The.idea you can do 90% of mega watts is bad.
You may be able to do 90% for 5 to 7 months in colder weather.
And 80% in warmer weather.
That is standard practice for all power systems. The 80% rule applies to everything from the gensets and all wiring & main distribution circuit protection devices down to the individual breakers and wiring feeding the individual miners. The reason is to allow the equipment to safely handle worst-case conditions and loads for extended periods of time.
Whoever is designing your power infra should be well aware of this and from a professional standpoint strongly resist all outside pressure (usually costs) from a customer to push the limits. Back in the day when I did power design work, for liability reasons I always put my objections in writing and required the customer to sign off on it if they insisted on doing it.
80% is the defacto industry rule of thumb.
In fairness we're only pushing at an actual rate of 85.5% (95% uptime X 90% Capacity = 85.5% actual). Aggressive, but realistic ~ presuming we stick the uptime. (Jenbacher J6 series are rated for 100% continual load. Of course real world results vary.)
Also True: Running at 85.5% continually causes higher internal temperatures → accelerated wear on windings, bearings, pistons, turbochargers, exhaust valves, etc. thereby shortening major overhaul intervals of every 10-15K hours by 10-20% sooner. Alberta cold does help counter this some (see image).
https://cdn.hikb.at/charts/average-monthly-temperatures/alberta-average-monthly-temperatures.pngOf Note: Professional operators this is modeled after (Bitfarms, Hut 8, CleanSpark) target 75–85% genset utilization.
Accelerated Miner wear / shortened lifespans: The project has a 10% sinking fund for rolling miner upgrades to maintain competitive edge and hedge against difficulty increases. So while implemented for a different reason, it does actually vicariously address this issue too.
The Reasoning and Rational:There's another aspect in the modeling which isn't mentioned herein yet, but is critically applicable to feasibility, and that's the paper (debt) (servicing interest, Investor returns, etc.). It runs on 12, 24, and 48 month terms (plus revolver facilities).
The minor increases in maintenance by shortening equipment intervals between maintenance services (or replacement for miners) even when factoring in the down time for the same ~ is far eclipsed by the production benefit of the additional output.
Works out to about +$2.6M net over 4 years after extra maintenance (~$300K), with the same number of major overhauls (2).
So the actual costs / effects of running it this way (a bit "hot" as it were) don't arrive and materialize until after the initial 48 month debt / Investor cycle is complete. Providing additional capital when the demands for it / on it are at the highest.
{I know it runs counter to the broader crypto culture to design a project in alignment with traditional capital confines and constraints. But the reality is Investors want their money back, and they want a return on it too, and reasonably so.
As I've also mentioned ~ this isn't an altruistic or ideological endeavor. It's boring boring steady predictable production output with calculable returns.}
P.S. Lol - I expect we'll get a similar waiver you used to make your clients sign from our electrical / mechanical engineers when the time comes ;-). Seems only prudent.
