Tek-Tips is the largest IT community on the Internet today!
Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!
-
Congratulations cowski on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!
PLC Failure Probability 1
- Thread starter mahsaman
- Start date
Most PLC manufacturer's have MTBF information either in studies or usually via calculations. These typically are not done with a specific industry targeted but do usually include environmental information. One could likely calculate a 'failure probability' based on the MTBF data.
I do know that GE-IP has a lot of hardware in oil & gas industries in both production and transportation of each. I suggest that you contact them and ask.
I do know that GE-IP has a lot of hardware in oil & gas industries in both production and transportation of each. I suggest that you contact them and ask.
- Thread starter
- #3
dV8r thanks for your response. I found MTBF as one of the factors. What I mean by this study is to consider all the environmental and hardware factors and give them a probability factor. For example what is the probability of this fact that "dust" causes a PLC to fail.
Thanks,
Thanks,
Dust, per se, should not be a failure factor, UNLESS, there's so much of it that the PLC can no longer maintain a safe operating temperature. MIL-HDBK-217 identifies the typical electronics failure mechanisms
TTFN
faq731-376
Need help writing a question or understanding a reply? forum1529
Of course I can. I can do anything. I can do absolutely anything. I'm an expert!
There is a homework forum hosted by engineering.com:
TTFN
faq731-376
Need help writing a question or understanding a reply? forum1529
Of course I can. I can do anything. I can do absolutely anything. I'm an expert!
There is a homework forum hosted by engineering.com:
- Thread starter
- #5
The main factors are temperature, vibration, time, and voltage. These are all used to some degree to accelerate electronics failures; whenever you see Highly Accelerated Life Test (HALT) or Highly Accelerated Stress Test (HAST), that's what's being manipulated.
Actually, in 217, all the factors but voltage are listed. Section 3.4.3 Table 3.2 lists a series of environments that encompass certain vibrations and temperatures. Time is inherent in the failure rate λp since that's what goes into the exponential failure rate equation. The last page in section 5.6 lists the acceleration factors for the different environments.
TTFN
faq731-376
Need help writing a question or understanding a reply? forum1529
Of course I can. I can do anything. I can do absolutely anything. I'm an expert!
There is a homework forum hosted by engineering.com:
Actually, in 217, all the factors but voltage are listed. Section 3.4.3 Table 3.2 lists a series of environments that encompass certain vibrations and temperatures. Time is inherent in the failure rate λp since that's what goes into the exponential failure rate equation. The last page in section 5.6 lists the acceleration factors for the different environments.
TTFN
faq731-376
Need help writing a question or understanding a reply? forum1529
Of course I can. I can do anything. I can do absolutely anything. I'm an expert!
There is a homework forum hosted by engineering.com:
IMHO "probability" can only be calculated for a given set of criteria i.e.
"What is the probability that the system will get destroyed by lightening on July 4, 2021?"
What are the facts that need to be known to calculate this, the more facts one has the more accurate one can be?
Where is it located? Central Florida or Death Valley?
How is it protected from surviving a lightening strike or near by strike?
etc.
YMMV
"What is the probability that the system will get destroyed by lightening on July 4, 2021?"
What are the facts that need to be known to calculate this, the more facts one has the more accurate one can be?
Where is it located? Central Florida or Death Valley?
How is it protected from surviving a lightening strike or near by strike?
etc.
YMMV
I'd take the semantics a step further than dV8r in that the "probability" of failure is 100%, eventually. That's why we use MTBF, because that failure IS eminent, you just want to understand your risks as to WHEN it becomes more likely.
That's why when I get asked this question, I like to also bring in the concept of MTTR, Mean Time To Repair. MTBF numbers are important, but WHEN your PLC fails, if the manufacturer is gone or has moved on to a new model and no longer supports yours, the MTTR goes out the roof!
"You measure the size of the accomplishment by the obstacles you had to overcome to reach your goals" -- Booker T. Washington
That's why when I get asked this question, I like to also bring in the concept of MTTR, Mean Time To Repair. MTBF numbers are important, but WHEN your PLC fails, if the manufacturer is gone or has moved on to a new model and no longer supports yours, the MTTR goes out the roof!
"You measure the size of the accomplishment by the obstacles you had to overcome to reach your goals" -- Booker T. Washington
In a proper setting, replacements in the face of discontinuation or obsolescence of components is planned for through life cycle analysis. Traditional military programs from the pre-1990s took those sorts of things into account. The US Army's Mast Mounted Sight (MMS) (NOT Site!) was planned and budgeted for a standard service life, but the logistics infrastructure and funding were planned so well that we are able to keep systems operationally functional and available nearly 20 years after the first time the program was supposed to be terminated because of imminent production of the Commanche helicopter. MMS budgeting allowed for multiple obsolescence reduction campaigns during the entire time the Commanche was delayed until it was canceled altogether. MMS was still chugging along 10 years after that.
