Power failure planning is the set of decisions made before a cold store loses mains power, not during it: what backup power exists, how long the building holds temperature without it, which loads get protected first if backup capacity is limited, and how the site keeps watching temperature while the lights are out. A cold store with no plan finds out the answers to all of these questions during the outage itself, at the worst possible time to be improvising.
The core physical fact behind the whole plan is simple: a well-insulated, fully loaded cold store holds its temperature for far longer without power than an empty or poorly sealed one, because the product itself and the cold air trapped inside are both thermal mass working in the store's favour.
Generators and uninterruptible power
Two different technologies cover two different gaps. An uninterruptible power supply, a bank of batteries, bridges the seconds to minutes between mains failing and a generator starting, keeping refrigeration units and monitoring systems running through that gap rather than restarting from cold. A generator covers everything after that, for as long as it has fuel, and needs to start automatically rather than depend on someone driving to site to switch it on. The combination matters because a generator alone still has a start-up delay, seconds for some units, longer for others, during which the store is unprotected without the battery bridge. Fuel supply is its own planning question separate from the generator itself: a tank sized for a few hours covers a short outage but says nothing about a multi-day one, and a refuelling arrangement agreed only after the generator is already running is not a plan, it is a hope.
Hold-over time
Hold-over time is how long a loaded cold store stays inside its temperature band with no active refrigeration at all, and it depends on insulation quality, how full the store is, how often its doors open during the outage, and the outside temperature. A well-sealed freezer, kept closed, can hold for many hours; the same room with doors opening regularly to check on it, or with thin or damaged insulation, holds for a fraction of that. Knowing the actual hold-over time for a specific store, rather than assuming a generic figure, is what tells a site how much time it genuinely has before an outage becomes a temperature excursion.
Priority loads when capacity is limited
Not every site's generator can run the entire building at once, especially where refrigeration sits alongside lighting, office power and other equipment on a shared supply. Priority planning decides, in advance and written into the site's cold chain SOPs, which loads get power first if total capacity is constrained: refrigeration and freezer units ahead of general lighting, the highest-value or least-replaceable product ahead of lower-priority stock if a facility has to choose between rooms, a decision that matters most acutely in a pharmaceutical cold chain storing product that cannot be replaced on short notice. That decision is far easier made calmly in advance than during an actual outage, when the pressure is to keep everything running and the generator cannot actually do that.
Monitoring through the outage
Temperature monitoring needs to keep running on the same backup power as the refrigeration itself, not on mains alone, because a store that loses monitoring at the same moment it loses cooling gives staff no way to know how close it is getting to its limit. Sites relying on manual checks during an outage need a schedule for how often someone physically checks each affected room, since the whole point of the plan is knowing the store's status continuously, not discovering a problem only once power is restored and the record can finally be read back. A monitoring system that depends on an internet connection carried over the same mains supply that just failed needs its own independent backup, or it goes dark at exactly the moment it matters most.
Testing the backup
A generator that has not been started and load-tested recently is an assumption, not a plan. Backup power needs a real test on a schedule: starting the generator under an actual electrical load, not just running it idle, and confirming the automatic transfer switch actually moves the building onto backup power without anyone flipping a manual switch. The same applies to the battery bridge and the monitoring system's own backup power. A cold storage warehousing operation that tests its emergency procedures on paper but never load-tests its generator has rehearsed the wrong emergency. The test itself belongs on a fixed schedule, not left to whenever someone remembers, since a generator with a fault only reveals it under an actual load, and the middle of a real outage is the worst possible time to discover one.