When an 'out of course' failure bites you in the bum!
Did you notice the 'in' rather than 'on', very pertinent for this little discourse!
Many moons ago in my first job when dinosaurs roamed the earth and R12 was as plentiful as water and treated in the same cavalier fashion it was a 'thankless' part of my role to conduct reliability studies for all of the HVAC systems that we supplied to BREL (remember them?) and their ilk. That said it was a good grounding in how to design out superfluous bits of kit which added little to the system functionality but jacked up the FRPMH (Failure Rate Per Million Hours).
The thrust here isn't really to dawdle down memory lane but to apply some first hand experience to how some of the ridiculously small FRPMH were arrived at in the source of all truth, a weighty tome titled 'Non-predicted Reliability Data' or the version which I was using which was snappily titled NPRD 3. When you were looking for data it seemed apparent that most data was gleaned from the military as your data could be selected from a number of environments.....aerospace, static, wheeled road vehicles and the give away tracked off-road vehicles. The failure rates obviously varied greatly between applications with the same component in static application being many more times reliable than in tracked off-road vehicles, no surprises there! Oh the electronics 'bits' were calculated using data contained in another riveting read, MIL-Hdbk-217E again a Military Handbook using data collated from military sources!
That I can quote the sources must give some indication of just how scarred I was by the work that took place 30 years ago.
Before getting on to the thrust here is another little digression, operating environment and the importance thereof! One application where we saw a ridiculous failure rate was on the fan belt drive system on the under frame mounted condenser section of the Class 158 DMU HVAC system. It was simply ridiculously high for a system used only in summer with one belt failure per year a given and more usually two or three failures per annum. It got to the stage where the service chaps took to cable tying two or three spare belts in position to easily fit them next time they were called out. Bear in mind in this era there were no toilet retention tanks so changing this belt was indeed a shitty job. Oh and the reason for the horrendous rate, far above any application we could find? Well quite simply the promised soft start where the vehicle hydro-static system ramped up simply wasn't there and the system went from dead to full on instantaneously causing all manner of train issues.
So where was I, oh yes failure rates and how they impact you in real life!
Lock-nuts are reliable and used in many safety critical applications but they do fail, indeed the failure rate is quoted as between 0.054 and 0.084 Failure per Million Hours, so you would hope that any failure would be meteorite strike territory for most oily mechanical engineers....wouldn't you?
Well perhaps not this is my lower back and you can see the nut that has come adrift, apologies for the direction, I don't often lie face down but I'll be damned if I can edit the shot!
So I got to thinking, what is the probability of this nut failing in a reasonably benign environment? I mean I don't do anything that puts a great deal of stress on my back, I gave up the tower-jumping and skydiving when they nailed everything together so it has a relatively easy life.
Using the figures from NPRD3 and assuming that people have 4 nuts fitted at any time then in the 'time' it took my nut to fail there would need to be 10 of me to reach 1 million operating hours, so to see a fault you would need between 120 and 185 people to have four nuts fitted in that time span to see a single fault.
All of a sudden there appear to be an awful lot of meteorites headed in our direction, especially when you factor in people who have more than 4 nuts fitted and have them in place for a longer time!
Where is this leading in a 'practical engineering application'? Well remember all of those secondary retention devices fitted to the under frame mounted kit, the reliability figures don't seem quite so assuring when you start looking at real world operation so please remember to plan in regular maintenance checks or the equipment will plan them in for you at the most inopportune time and in the most public way, stand your ground and remember this lesson when bean-counters are attempting to rationalise the maintenance regimes!
So endeth todays lesson!