I like to use the term condition monitoring because the term Predictive Maintenance excludes the very important part of basic inspections that includes See, Listen, Smell, and Touch. Condition Monitoring includes all tasks you do to discover problems early—basic objective inspections, basic subjective inspections, vibration analysis, infrared measurements, wear particle analysis, ultrasonic material testing, acoustic emission testing and other methods.
Teaching operators and maintenance craftspeople basic inspection techniques and then deciding who is going the do the inspections in a route-based system is a very cost-effective way as a first step to reduce reactive maintenance. In several studies we have found that most problems are, in fact, detected through basic inspections.
Some basic inspection tools include:
- Portable high intensity flashlights improve visual inspections.
- Inspection mirrors and fiber optics for areas difficult to reach.
- Ultrasonic leak detectors are especially useful to detect vacuum, and other leaks.
- Ultrasonic material testing discovers cracks or thinning in high-wear material.
- Stroboscopes enable inspections while equipment is in operation.
- Dye penetrantcan be used to discover micro cracks in material.
- Tensiometer to check and properly set tension of V-belts.
- Infrared thermometers for detection of misaligned couplings, leaking valves and other components in hydraulic, and media systems.
Believe it or not, basic inspections are very important and often overlooked. Several years ago I visited all sawmills owned by a company in New Zealand. During training sessions I showed a well-known method used to measure wear of chains. You measure the distance between five to ten chain links and as the distance increases you record the wear to a point where you have decided to replace the whole chain. Logs arriving to the sawmills were loaded on a log deck with four to six parallel chains that lifted the logs into the sawmill. Their maintenance practice was to change the chains once a year in their low season, because “That’s what we’ve always done.”
During a conference in Auckland, seven years later I ran into one of the maintenance managers who had listened to my idea to measure the chains and change them when needed. He told me they had abandoned the old approach and by measuring the wear of the chains they now replaced the chains every eighteen months or so. In addition to that, he told me, a sales person had offered them a chain for nearly half the price so they bought three to run a test and they ended up lasting for more than two years. So “Double life at half the price!” was the result of using this very simple method. I havehundreds of similar examples as to what results basic inspections can generate in addition to early detection of problems and increased time to plan and schedule work.
A rather major result of what basic inspections can yield in is a scenario involving a patent dispute, years ago.
In the early 1970s we worked in a steel plant in northern Norway where we suggested marking some type of couplings with a line across the coupling halves. By using a stroboscope they could then inspect if the distance between the marks increased. If it did, it indicated wear of coupling bolts and, or, rubber bushings, and it was likely due to misalignment.
About 20 years later, a manufacturer of these types of couplings contacted me because they were in a dispute and a competitor had filed for patent to use this method on the same types of couplings. They had found documentation from a training I had done in China, describing the very same method and they wanted me to verify this.
Condition Monitoring tools include:
As we know, vibration analysis, which is mostly used to analyze conditions of bearings and imbalances in rotating equipment, is part of predictive maintenance. So is wear-particle analysis, including ferrography and spectrometric oil analysis, acoustic emission used to discover early cracks in material,and infrared cameras used to visualize temperature patterns in equipment. This is an area where technology is developing very fast with many online systems installed on equipment for continuous monitoring and data collection.
The PM System
For any fixed-time PM, work orders can be used, but for any tasks to be done while operating (lubrication and basic inspections) a route-based system is more efficient. In setting up a new PM system it can be very useful to have some general guidelines to follow. As a principle you first decide what task must be done, then you decide who is going to do it. First decide which categories of employees will be involved in executing the PM. I often recommend following categories by priority in this order: operators, maintenance craftspeople, in-house expert, and lastly, external expert.
Operators should be the first choice if the task is close to the area they work in and they know how to do the task safely. If not, consider how fast they could be trained to do the task safely. If they can be trained to do it in less than 30 minutes, you train them and then document frequency and description of what needs to be done in a system. If the answer to these questions is no, the task will end up with maintenance craftspeople. This includes tasks that require more skills and/or are done with a longer frequency. When vibration analyses, ultrasound, or other predictive and specialized tools are needed, enlist an in-house expert. As a last resort, turn to an external expert who can bring expensive tools and skills not available in the organization.
Many existing PM systems were set up many years ago, and it’s important to update and review our systems for optimized performance. Technology for Predictive maintenance has improved and become much more affordable over the years. The use of these technologies has also increased oftentimes without taking into consideration the existing PM tasks. Too much or convoluted PM should be reduced, and I’ve found that it is possible to reduce total PM work up to 30 to 50%, which also results in much better PM.
FTM can be challenged and often switched to condition-based tasks on-the-run inspections. Many of these tasks can also be merged between mechanical and electrical and operators.
In addition to equipment and component knowledge, one needs to understand the basics of Reliability Centered Maintenance (RCM) methodology, in order to set up or optimize a PM program. I learned this from professor Horst Grothus already back in 1968. We did a training seminar together and he explained how failures are distributed in time as well as the different ways failures develop into breakdowns. Since then I have used this as a guide when helping to set up and optimize PM systems.
Full blown RCM analyses might be needed, but in most systems, critical failure symptoms and required PM tasks are quite obvious for the experienced. In my opinion full RCM analyses are most beneficial in the early design phase of systems and equipment.
Many suggest that you need to do Failure Mode and Effect (FMEA) analyses on all components when you set up a PM system. It can be necessary during the design phase andon complex and unknown systems but in most cases it is an over-complication. Instead of FMEA I propose you think of what symptoms you might notice on a failing component and what method you can use for early detection.
Most common failures are random. Others are more regular or predictive. Regardless of experience or what statistics show, you cannot predict when random failures will occur. A drastic but clear example is how a valve can show a time between failures ranging from 1 to 12 years. If we combine random and regular with the estimated failure developing periods, which is time between failure and break down of component function, we can use this as a guideline to decide what’s the best maintenance procedure is: Operate to breakdown (OTB), do FTM (fixed time replacement or overhaul) or condition based maintenance (CBM) If the failure is random and there is a very short developing period and you cannot find any method to prolong this period, your choices are to apply redundancy of component or decide to run to breakdown and prepare in advance for corrective action, making sure you have needed spare parts and standard corrective job instructions. For all scenarios that have a failure developing you can use condition monitoring, regardless of if they are distributed randomly or regularly in time.
I am a big fan of case studies when we have large and/or complex issues to put theory into practical and tangible examples, when deciding which maintenance procedure is most cost effective. I find many PM programs containing too much Fixed Time Maintenance based on no more facts than “We’ve always done it this way.”
______________________
Christer Idhammar is the founder of IDCON, Inc., a management consulting firm (idcon.com). This article was excerpted from a recent book authored by Mr. Idhammar entitled Knocking Bolts. More information can be found on this book at https://www.idcon.com/reliability-and-maintenance-books/
