the true condition of the grease inside the bearing and may also contain particulate and contamination picked up during the sampling.  Ultrasonic analysis and advanced vibration signal processing are viable methods of monitoring grease condition since they measure the stress energy caused by friction between the rolling and sliding elements in the bearings. 

This topic of study will look at the application of statistical distributions, primarily the weibull curve, using it in a probability cumulative distribution function to determine how long an event must "happen" to be able to claim a new distribution exists. For example, if a seal improvement is made to a pumping system, how long does the "improved" seal system need to run before we know it's really improved? What's the confidence level? Is one point on the curve enough to make a decision? If the new seal lasts 3x as long, or 7x as long, or 10x as long as the previous design what is the probability of those events occurring if the design is no better? Furthermore, if the new seal design fails prematurely, or within the same range as the old seal system, was it because the design is no better or was it mis-installed. Did infant mortality occur because of some maintenance personnel induced failure?

The Weibull distribution is valuable because it takes on different characteristics to fit real world data. It handles skewed distributions, often found in the reliability world, with ease. We'll also explore several actual cases involving pumps, centrifuges, gaskets, filters, and others looking at how these statistics were applied, and the end results. Smart decision making obviously involves more than just statistics, but it's impossible without it.

(How we went from mean to lean to green and beyond)

Every building has one, a roof that is, and they don’t get the attention they deserve until they fail. Though roofs are in and of themselves a complex building system you seldom hear them mentioned when maintenance technologies and approaches are discussed and presented at conferences and symposiums. Roofing isn’t a glamorous subject to demonstrate predictive technologies; for instance there are no really cool multi-color waterfall charts you see with vibration analysis and no imminent catastrophes in vivid color from a thermo-graphic image of a breaker “about to explode,” just a few images of wet insulation big-deal! A roof doesn’t attract attention by spectacular failures; no explosions, fires or flying shrapnel, just a steady increase of stained ceiling tiles to repair and ruined carpet to replace (whose cost of repair is rarely attributed to the failed roof) ho-hum.

In 1994 Sandia National Laboratories roofing program was 100% reactive and we were using the “Able-Body” approach to roof maintenance (after a rain every “Able-Bodied” maintenance worker was handed a can of “pookey*” and sent into the field to stop the hemorrhaging roofs until a more permanent repair could be made. After a typical rain an average of 115 leaks were reported. We had a multitude of roofing systems to maintain (bitumen, modified bitumen, built-up, single-ply, PVC, shingle, metal …). In addition to the “Able-Body” crew, we had a sub-contract crew of 10 to 12 roofers on site daily, a dedicated construction inspector and several inspectors and support personnel. Over 25% of our roofs were considered to be in the failed state.

Today, by applying good maintenance technologies, judicious capital reinvestment, iron-clad construction specifications and persistence, our roofing program is outstanding. Despite a nearly doubling of the square footage of roof surface the cost of maintenance is less than ¼ of the previous expenditures; we now average fewer than 5 roof leaks per rain event and the staff consists of one engineer, one roofing specialist and a part-time crew of 2 – 4 contract roofers. The most recent Roof Condition Index (RCI) showed zero roofs in a failed state and over 95% are rated excellent. In addition we have been able to incorporate many green concepts: we now recycle old roofing materials (e.g. ballast-gravel is used for landscaping); we use low VOC products; we specify recyclable roofing materials; and we specify energy star products. Our most recent roofing project has taken us beyond green to the leading edge of technology by incorporating thin-film photo-voltaic technology into our roof and producing approximately 3 kw of electricity daily. The road to success is long, but the steps are clear. We would like to share our success story and show people a path to excellence.

(* “pookey,” is a technical roofing term used to describe the various liquids, powders and elixirs applied to a leaking wet roof to temporarily stem the flow of water into a building)

Piezoelectric self-sensing actuators used for vibration control have many advantages over non-collocated systems as they are lighter, less costly, and unconditionally stable for velocity feedback control. Self-sensing actuation allows a single piezoelectric element to be utilized as both a sensor and actuator. Since the control and sensing voltages both exist simultaneously in the piezoelectric material, a specially designed electric circuit, referred to as a bridge circuit, is required to realize the concept. However, precise equilibrium of the bridge circuit is difficult to maintain because the piezoelectric material properties are influenced by changes in environmental conditions. Loss of vibration control performance and stability results from bridge circuit imbalances. In this study, an array of mechanical thermal switches are used to passively control self-sensing bridge parameters in a piecewise fashion to maintain vibration control performance and stability over a wide range of environmental conditions. The original and modified self-sensing circuits were modeled analytically to simulate the effects of temperature changes on vibration control stability and performance. Compared to traditional self-sensing circuits, the addition of nine thermal switches was analytically shown to extend the stable operating range by 95 Cº while maintaining vibration control performance. The analytical simulations are being experimentally validated on a cantilevered beam system to show the addition of thermal switches to self-sensing circuits increases the stable operating range while retaining control system performance.

