For many commercial facilities, the biggest line on the electricity bill is not energy — it is the demand charge, billed on the single highest 15-minute draw of the month. A battery that discharges for a few minutes at exactly the right moments can cut that peak dramatically, and unlike energy arbitrage, the savings arithmetic is simple enough to verify on one page.
How demand charges work
Commercial tariffs typically bill two components: energy ($/kWh consumed) and demand ($/kW of maximum draw, measured over 15–30-minute windows, per month). Demand rates of $10–25/kW/month are common — a facility peaking at 500 kW can pay $60,000–150,000 a year for capacity it touches a few hours monthly.
What the battery does
The BESS monitors site load in real time. When draw approaches a set threshold, it discharges to hold the metered peak down; it recharges during off-peak hours (or from solar). Flattening a spiky profile means the utility bills you for the plateau, not the spikes.
Worked example: 500 kW facility
- Monthly peak: 500 kW, driven by 90-minute midday spikes; base plateau 350 kW
- Demand rate: $18/kW/month
- BESS: 150 kW / 300 kWh LFP (2-hour) shaving the peak to ~360 kW
- Saving: 140 kW × $18 × 12 = $30,240/year
- Installed cost at ~$280/kWh: ~$84,000 → simple payback ~2.8 years
- Stack self-consumption of solar surplus and time-of-use arbitrage on the same asset, and effective payback drops below 2.5 years
The sizing logic
| Question | Determines | How to answer |
|---|---|---|
| How many kW must be shaved? | Power rating (kW) | Interval data: peak minus target plateau |
| How long do peaks last? | Energy rating (kWh) | Longest continuous excursion above target × shaved kW × margin |
| How often do peaks recur daily? | Cycling duty & chemistry | LFP handles daily deep cycling — see our chemistry guide |
| What if the battery misses one event? | Control quality | One missed 15-min window resets the whole month's charge — controller reliability is everything |
The failure mode to respect: demand charges are ratchet-style — a single missed peak forfeits that month's savings entirely. This puts a premium on accurate load forecasting, adequate energy reserve, and controllers that hold headroom for back-to-back peak days.
Which sites benefit most
- Spiky loads: crane and hoist operations, EV fleet depots, cold storage defrost cycles, injection moulding, welding shops — short sharp peaks over a modest base are ideal.
- High demand rates: above ~$12/kW/month the arithmetic usually works; below ~$8 it rarely does on demand savings alone.
- Flat 24/7 loads benefit least — there is no spike to shave; their storage case rests on tariff arbitrage or backup value instead.
The Sungrow ST255CS-2H (257 kWh liquid-cooled LFP cabinet) is designed exactly for this duty — daily cycling, modular parallel expansion, integrated PCS. Send us 12 months of interval data and we'll size the shaving system and model the payback for your tariff.