In the spirit of simplifying and quickening our routes, I'd like to know which one of these will give us the earliest detection of problems. This is for a power plant environment, so the bulk of our inspection will be pumps, compressors, motors, and heat exchangers.

I know there is alot of overlap in the usefulness of the technologies, but I'd like to just have the guys do a quick run through with a combination of IR and Ultrasound on the above equipment to get a qualitative, "yes/no", "is there a problem?" answer. And THEN we can go back and invest the time into an MCA/ESA/Vibro analysis of the situation to dig in and try to figure out the problem if the situation warrants it.


I keep asking the vendors I've been in contact with about which PdM technique gives the earliest detection, and they each say their own product does. I know I'm probably oversimplifying the situation, but I guess my question is, will those Ultrasound and IR give reliable and early enough warnings for the equipment I mentioned above in a power plant setting? What is your experience?

Sam,

Talk about a loaded question!

First of all, I'd suggest input from non vendor specific folks like Jim Hall for ultrasound. He has 15 years as a user and several years as a trainer. I've used airborne ultrasound for finding natural gas leaks with excellent, almost pin point accuracy as far away as 20-25 feet! These are low pressure gas regulator trains so you know the leaks are small. I've had to wait for Snoop bubbles to show after finding leaks easily with airborne ultrasound, as I always try to confirm any suspicioned problem with another instrument or method.

I've also used UE and IR prior to dynamic/online 460 VAC motor circuit testing, mainly as a safety precaution, just to make sure nothing sneaks up on me. I don't know of any service like Jim Hall's offered in IR, vibration and motor testing, with the exception maybe of Howard Penrose for motor testing. Are you out there Howard?

As a ten year predictive maintenance user of IR, UE and motor circuit testing I would have to say that you need all of the technology's working together for you because for any circumstance or application, one or the other may well work better than another. A lot of it is timing, whom ever gets to it first and which one has the most opportunity to test it.

As in the case of offline motor circuit testing, IR has a hands down advantage to find high resistance problems that can be viewed with the IR scanner while the equipment is running, as long as it can be seen. But what about the rest of the circuit, is it not equally as important? In the case of equipment that is only jogged or ran for adjusting purposes (intermitently) offline motor testing takes the upper hand.

And of course the given is that only motor circuit testing can tell you about the things that the IR camera cannot see like power quality, resistance to ground, rotor bar problems, load related problems like belt, bearing, torque problems, etc.

I guess my point to make after saying all of the above would be: If you're going to do predictive maintenace, I would recomend doing them all in orchestration. This covers all the bases, as frequently happens, one technology can ask another for a second opinion, getting that 2nd confirmation of a problem just gives anyone more confidence to write the work order. There is nothing worse in Predictive Maintenace than an incorrect diagnosis. The other side of that coin is when the decision is made to ignore the PdM work order, the failure occurs and then of course you have the "I told you so" opportunity that makes the technology shine. This all to frequently is the case but is necessary to get some folks to buy into the new way of doing things. Everyone resists change to some degree.

As an intersting point of view from where I sit, I really feel confident when confirming a rotor bar problem with the vibration folks, even when the potential problem is still in its infancy, which in those applications where the vibration folks can't test (like submersible pumps, machine guarded equipment and other inacessable places) I can rely on my technology with even more confidence.

In an attept to answer your question about quick and easy technologies pointing to other problems, it's not going to happen. If you want to cover all the bases, you need to do them all, or pick what you feel is important and let the rest go.

I'll leave you with this question, how many routinely scheduled preventative maintenance work orders do you intend to replace with predictive maintenace routes or standard practices like quarterly outages or other interval based maintenace? There is an opportunity here to reassign some of your existing work force to do PdM rather than PM work.

