Page 1 2 3 
Go
New
Find
Notify
Tools
Reply
  
The difference between reliability and availability Login/Join
 
posted
What is the difference between reliability and availability?

Terry O
 
Posts: 989 | Location: Southwest Florida Gulf | Registered: 03 April 2004Reply With QuoteReport This Post
posted Hide Post
 
Posts: 4285 | Location: Borneo | Registered: 13 February 2005Reply With QuoteReport This Post
posted Hide Post
Thanks Josh

That article was the inspiration for my question.

What do you think of the points made in the article?

Terry O
 
Posts: 989 | Location: Southwest Florida Gulf | Registered: 03 April 2004Reply With QuoteReport This Post
posted Hide Post
If it's reliable, it's available. However, if it's available it's not necessarily reliable.

Reliability is a state of knowing where the machine is along its failure curve. A spare can be available but without knowing its reliability it may not be available very long. However, if the spare gets you through the 'out of service equipments' repair then the pumping system is available; therefore, the process is available. Although this presents available, the MTBF will only be extended through a reliability program or so I think.


Cordially,
Sam Pickens
pdmsampickens@gmail.com

 
Posts: 1872 | Location: Eastern USA | Registered: 04 August 2004Reply With QuoteReport This Post
posted Hide Post
Although RAM is important, I guess not many do RAM modelling and simulation.
 
Posts: 4285 | Location: Borneo | Registered: 13 February 2005Reply With QuoteReport This Post
posted Hide Post
Terry,
PaulB is a well respected authority in the field.

quote:
That article was the inspiration for my question.

What do you think of the points made in the article?


I am not sure if the majority of us need the maths lesson to understand, or even if it will actually help understand the terms.
Availability is simply the time an item is able to operate (operating time+standby time) as a ratio to the time in service. It tells us how much the cash machine can yield if we feed it with the required inputs e.g. raw material, power etc. and remove the product as it is made.
Reliability is the probability that at any point in time, it will operate correctly for a further specified length of time. For example, a pump has been running well for 3 months; what is the probability that it will run for 6 more months, 12 more months etc.
The two are related; availability depends on reliability.

PaulB says (incorrectly) that the two are never the same. He ignores a class of items called non-repairable, which are replaced with a new (or as good as new or AGAN) item when a failure takes place. Think of light bulbs, ball bearings, smoke/gas/fire detectors, emergency shutdown valves etc., and you will recognize what I mean. In all these cases, once we know it has failed, we will simply replace it with a new or AGAN item. In this specific case, the reliability and availability are numerically the same.

But for the majority of items, which are repairable, his comment is perfectly true.

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


Regards,
V.Narayan (Vee)
Lead Author, Case Studies in Maintenance and Reliability: A Wealth of Best Practices, 2012, Industrial Press.NY ISBN-13: 978-0831102210
Author, Effective Maintenance Management: Risk and Reliability Strategies for Optimizing Performance, Second Edition, 2011, Industrial Press NY ISBN-13: 978-0-8311-3444-0
 
Posts: 1374 | Location: Scotland, UK. | Registered: 16 May 2004Reply With QuoteReport This Post
posted Hide Post
Terry

Here's my 2 cents on it.

Reliability is all to do with the probability that the system will function correctly and is usually measured by the MTBF.

Availability is the percentage of time that the system is capable of functioning correctly. This is usually measured by calculating uptime divided by the total hours for a given period (i.e. 8760 hours in a year)

You can always dig into the data to differentiate between planned and unplanned downtime to get a clearer picture of causes. When MTBF and availability are reviewed together it can give a better analysis of what is going on. For example, you could have a low MTBF but if the downtime is only a few minutes each time the failure occurs then the your availability could still be be very high.

If the system was switched off during a shutdown to allow maintenance or repair etc of other equipment / systems and is successfully started up again when the shutdown is over was it available? I think so but interested to see if others agree.
 
