Belt, Rope and Suspenders? Statistical Insanity for Reliable Power

Redundancy is a factor of the critical nature of the system. I worked on a similar system for the military where they were powering air traffic control radar sites. The nature of the task may have dictated MORE of what you are calling "insanity" than they actually implemented, especially when I wad on the plane heading home, guided by that very system!

I just attended a power systems seminar at Eaton in Pittsburgh. I was really impressed with their state of the art UPS and active power factor correcting equipment. Their UPS can go from standby to full power so fast that it doesn't have to run until needed. I have never seen anything like it.

Without sounding like an endorsement, you should look into what you can learn from them. Growth is not a problem and neither is redundancy if you pick your equipment correctly. Like I tell others, you will have to learn enough to know what you want. Not only will they tell you, they will demonstrate it. Of the nearly 10,000 people that have visited their labs, a common comment went something like "I learned more in 2 hours than I learned an entire semester in college."

The hardest part of finding an answer is knowing where to look!

Their UPS can go from standby to full power?

Maybe you mean emergency gen set or something else?

No, I mean what I said. Using IGBT transistors and fully charged ultracaps, the switch over time is less than 2 msec to full power.

Ah yes, Eaton has been a player in this design. We are in the midst of trades of floor load, space claim (swtich gear access), wiring simplicity (wireway clutter in small area), air flow/heat dissipation and satisfying "What ifs". What if both generators and the utility power fails? (motorized damper ventilation system to keep UPS from cooking). What if the whole UPS fails? (bypass?). What if the transformer fails?

There are a few techniques that may help you, Fault Tree Analysis (FTA) and Failure Mode and Effects Analysis (FMEA).

http://en.wikipedia.org/wiki/Failure_mode_and_effects_analysis

http://en.wikipedia.org/wiki/Fault_tree_analysis

The hardest part is finding reliable failure rate data, but at least you'll have some handle on the relative magnitudes of the probabilities involved, and it may help point out some cascading failures that nobody had thought about. Ultimately it will come down to cost, scheduling, and the law of diminishing returns....and don't forget those events known as Black Swans....

"...The Black Swan Theory or Theory of Black Swan Events is a metaphor that encapsulates the concept that The event is a surprise (to the observer) and has a major impact. After the fact, the event is rationalized by hindsight..."

http://en.wikipedia.org/wiki/Black_swan_theory

Good luck!

If you have n pieces of equipment to function/malfunction, there will be 2ⁿ scenarios to consider....

There IS a system that does it all. It is not even new.

It only requires, that the non-interruptible equipment be DC powered. That is common in today's 48VDC medium voltage distribution fed from batteries / rectifier. It takes a generator / flywheel / diesel combination. In a power failure it takes less than a second to have it up and running. AC etc. can be sequenced on later. Odd equipment needing other power handled with inverters at all time on from the batteries.

Obviously, one can overcomplicate: what if battery fails, what if diesel fails, etc. Fukushima reminds us, perversely, that there is no end to the what ifs's. One has to terminate the what if chain at a certain point, firmly.

Belt, rope, suspenders?--you forgot the skyhook.

I heard Siemens made them. Special order only though.

They come from the same shop as glass to wood welding electrode I believe. Wonderful stuff!

Yep, mine was a Siemens one. Batteries rated for submarine use. Considering the savings of not having service outage, pretty reasonable at the price. Two New York blackouts with no interruption. Not bad.