Peak Shaving & Solar HVAC — Demand Charge Management & Solar-Augmented Cooling
40–50%
Share of commercial building electricity consumed by HVAC
30–70%
Portion of utility bill that can be demand charges (peak kW)
20–40%
Typical demand charge reduction achievable with solar + battery
3–7 yr
Typical simple payback period for commercial solar + BESS systems
$5–$50
Utility demand charge range per kW of monthly peak (varies widely)
What Is Peak Shaving?
Peak shaving is the practice of reducing a facility's maximum electricity demand during periods when grid power is most expensive or congested. The "peak" refers to the highest power draw (measured in kilowatts, kW) recorded during a billing period — often just a single 15-minute interval that sets the demand charge for the entire month.
Utilities charge demand fees to recover the cost of maintaining generation and transmission capacity sized to serve all customers at their simultaneous maximum demand. Even if a building hits its peak for only 15 minutes on one afternoon, that spike determines a demand charge that may represent 30–70% of its electricity bill for the whole month.
Peak shaving strategies aim to flatten the load profile — reducing the spike without necessarily reducing total energy consumption. Key tools include:
- Solar PV — generate power on-site during peak solar hours (which often align with HVAC peaks)
- Battery Storage — charge during off-peak, discharge during peak demand intervals
- HVAC Pre-conditioning — cool the building before the peak period, then reduce compressor demand during peak
- Demand Response — voluntary load reduction during grid emergencies in exchange for bill credits
- Time-of-Use Shifting — reschedule flexible loads (EV charging, ice storage, DHW) to off-peak hours
Typical Commercial Building Load Profile (Summer Weekday)
NREL ComStock commercial building simulation data; EIA Commercial Buildings Energy Consumption Survey (CBECS) 2018; EPRI demand charge analysis.
The 15-minute rule: Most US commercial utilities measure peak demand as the highest average power draw over any 15-minute interval in the billing month. A single brief spike — a rooftop HVAC unit starting on the hottest afternoon of the year while the kitchen exhaust fans, elevators, and lighting are all at full load — can set the demand charge for the entire month. This asymmetry is why peak shaving has such a high return on investment: preventing one spike per month eliminates $5–$50 per kW of that spike, every month.
How Demand Charges Work
A commercial electricity bill typically has two main components:
| Charge Type | Unit | What It Measures | Typical Range |
|---|---|---|---|
| Energy charge | $/kWh | Total electricity consumed (area under the load curve) | $0.06–$0.20/kWh |
| Demand charge | $/kW/month | Peak power draw during any 15-min interval in the month | $5–$50/kW/month |
| TOU energy charge | $/kWh (by period) | Energy consumed during on-peak, mid-peak, off-peak windows | On-peak: 2–3× off-peak |
| Ratchet clause | % of annual peak | Minimum demand charge set by highest month in prior 12 months | 50–100% of annual peak |
Ratchet clauses are especially punishing: many industrial and large commercial tariffs bill a minimum demand charge equal to 50–100% of the highest demand recorded in the prior 11 months, regardless of actual current demand. A single anomalous spike in July can affect bills through June of the following year.
For a 500 kW peak demand building paying $15/kW/month, the demand charge alone is $7,500/month — $90,000/year. A system that reduces peak demand by 100 kW saves $18,000/year before any energy cost savings are counted.
Demand Charge as Share of Total Bill by Customer Class
Rocky Mountain Institute; LBNL "Demand Charges and the Adoption of Distributed Energy Resources" 2016; EIA Form EIA-861; utility tariff analysis by customer class and climate zone.
Demand Charge Rates by US Utility ($/kW/month, large commercial)
NREL "Demand Charge Rate Database" 2024; Utility tariff filings. Rates shown are blended all-in demand charges; some utilities have separate transmission and distribution demand components. Hawaiian Electric has highest due to island grid costs; Southwest utilities high due to summer AC cooling peaks.
Why HVAC Drives the Peak
Heating, ventilation, and air conditioning systems are the dominant driver of commercial building peak demand for two reasons that align perfectly to create a double problem:
- HVAC runs hardest on the hottest days — when outdoor temperatures are highest, compressor efficiency drops (coefficient of performance falls) and run-time increases. The hottest afternoon of the year is usually the day with the highest building peak.
- HVAC coincides with occupancy peak — maximum cooling load from outdoor heat, internal gains (people, equipment, lighting), and solar heat gain all peak simultaneously in mid-afternoon, exactly when buildings are most occupied.
A typical office building profile shows HVAC consuming 50–60% of total power during the 2–5 PM peak window on summer weekdays. For a building trying to shave its demand peak, HVAC is the primary target — but also the most complex, because comfort cannot be sacrificed.
Compressor inrush current: Beyond steady-state demand, large rooftop units and chillers draw 3–8× their rated current for a fraction of a second when starting. Building energy management systems (BEMS) use "soft-start" controllers and compressor sequencing to stagger starts and prevent inrush spikes from creating demand charge events.
HVAC Load Breakdown (Typical Office, Summer Peak)
ASHRAE 90.1 energy standard reference buildings; DOE EnergyPlus simulation database; EPRI commercial building DR potential study 2020.
