• Badabinski@kbin.earth
    link
    fedilink
    arrow-up
    4
    ·
    8 days ago

    Yeah, the lifespan and ability to leave a flywheel “discharged” makes me wish I could have one for my homelab (as unrealistic as that might be). I have a solar generator as a battery backup, but it’s not a true UPS with a fast transfer switch (I needed at least 3kWh of capacity for long power outages, my max draw is like 600 watts before I finish load shedding). Most of my servers can tolerate the brief voltage sag, but my R640 chokes and dies. My battery is hooked up to one of my PDUs, and I’d love to have a flywheel hooked up to the other PDU. The battery would be fully transitioned by the time the flywheel was discharged.

    On the point of safety, I have a question. I feel like it’s probably easier to prove that a flywheel system is deenergized, but there is the very slight risk of confinement loss. With a chemistry like Lithium Iron Phosphate that can’t sustain a flame and doesn’t produce flammable gasses, do you feel that batteries might begin to approach the safety of flywheels? It sounds like you have actual experience with flywheel systems, so I’m quite curious.

    EDIT: holy shit, someone is actually selling a 300 KVA flywheel system on eBay for $30,000. I wonder who the hell would buy something like that used.

    EDIT: I said “very slight risk” of confinement loss, and I should probably correct myself. The risk is ridiculously, stupidly small for a system like I linked above. Maybe the bigger systems that get buried and have concrete poured on them are riskier, but I don’t know if people even do that anymore for datacenters.