How to size battery storage for your critical loads
Backup hours, load priority and the trade-offs that drive capacity. Why the right answer starts with your load list, not a battery catalogue.
"How many kWh do I need?" is the question we get first, and it is the wrong place to start. A battery is sized by what it has to carry and for how long. Get those two inputs right and the capacity falls out of the arithmetic. Get them wrong and you either pay for storage you never discharge, or you find out mid-outage that the system was never designed to hold the load you actually care about.
Step one: sort your loads into tiers
Almost no site needs to back up everything. The first thing DE does on a backup assessment is critical-load identification: mapping which loads must never drop, which can ride out a short interruption, and which can simply wait for the grid. That sort is the single biggest lever on system cost.
Tiers are illustrative. Which loads land where is site-specific, and it is a conversation between your plant team and ours.
Step two: two numbers, not one
Storage is specified by power and by energy, and they are sized by different constraints. Confusing them is the most common error in a backup spec.
What it can hold at once
Set by the peak simultaneous demand of your backed-up tiers, including motor starting surges. Undersize this and the system trips exactly when you need it.
How long it can hold it
Set by that demand multiplied by the duration you need to cover, plus usable-depth and round-trip losses. This is where cost accumulates.
Step three: how long do your outages actually last?
Duration should come from your own outage history, not from a rule of thumb. Two sites with identical loads can need very different systems: one that sees frequent short dips needs power and instant response, while one that sees long scheduled cuts needs energy. DE analyses load profile, outage history and site constraints before proposing anything, precisely because these two cases look nothing alike on paper.
Every additional hour of backup duration costs energy capacity that sits idle most of the year. Beyond a point, a generator retained as a deeper backstop is the cheaper way to cover the long tail with the battery carrying the first, most damaging seconds.
Step four: let solar and tariffs pay for the capacity
A battery sized purely for outages earns nothing on the days the grid behaves. The same asset can discharge during peak-tariff windows to cut demand charges and high time-of-day rates, then recharge from solar the next day. For many sites, pairing solar generation with storage is what makes the sizing affordable: everyday savings and resilience come out of one investment rather than two.
This changes the sizing question. Instead of "what is the smallest battery that survives an outage", it becomes "what capacity clears both the critical-load requirement and the daily peak-shaving opportunity", usually a slightly larger system with a materially better return.
Step five: specify the switchover, not just the capacity
Capacity is useless if the transfer is slow. Batteries take over critical loads in under 20 milliseconds, before equipment notices the outage. A diesel generator needs 30 to 60 seconds to start and transfer, and can take up to two minutes to pick up load. If a batch, a procedure or a cold chain is mid-run, that gap is often enough to lose it. Put the transfer time in the specification alongside the kW and the kWh.
What DE needs to size your system
Sizing is an assessment, not a quote generated from a form. Bring these and it takes days rather than weeks:
From there the sequence is the same on every site: critical-load ID, site assessment, proposal and design, installation, then 24/7 AI remote monitoring through DE Atlas, which is also how you find out whether the sizing assumptions held. Battery health, state of charge, consumption and savings are tracked in real time, so the next system on the next site is sized against evidence rather than estimates.