By Jane Alexander, Managing Editor
When it rains, it pours. At least that’s what happened with April’s Reader Panel questions. They triggered an outpouring of responses—including several extremely detailed ones. In fact, we received so many thoughtful replies that, to fit them in, we’ll need to run them over two months. The questions we asked were:
What triggers our panelists’ maintenance scheduling, or if they are consultants or industry suppliers, that of their client(s) or customer(s)? Sensors? OEM recommendations? Daily walks/PdM tool data? Word of mouth? A combination?
Which approaches work best for them, and why, and vice versa?
Would panelists (or their clients or customers) want to change their current maintenance-scheduling process(es), and could they? If so, what would they do?
As always, we’ve edited this first wave of responses for brevity and clarity.
Maintenance Supervisor, Process Industries, Canada…
We use a combination of approaches. We have maintenance [personnel] and operators that use handheld devices with routes for regular inspections. Data is uploaded and emails are automatically sent out. Benchmark work orders in our CMMS are set to generate area-shut/routine work. We also have completed most of the areas on RCM. There’s still work to do to get PM work orders in the system, but that’s a continuous work in progress.
The handhelds are great if the operators/maintenance guys give us the correct information. The downside is the handhelds typically add to the huge list of emails that not everyone can read and some things fall through the cracks, i.e., minimum manning/new planners, and supervisors’ inexperience. [Other things that work well include] identifying critical assets, looking at types of failures most likely to happen, determining inspection frequency, and then “training the guys out in the field on what to look for.” Training on the sense and meaning of what can go wrong and what that looks like is critical for getting good data to act on.
Currently, work orders that are being generated are “go look at stuff”—they don’t identify or convey what should be getting done. A review of basic PMs needs to be done, as should a site/area audit to look at what is actually being inspected.
[I would recommend] providing some instruction for the operators and maintenance staff on the sense and meaning behind the PM program and ensuring there is feedback with follow-up discussions when reports coming in.
Maintenance Leader, Discrete Manufacturing, Midwest…
Our main scheduling is actually handled by our PM coordinators—who do an outstanding job handling several hundred machines per plant. Each one of our plants has a PM coordinator. If I had to choose a main trigger, it would be sensors. Our Maximo system also sets off triggers if we find an abnormality on a machine. Scheduling is generally [based on] an annual, semi-annual, or quarterly check on the machines. In our departments, we have multiple sets of machines, so we’re really doing one PM after another from cell to cell.
I really can’t say that our system has one approach that works better than the other. The biggest obstacles we run into are machines not being released from production to do the work.
The only thing that will correct that problem in our plants would be to have maintenance departments treated as a separate business and be given priorities when PMs are scheduled. Until then, we will still struggle to hit 100% compliance consistently.
Industry Consultant, Northeast…
To me, a combination of [maintenance-scheduling] triggers is best, but that requires a very dedicated planner who really understands how the world works. The most intelligent approach probably combines some fixed-time replacements, i.e., re-lamping an area or cleaning air handlers in late fall, with data supplied by predictive tools like vibration and infrared scanning.
Vendor recommendations can be a gamble. I know of a large OEM that suggests replacing bearings before they reach their design life, and the company designs around the L10! While reliability pros recognize that type of replacement practice actually reduces the reliability of the product, many of the OEM’s customers think the vendor is always right. (Along those lines, [reliability icon] Charles Latino used to say, “Never have a vendor do your failure analysis unless you have an experienced professional looking over the vendor’s shoulders.”)
There are many very competent vendors, but unless they know exactly how you are using their equipment, how can they do a good job suggesting maintenance procedures.
Sr. Maintenance Mechanic, Process Industries, South…
We use most of these [sensors, OEM recommendations, daily walks/PdM tool data, word of mouth] to some extent. We rely most heavily on PdM tool data for planned repairs with daily walks becoming a distant second. We use OEM recommendations for our newer equipment, but still verify the intervals with PdM tools. I don’t believe you can ever entirely eliminate walk-arounds. The senses that most of us possess, coupled with the desire for zero unplanned shutdowns, are still some of the best diagnostic tools available.
I think any of these methods can be very good, depending on the individual that uses them. PdM tools, in my opinion, are the best trigger available, provided they are in the hands of a trained person with the desire to learn and continue to improve their skills. Daily walks are also great if done by the right personnel. OEM “recommendations” are just that—and should be used with the backdrop of experience your company has with similar machines. As I alluded to earlier, I find OEM recommendations are very helpful with our new equipment, with which our experience is very limited.
I think ours [approach to maintenance scheduling] is heading in the right direction: away from time-based and more to condition based with the proper tools and training. I hope we continue to proceed this way.
The only downside I can see is the maintenance of the actual PdM tools. Training for employees, calibration, and repairs of the tools can be expensive and a great temptation to forgo during a tight budget cycle. Once this is done with no immediate consequences, it is very easy to repeat, sending your maintenance program a giant leap backwards. Sometimes “bean counters” don’t understand the value of trending data.
Industry Consultant, West…
Most of my clients use time-based maintenance plans, generated from experience combined with known best practices from other facilities within their companies. One of these clients has attempted to use counters within SAP, but at this point, that has not worked well for them.
All [of my] clients are using predictive techniques, with some success, and most count on the tribal knowledge of the facilities. Most of the technicians with that knowledge are close to retirement, and the knowledge is not being shared well. MT
Editor’s Note: Part 2 in the May issue.
About the MT Reader Panel
The Maintenance Technology Reader Panel includes approximately 100 working industrial-maintenance practitioners and consultants who have volunteered to answer monthly questions prepared by our editorial staff. Panelist identities are not revealed and their responses are not necessarily projectable. Note that our panel welcomes new members. To be considered, email your name and contact information to email@example.com with “Reader Panel” in the subject line. All panelists are automatically included in an annual cash-prize drawing after one year of active participation.