Case studies where Monte Carlo Simulation applied to Maintenance, uptime prediction

[Ravi Khare]

Does anyone have case studies where Monte Carlo Simulation is applied to Maintenance, uptime prediction and System Reliability? I would appreciate if there are any pointers to measured benefits at the bottom line.

 

[Ravi Khare]

No replies?

Looks like not too many share my interest in Monte Carlo simulation. I am afraid this thread is simply going to scroll down in obliviance without exploring the power of Monte Carlo Simulation.

Let me now invite you to discuss other applications and cases of Monte Carlo Simulation... not just maintenance this time.

Thanx

 

[David Mullins]

Is that like the China Syndrome?

Seriously, I'm assuming this is an alternative to risk assessment - the probability / likelihood of something haooening? Only risk assessment also looks at severity, remediation and reassessement

 

[Kevin Mader]

For an interesting application of the Monte Carlo Simulation, you might browse the meteor strike pattern on the planet Mercury. It is about the only application of Monte Carlo I know of. Sorry...

Kevin

 

[Ravi Khare]

Statistical Tolerancing is one very important application I know of Monte Carlo Simulation. You could relax tolerances on critical proceses without affecting Quality. You could also evaluate compensating tolerances to reduce overall costs.

Another application is the one I had written about initially. If you were able to predict the types of failures in equipments based on component reliability data, your spare stocking policies would yield more uptime, and at a lower cost too. The more you think about it, the potential appears to be enormous.

Any thoughts?? Or is it only my flight of fancy?

 

[D.Scott]

I don't think it is a flight of fancy. From what I've learned, the Monte Carlo Simulation is a powerful tool when you use it right. Trouble is, I don't know enough about it to actually discuss it intelligently. I wish I did. If you have a paper on it, or know of a link to a good explanation, please post it. I would really like to know more.

Dave

 

[Kevin Mader]

Ravi,

I'm like Dave: short on knowledge on this topic. I am only familiar with it through my exposure while studying for the CQE. Before then, I hadn't heard of it. Since, I have heard it raised a few times, mostly in relation to scientific applications, the most memorable being a Nova program where the Monte Carlo Simulation was used to determine if the surface features came from the inside out or whether they came from the outside in (the study determined that the pattern of features best fit the Monte Carlo Simulation in relationship to meteor showers).

However, if you have written a paper on it, I would be interested in reading it (as would Dave). Mostly, organizations never get to doing basic statistics so getting to advance statistics is a pipe dream. There is plenty of power in these tools, just a limited exposure to them by business rendering them mostly impotent.

Your comment on relaxing critical tolerances without compromising Quality has really gotten my attention (recent, current need). We should continue this dialogue and allow you to expose/teach us a little more about this technique.

Regards,

Kevin

 

[Dave W 2005]

I work in a company that wouldn't know basic statistics if they bit them on the arse, but we are using monte-carlo simulation (even if in a limited application).
We are modelling how structural assemblies come together and applying nominal tolerances to the component parts. Several thousand computerised simulations later where we try and replicate the probabilities of achieving the tolerances; we have an output that tells us what the most likely tolerance stack will be (in most cases it tells us the parts were all dimensioned wrong in the first place).
It becomes a very powerful tool to demonstrate to Designers that their tolerancing strategies will not achieve a satisfactory finished assembly; and then there is the battle to get tolerances changed / redistributed.

Dave

 

[ben]

A very practical and pragmatic book that might be helpful is "Decision Making Under Uncertainty" by John X. Wang.

He goes into depth on using Monte Carlo simulations in Excel.

 

[Bill Ryan - 2007]

Your comment on relaxing critical tolerances without compromising Quality has really gotten my attention (recent, current need).

It piqued my curiosity too - until it was explained to me that the <ST> on a drawing can only be applied if you are in statistical control in the first place. If you are in statistical control, why do you need additional tolerance? Also the statistical approach for relaxing tolerances assumes a "zero mean shift" (which doesn't help my company - die casting/machining - at all!!)

I'm not trying to say there is no application for statistical tolerancing - I just don't see much application in our side of the manufacturing business. However, maybe I'm missing something.

Bill