Battery Energy Storage System (BESS) Sizing — Industrial Heuristics

The Five Use Cases (and Why They Drive Different Sizes)

BESS is not one product. The same chemistry and form factor can serve completely different purposes, and each demands different sizing logic. Don't size a battery without first knowing which of these you're solving for.

The C-rate trap: If your application demands >1C continuously (1 hr discharge), most LFP commercial systems will derate or void warranty. Talk to the OEM about thermal limits before promising the rate to operations.

Step 1: Build the Load Profile

Pull 12 months of 15-minute interval data from the utility meter (most utilities provide CSV via Green Button or AMI portal). Don't guess based on monthly bills - the demand peak is hidden in 15-minute spikes you can't see at the monthly level.

Compute these statistics:

• P95 demand (95th percentile of 15-min readings)
• Peak demand (max 15-min reading) and when it occurred
• Top 100 peak intervals - if these cluster in 2-3 hours per day, you have a great peak-shaving opportunity. If they're random, peak shaving will not work reliably.
• Coincident demand on utility peak windows (often 4-9 PM weekdays)

Step 2: Pick the Sizing Method

Method A: Peak Shaving

P_battery (kW)   = Demand_target_reduction
E_battery (kWh)  = sum of (P_load(t) - P_target) over peak window
                 / round-trip efficiency (RTE, ~0.85 for LFP DC-DC)
                 / depth of discharge (DOD, ~0.90 for LFP)
                 / 0.95 (degradation margin to end of warranty)

Rule of thumb for a typical 4-hour utility peak window: E/P ratio of 2.5-3.0 hours covers most cases. Going higher than 4 hours rarely pays back on demand-charge alone.

Method B: Energy Arbitrage

Useful only where the spread between off-peak and on-peak energy exceeds about $80/MWh on a daily basis. With LFP at ~$300/kWh installed, ~85% RTE, ~5,000 cycle warranty, the breakeven spread is roughly:

Spread_breakeven ($/MWh) = (Capex_$/kWh / Cycles) / RTE / DOD * 1000
                        = ($300 / 5000) / 0.85 / 0.90 * 1000
                        = ~$78/MWh

Most US markets don't sustain this. CAISO, NYISO, and ERCOT can during summer. PJM and MISO usually can't.

Method C: Frequency Regulation / Ancillary Services

The product here is capacity (MW) and responsiveness (ms), not energy. Most batteries sized for FCAS use E/P = 0.5 hr (15-30 min), because the product doesn't pay for energy throughput. Regulation/RegD in PJM is the canonical case. If your industrial site doesn't have FERC market access via an aggregator, skip this.

Method D: Backup / UPS

Size for the longest tolerable outage. For diesel-genset replacement, this is usually 8-24 hours, which makes batteries economically uncompetitive unless you also stack other use cases (peak shaving + backup is a common combo).

For UPS-class outages (1-10 minutes), batteries make sense at the IT-load scale only. At industrial process scale (MW), look at flywheel + diesel before BESS.

Method E: PV Firming / Renewable Smoothing

Two flavors:

• Smoothing (cloud-passage ramp limit): size for ~10-15% of PV nameplate, 30-min duration. Cycles many times per day - degradation matters.
• Firming (commit to a fixed energy block): size for the deepest expected dip. Typically 25-50% of PV nameplate, 4 hours.

Chemistry Selection

Code, Standards, and the NFPA 855 Wall

NFPA 855-2023 (Standard for the Installation of Stationary Energy Storage Systems) is the dominant US code. Key items that drive BESS layout and cost:

• 20 kWh per unit, 50 kWh per group max for indoor non-dwelling occupancies without a fire-rated enclosure (Sec 9.6).
• 3 ft separation between battery cabinets unless tested otherwise (UL 9540A).
• 10 ft setback from buildings/lot lines for outdoor systems >50 kWh (varies by AHJ).
• Explosion control required for indoor systems >50 kWh (Sec 9.6.5) - either deflagration vents or NFPA 69 inerting.
• Gas detection for hydrogen (lead-acid) or flammable electrolyte vapors (Li-ion).

Cost Structure (2025 Order of Magnitude)

Note: tariffs on Chinese cells (Section 301 + AD/CVD) added 25-50% to LFP cell cost in 2024-2025 depending on origin; verify current pricing with an EPC quote.

Heuristics That Save Time

• Always start with the load profile, not the battery. 80% of bad BESS deployments came from sizing to a single bill peak instead of the full interval data.
• Stack revenue streams - a single use case rarely pays back. Peak shaving + backup + (where available) ancillary services is the typical stack.
• Plan for 2.5%/yr capacity fade over 10 years. Size 25-30% above year-1 need to hit warranty endpoint.
• Get the AHJ engaged before pad design. Setbacks, gas detection, and fire suppression vary by jurisdiction and can change layout meaningfully.
• If C-rate is ambiguous, derate. Continuous discharge above 0.5C accelerates degradation in most LFP packs. The OEM rating sheet is what matters, not the chemistry headline.

References

• NFPA 855-2023 - Standard for the Installation of Stationary Energy Storage Systems
• UL 9540 - Energy Storage Systems and Equipment
• UL 9540A - Test Method for Thermal Runaway Fire Propagation in BESS
• IEEE 2030.2.1 - Guide for Design, Operation, and Maintenance of BESS
• NREL 2024 Annual Technology Baseline (ATB) - Battery cost projections
• DOE Energy Storage Grand Challenge Roadmap (2020, updated 2024)
• EPRI Technical Update 3002025610 - BESS Procurement Guide for Industrial Customers
• Lazard 2024 Levelized Cost of Storage Analysis (LCOS v9.0)