Posts: 16 | Location: Midwest US | Registered: 13 April 2004Reply With QuoteReport This Post
posted Hide Post
It's easy to forget the intent is to monitor the effectiveness of the maintenance effort. If you're lucky enough to have the data readily available, both availability and reliability can be tracked over time as a measure maintenance value. If the two do not track together, the relevant factor will be downtime. As an example, suppose maintenance becomes more proficient at repairing a recurrent failure, reducing downtime. Availability will improve but reliability will not be improved. Certainly, there is value to the improved maintenance efforts even though the root cause of the failures has not been corrected, and the reliability remains unchanged.
 
Posts: 6 | Location: Missouri | Registered: 12 October 2006Reply With QuoteReport This Post
<Rui Assis>
posted
Availability is the probability of having a piece of equipment available to work at any moment within a certain time span (long enough to comprise several episodes of failure and repair).

Reliability is the probability of a piece of equipment not failing until a certain moment in time (mission).

Availability depends on Reliability and Maintenability (time to recover following a failure) but the opposite is not true. On the other hand, Reliability and Maintenability are independent of each other. You can have a machine with high availability although its reliability is low if it takes little time to repair or still you can have a machine with low availability despite its reliability is high in case the times to repair are long.

The attached spreadsheet is a Monte Carlo simulation model that allows you to simulate 1,000 episodes of failures and repairs (moments in columns E and F) of a critical component pertaining to a piece of equipment. The renewal of components is supposed to be a homogeneous Poisson process, that is, the failure behaviour of each component is exactly the same as the one that preceded it. The failure behaviour is according to a Weibull distribution (the three parameters are in cells C4 to C6) and the time to repair follows a LogNormal distribution (the two parameters are in cells E4 and E5). The mission is entered in cell C9. In cells O4 through X103, 1,000 moments of need of the equipment are generated at random between 0 and the latest simulated event (cell F2013).

After each iteration (by pressing the key function F9), you can observe:

 The reliability that results in cell H10. If you compare this to the theoretical value in cell H8, you notice that they are quite close – the former tends to the later as the number of iterations increase or the number of repetitions increase by calculating the expected value of cell H10;
 The availability (as a probability) that results in cell J10. If you compare this to the theoretical value in cell J8, you notice that they are quite close – the former tends to the later as the number of iterations increase or the number of repetitions increase by calculating the expected value of cell J10;

In cell J11 you have the availability that also resulted from the simulation but as a natural frequency instead.

In conclusion, the Monte-Carlo simulation model presented here proves my definitions of Reliability and Availability described above.

Regards,

Rui

This message has been edited. Last edited by: <Rui Assis>,

Excel SpreadsheetReliability_Availability.xls (460 Kb, 110 downloads)
 
Reply With QuoteReport This Post
posted Hide Post
Let me just rephrase your question

1st Question : If the equipment's Availability is 100%, is the equipment said to be at its peak & reliable ?

Answer : NO, The equipment may be 100% AVAILABLE but have a low UTILIZATION or have not been utilized at all, hence availability is
not a measure of reliability but only when the machine is used. You drive your car to work and then you park it in your building, your car is 100% available but it was only utilize for a couple of hours, 1 hour going to your work and 1 hour going home

If the equipment's Availability and Utilization is 100%, is the equipment said to be at its peak and reliable ?

Answer : NO, The equipment may be 100% Utilized and Available but machine may suffer from errors and speed loss. Example a machine
speed need to be reduce from 1500 uph to 1200 uph so as not to deliver problems in quality.

In laymans terms, you might be 100% utilizing your car and using it 24 hours of driving, but you cannot speed up since if you do then you will hear a lot of unwanted sound in your engine, transmission, tie rod bearing etc and you know soon that it will break down if you continue to travel at the speed you want.

If the equipment's OEE is 90 % or more, is the equipment said to be at its peak & reliable ?

Answer : NOT EXACTLY, the real objective of OEE is not about achieving 85% or more but about satisfying the 3 components, in the case below
the equipment is suffering from a large rate of defects.