Many plants and businesses have been experiencing the affects of the dwindling supply of skilled and knowledgeable maintenance personnel. When we look at the trends and the demographics all signs are 'it will get worse, before it gets worse!' Now is the time for action using maintenance and reliability strategies that are quite different from traditional approaches. Maintenance can no longer do it alone. Learn how to rethink your maintenance strategy, engage the entire organization, overcome the traditional barriers, and show fast and sustainable results using proven principles.

 The Army’s focus for future forces is to achieve lowest possible logistical footprint while maximizing combat power. Low logistical footprint represents billions in operations and support cost savings as well as a less number of soldiers’ lives put at risk on the battlefield. The Future Combat System (FCS) is the Army’s center piece for its Future force. It is a multi-billion dollar acquisition program lasting over the next 10 years. The program is managed by an Army PM and Lead System Integrator (Boeing – SAIC partnership) and consists of simultaneously developing 18 systems integrated through a wireless network. Maintainability requirements for the program are challenging yet the achievement of these requirements will be a major determinate of whether or not the Army can achieve the reduction in logistical footprint it desires. Most of the FCS platforms call for maintenance ratios on the order of 2-4 times less than Army current force systems and mean times to repair of 30 minutes or less for complex, high density, and light weight systems. As such, an aggressive approach to design for maintainability termed “Pit Stop Engineering” is being employed.

A panel of Army, Boeing and key FCS suppliers will present their current and state-of-the-art methods for achieving these challenging maintainability requirements and share their success stories as well as lessons learned. Boeing and the Army will provide a short 20 minute presentation of the program updates since last year including current schedule, description of the strategy, requirements, modeling and metrics including a Pit Stop Engineering overview. Each panel member will then be given 15-20 minutes sharing some of their maintainability efforts including Pit Stop Engineering success stories. Systems covered will include Manned Combat Ground Vehicles, Unmanned aerial vehicle and autonomous navigation for robotic vehicles.

Categories to be covered by Panel members: Maintainability Requirements in Product Design, Design Tools and Best Practices, and Resultant Design Changes

The detection and isolation of faults in a large industrial system, such as a fossil power plant, is generally complicated due to measurement limitations and uncertainties, various types of devices and equipment, and the interaction among the sub-systems. An Anomaly Detection and Isolation (ADI) system is expected to be able to predict incipient equipment degradation and optimize maintenance schedule, which helps to reduce the operating cost and to enhance the safety. The new techniques being developed are based on the comparison of the measured variables with information derived from the process empirical models. The ADI module proposed in this work is a two-stage approach: (1) Development of empirical models for fault residual generation. (2) Detection and isolation of incipient faults based on these residuals. It is usually assumed that during normal operation the residuals are small, comparable to the noise levels of the signals. When a fault occurs in the system or a device, the model residuals deviate from normal allowable values, and generally indicative of an abnormal process situation.

The focus of the current research includes analyzing historical data obtained from several operating coal-fired units and constructing high fidelity empirical models for a variety of process units such as coal pulverizers, feedwater heaters, and condensers. Two data-based modeling techniques have been developed and implemented. These are (1) Group Method of Data handling (GMDH) for characterization of steady-state data, and (2) Multiple-Input Single-Output (MISO) time-series models for transient data analysis. The GMDH generates models by successively minimizing the error between the measured and the model-predicted output using a multivariate polynomial fit to a given set of inputs. The MISO model is a discrete-time model that generates a linear model of an output variable as a function of a specified set of inputs. The proposed empirical techniques model the normal process effectively, and are capable of detecting impending faults in an on-line fashion. Different fault types are classified based on the patterns of residuals, each of which is unique to a given failure type.

