SRC Forum - Message Replies
Forum: Reliability & Maintainability Questions and Answers
Topic: Reliability & Maintainability Questions and Answers
Topic Posted by: Reliability & Maintainability Forum
Organization: System Reliability Center
Date Posted: Mon Aug 31 12:47:36 US/Eastern 1998
Posted by: Tim Daniels
Date posted: Tue Oct 26 10:09:55 US/Eastern 2004
Subject: Spares level over set period of time
Sorry this may be a silly question but it's the first time I've done anything like this.
I need to calculate the number of spare LRUs required for a product to have 95% confidence level (of having a spare available) when the repair time is 10 months. I have failures/million hours figures and thought the RAC Poisson calculator for sparing was the answer! But now I have to consider a mission time of 18 months and calculate spares for this time only. I cannot see what will change. Can anybody give me any clues?
Any ideas will be very much appreciated.
Subject: Spares for finite mission time
Reply Posted by: Larry George
Organization: Problem Solving Tools
Date Posted: Sat Oct 30 19:18:09 US/Eastern 2004
You don't need as many spares for a finite mission time as you need for continuing operation. Compare the following probability statements to see if the last suits your desired 95% confidence of no stockout:
1. 1.8*P[Demand > Spares in 10 months] < 0.05 (The 1.8 is because the mission is 18 months.
2. P[Demand <= Spares in 10 months]*P[No Demand in 8 months]+
P[One demand in 10 months]*P[Repair is done before next demand in remaining 8 months]+
P[Two demands in 10 months] and so on > 0.95
I'll try to work out the computation of the latter and send it, assuming constant demand rate, which implies demand is Poisson.
Try to estimate age-specific failure rates, because failure rates aren't constant. Infant mortality could quickly consume your spares and you'd have to hope repairs are done in time. Don't rule out premature wearout in 18 months that could wipe out spares.
Generally accepted accounting principles require ships and returns counts. These data are statistically sufficient to make nonparametric estimates of age-specific failure (actuarial) rates, actuarial forecasts, and estimates of the demand distribution. Send data, and I will do that, free of charge.