OEE = 100% (Utilization) x 97% (Efficiency) x 93% (Yield) = 90.21%

In this case the machine is failing

Hope this helps,

Rolly Angeles


Rolly Angeles
Teacher
www.rsareliability.com
 
Posts: 330 | Location: Philippines | Registered: 09 December 2005Reply With QuoteReport This Post
posted Hide Post
Vee,

You quote :

Availability is simply the time an item is able to operate (operating time+standby time) as a ratio to the time in service.

I think you are speaking about utilization and not availability, these 2 are different in my own perspective and from what I am teaching.

The machine may have been available but not utilized. But the machine can only be utilized if its available.

Rolly Angeles


Rolly Angeles
Teacher
www.rsareliability.com
 
Posts: 330 | Location: Philippines | Registered: 09 December 2005Reply With QuoteReport This Post
posted Hide Post
Josh,

In your article, it states MTBF or MTTF are these two the same or they are different ?

Appreciate any feedback from this,

In my opinion and understanding these two Mean indicators are entirely different but related in a way with each other, I'll explain later.

Rolly Angeles


Rolly Angeles
Teacher
www.rsareliability.com
 
Posts: 330 | Location: Philippines | Registered: 09 December 2005Reply With QuoteReport This Post
posted Hide Post
Rolly,
I am not sure I follow your query, as I think we are saying exactly the same thing. You say,
quote:
You drive your car to work and then you park it in your building, your car is 100% available but it was only utilize for a couple of hours, 1 hour going to your work and 1 hour going home

How does this differ from
Availability is simply the time an item is able to operate (operating time+standby time) as a ratio to the time in service
As you correctly point out, Utilization is the ratio of operating time (read - driving for 2 hours) to total time, and availability is as I defined and you illustrated (read - driving + parked time).
As far as MTTF is concerned, it is only applicable when the reoair is to As Good As New (AGAN) condition. It is always applicable to non-repairable items, and for those cases of repairable items done to AGAN standards. Some people (incorrectly) bring in the issue of repair time, which is NOT the criterion to use. The point is that AGAN repairs bring the reliability back to 100% at the start of the new run, while other repairs do not bring it to 100%, so the item starts off its new run with a lower reliability.
MTBF, on the other hand is applicable to repairable items (I gave some examples in my earlier post).
The formula used for computing MTTF and MTBF is the same, but they are applicable in different situations.


Regards,
V.Narayan (Vee)
Lead Author, Case Studies in Maintenance and Reliability: A Wealth of Best Practices, 2012, Industrial Press.NY ISBN-13: 978-0831102210
Author, Effective Maintenance Management: Risk and Reliability Strategies for Optimizing Performance, Second Edition, 2011, Industrial Press NY ISBN-13: 978-0-8311-3444-0
 
Posts: 1374 | Location: Scotland, UK. | Registered: 16 May 2004Reply With QuoteReport This Post
posted Hide Post
Vee,

I know your good but I got u this time, MTBF and MTTF are not the same, I'll explain later. The answer is in your reply. These 2 are different in my own understanding and humble opinion.

This message has been edited. Last edited by: Rolly12,


Rolly Angeles
Teacher
www.rsareliability.com
 
Posts: 330 | Location: Philippines | Registered: 09 December 2005Reply With QuoteReport This Post
<Rui Assis>
posted
I thought availability and utilization were two concepts long mastered by us all but after reading these last few posts, I myself became a little bit confused. Would you please point me out any desagreement in the following example:

One piece of equipment is available to work for typically 40 hours per week, that is, its capacity is 40 hours/week. Suppose two situations:

1. At one particular week it was loaded with 36 standard hours. During the course of the week the equipment failed while it was working and it took 2 hours to repair. Because (40 – 36) > 2, the work could all be accomplished and it took 36 standard hours or 37.5 actual hours (efficiency = 37.5/36 = 96%). In this instance, one may say that the availability was (40 – 2)/40 x 100 = 95% and the utilization was 37.5/40 x 100 = 93.75%;
2. At another week the equipment was loaded with 38 standard hours. During the course of the week the equipment failed while it was working and it took 8 hours to repair. Because (40 – 38) < 8, the work cannot all be accomplished and it took (38 – 8) + (40 – 38) = 32 standard hours or 30.476 actual hours (efficiency = 30.476/32 = 105%). In this instance, one may say that the availability was (40 – 8)/40 x 100 = 80% and the utilization was 30.476/40 x 100 = 76.19%.