The results of the current research and development demonstrate the capability of these techniques for modeling long-term process data, characterization of their relationships, and for continuous monitoring of process units. As an example of application, the GMDH model prediction of pulverizer bowl differential pressure (DP) is shown in Figure 1a and the residuals between predictions and actual measurements are illustrated in Figure 1b. The three variables closely related to the Bowl DP are selected as inputs of the GMDH model; these are Primary Air Mass Flow Rate, Feeder Speed, and Hot Air Damper Position. This user-defined choice of the model inputs might be based on the knowledge of an expert, from physics model simulations, or from mathematical tools such as correlation coefficients.

The Society for Maintenance & Reliability Professionals (SMRP) has taken the initiative to standardize Maintenance & Reliability (M& R) terminology based on the needs of our members and industry. This effort is being carried out by the Best Practices Committee of the SMRP. The committee has been developing definitions for key M&R performance measures, commonly referred to as “metrics”. Through group consensus and an extensive review by subject matter experts, including the use of web-based surveys, these metrics are becoming SMRP standards. As such, they can be used in benchmarking processes and when searching for best practices.

This presentation will outline the SMRP metrics development process and provide an update on the current status of the SMRP metrics development. Examples of metrics will be put forth, and the audience will have the opportunity to provide feedback.

To date there have been a plethora of methodologies that address several key requirements of monitoring, diagnostic, and prognostic systems. While the proposed methods are beneficial to the scientific and engineering community at large, they do not address the issue of being readily integrated into a real world system. For example, recent work by Whisnant et al. [2005] describes a monitoring system that uses a nonparametric prediction algorithm to estimate the state of the system and then applies a statistical test to the prediction residuals to determine if a fault has occurred. Next, recent work by Yan et al. [2006] describes a diagnostic system that uses multiple classification algorithms to diagnose faults. Finally, recent work by Vichare and Pecht [2006] provides a survey of different prognostic algorithms, which range from built-in-tests (BIT) to cumulative damage modeling. While these three examples represent significant steps in advancing the systems that address the monitoring, diagnostic, and prognostic fields respectively, they do not provide insight into how to bring the “pieces” together into an integrated system. This paper will address this issue by describing a data based fault detection, diagnosis, and prognosis system. In addition to describing the algorithmic framework, this paper will also present results of applying the proposed system to detect, diagnose, and prognose faults in the steering system of an oil drill.

Asset performance translates to company performance, and reliability groups can illuminate and improve physical asset capabilities to help drive margins, growth, and capabilities of a manufacturing firm. The Reliability Technology Department at Eastman Chemical's Tennessee Operations is finding success in building partnerships with operations, business unit, and finance stakeholders, working together to deploy improvement strategies that are more holistic by leveraging knowledge from these multiple perspectives. In focusing on asset performance, this provides common ground upon which to discuss challenges, opportunities, and priorities for the company. Functional organizations sometimes speak different languages, but at the end of the day, all have focus on the people and the physical plants producing the goods. Companies can make better strategic decisions with more comprehensive understanding of manufacturing performance and potential capabilities. More broadly, this places emphasis on overall asset effectiveness. Using reliability tools and analysis information in concert with operations goals, desired business results, and financial systems allows companies to apply better business strategies to achieve desired business results. For example, this collaborative approach starts to highlight organic growth immediately available with current manufacturing capabilities as well as reasonably accessible through process and reliability improvements to physical plant assets. Additionally coming into view more clearly in this collaborative process is where large capital expenditure to build new capacity truly is required as compared to when it is not justified because of the hidden plant illuminated as available for use with improvements to process variability. In summary, this paper illustrates Eastman Tennessee's journey with corporate reliability and intra-company collaboration across different functional organizations to drive company performance. It touches on the working relationships and work processes developed across employee groups to achieve success. Technical content covers the reliability information systems, reliability-based operations and maintenance programs, system reliability modeling, reliability analyses in application, and asset effectiveness metrics employed at the plant site supporting the manufacturing and business performance improvements.

One of the largest problems faced in analyzing electrical equipment lies in accurately identifying where the true problem exists.  By conducting a Fault Zone Analysis, we are able to pinpoint the actual root cause of the problem, ensuring that the problem is properly identified rather than misdiagnosed or overlooked.  When analyzing the diagnostic processes of electric motors, we focus on six specific fault zones: 

a.         Power Circuit: A demonstration project on industrial power distribution systems identified connectors and conductors as the source of 46% of all efficiency-reducing faults.  Often a motor, although initially in perfect health, is installed into a faulty power circuit, causing problems with harmonics, voltage imbalances, current imbalances, etc.  As these problems increase in severity, the horsepower rating of the motor declines, ultimately causing temperature increases and insulation damage.  The motor is replaced, and the failure cycle begins again.

