Peak Shaving Battery Storage: How C&I Demand Charges Work
Two different things get called peak shaving. One cuts your energy bill by moving cheap kWh into expensive hours. The other cuts your demand charge by flattening a short spike. They use the same hardware and reward opposite sizing — and confusing them is the most common reason a C&I system underperforms.
The Bill Has Two Parts
A commercial electricity bill is not only kilowatt-hours. There is usually a second line: a demand charge, billed on the highest average power you drew during any single interval of the month — often 15 or 30 minutes.
That number is set by your worst window, not by how much energy you used overall. A plant that runs steadily can have a lower demand charge than a plant that uses less energy but spikes hard once a day, because spikes are what get measured.
A battery can attack either line of the bill. It rarely does both optimally at once.
Demand Charges: You Are Fighting One Bad Window
If demand charges dominate your bill, the job is not to store cheap energy. It is to shave a short, sharp spike — a morning startup, compressors cycling together, a forklift charger bank all drawing at once.
What that actually needs:
- Power, not duration. The system has to discharge hard for 15 to 60 minutes, right when the spike happens.
- A controller that reacts inside the interval. If the EMS responds after the 15-minute window has closed, the peak is already recorded and the discharge saves nothing.
- Modest energy. A 1-hour or 2-hour battery is often plenty.
The sizing trap here is buying a 4-hour battery to solve a 20-minute spike. You pay for energy you never discharge.
Energy Arbitrage: You Are Moving Volume
If your tariff has a wide spread between off-peak and on-peak rates, the job flips. Now you want to buy low and displace high, which means moving as many kWh as the peak window allows.
What that needs:
- Energy (kWh), because volume is the revenue.
- Enough power to empty the battery inside the peak window. This is the check people skip — a big pack behind a small PCS never delivers its nameplate.
- Cycle life, because every day is a full cycle, and twice-daily cycling doubles the burn rate.
A 2-hour to 4-hour system is the common choice in this mode.
Why 2-Hour vs 4-Hour Falls Out of This
Duration is simply kWh divided by kW. It is not a spec you pick first — it is a consequence of which revenue stream you are targeting.
| Primary target | Typical duration | What the system must do well |
|---|---|---|
| Demand charge reduction | 1–2 hours | Discharge hard and fast inside a short interval |
| Tariff arbitrage | 2–4 hours | Move maximum kWh within the peak window |
| Both, seasonally switched | 2–4 hours | EMS that changes strategy as the tariff changes |
A factory that asks which revenue stream you are targeting before quoting a duration is a good sign. One that leads with a duration and a price per kWh is guessing.
What the Battery Has to Survive
- Cycles. LiFePO4 cells are commonly rated 6,000 to 8,000 cycles. Demand-charge duty is shallow and infrequent, so it burns few full cycles. Twice-daily arbitrage burns far more.
- Depth of discharge. Usable energy is typically around 90% of nameplate, not 100%.
- Round-trip efficiency. Expect roughly 90%. Energy in is not energy out, and that gap sits directly on your arbitrage margin.
- Response time. For demand charges specifically, the EMS has to act inside the interval. This is a control problem more than a cell problem.
What to Ask the Factory
- Within my peak window, how many kWh can the system actually deliver — after depth of discharge and round-trip efficiency?
- What is the continuous discharge power in kW, and can it cover my worst spike?
- At one full cycle per day, how is cycle life budgeted, and is the warranty written in years or in cycles?
- Who configures the EMS strategy, and can it be set against my tariff and measured load profile?
Specific numbers plus an explanation of the derating logic is the answer you want. A nameplate capacity with no derating is not.
FAQ
Is peak shaving the same as energy arbitrage?
No. Peak shaving targets the demand charge, billed on your highest average power draw in one interval and measured in kW. Arbitrage targets the spread between off-peak and on-peak energy rates, measured in kWh. Same battery, different sizing.
What size battery do I need for demand charge reduction?
Usually shorter and punchier than people expect. You are shaving a spike that lasts 15 to 60 minutes, so a 1-hour or 2-hour system with enough discharge power often covers it. A 4-hour battery for a 20-minute spike means paying for energy you never use.
How do I choose between a 2-hour and a 4-hour C&I system?
Duration is a result, not a starting spec. Demand-charge-driven bills suit 1 to 2 hours; tariff arbitrage across a long peak window suits 2 to 4 hours. Pick the revenue stream first and the duration follows.
How does peak shaving affect battery cycle life?
Demand-charge duty is shallow and infrequent, so it consumes relatively few full cycles. Twice-daily arbitrage consumes far more. With LiFePO4 commonly rated at 6,000 to 8,000 cycles, the strategy you run matters as much as the size you buy.