Management of Condition Monitoring
From Imrtwiki
Return to Condition Monitoring - Predictive Maintenance Techniques
Changing the Perception of Condition Monitoring
By Michael Tansley - Mt Thorley Warkworth Mine
In the last few decades Condition Monitoring (CM) has become an accepted part of the maintenance regime on any equipment. However the author believes that in many situations the full benefit of CM is not realised. Ask most people why they do CM, and the answer is “to find problems before there is a failure of the component”. The real reason we should be performing CM is to help manage the risk of operating an item of equipment, by providing information to maintenance, to enable a qualified decision to be made on the maintenance required. Maintenance is about managing the risk of operating equipment in a certain condition and under certain operating conditions. This can be from a “duty of care” perspective in the provision of safe equipment to be used on our sites, or from an operational perspective, in ensuring that equipment is available and reliable.
CM is touted as a predictive and proactive tool however in many situations CM is still used in a reactive mode. The CM program will find a problem, and then the maintenance department will repair the problem. The long term benefits of CM are to use it in a proactive mode. CM is closely linked with reliability, and the long term benefits are to think this way. What CM provides is a detailed failure mode pattern for our equipment. What we need to do is to use this information to plan the maintenance philosophy on that equipment. Even to the stage that design improvements can be done to new equipment to eliminate or better manage known faults or weaknesses. CM data needs to be used to improve the planning function of maintenance. It is a fact that machinery will wear out, and depending on the type of machinery, it will have different wear modes. We need to identify these wear modes and failure modes, and perform maintenance or CM actions to manage the issues occurring with that particular equipment.
CM is still seen in some industries as a luxury item, which needs to be justified in its own entity. While every process or system used in industry needs to be justified, we need to change our perceptions of CM as a stand alone, and instead consider as an integral part of the maintenance function. CM is often performed by contractors or maintenance / technical support groups and is perceived as a specialist task which does not involve the maintenance personnel on the floor. One of the major obstacles is the key performance indicators (KPI) used for CM. These are nearly always the number of wins versus the number of misses. The basic function of CM is to detect problems. Surely we have progressed beyond the point that we need to justify its existence. CM is an integral part of the maintenance function, and will be so in the future.
We perform maintenance to ensure the availability and reliability of equipment. It is more than the misguided believe of a “fix it” function. Any maintenance function needs to be able to guarantee that the task of maintenance has restored the machine to the same or better state than before maintenance was performed. This is one area where we rely heavily on the experience of the maintenance personnel. The maintenance people need to be able to perform a mini failure analysis as they are working, so that they can adjust their maintenance to eliminate the cause of failure, and work with the CM to manage the defects.
Our maintenance philosophy for an item of equipment needs to be designed around the expected life and failure modes of that equipment or component. All equipment needs to have clearly understood life management plans. These plans need to include operational expectations and plans as well as maintenance proposals to manage the equipment. Once we know how a component fails or lose performance we can start setting maintenance plans. Components which wear out in a rate proportional to hours used are typically managed by scheduled time based maintenance. Items which have more random failure modes rely on inspections and or CM.
In industries which have tightly controlled operating and maintainability methods, setting the maintenance plan becomes more accurately determined. In industries such as mining, in which there are extreme variables in how equipment is operated and how it is maintained, there is less ability to accurately determine what maintenance intervals are required. Maintenance plans are normally recommended by manufacturers, and modified over time by the maintenance department as experience is gained on that item of equipment. This can be done in a structured manner, but will normally end up been refined over time, as experience is gained. With maintenance philosophies, one must be careful of not falling into the trap of striving for the ultimate solution, when in many cases the basics have not been covered. Organisations often set standards or goals for maintenance to achieve, but in reality these goals are not practical for the equipment or operating conditions.
The life cycle plans for equipment is normally based on one of the classic failure patterns of machines, from infant mortality, through to progressive wear out. (Figure 1). However from the authors experience there is a variation to these classic failure modes. Machines or components often fail in step changes. Steps can be from the initial root cause, through different stages, until eventual failure. The importance of these steps is that if the situation can be managed, then the machine can run with a defect condition. An example could be gearbox, which can often run with damaged gears for many years, with no further wear occurring to the gear teeth. If one can remove the initiation cause of the step change, then the life can be extended indefinitely. Rolling element bearings are a classic example. In ideal conditions, these can run for 15 to 20 years or longer, but if not maintained, then the life can drop to a matter of weeks. The initiation of a step change can be from a maintenance practise to an operational change. To the example of the gearbox again, if the load on the gearbox is increased it may trigger a start of the previous wear. Another example is cracks in structures which are very susceptible to operational changes.
Figure 1 - Classic Bath Tub Curve
Figure 2 – Step changes in component degradation
To manage these step changes, we need to know when we have dropped out of our ideal steady state operating condition, into the start of another failure mode cycle. This is where CM informs us of the state of the life cycle we are in at present. To be able to accomplish this, we need to have CM performed to a very high level. CM has to be able to identify the earliest defects, and be able to track those defects, and be able to identify where that component is within its failure mode. Components or equipment not achieving their design life is normally due to some contributing condition. The condition monitoring person needs to be looking for these contributing factors. Failure analysis needs to pin point the cause of these changes from steady state condition. To the example of rolling element bearings, these have been known to run for years with damage, but then a slight change can cause instant failure. How to recognize this?
Back to reliability. How do we know when we are moving out of the steady state condition? First what causes the changes? The two obvious answers are maintenance and operations. Maintenance and Operations need to know when the equipment has dropped out of the steady state operation, so that the situation can be managed. Information such as overload conditions or poor maintenance needs to be recorded for future use. This information can then be used by the CM or reliability engineers to understand the true state of the equipment. We need to know what the risks are of running in a certain condition. Knowledge is the key here. The more information we have the better we can manage the risk. This may involve running under a risk assessment condition for some critical equipment.
One often sees CM inspection frequencies been increased, to compensate for lack of maintenance. CM is seen as an easy replacement to a permanent fix to a problem. The only time it is acceptable to reduce the interval between inspections, is if it is not possible to access the machine for a proper repair.
CM information needs to be able to be broken down into problem types and components to allow for analysis of the data. This is easier with CM such as VA or Oil analysis, but NDT has less variables. NDT is always finding cracks in structures, and the cause is normally fatigue. With VA and oil analysis there are very specific problems to specific machines, with specific causes. An example would be abrasive wear in conveyor pulley bearings due to dust contamination, or low viscosity due to fuel dilution. Maintenance data needs to be complete as well, to cross reference data. Most CMMS systems have a reliability field, where comments can be made regarding causes of the problem, etc. Having detailed data enables detailed analysis to be performed. All too often the reliability engineer is left to make decisions based on perceived ideas and verbal communication with the relevant maintenance personnel.
By improving the level of the CM activities, and by integrating the CM into all of the maintenance activities, we will be better prepared to manage our equipment into the future.
Paper by Michael Tansley - Mount Thorley Warkworth