This message has been edited. Last edited by: Jim Zuidema,

In my opinion you are trying to over simplify the whole pdm process. Are you short on time or money so that you can not do it properly? If you only apply vibration to a problem found by IR or UE , how are you trending your vib data, comparing on machine to the next, setting proper alarms? If you are doing things by the shot gun method, you will get speratic success and then your PDM program credability will suffer and the program will fail. If you are short on resorces you need to do vibration and MCE on the most critical mechanical assets, IR on all electrical systems, and Ue for air leaks. This will give the most bang for the buck.

At our plant, we don't do thermography or ultrasound on a routine basis for rotating equipment, but we do perform vibration monitoring on a routine basis(quarterly for most machines in the program). So I guess the answer is that if we had to choose one of the three for routine monitoring (we already have), we choose vibration.

The justification for not performing thermography: critical machines have motor winding temperature and bearing temperature monitored on plant computer anyway.

The justification for not performing U.E. - We believe we can detect most faults early enough using vibration.

In addition to routine vib monitorign, we do perform perioidic current monitoring on a fairly infrequent basis for critical machines only (measure one phase of current and look for pole-pass sidebands)approximately yearly and off-line electrical tests for a bigger group of machines at approximately 2-year interval.
I'm not saying that's right or wrong but it's what we do, it makes good sense to me, and I believe it's not too far out of line with what most plants do.

As far as heat exchangers I'm not sure.\

This message has been edited. Last edited by: electricpete,

Recently I have to justify for a lube oil analysis program to cover more equipment and after some readings, I tend to think that a proper lube oil analysis (where applicable)should be able to detect problems earlier than vibration monitoring. Is this thought right?

For motor monitoring, is it sufficient to monitor their power consumption as the first line of condition monitoring?

It appears to me there has to be a complimentary combination of condition monitoring technologies being applied across the installation and each CBM technology has to be applied where it's most effective in giving the earliest indication of equipment conditions.

Having said the above, we should also focus on hidden failure detection which can only be detected by failure finding task.

This message has been edited. Last edited by: Josh,

Thank you for all of your responses.

I'm new to this ballgame and just trying to get a feel for what a PdM program requires to be successful. I've been given the project of implementing the program at the plants, but I generally will not be the one performing routine inspections. Our plants are by no means large, and so I'm trying to find the balance between an effective program and a quick/simple one. Keeping tabs on a PdM program will not be my sole job, supervisors at the plant will be the frontline on this. But for now this project is the main thing on my desk.

I realize that to be truly effective, you need the whole smorgasboard of tools, and be able to integrate/corroborate results and single out the root cause of a potential failure. But since we're (slightly) constrained on budget and personnel/manhours, my question boils down to:

If we were to phase a fully functional PdM program in over a period of two or three years, which technologies might we invest in to get the best results right away, with the intention of filling out our tool belt as time goes by?

-----------

Recently I read an excellent article (URL below) by a guy named Jim Taylor, a maintenance consultant our company hired 1.5 yrs ago (before I was with the company). He stresses the importance of basing a PdM program around your machinery, not your tools & toys (IR, VA, UE, et al). This made alot of sense, so what I'm doing is making a grand master table with all our systems types--generally stated (ie, "pump system", "fan system", "switch&breaker system" etc) because we have several instances of each that vary slightly--and then the table:
1. breaks each system down to components
2. assesses the functions of each component
3. lists how each component could fail to perform those functions
4. lists what the possible causes of that failure could be
5. recommends a tool or measurement (IR, VA, lube etc.) needed to recognize imminent failure behavior before it actually occurs.
I'll then apply that table to the high level list of critical systems I've made, and see which PdM tools look to be the most useful for us right away.

What do you guys think?

Jim Taylor Article

~Sam
CIPCO

Sam,

Read up on RCM (Reliability Centered Maintenance). I think that is what you are getting to.

Dave

I would say, try to get hands on "Introduction to Predictive Maintenance" by Keith Mobley.

We started with ultrasound in 2000, mainly to control steamtraps. It was the cheapest one at that time, 2 years ago we managed to buy vibration monitoring equipment, now management is jumping for thermography. Four years ago I didn't dare to pronounce the price of a Flir camera