Are predictive maintenance technologies being used more or less than thay has been in the past?

Terry O

In my opinion the technologies are more used, simple because of the fact that the equipment became cheaper and accessable to more firms/people. The same revolution we have seen in the computer/micro processor area. When I was studying in the 80's Bruel & Kjar was the equipment for predictive maintenance. An Agema thermacam (the one looking like a rocket launcher) was the brand for thermal imaging.
Nowadays with a modest budget of $15.000 you have already suitable equipment. In 2000 a Flir P-60 was around $60.000, nowadays similar equipment is below $20.000 including software, voice recording and video capabilities.

I think that predictive technologies are defintely being used more both in the USA and worldwide due to applicability, accessibility, cost and user friendliness although I often wonder if they're being used any more effectively. Using the tools is one thing...effectively using them to improve your "environment" is another. I think there's a significant number of locations who may not be meeting there objectives.
Den

Many larger companies are becoming more involved in PdM or CBM, depending upon what they wish to call it. However, most of it is ineffective and is actually being used as periodic testing (very little trending). The 2003 MDMH study bore this out, as well (see attached executive summary).

In effect, most companies are generating more work, or even testing and not following up, resulting in a less effective program. Where PdM/CBM have their strength is the reduction or improvment of PM's (which rarely, if ever, happens).

So, while there is an increase in the sale of instrumentation, there has been little increase in PdM/CBM, to date. And that needs to change.

Howard
New Email Address: howard@motordiagnostics.com

MDMH_Study.pdf (125 Kb, 158 downloads)

MotorDoc,

You hit the nail on the head.

Trending is an important part of it all but, before that one has to have the diligence to sit down and identify all that is needed to identify the faults that may be detected with the equipment at hand. Then configure hardware and software setups and schedules based on this. Part of this includes setting trend bands and alarms so that they have meaning.

IMO this comes from not only can it be tedious and boring but difficult to collect and evaluate all this data, then correlate it to bin width, window noise factors, lines of resolution, data collection types, data collection time, band alarms and band widths.

This is where the problem is. It just doesn't get done. However, this is what makes it work. The information is in the detail. The detail is in the data.

We also must document and present at some point the business case as well. What value are we providing in this effort?

Most don't have or take the time to do the above. There are many reasons I am sure.

Couldn't agree more!
Den

Awareness about predictive technologies is at an all time high...even at the management levels of industry. The most obvious change in attitude came about when companies began using and enforcing performance expectations tied with pay for mid and upper level managers. With part of one's pay on the line, the decision to use predictive/diagnostic technologies to improve uptime and verify the mechanical health of equipment was an easy one to make.

I don't believe the majority of companies know how to manage and grow a predictive maintenance technology program. One of the biggest problems is related to pay for the PM technician. Few companies have good written performance expectations and appraisal reviews for PM technicians. PM techs are like square pegs forced into a round hole. They are under paid; rated somewhere between a trades person and an engineer with no good way of tracking improvements in technical ability or quality and quantity of work completed. As a result, PM technicians come and go and PM programs lack growth and often die.

I have to agree with Tony. I support NASA by maintaining their test labs. We aren't responsible for bricks and mortar but rather the very expensive and one of a kind test equipment. The contractor responsible for bricks and mortar has a vibration route but I can't get buy off for any CBM technologies for the labs. Management doesn't understand what it can do for them nor do they have any idea of how to organize it.

So.....what you're saying is that even REAL rocket scientists can't get to grips with this technology.

PdM is "rocket science" technology, but every activity has his dull and boring side.
Looking at the graphs, giving "intelligent" answers, taking spectacular readings on exotic equipment is the glamour part.

The dull side is: looking at piles of equipment documentation trying to find out what is in it. Or trying by interrogating the craftsman have an idea about the history without being accused of witch hunting.
The tendency is to replace people with machines/computers, but like someone stated in another thread "I have not seen computers pulling any wrench lately".


But someone has to do it, for the program become effective.

My opinion is that predictive maintenance is growing and will continue to grow. It may be at different levels of competency but none-the-less someone will be collecting data and attempting to head off catastrophic failures. This will happen until machinery has self-diagnostic technologies that are reasonably priced.

Without a doubt, predictive programs need management support or they become stagnant or even die. The analyst/technician has to be a good sales person and even then the management has to want the program for it to advance. Value $$$ has to be recognized or management support fades.