Telecom companies in the late 1980s paid premium dollar to maintain production of 1970s-era parts that should have gone obsolete. We were outsourcing 10-micron metal gate PMOS parts when 2-micron CMOS parts were being mass produced, because the original PMOS parts were thoroughly characterized and qualified, and the re-qualification of replacement parts was more expensive than the premiums paid for continued manufacturing of the obsolete parts. One method of maintaining adequate repair inventory is to make "lifetime" buys, i.e., you buy additional inventory in the present to cover future attrition and failure. Of course, that requires you to have extremely good reliability analysis and data.
TTFN
faq731-376
Need help writing a question or understanding a reply? forum1529
Of course I can. I can do anything. I can do absolutely anything. I'm an expert!
There is a homework forum hosted by engineering.com:
Telecom companies in the late 1980s paid premium dollar to maintain production of 1970s-era parts that should have gone obsolete. We were outsourcing 10-micron metal gate PMOS parts when 2-micron CMOS parts were being mass produced, because the original PMOS parts were thoroughly characterized and qualified, and the re-qualification of replacement parts was more expensive than the premiums paid for continued manufacturing of the obsolete parts. One method of maintaining adequate repair inventory is to make "lifetime" buys, i.e., you buy additional inventory in the present to cover future attrition and failure. Of course, that requires you to have extremely good reliability analysis and data.
TTFN
faq731-376
Need help writing a question or understanding a reply? forum1529
Of course I can. I can do anything. I can do absolutely anything. I'm an expert!
There is a homework forum hosted by engineering.com:
- Thread starter
- #10
No, because it has to do with both intensity and duration. Read 217 and study the environmental acceleration factors
TTFN
faq731-376
Need help writing a question or understanding a reply? forum1529
Of course I can. I can do anything. I can do absolutely anything. I'm an expert!
There is a homework forum hosted by engineering.com:
TTFN
faq731-376
Need help writing a question or understanding a reply? forum1529
Of course I can. I can do anything. I can do absolutely anything. I'm an expert!
There is a homework forum hosted by engineering.com:
Skogsgurra
Electrical
I would say that, in general, the answer to "Can anybody please list me the factors for PLC failure and the rankings from most harmful factor to less?" does have a simple answer.
I have met many failures in analogue and digital systems and, depending on what kind of system it is, you usually have:
1. Connectors and cables.
2. Vibration.
3. Temperature cycling and condensation.
4. ESD.
5. EPROMs that lose their contents.
6. Dust,Corrosive fumes, flooding.
7. Theft.
And, very often, documentation and know-how that has disapeared
The order may vary, but this is what I have seen in my active life in industrial automation and drives for more than 50 years.
Gunnar Englund
--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.
I have met many failures in analogue and digital systems and, depending on what kind of system it is, you usually have:
1. Connectors and cables.
2. Vibration.
3. Temperature cycling and condensation.
4. ESD.
5. EPROMs that lose their contents.
6. Dust,Corrosive fumes, flooding.
7. Theft.
And, very often, documentation and know-how that has disapeared
The order may vary, but this is what I have seen in my active life in industrial automation and drives for more than 50 years.
Gunnar Englund
--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.
I've seen more I/O module failures than processor failures. I've seen very few I/O module failures but conditions and installation were very controlled. Most of Skogsgurra's items, to me, are not inherent to the equipment but the installation.
Pamela K. Quillin, P.E.
Quillin Engineering, LLC
Pamela K. Quillin, P.E.
Quillin Engineering, LLC
- Thread starter
- #14
Connectors and cables generally fail only because of handling or excessive vibration. There are few inherent failure mechanisms for cables; they're basically wires and packaging, and given proper design and margins, there's no reason for them failing, which, is consistent with all reliability prediction models, which given them VERY high MTBF. That's not to say that cables can't fail, but those failures are not reliability failures in the strict sense, but more akin to lacajun's assertion that it's design, fabrication, installation, or handling. Cables chafing until insulation is damaged is a customer perceived reliability problem, but there's nothing inherent in the cable that could have used to anticipate such a failure.
Can one rate vibration higher than temperature without knowing some actual values? Temperature impact on electronics is typically characterized by a doubling of the failure rate for a mere 3 deg C increase in temperature. In fact, accelerated life testing for integrated circuits was typically done solely with temperature, reducing 20 years of operating life to 2000 hrs simply by running the parts at high temperatures. MIL-HDBK-217's acceleration factors only go to about 20x for for a combination of temperature and vibration, while for discrete components, it has factors up to 40, driven mostly by temperature.
TTFN
faq731-376
Need help writing a question or understanding a reply? forum1529
Of course I can. I can do anything. I can do absolutely anything. I'm an expert!
