Peak Shaving vs Backup Power vs Load Shifting: One ESS, Three Use Cases
Key Takeaway
Peak shaving, backup power, and load shifting are three different jobs for one C&I ESS. Map each job to kW, kWh, reserved SOC, and EMS mode priority before you buy. Stacking all three without a hierarchy is how projects overbuy energy or underbuy power.
Table of Contents
- Why “One Battery Size” Breaks Multi-Use Projects
- Three Use Cases Defined
- Sizing and Control Comparison Table
- Peak Shaving: Power-First Design
- Backup Power: Reserved Energy and Transfer
- Load Shifting: Schedule and Tariff Energy
- How to Stack Modes Without Double-Counting
- Selection Boundaries
- RFQ Checklist Before PO
- Frequently Asked Questions
Why “One Battery Size” Breaks Multi-Use Projects
Buyers often ask for “a 1 MWh system that does peak shaving, backup, and TOU shifting.” That sentence hides three different engineering problems. Peak shaving is driven by the worst demand spike; backup is driven by critical-load hours after an outage; load shifting is driven by how much energy you can move across tariff windows or solar surplus.
If you size only on nameplate kWh, you may still fail a demand-charge day because PCS power is short. If you size only on peak kW, you may empty the battery before a planned outage window because SOC was spent on arbitrage. The fix is to name a primary use case, list secondary modes, and size power, energy, and reserved SOC as separate decisions.
Weltrus C&I ESS covers roughly 50 kW to 5 MWh cabinet and containerized classes with BMS, EMS, fire suppression, and thermal options—see the Weltrus C&I energy storage product line overview. Factory-scale solar-plus-storage checklists that mix these modes: solar-plus-storage for a 500 kW factory load.
Three Use Cases Defined
Peak shaving discharges to keep site demand under a kW target—demand charge, transformer limit, or contractual cap. Events are often short and sharp.
Backup power (resilience) supplies critical loads when the grid is unavailable. Design centers on islanding or UPS-class transfer, critical panel design, and guaranteed usable energy for a stated duration.
Load shifting (including many TOU / arbitrage strategies) charges when energy is cheap or solar is surplus, and discharges when energy is expensive. Events are longer; power may be moderate while energy throughput matters more.
These are not marketing synonyms. They change which datasheet line item binds first.
Sizing and Control Comparison Table
| Factor | Peak shaving | Backup power | Load shifting |
|---|---|---|---|
| Primary metric | kW above target | Critical kW × hours | kWh moved per day/cycle |
| Typical duration | Minutes to a few hours | Hours (stated ride-through) | Multi-hour tariff windows |
| EMS priority | Demand ceiling / peak forecast | SOC reserve + island mode | Price / solar forecast schedule |
| PCS behavior | Often grid-following | Often grid-forming / fast transfer | Grid-following + schedule |
| Common failure | Underbought power | SOC spent before outage | Overbought kWh, weak schedule data |
Peak Shaving: Power-First Design
Start with interval load data (15-minute or better). Define the demand ceiling you will not exceed. Required discharge power is the maximum of (load − target) over the year—or over the tariff season that matters. Energy is the integral of that excess over each peak event, then converted to nameplate kWh with usable DoD, discharge efficiency, and reserve.
Worked sizing logic belongs in the RFQ: worst peak day, target kW, event duration, and product band. Pair this with an energy audit so the target is based on measured demand, not a guess.
Honest boundary: peak shaving alone does not guarantee backup. If the EMS empties the pack every afternoon to chase demand charges, there is no reserve when the grid fails at night.
Backup Power: Reserved Energy and Transfer
Backup sizing starts from the critical load list, not the whole factory. Multiply critical power by required hours, then add PCS and battery path losses and a margin. That usable energy must sit in a reserved SOC band the EMS will not spend on peak shaving or shifting unless you explicitly allow “release on grid loss only.”
Also specify transfer time, whether seamless UPS behavior is required, and whether the PCS must form a local grid for motors or sensitive IT. A peak-shave cabinet that cannot island is the wrong product for a backup-primary story—even if kWh looks large on paper.