b.         Insulation Condition: High temperatures, age, and moisture & dirt contamination are just a few elements that ultimately shorten insulation life.  Although frequently involved in a failure, this fault zone is heavily influenced by other contributing factors. 

c.         Stator Condition: EPRI studies show that 37% of motor failures occur due to stator faults.  Age, overheating, vibration, and circulating currents contribute to turn-to-turn and phase-to-phase problems.   

d.         Rotor Condition: When left unnoticed, rotor bar or lamination damage can easily cause overheating and even catastrophic failure of the stator winding insulation.  If the rotor is not repaired or replaced after rewinding the motor, the failure cycle begins again. 

e.                   Air Gap: Air gap deficiencies driven by a bowed rotor, warped stator, cocked bearing, endbell misalignment, among others, can cause uneven magnetic fields and high vibration.  These contributing factors ultimately result in both winding and bearing failures. 

f.                   Power Quality: Power quality refers to the condition of the voltage and current signal. When a motor circuit is supplied with poor voltage quality, the insulation system can overheat as a direct result of excessive harmonics on the distribution system.  This excessive heat causes electrical losses, decreases the efficiency of the motor, and ultimately causes failure of the insulation system.

 There seems to be confusion within the reliability engineering community. “We want to improve reliability by performing predictive maintenance.” This phrase, or similar phrases, has become a mantra with various manufacturing industries. Conventional wisdom is that predictive maintenance does not improve reliability. Why the difference? There is a magazine called “Reliability” which focuses on predictive maintenance, computerized maintenance systems, reliability-centered maintenance, and other aspects of reliability. The sub-title is “The Magazine for Improved Plant Reliability”. While the focus of the magazine is on reliability of operating equipment, it does not make that distinction. University reliability engineering courses tend to focus on design reliability. Effects of reliability caused by operating the equipment are not identified as such. The American Society for Quality (ASQ) has a Reliability Engineering certification program centered on design reliability principles. Those who pass the test are given the designation of “Certified Reliability Engineer”. The Society for Maintenance and Reliability Professionals has a certification program centered on maintenance and reliability of operating equipment. Those who pass the test are given the designation of “Certified Maintenance and Reliability Professional”. While both organizations use the term reliability, they are using it in a different context. By not making the distinction, the meaning of the term “reliability” becomes context driven. We are faced with two different definitions of reliability. The person saying that they want to improve reliability with predictive maintenance is referring to operational reliability. The one who says that predictive maintenance cannot improve reliability is referring to design reliability.

In general, reliability is defined as “The probability that an item will perform its intended function for a specified interval under stated conditions.”(“Reliability Analysis Center”, p.36) For design reliability, the intended function is set by the designer. From an operational perspective, the intended function is determined by how the user operates and maintains the equipment. The life of the equipment can be much different depending on which Reliability context is used. We propose to explore the differences between traditional Design Reliability Engineering and Operational Reliability Engineering and the effect predictive maintenance has on each.

Bearing failure in rotating equipment of any kind can have catastrophic and costly effects. The consequences of failure of rolling element bearing in high performance aircraft engines can be particularly catastrophic resulting in the loss of aircraft and putting aircrew at risk. Ideally, eliminating root cause provides the most effective mitigation for this type of failure. When root cause(s) is not readily identifiable, other steps must be taken to mitigate the consequences of failure. In November 2001, within a nine-day window, the US Navy and Marine Corps EA-6B Prowler fleet experienced four in-flight incidents, two of which resulted in loss of the aircraft. All four incidents were ultimately attributed to failure of a rolling element bearing in the J52 engine. The occurrence of these incidents in the wake of the September 11, 2001 terrorist attacks on the US military’s only fleet of tactical jamming aircraft drew attention at the very highest levels in the Department of Defense. The J52 Engine Team needed solutions that could be quickly and effectively implemented. An analysis of mitigation options using Reliability-Centered Maintenance highlighted the need for an “On-condition” failure management strategy. The location and configuration of the bearing drove the development of “on-condition” inspection tasks toward wear debris analysis. Existing spectrometric oil analysis capability was refined and an additional capability to analyze filter debris using Energy Dispersive X-ray Fluorescence (EDXRF) was incorporated with positive results. However, in-flight incidents continued to occur. Adjustments were made to the decision criteria used to remove engines from service and some improvement was achieved. Key Performance Indicators (KPIs) were established to monitor the effectiveness of the oil and filter debris analysis. By monitoring the KPIs, the team was able to evaluate the process and initiate engineering studies to investigate ways to improve the accuracy of the wear debris analysis in diagnosing bearing failure. These investigations resulted in significant revisions to earlier assumptions about diagnostic equipment performance in measuring filter debris as well as the methodology used in developing the decision criteria for removing the engine from service. The revised assumptions formed the basis for changes that resulted in a significant reduction in incorrect decisions to remove engines from service (false positives) and a higher confidence that the risk of continued in-flight incidents (misses) has been effectively reduced to an acceptable level.