There is a homework forum hosted by engineering.com:
Can one rate vibration higher than temperature without knowing some actual values? Temperature impact on electronics is typically characterized by a doubling of the failure rate for a mere 3 deg C increase in temperature. In fact, accelerated life testing for integrated circuits was typically done solely with temperature, reducing 20 years of operating life to 2000 hrs simply by running the parts at high temperatures. MIL-HDBK-217's acceleration factors only go to about 20x for for a combination of temperature and vibration, while for discrete components, it has factors up to 40, driven mostly by temperature.
TTFN
faq731-376
Need help writing a question or understanding a reply? forum1529
Of course I can. I can do anything. I can do absolutely anything. I'm an expert!
There is a homework forum hosted by engineering.com:
8. RAM or Real Time Clock battery failure following repeated ignoring of warning messages to that effect.
9. PPP (Pi55 Poor Planning) of the design and use considerations. I'm guilty of that myself; had one where the user ignored 48 flashing red pilot lights lights meant to warn them of a temporary manual override. He ran it that way for 6 months with all 48 red lights flashing once/sec. (he put on sunglasses, I'm not kidding). I had used relay outputs in the I/O cards, it wore them out and the contacts welded.
"You measure the size of the accomplishment by the obstacles you had to overcome to reach your goals" -- Booker T. Washington
9. PPP (Pi55 Poor Planning) of the design and use considerations. I'm guilty of that myself; had one where the user ignored 48 flashing red pilot lights lights meant to warn them of a temporary manual override. He ran it that way for 6 months with all 48 red lights flashing once/sec. (he put on sunglasses, I'm not kidding). I had used relay outputs in the I/O cards, it wore them out and the contacts welded.
"You measure the size of the accomplishment by the obstacles you had to overcome to reach your goals" -- Booker T. Washington
Failure probability is categorized like this (in order of most likely to least likely.
[ul]
[li]Mechanical[/li]
[li]Electro-Mechanical[/li]
[li]High Current[/li]
[li]Low Current[/li]
[li]Solid State[/li]
[/ul]
As you eluded to, environment can be an exception to the general rule.
(Relay card most likely, and MTBF is useless unless solid state, as frequency PLC program activates, dictates failure rate as well as environment.)
The above probability guide translate to PLCs as ...
[ul]
[li]Output Cards[/li]
[li]Input Cards[/li]
[li]Power Supply[/li]
[li]Backplane[/li]
[li]CPU (processor Card)[/li]
[/ul]
Environment can change basic order above. a couple examples, in a foundry, the metal dust causes backplane to fail more frequently. In the machining industry, penetrating fluid getting into sensors causes input cards to fail more than output cards because the liquid causes input card to sink more current than designed to. In wire manufacturing where a lot of high voltage insulation testing is going on, CPUs fail more often because control voltage for PLC does not have line filter in line with it. etc. etc. Other variables that effect the failure rate are improper design, alight not allowing the 20% extra when calculating relay card or power supply current, etc etc.
Don
BIN Industrial & PLC Training
[ul]
[li]Mechanical[/li]
[li]Electro-Mechanical[/li]
[li]High Current[/li]
[li]Low Current[/li]
[li]Solid State[/li]
[/ul]
As you eluded to, environment can be an exception to the general rule.
(Relay card most likely, and MTBF is useless unless solid state, as frequency PLC program activates, dictates failure rate as well as environment.)
The above probability guide translate to PLCs as ...
[ul]
[li]Output Cards[/li]
[li]Input Cards[/li]
[li]Power Supply[/li]
[li]Backplane[/li]
[li]CPU (processor Card)[/li]
[/ul]
Environment can change basic order above. a couple examples, in a foundry, the metal dust causes backplane to fail more frequently. In the machining industry, penetrating fluid getting into sensors causes input cards to fail more than output cards because the liquid causes input card to sink more current than designed to. In wire manufacturing where a lot of high voltage insulation testing is going on, CPUs fail more often because control voltage for PLC does not have line filter in line with it. etc. etc. Other variables that effect the failure rate are improper design, alight not allowing the 20% extra when calculating relay card or power supply current, etc etc.
Don
BIN Industrial & PLC Training
-
1
- #18
controlsdude
Electrical
MTBF = Its really how long before the controls design goes obselete, which is usually 20-25 years. Hardly ever see PLC fault on hardware as long as the cabinet is not full of dust or corrisive fumes.
Or could be the time that company goes bankrupt, ships all the equipment overseas, or whatever corporate decision is made to make it obselete due to different process or new product being made.
Or could be the time that company goes bankrupt, ships all the equipment overseas, or whatever corporate decision is made to make it obselete due to different process or new product being made.
- Status
Similar threads
- Locked
- Question
- Locked
- Question
- Locked
- Question