It still makes me appreciate robot engineers and system designers when I see a robotic implementation working flawlessly. And really, they do work flawlessly most of the time. FANUC (Rochester Hills, MI) estimates that the mean time between failure (MTBF) for their robots now is in a range of:
- 40,000 to 100,000 hours; which is
- 1,667 to 4,167 days; which is
- 238 to 595 weeks; which is
- 56 to 139 months; which is
- 6 to 11.4 years.
That’s really terrific, considering a robot is a compendium of parts, by definition they are in motion much of the time, and each part has its own reliability. In fact, if you’re of an engineering mind, you remember that system reliability is based on multiplying the reliability of each component. Suppose you have eight parts, each with a reliability of 0.98. Multiply that number by itself eight times, and you get a reliability for the system of only 0.85.
Let’s increase the numbers to be as reliable as Ivory Soap is pure—ninety-nine and 44/00 of a percent pure. Using that improved number, we have a system reliability of roughly 0.96. Much better, but far from perfect. Cheers to today’s component quality, and cheers to the designers of these robots.
FANUC has a software suite called ZDT for Zero Down Time. This software monitors the robots and can find common errors and show hidden problems. The whole idea is to reduce or eliminate down time and optimize robot performance. It uses the following methodology to do so:
Analyze à Predict à Maintain à Optimize
It’s really a good recipe and the system calls for, and gets, a goodly amount of data to analyze so it can make the predictions and make recommendations for maintenance before it is critically necessary. We must remember that maintenance and optimization are two very important parts of reliability, and will have an effect on your company’s reliability.
An Exercise
What is your company’s reliability anyway? At its highest and most low-resolution level, the parts going to the shipping department should about equal the materials coming in from the dock. Any faster on the parts and you’ll be waiting for materials at some point in the serpentine route your parts take through your business.
Like a fractal, the structures in your business repeat at larger and smaller scales. For example, you could take a particular laser cutter and think of it like the paragraph above. Your laser cutter has inputs and outputs—and so does an individual welder, and for that matter an entire welding department does as well.
A caveat: Putting all of this into a reliability equation is much harder than doing it for a few parts in a machine. There are mathematical tools you can use, though, and some of these you can do by hand, such as reliability block diagrams and fault tree analysis. Other tools are best used as part of a software strategy, and there are many such programs and modules for other programs that can help here.
Just as an exercise, though, you may find it interesting and eventually enlightening to do such an analysis in a thumbnail fashion. It might take you a day or two, depending on how big your company is, and how many processes are going on. It’s possible that you will be interested, enlightened, entertained even, and maybe shocked when you are done. The payoff is the information you glean.
If you have already done such an exercise, please do get in touch with me at dave@fifthwavemfg.com, I would love to hear your results and discoveries. Anonymous submissions are welcome.