SIRF Roundtables (SIRF Rt) develops and facilitates shared learning networks across Australia for operational excellence in industry. The Strategic Industry Research Foundation was originally an Australian government sponsored organization created in 1988. The key themes for the 15 networks currently facilitated by SIRF Rt are Industrial Maintenance, Manufacturing Excellence, Knowledge Management and Sustainability. The Industrial Maintenance Roundtable services capital intensive industry in particular where high plant reliability at low cost is essential. 300 member companies range from airlines to railways, power generation and distribution, mining and refining, manufacturing and processing of everything from petroleum to food. For these industries maintenance is a high leverage business activity and is often the highest controllable operating cost. The objective of the Industrial Maintenance Roundtables is to provide a forum and support network for members to share learnings and to accelerate their progress on the path to maintenance excellence. Key focus areas are defect elimination, teamwork, reliability principles and front line engagement.
The R in SIRF stands for “Research” which compels the organisation to seek to quantify and not just facilitate. This has led SIRF Rt to conduct more than 130 formal benchmarking studies within Australia and in Asia, Africa and North and South America. These combined with surveys and observation from the 250 events per year where members present and share their approaches and levels of success have provided a deep insight into what does and does not work. Organisations which have been able to engender good and appropriate defect elimination practices in their workforce and use appropriate tools to their particular environment to select the correct countermeasures have demonstrated remarkable levels of success. While companies that have tried to enforce an inappropriate approach have tended to not only fail in the task but also to de-motivate their personnel. For example formal RCM is known to be extremely successful in the aviation industry but has been seen to be a “Resource Consuming Monster” in some other industries. It seems as though people who have success with a particular approach in one scenario can come to believe that the approach is the only one to take in all scenarios and can become zealots for their chosen approach with less than satisfactory results.
The observation that there are in fact many different approaches that can be successful in different circumstances has led SIRF Rt to refrain from advocating one particular approach. Instead SIRF Rt would recommend that leaders understand the strengths and weaknesses of the different approaches and be in a position to select and use the particular approach that suits the particular set of circumstances. Facilitators in the SIRF Rt networks are themselves drawn from industry and tend to be people who want to “do” something. So it has been inevitable that tools and processes would evolve out of the observations. The common thread in these tools has been that they are frameworks that are flexible enough to engage and support different approaches. For example;
- The observation that Toyota has found 5 Whys to be enormously powerful, while others have demonstrated success with Six Sigma and DMAIC illustrates that different approaches to problem solving are successful in different environments. Our observation has been that organisations can benefit greatly by engaging personnel at the front line asking “5 Whys” as they do their normal work while others in the same organisation apply statistical approaches in a six sigma approach. SIRF Rt’s approach has been to create a framework that provides a continuum from 5 Why through to Six Sigma. We see Root Cause Analysis (RCA) as a mid point between 5 Why and Six Sigma and is the bridge or framework that allows a problem to surface through 5 Why, be elevated through to RCA if appropriate and then further refined into Six Sigma if justified. The continuity allows personnel to stay involved throughout. The framework is called RCA Rt and seeks to provide a simple framework that shop floor personnel can fully engage with that will help them stay engaged from 5 Whys to DMAIC.
- The work of Nolan and Heap showed that RCM is invaluable in the aviation industry for defining the best mix of work procedures. At the same time case studies from power distributors maintaining hundreds of thousands of similar power poles have shown that streamlined approaches have a place. And again case studies from low consequence of failure, “brown field” processing plants made up of innumerable and different items of equipment have shown that PMO can significantly improve cost and reliability with a significantly lower investment than full RCM. In a large facility the underlying principles of RCM might be applied rigorously to some critical equipment, while PMO is applied to a broad range of equipment and streamlined or template might have a place in another area. SIRF Rt’s approach has been to create a framework that can support the full spectrum of approaches, highlighting the fundamental core principles so they can be easily applied and documented.
Insights from Ledet, Johnson and Moore
SIRF Rt has itself been on a journey of discovery over the last two decades and has learnt from many sources. Our greatest teacher has been the case studies from members who illustrate the reality of what works, which does not always align with theory. This paper is being delivered in the United States so we thought it might be appropriate to thank three Americans who have visited us and left their mark on Australian industry.