Document which loads are on the critical panel, who may add loads later, and what happens if the outage exceeds the reserved hours (shed priority list).
Load Shifting: Schedule and Tariff Energy
Load shifting needs a credible charge source (grid off-peak, PV surplus, or both) and a discharge window with enough tariff or process value to justify cycling. Energy throughput and cycle-life assumptions matter more than a single spike kW—unless you also peak-shave on the same day.
On solar-plus-storage sites, shifting often means absorbing midday PV that cannot export and discharging into evening peaks. Design checklists for ~500 kW factory loads: solar-plus-storage factory example.
Honest boundary: without accurate tariff rules, export limits, and forecast quality, “arbitrage” becomes random cycling that burns warranty throughput without IRR.
How to Stack Modes Without Double-Counting
Stacking is allowed when the EMS has a clear priority stack, for example:
- Safety / grid-loss — enter backup; protect reserved SOC
- Peak shaving — enforce demand ceiling with remaining usable energy
- Load shifting — fill remaining schedule with tariff/PV optimization
Size for the binding constraint of the primary mode, then add energy (or power) for secondary modes instead of assuming one number covers all peaks. Example pattern: size PCS for the larger of peak-shave kW and critical backup kW; size energy for peak-shave events plus reserved backup hours plus planned daily shift—if those energies can overlap on the same calendar day, take the union, not the sum of marketing slogans.
Write the priority stack into the purchase order and FAT script. If the vendor cannot demonstrate mode transitions on a test load profile, you do not have a multi-use ESS—you have three promises.
Selection Boundaries
This multi-use framing is not suitable when:
- You refuse to name a primary use case and want “maximum of everything” at minimum capex
- Backup must be UPS-class but the quoted PCS is grid-following only
- Tariff or demand data are missing and the RFQ still demands guaranteed payback
- Reserved SOC for backup would leave too little energy for the stated peak-shave target—pick one or upsize energy honestly
- Export rules or transformer limits make charge windows unreliable for load shifting
In those cases, split systems, reduce scope, or buy more kWh/kW—do not hide the conflict in a single brochure SKU.
RFQ Checklist Before PO
- Primary use case and ranked secondary modes
- Demand target (kW) and worst-peak interval data for peak shaving
- Critical load list, hours, and transfer-time requirement for backup
- Reserved SOC (%) and EMS rule: when reserve may be used
- Tariff / PV / export assumptions for load shifting
- PCS modes: grid-following, grid-forming, black start—checked against needs
- FAT script that runs stacked priorities on a sample load profile
- Product class band (cabinet vs container) matching pad and power
Incomplete answers mean the one-line diagram is not ready for OEM comparison.
Share Your Primary Use Case with Weltrus
Send interval load samples, critical-load hours, and tariff or PV constraints. We can help map peak shaving, backup, and load shifting to a realistic kW/kWh and EMS priority stack for C&I ESS.
Frequently Asked Questions
Can one C&I ESS do peak shaving, backup, and load shifting?
Yes, if you size kW, kWh, and reserved SOC together and rank EMS priorities. Without a hierarchy, one mode will starve the others.
What is the difference between peak shaving and load shifting?
Peak shaving cuts short demand spikes under a kW ceiling. Load shifting moves energy across hours for tariff or solar value. One is power-first; the other is energy-and-schedule-first.
How much SOC should I reserve for backup?
Enough usable energy for the critical load duration, including losses and margin. Keep that band out of day-ahead shaving or shifting unless the EMS releases it on grid loss.
Do I need a different PCS for backup than for peak shaving?
Often yes in behavior: backup may need grid-forming or fast transfer. Confirm the same PCS can meet both mode requirements before combining them in one SKU.
Should I buy more kWh or more kW first?
Depends on the primary mode. Peak-shave failures are often kW-short; long backup and deep shifting failures are often kWh-short. Size the binding constraint first, then add for secondary modes.