This work considers setting an optimal policy for inspection, repair, and replacement (IRR) of a protective system. In the context of this work, a protective system exists to operate in the case of an emergency incident and may prevent undesirable consequences, such as creating danger to human, environment or downstream equipment. Failures of
protective systems are hidden and can only be discovered during inspections or an actual case of emergency. The model proposed here considers maximization of expected interval availability of a protective system during a given interval of time. In this model, at any decision epoch we use dynamic programming to determine the optimal time to next inspection, and the type of action to be undertaken, depending on the observed state of the device. During a perfect inspection (all failures can be detected) and in case failure was detected, an immediate repair or replacement is performed and optimal time to next inspection is determined. In case when the system was found in operating condition, immediate repair or replacement can be performed or an optimal time to next action found. It is assumed that repairs are imperfect, i.e. they change the distribution of times to failure. The IRR policy is constrained by a maximum number of repairs and a maximum operating age limit.

Our case study considers inspection setting for safety pressure valves from a real refinery plant. The sample of interval-censored data representing times to failure has been analyzed. Based on the results, two theoretical probability distributions have been fit: one for times to first failure and another for times to other failures. The model provides IRR strategies that significantly improve the current practice. Results of sensitivity analysis with respect to the model parameters are presented. The possibility of extending the model to optimization of expected cost is discussed.

• Preventive Maintenance Implementation

The presentation will focus on a discussion around these six areas.

The Y-12 National Security Complex Condition Monitoring Program consists of the following technologies: Vibration Analysis, Fluid Analysis, Ultrasonic Analysis, and an Infrared Thermography Program.

Vibration analysis is the most mature condition monitoring program implemented here at Y-12. This program started in 1984 with about 200 machines (driver and driven rotating systems) and matured to about 700 machines in the mid 90’s. We currently monitor about 515 rotating systems throughout the complex.

The Vibration Analysis Program monitors the condition of rotating equipment by focusing on the mechanical vibration transmitted by the equipment. Vibration analysis can also be used on fixed systems as well. The program at Y-12 collects data on fans, pumps, motors, compressors, chillers, turbines, and gearboxes. Wireless data collection of vibration data is currently being developed to enhance the walk around routes currently being done.
Problems that can be detected by vibration analysis are bearing problems, bent shafts, gear problems, mechanical looseness, misalignment, resonance, broken rotor bars, loose stators, and unbalance.

Fluid analysis focuses on the condition of a fluid by determining the amount of contamination and degradation exist in a fluid. Oil is the predominate fluid we analyze at Y-12. We analyze the oil in compressors, standby power generators, and transformers. We also analyze the hydraulic fluid of certain critical pieces of equipment.
Infrared thermography, also called Infrared Imaging, is a technology that uses an infrared radiation detector to measure heat radiated from an object. The imager then takes the infrared measurement and converts it to an image that the human eye can see.

We use infrared thermography for electrical equipment inspections, rotating equipment inspections, and roof surveys at Y-12.

Ultrasonic analysis is used to measure. At Y-12, we have used ultrasonic analysis to do steam trap surveys and glove box leak tests. For steam traps, the ultrasonic detector can determine if the steam trap is performing properly. For glove box leaks, a sound generator is placed inside the glove box. The ultrasonic gun is then used to detect breaches in the seals, and gaskets that surround the glove box.

The Condition Monitoring Program includes four elements – monitoring, analysis, reporting, and repair. Data is collected while the machine is running or electrical component is energized. Data obtained from monitoring is analyzed to determine if the condition of the component is deteriorating and repairs are needed or if the component is in good condition. A report of repairs needed is generated and sent to the equipment owner. After repairs are made, the loop is closed by taking follow-up readings either as a post work test or on the next collection schedule that shows component condition improvement.