Winston Ledet, formerly of DuPont and developer of The Manufacturing Gamei came to Australia as a guest of SIRF Rt in the mid 1990’s and met with a number of members. He delivered many valuable insights some of which are embodied in his concept of “paradigms of maintenance” shown in the Journey to Excellence figure below. Many companies struggle to break out of a reactive maintenance environment. Many recognise and can visualise a planned paradigm but find it harder to understand how to create a proactive culture.
Winston helped us see the importance of defect elimination and set us on the path of looking for better ways to solve problems. We looked at particular approaches to problem solving looking for the “best” approach but ultimately concluded that there are “horses for courses” and that it may be better to apply different techniques for different problems and find a way to being it all together.
Winston’s paradigms also highlighted the importance of planned work. Subsequent benchmarking studies by SIRF Rt showed us how sites with strongly developed repeating work orders have a distinct advantage. They find that a significant part of the work that can be planned, is already planned, with repeating benefits to pay for the development. The benchmark studies also revealed sites whose maintenance programs had grown organically over time, with little rigorous analysis, so that a large number of repeating tasks were a waste of resources as they were duplicating other tasks, were done too often, or too late, or were intrusive or simply served no purpose at all. In this case improved planning will only efficiently allocate resources to too many tasks and no real gain is achieved. Also without a good defect elimination program then improved planning will only allocate resources to fixing recurring problems more efficiently.
The increased awareness of the importance of repeating work orders and feedback from member organisations that traditional RCM was not always the answer caused us to search for other solutions. In 1995 Larry Johnson’s paper iiat the 1995 SMRP conference documented the use of Preventive Maintenance Optimization (PMO) at nuclear power plants. Larry visited Australia for SIRF in 1996 and presented a number of workshops on RCM and PMO for Roundtable members. This culminated in a benchmarking exercise where three member sites conducted improvement programs on their preventive maintenance schemes using different approaches. In the case of brown field equipment it was found that the PMO approach yielded similar results but was a third of the time required by the traditional RCM approach. But at the same time there is a concern with PMO that a key fault may be missed so the conclusion was that there is a place for both approaches.
Ron Moore has been another guiding light for SIRF Rt and has visited members throughout Australia and New Zealand during his many visits. His insights into “keeping it simple”, the operators taking responsibility for reliability and the importance of doing the basic things right all the time have been invaluable. Ron’s bookiii Making Common Sense Common Practice provides an understanding of the human issues that underpin any improvement program.
Insights from these visiting “gurus” together with insights from member experience have provided clues to how organisations can find the time and resources to both fix things that break and make improvements to break out of reactive paradigms. If organisations can inspire their personnel to eliminate defects through problem solving and improve their repeating work processes then they reduce their workload. This provides more time for eliminating defects and improving repeating work which becomes a virtuous feedback loop.
SIRF Rt’s Framework for Defect Elimination
Central to all that we have learnt is that there are different solutions that are suitable in different situations. Our focus has been on helping to create a framework that can bring together diverse notions. Our initial focus was on problem solving. We wanted to create a bridge between 5 Why and Six Sigma. We felt that Root Cause Analysis is close to the centre between these two poles. So we have developed a process centred on Root Cause Analysis called RCA Rt which has software support that can cater for 5 Whys, through RCA and Six Sigma’s DMAIC.
Remembering Ron Moore’s advice to keep it simple the Root Cause Analysis tool makes use of the fundamental mechanism of asking why over and over then representing this with a cause tree diagram. Most people gain a very quick understanding of the 5 Whys strategy and can adapt it quickly applying it to almost any problem instantly. It therefore is one of the strongest preliminary steps that can be taken to introduce a problem solving culture in any business. It establishes a culture of looking past the symptoms of problems and establishing what might be the underlying cause.
The two main benefits of the 5 Whys include:
- It helps to quickly determine the root cause of a problem
- It is easy to learn and apply
Experience has shown that this type of introduction to cause trees is very easy to understand and ensures that all personnel can actively participate. It is more important in the initial stage of cause tree development to keep the process simple, avoid overly complex rules and syntax between causes in order to keep all engaged.
People inevitably find that there are alternative “whys” and naturally progress to develop cause trees. There are inherent weaknesses in building cause trees only using 5 Whys however there are skills that can easily be learned to overcome these weaknesses.
