Demonstration Test Plan for Reliability of Big Equipment

AgaWr

Registered
Hi, I've found information on this forum very valuable. It helps me finding answers to topics I've been struggling for a while.

I have concern similar to bcoolnow. (see this thread starting at post #45) I want to perform the demonstration test for reliability but I am no longer sure that what I am thinking is wright. There is a component I want to test for reliability. Its working time is 10 years and I need to perform a test for a 90% reliability at 30 weeks time. A confidence level required is at least 90% and I don't want to have any failures. The sample size I am getting are high and I don't think it's feasible for my test. Well, maybe it is feasible and I understand the need for it but if I have time limitations and have only few samples available it looks like impossible to complete! :) Is there a rule for big machinery components like motors, pumps, etc...?
Thank you very much for any help.
 
Last edited by a moderator:

Miner

Forum Moderator
Leader
Admin
Re: Confidence & Reliability - Reference to a 90/95 Confidence & Reliability Level

Do you have any historical data regarding the shape parameter for these components? You will need this for your demonstration test plan.

Normally the test plan is a balance of samples and test time. If you are constrained on sample size, you test longer. If you are constrained on time, you test more samples. If you are constrained on both time and samples, you punt :). You can try an accelerated demonstration test plan. This requires more assumptions about the acceleration model. Some typical acceleration factors are duty cycle, temperature, load applied, etc.
 

AgaWr

Registered
Re: Confidence & Reliability - Reference to a 90/95 Confidence & Reliability Level

Thank you for such a quick reply!

Yes, I believe I am struggling with the balance. I don't have historical data for the shape parameter but I chose 1 as I want to simulate the time up to the wear point when the curve starts to increase.
But I think it may be a good option to perform the accelerated demonstration test plan as you suggested. The machine is being used in time intervals and the service life is planned for 10 years. I want to perform the accelerated test for reliability and I have 26 weeks for that. I plan to apply an increased load (60%) as a stress factor. Do you think it can work? Are there other factors I should consider?
 

Miner

Forum Moderator
Leader
Admin
Re: Confidence & Reliability - Reference to a 90/95 Confidence & Reliability Level

What type of failure modes do you experience? A shape factor of 1 implies 100% random failures due to overload stress. If your product fails due to any type of wearout, you should use a shape factor > 1.

Load is definitely one means of accelerating failures. Temperature is another. If your product is not used 100% of the time, you can accelerate by increasing the duty cycle. You can also combine 2 or more means of acceleration. You will also need to define the accelerating relationship for each. For example, duty cycle is usually linear while temperature is often Arrhenius.
 

AgaWr

Registered
Re: Confidence & Reliability - Reference to a 90/95 Confidence & Reliability Level

Hi,
This is a new product, however I can expect that potential failure modes would be diaphram deformation, motor failure, bearing failure.They are in order to what is expected to fail first.
I will definitely use shape factor >1 as the probability of wear increase over the time.

I have few option how to accelerate the test:

1. Duty cycle,
2. Load (motor speed),
3. Temperature (result of increase first two)

Can you advise what are the rules to define the acceleration relationship for various accelerations?

Thanks!!
 

Miner

Forum Moderator
Leader
Admin
I split these posts off into a separate thread. The two intertwined topics were getting confusing.

Duty cycle is usually a linear relationship. That is if normal usage is 2 hours per day, and you run it 12 hours a day, your acceleration factor is 6 (12/2 = 6). However, you must be careful if your product is intermittent duty instead of continuous duty. I doubt this is the case in your scenario.

Load is product dependent. Two likely models are Arrhenius and Power models. The former is temperature driven, the latter is voltage, current, speed, or power driven. Motor insulation is often Arrhenius. Bearing failures are often power models.

Increasing temperature is an Arrhenius relationship.

What is the diaphragm and how is it deformed?
 
Top Bottom