The primary benefits include prevention of catastrophic failure, eliminates unnecessary periodic maintenance (i.e. changing oil based on condition and not a set schedule), and more effective planning and scheduling of maintenance work and resources. All of these benefits result in cost savings as well as effective facility operations.

Integral reciprocating engine compressors designed for a throughput of 12 mmscfd is currently producing an average of 9 mmscfd. This has negatively impacted on the company ability to meet their contractual obligation to supply natural gas for gas-lifting hence; there is a need to improve the performance of these compressors.

Throughput of reciprocating gas compressors is linked to the performance of the engine and compressor. The capacity of the compressor would have to be reduced if the engine does not develop the required horsepower to drive the compressor. Additionally the compressor will incur losses from leaking piston rings or leaking suction and or discharge valves. How can these losses be reduced to meet contracted volumes for gas lifting? A holistic approach is needed to develop a maintenance strategy that ensures the engine develops the required horsepower and the compressor operates at design capacity with losses that are within design parameters.

The goal of this case study is to develop a maintenance strategy that will to keep the engine in a balanced state (all peak firing pressures within specification) and optimize the life of the compressor suction and discharge valves. The engine peak firing pressures are maintained within specifications through balancing of the power cylinders at optimum frequency with a Beta Balancer. On the compressor end statistical methods using both condition data and historical data are used to develop an optimum replacement policy for the suction and discharge valves. The RecipTrap Engine Compressor Analyzer is used to collect the condition data of the suction and discharge valves while the historical data is extracted from the CMMS.

The condition data, historical data, replacement cost and failure costs of the suction and discharge valves are connected through the use of EXAKT, a condition-based maintenance optimizing software. A model is developed to optimize the life of the compressor valves by extending their useful life and at the same time minimize the risk of failure without loss in performance. This would allow the exploitation of the margin from when a problem is detected to the time before real risk and real commercial risks occur. It is anticipated that this approach would lead to a 20% increase in throughput of the compressor.

This Company was operating equipment with no routine maintenance simply because it did not affect plant availability.  The fact is that the hidden failure costs of this equipment were misunderstood.  By exploring the
real hidden costs of failure it was determined that efficiency losses were dramatically affecting plant output.  By undertaking a RCM study, maintenance plans were developed to reduce unplanned equipment breakdowns.

By reducing unplanned breakdowns the predicted reduction in efficiency losses was 45%.  This case study presents an approach where the RCM study was used to identify maintenance plans primarily focused on reducing efficiency losses.

MARCON is pleased to host RMG’s Maintenance Planning and Scheduling Game & Workshop.  This interactive game is designed to illustrate the impact of communication and teamwork on costs and reliability. This game encourages your company to openly discuss ways to enhance strategies on implementing chosen maintenance and reliability goals. 

• Demonstrates the importance of good planning & scheduling practices

• Illustrates the impact of communication and teamwork on costs and reliability

• Show good planning & scheduling processes support production goals and schedules

• Identifies opportunities to fine tune your planning & scheduling through game play experience

• Provides forum to discuss and enhance strategies to implement your chosen maintenance and reliability goals

• Illustrates measurable improvement potential through game play 

Objective 

The objective of the Reliability Management Group (RMG) Planning & Scheduling Game and Seminar is to provide an experiential learning and illustrate potential and measurable improvement opportunities through a game play experience. This game encourages your company to openly discuss ways to enhance strategies for implementing chosen maintenance and reliability goals. 

The Planning & Scheduling Game and Seminar will:

·         Demonstrate the importance of good Planning & Scheduling practices

·         Illustrate the impact of communications and teamwork on costs and reliability

·         Exhibit how good Planning & Scheduling processes support production goals and schedules

·         Identify opportunities to fine tune your Planning & Scheduling implementation through the game play experience 

About The Planning & Scheduling Seminar/Game

• Participants set up in five-person crews (Foreman, four craftpersons)
• Additional participants assigned to Operations and Stores
• Object is to repair “broken” machines constructed from Lego© blocks
• Two rounds or shifts per game.  Round One utilizes “typical” planning and scheduling and Round Two demonstrates the value of an improved planning and scheduling process.  Dramatic improvement in work accomplishment is the norm.

Who should attend?

• Corporate management
• Site management
• Reliability personnel
• Engineers
• First-line Supervisors
• Crafts  (Operations, Maintenance, I&E, Stores)
• Anyone wanting to learn and experience the value of Planning & Scheduling