Do you agree?

Regards,

Rui
 
Reply With QuoteReport This Post
<Rui Assis>
posted
MTTF means Mean Time To Failure or "mean or average life" of a component regardless it is replaced or not immediately after a functional failure.

MTBF means Mean Time Between Failures and is the same as "the average time between two successive failures", therefore comprising the time to repair every failure, that is MTBF = MTTF + MTTR (Mean Time To Recover, Replace or Repair).

Suppose 3 failures of a component in a row (failure 1, failure 2 and failure 3) extending over a time span where event 0 is the origin of time and T is today. At moment 0 the component was working properly, at moment 1 failed and it was replaced by a new one, the same happened at moments 2 and 3. Today (moment T) the component is still working fine.

Call 1´, 2´ and 3´ the moments when the replaced components started again their work. Suppose that events (failures) 1, 2 and 3 took 1, 2 and 3 calendar hours to recover, that is, moments 1 and 1´ are 1 hour distant in time, 2 and 2´, 2 hours and 3 and 3´, 3 hours. Suppose furthermore that events 0 and 1 are 500 hours distant in terms of running time, 1´ and 2 are 700 hours and 2´ and 3, 600 hours.

In this instance, and from a management point of view, MTTF is equal to (500 + 700 + 600)/3 = 600 hours, MTTR = (1 + 2 + 3)/3 = 2 hours and, finally, MTBF = 600 + 2 = 602 hours. The most important issue here is to track how this indicator is behaving over the time (increasing, which is good, or decreasing).

Please note that, from an engineering perspective, the first interval (500 hours) and the last one (600 hours) should be censored.

If the system is not repairable, it doesn't make sense to address MTBF, of course.

In order to avoid confusion, I always refer a system as "being repairable in service" (the most frequent circumstance) or "not being repairable in service" (the case of a missile or a satellite after having been launched or still a parallel (redundant) arrangement of batteries located in a remote place where you go only from time to time on a regular basis). The missile and the satellite are "repairable" while they stay at the stores but they are not as soon as they are "in service". The batteries are not repairable in service, that is, while the maintenance person in charge is away, if one unit fails (the remnant batteries split the load), it will keep this way until somebody comes, notices and corrects the failed situation.

Regards,

Rui

This message has been edited. Last edited by: <Rui Assis>,
 
Reply With QuoteReport This Post
posted Hide Post
Rui, in a plant environment, I think instruments are normally not repairable so replacement is a common policy eg maybe transmitters, thermocouples, pressure gauges, sensors, etc. However, worn out trims of control valves of punctured diafragam of their actuators can be replaced.

This message has been edited. Last edited by: Josh,
 
Posts: 4285 | Location: Borneo | Registered: 13 February 2005Reply With QuoteReport This Post
posted Hide Post
Rolly, If I remember correctly from what I read, in most cases, MTBF is different from MTTF but there is a special case where MTBF is approximately equal to MTTF. Maybe Vee is pointing out this special case where MTBF~MTTF.
 
Posts: 4285 | Location: Borneo | Registered: 13 February 2005Reply With QuoteReport This Post
posted Hide Post
I think there is an IMechE publication addressing this subject of repairable and non-repairable items quite in detail. I don't have the tittle offhands now.

This message has been edited. Last edited by: Josh,
 
Posts: 4285 | Location: Borneo | Registered: 13 February 2005Reply With QuoteReport This Post
posted Hide Post
I saw Paul Barringer has registered but not commenting?
 
Posts: 4285 | Location: Borneo | Registered: 13 February 2005Reply With QuoteReport This Post
  Powered by Social Strata Page 1 2 3  
 


Copyright © 2004-2013 Reliabilityweb.com All rights reserved.