Failure Modes 3

[nnjunger]

We are beginning to establish our failure modes, and I am sure that libraries of standard failure modes exist for common pieces of equipment (pumps, fans, gearboxes, etc.). Does anyone know how to obtain this information as a starting point rather than reinventing the wheel. I understand that there is value in working through the exercise of analyzing, what I am looking for is a starting point. Thanks.

 

[Pressed]

[URL unfurl="true"]http://www.uscg.mil/hq/gm/risk/e-guidelines/rbdm/html/vol3/07/v3-07-03.htm[/URL]
[URL unfurl="true"]http://www.sverdrup.com/safety/fmea.pdf[/URL]
[URL unfurl="true"]http://www.plant-maintenance.com/maintenance_articles_failure.shtml[/URL]   (many examples here)

==========================================
Business Card

http://mech.e.tripod.com

__________________________________________
Cycle Heaven.......

http://www.tailofthedragon.com

 

[jeez]

nnjunger, There really IS value in working through your own failure data, preferrably in a group including subject matter experts (SME) on the equipment maintenance history, if for no other reason than all of that focused thinking improves the awareness of and "intimacy with" the specific failure issues for you and your group. There are proven methods like RCA, FMEA, RCM for pure analysis approaches. (FMEA is best for identifying the failure modes, consequences and action to prevent failures.)

However I think in your case you're looking for a jump-start to selecting appropriate failure codes, right? First thing coming to my feeble brain: What form is your failure data in? Do you have a CMMS and are you capturing the important maintenance info for all the breakdowns, e.g. repairs done; time to repair; subcomponents replaced; etc. WHY are you needing to develop failure modes? IF you don't have the quality of maintenance performance data to start with, you are likely to wind up picking "generic" failure and/or repair codes that ultimately won't always apply to your specific asset sets. Although I think you're right in a general sense that "generic" equipment have some common failure modes...BUT, not everything you do at your plant is generic, I would bet. I further suggest that you likely have very specific failures unique to your equipment/process/operations environment. So, if true, this is an argument for you to "do your homework" and don't bank on getting failure modes from "generic" sources.

There is an ISO standard for the Oil & Gas Industry that has suggested codes for repair, failure mode, other codes used in that industry: ISO 14224. That might help you a bit on equipment that you have matching or nearly matching the assets described in there, such as pumps, heat exchangers, valves, etc. I assume you don't work in the Oil & Gas industry.

I happen to work for a reliability consulting firm now that has compiled a huge set of databases (over 3 million records) that include, among other sets of reliability-related data, failure modes and codes for industry equipment across basically all major industries. I don't think they'll release that data without a purchase order, however. Lemme know if you're interested.

Unfortunately I have found in my own experience (prior to current employer) that the best set of failure codes you can develop for your asset sets and inserted into your CMMS is the one you build, one step at a time, from a tedious review of a good representative set of the WO history and resident braintrust. From either the CMMS closed work orders module ('WO History') or, if you're still in The Dark Ages with paper-based Work Orders (nothing personal, dude!) you will have to sift through those puppies. For either the data-mining of the CMMS or the paper shuffle approach, I recommend a few techniques to make it a more manageable task:

* If they aren't already so categorized, separate your WO's by equipment Family, e.g. all Pumps, Centrifuges, Chillers, etc. together. Because you want to determine failure modes that apply to familial equipment. For unique one-of-a-kind equipment, you'll have to develop failure modes applicable to just that asset.

* If you haven't established equipment Criticality Rating yet for your assets, try to do so in terms of cost impact of downtime; cost to repair (spare parts, labor etc.); availability constraints (only one exists, etc.). Once you define Criticality Rating CR for all equipment/asset (and once again, get that precious info into your CMMS), you can then separate your WO into sets prioritized by the CR. THEN you focus on just the top priority CR equipment and develop THOSE failure codes first, on down to next highest CR, etc. (sidebar note...It goes without saying that every action to improve reliability, reduce costs, enhance uptime, etc. etc. should be DRIVEN by a proper CR analysis!)

* If you're sorting through the WO History or closed paper WO's, make sure you get a "representative sample" (the statisticians hate that term) going back in time far enough to capture the nominal events of your plant operations cycles. By that I mean, seasonal ups and downs; product changes, re-tooling; annual PM's and PdM's are represented. Hint: When I did the paper shuffle for a former Dark Ager a few years ago to define Repair, Failur Mode codes etc., I went back THREE YEARS...about twelve thousand WO's for those guys...

* Corner your most knowledgeable Maint. Techs all crafts mech/elect in a friendly way, just interview them informally "Hey Jerry, on those cooling towers, what are the repairs you're doing, and how often?" or "Bill, if I could remove the equipment failures around here that bug you the most, which would they be?" Jerry and Bill's answers aren't necessarily the same equipment as the highest Criticality Rating set, but you get a chance to understand the failures that those guys deal with the most, you'll learn new failure modes, guarantee it. And I would then argue that you do failure codes on these right away if possible as they will be recurring. Plus you win credibility with Jerry and Bill.

Anyway once you (hopefully have a CMMS and) insert those failure mode codes in, you associate them with the equipment Family or, if a one-of-a-kind equipment, with that one. Then when closing out a WO, you must have the person doing the WO closing select the Failure Code, the Repair Code, Labor Time to Repair, all that stuff.

THEN you can have the basis to do filter inquiries of the CMMS and get Pareto charts of the Failure Modes you worked so hard to build. Then, you can really start making hay.

Hey I drank too much coffee for this one...But, hope you get something out of it.


Also, Pressed, thanks for your links. Good shtuff.

 

[unclesyd]

There are a few important points that I would like to add to the excellent responses above. The mention by jeez of contact with the mechanics and technicians is essential. You need to establish a rapport with all and visit the maintenance shops frequently and listen to their problems, along with there trials and tribulations.

While in the shops tour the waste bins look for repeat parts failure. This is very necessary in that there is a propensity to quantify things as routine if the fail all the time.

If you have a central stores for maintenance items make sure that the inventory is maintained. A stock out bearing will cause a questionable bearing to be reused and items such as gaskets and Orings to be substituted. The mechanics and technicians will let you know as most don't like to reuse questionable parts.

If you are charged with and given the time make sure up get in on a failure or outage from the get-go. Second hand information isn't the most reliable. You will learn a lot of critical information that will not get recorded or transmitted.

Get yourself a good digital camera.

 

[SRC3]

The System Reliability Center Offers, for a fee, a database of Failure modes and mechanisms for thousands of different part types.
REfer to

http://src.alionscience.com/spidr/

 

[rich321]

FMD-97 contains the information that you want. It is based on data collected by the Reliability Analysis Center. It is available from both SRC and RIAC, although pricing may vary. It gives you the failure mode distributions for a wide range of part types. This link will take you to information on FMD-97:

http://www.fmea-fmeca.com/other-fmea-resources.html

Good luck!