Lean manufacturing means lean maintenance. Manpower has to be used effectively. Hard to predict a pipe gasket or a diaphram failure but at the very least hot spots in switch gear...excessive vibration from a location on a pump or motor...steam trap not blowing or continuously blowing all can be identified and scheduled for repair or replacement.

Those that wish to run to failure will just throw money at these failures and sooner than later not have a business to worry about. Do you think?

I think it's a paradigm shift that needs buy-in from everyone to be effective. That buy-in will only happen when someone sees the benefit; management must see the return on investment (i.e. I know it costs money up-front, but it saves me money down the line), and the person actually using the technology must see that it makes their job easier.

Both of these things will only happen once someone demonstrates a success, but it's a catch-22 where someone has to take the leap and struggle with it. It appears to me like the classic situation presented in Geoffrey Moore's Crossing the Chasm; there has to be a pragmatist that shows it works before others will buy in, but no one will buy in until that point.

Jeff

I´ve been in some sites in Venezuela, Colombia, Chile and Peru (South America). In general I can see an expansion of Predictive Maintenance programs, principally sponsored by international companies with local plants or bussinesses. However, in most cases I see that most PdM/CBM people just act like a data/report guy: test, analyze and report, and avoid the last steps which are to process the information coming from the program and to provide echonomical or improvement tools for managers, enhancing our value.

This limit the PDM/CBM success, and after a few years the honeymoon finish and we are no longer perceived as people adding value, just some guys who test machines.

So, finally I think that of course there are more companies doing PDM/CBM, but just a limited number of them have succesfull programs running. At least, on this side of the globe.

I think there is a wide gap in the application of predictive maintenance technologies between the leaders and the laggers. How many sites use all the 5 major predictive maintenance technoclogies ie lube oil analysis, vibration, thermography, ultrasonic & motor testing? How many of the laggers are still heavily stuck with fixed-time maintenance policies and predictive maintenance is minimal, only to few equipment? I guess this is the challenge for the maintenance community. How fast will we close the gap?

Rainy saturday, had to chance to read an article
http://www.mt-online.com/articles/moubrayseries.cfm
and decided to stir the pot "again"

1) When it comes to predictive maintenance, we focus exotic machinery, rocket science testing equipment etc.

We talk about preventive maintenance, predictive maintenance and condition based maintenance and mix a whole plethora of terms

quote:

OLD
If technicaly appropriate, fixed interval overhauls and replacements are usually cheaper and more effective than condition-based maintenance
New
If technicaly appropriate, condition-based maintenance is nearly always cheaper and more effective than fixed interval overhauls and replacements ....

We started with technology and we are landing in the procedures/policies , the soft area

Mr Moubray states also:

quote:

Predictive maintenance tasks can nearly always performed without moving the asset etc...

Where do we draw the line between preventive and predictive?
If I put on a periodic PM worksheet of a generator to replace the airfilter if the vacuum gauge reaches a certain point, or on the operator's logsheet I have indicated this warning point (nowadays they come color coded), but I am not taking vibration readings, oil analysis or infrared scan, will that automatically classify my organization in the camp of the troglodytes?

quote:

but I am not taking vibration readings, oil analysis or infrared scan, will that automatically classify my organization in the camp of the troglodytes?

Svanels,
I posted the reference to the Moubray article and would like to respond to you - I may not get the gist of your comment above. Maybe you could elaborate just a little.

Thanks Steve, very useful link.

My point is that we fall constant in the preventive x predictive discussion, while the technology silently is build in the equipment or components.

Another point is that when we talk about predictive maintenance we think only of vibration, ultrasound thermography, etc.. while the human senses combined with plain horse sense and knowledge is neglected.

Where do we draw the line between preventive and predictive?

I think once we blend in human knowledge into machines like expert rules, neural network etc, this will be called intelligent maintenance?

Intelligent Maintenance, that is a good one Josh! I suggest to ask patent on this, before some big company use it to promote their new silver bullet for maintenance.
Although I am a big fan of computers, I would rather take orders from Humans.

After reading : The knowledge, crashing computers from Michael Coleman I persist in my thinking that we must take noting for granted, even from a computer

i think predictive maintenance is being used more and more to increase productivity. If you look at Rolls Royce, their predictive maintenance technology for jet engines is hugely rewarding for them. Trasnmitting engine data via satellite and analysing keeps the planes flying when they dont need to emergency land. I think i read that here http://www.engineeringmaintenance.info but not sure perhaps on maintenancetalk.