It was recognized that problem solving is a key skill required for successful RCM /PMO as the real failure mode must be identified for functional failures and an error will lead to a faulty RCM / PMO analysis. The representation of the RCM / PMO analysis on a tree was also found to be very helpful in group exercises. This led on to redevelopment of the RCA Rt software to suit RCM / PMO analysis. From the experience SIRF gained from research into RCA techniques it was found that most people immediately understand a diagram and a simple approach without being burdened upfront by over technical analysis. Once the basics are captured then more detailed information can be added to the diagram.
A three phase approach has been adopted. The first phase focuses on gathering information. This is everything that is at hand about the system being reviewed. An understanding of the existing maintenance tasks is gathered at this stage. Then during the second phase these tasks are aligned with the functional failures they address. A representation of this is constructed in a “tree” which results in an understanding of all the possible functional failures of the system and the failure modes. During the third phase each failure mode is reviewed and using traditional RCM logic (Nolan and Heap) a decision and justification for a PM task can then be made.
- To begin collecting the information for phase one an excel spread sheet is used. Information is gathered into four columns. The first column captures information that relates to Counter Measures, (including existing PM's) Failure Modes, Functional Failures and Functions. This can be information gathered from interviews with operators and maintainers, existing tasks that are downloaded from a CMMS, Work Order history, any existing failure history or analysis or OEM information. At this stage it is important to involve as many people as possible who are actually engaged in maintaining the equipment under review. As the data is captured more information can be included in the other three columns under the headings Detail, Evidence and Source. Different information can be included under each of these columns depending on what is entered in the first column. For example if the data entered into the first column relates to a function then the detail may capture a complete statement of the function including the operating parameters and acceptable limits whilst if the data relates to a failure mode then the detail might be a list of the possible causes of the failure or an OEM recommendation, or it might be information that helps to determine the Failure Consequences are;
- Safety Environmental
- Non operational
- If the Failure Effect is hidden or evident
Perhaps the information in the first column relates to a functional failure then detail at this point may be an observation extracted from an operator interview. Again the same is true for evidence entered into the third column. It may relate to the OEM manual or simply state the name of the operator and the date of the interview depending on the data entry for the first column. The reliability of the Source can also be entered in the 4th column.
- The next step involves constructing a “tree”. The nodes of the tree include: the functions, functional failures, failure modes and counter measures of the system under review.
The tree can be built by filling information in the tree nodes from left to right or from right to left. It’s probably simplest to understand that in a traditional RCM approach, where no maintenance activity is assumed then the “tree” would be built only from the left to the right. However, when there are current maintenance activities being performed and it is a review of these activities that is the goal then the tree can be build from both left to right and right to left. This is done by using the information in the library to fill the some nodes across the “branches” of the tree. Sometimes more “thinking” might be required to complete a tree branch, for example there may be a number of counter measures that have been imported into the library so with a little thinking and manipulation the failure modes that these actions address can be identified along with the functional failures that are associated to the failure modes. In this way the tree can be built from right to left. The tree can be constructed by dropping the information in the into the nodes on the “tree”. When the information was imported into the library form the spread sheet the additional detail, evidence and source information was also included. This can be found included in the information capture for each node. More information can also be added to these fields as the tree is built.
- The final stage of the process is to test the logic of the tree. Each failure mode can be reviewed using the RCM decision logic of Nolan and Heap. The detail for the countermeasure can be used to capture the type of maintenance task recommended along with the justification for the decision. The action field documents the task and can be used to include the cost benefit of the task when necessary. Each node can also be assigned with a “risk” assessment. This will identify high priority tasks, critical parts and will help to determine stores.
Once the tree has been build and the logic rigorously reviewed anything that is left in the library and has not been included in the tree would have to be deemed unnecessary work, or redundant functions. The countermeasures identified can now be used to structure a new maintenance program.
The steps outlined obviously can be adapted and modified to suit. Any amount of detail can be included or precluded depending on the depth of understanding required and the results have been determined. SIRF Rt’s approach has been to create a framework that can support the full spectrum of methods, highlighting the fundamental core principles so they can be easily applied and documented.
i Dr Winston Ledet. The Manufacturing Game http://www.manufacturinggame.com/
ii Johnson L and Johnson S. "Improving Equipment Reliability and Plant Efficiency through PM Optimization at
Kewaunee Nuclear Power Plant." , SMRP 3rd Annual Conference, Chicago Illinois, October 1995.
iii Ron Moore. "Making Common Sense Common Practice: Models for Manufacturing Excellence" 2004