HVAC Peak-Reduction Strategies (Without Compromising Comfort)
| Strategy | How It Works | Peak Reduction | Best For |
|---|---|---|---|
| Pre-cooling / thermal mass pre-charge | Cool the building to 68–70°F before the peak window (6–10 AM); then allow temperature to float to 74–76°F during the 2–5 PM peak. Thermal mass of slab and contents absorbs heat, delaying HVAC restart. | 15–35% of HVAC peak demand | Buildings with high thermal mass (concrete, masonry); moderate climates |
| Ice storage / thermal energy storage (TES) | Chiller makes ice during off-peak hours (overnight). During the day peak, melt ice to cool the building — compressor off or at minimum during demand peak. | 40–80% of chiller demand during peak | Large commercial, hospitals, data centres with central chillers |
| Variable frequency drives (VFDs) on fans/pumps | Reduce fan and pump motor speed proportional to actual load. Fan power ∝ speed³ — a 20% speed reduction cuts power by ~50%. | 10–30% of HVAC auxiliary demand | Any building with central air handling units and chilled water loops |
| Demand-controlled ventilation (DCV) | CO₂ sensors modulate outside air intake based on actual occupancy. During low-occupancy periods (lunch, pre-occupancy), reduce ventilation and HVAC load. | 5–15% of HVAC energy; peak reduction modest unless timed to peak window | Conference rooms, restaurants, gymnasiums with variable occupancy |
| Compressor staging & soft-start | Stage multiple compressors sequentially rather than simultaneously. Soft-start controllers limit inrush. Prevent multiple large units from starting in the same demand interval. | 5–20% of peak demand (prevents spikes) | Rooftop units; multi-chiller plants; large commercial/industrial |
| Solar-integrated HVAC (see next tab) | On-site solar PV offsets compressor electricity during peak coincident solar hours; battery stores excess to cover late-afternoon demand when sun drops. | 20–60% of HVAC-driven demand charge | High solar resource locations; buildings with roof area; hot climates |
| Evaporative pre-cooling | Spray water on condenser coils or use indirect evaporative cooling to lower entering air temperature. Reduces compressor work by 5–15% on hot dry days. | 5–15% compressor demand reduction | Dry climates (Southwest US, Mediterranean, Australia) |
Why Solar and HVAC Peaks Align
Commercial solar PV generation and HVAC cooling demand have a natural alignment that makes solar an especially effective peak-shaving tool for buildings:
- Solar peaks around noon–2 PM — peak solar irradiance occurs 1–3 hours before the typical commercial building peak demand window (2–5 PM)
- Hot sunny days = high solar output AND high cooling load — the days when HVAC works hardest are precisely the days when rooftop solar is most productive
- Summer peak alignment — utility peak demand months (July–August in hot climates) match peak solar production months
- Battery bridges the 2-hour gap — solar generates heavily at noon; battery stores excess; battery discharges during 2–5 PM demand peak when solar output begins falling but cooling demand remains high
Solar "self-consumption" value: For peak-shaving purposes, what matters is not just whether solar reduces energy consumption (kWh), but whether it reduces peak demand (kW). Rooftop solar that exactly offsets the HVAC load at peak can have 3–5× the effective value compared to simple energy offsetting, because it reduces both the energy charge and the demand charge simultaneously.
Solar + Battery Peak Shaving — Load Profile Comparison
NREL SAM (System Advisor Model) simulation; DOE SunShot initiative commercial solar case studies; LBNL "Solar + Storage" economics 2023.
System Design for Peak Shaving — Key Sizing Principles
| Component | Sizing Consideration | Typical Ratio | Design Tip |
|---|---|---|---|
| Solar PV array | Roof area, shading, tilt; goal to cover HVAC load during solar hours | 1–1.5 kW solar per kW of HVAC cooling capacity | West-facing panels extend generation into late-afternoon peak window; dual-axis trackers add 15–25% yield |
| Battery (BESS) energy | How many kWh needed to cover the demand peak window | 2–4 hours of peak HVAC load offset | Size for the target demand reduction × duration of typical peak event; oversizing wastes capital |
| Battery power | Maximum discharge rate (kW) must equal or exceed target demand reduction | Battery C-rate typically 0.25–0.5C for peak shaving | 0.25C means 4-hour discharge; 0.5C means 2-hour; choose based on utility demand window length |
| Inverter | Must handle both solar input and battery charge/discharge; bidirectional for BESS | Inverter ≥ 1.1× peak solar output | Hybrid inverters (solar + battery) simplify installation; separate string inverters + battery inverter give more flexibility |
| HVAC demand target | Set a demand limit (kW setpoint) in the building energy management system | Reduce peak by 15–30% of baseline | Aggressive targets increase savings but risk comfort violation on extreme days; include weather forecast in controller logic |
| Peak detection algorithm | 15-minute interval demand tracking with rolling forecast | React 5–10 min before interval closes | Machine learning models that predict demand 30–60 min ahead outperform simple threshold-based controllers by 10–20% |
Net Energy Metering (NEM) interaction: Where NEM rates are low (e.g., California NEM 3.0 export rate ~$0.04/kWh), it is economically better to store excess solar in the battery and dispatch it during the demand peak than to export it to the grid. This fundamentally changes optimal battery dispatch logic compared to states with higher export rates.
Building Energy Management Systems (BEMS)
A BEMS (also called a Building Automation System, BAS, or Energy Management and Control System, EMCS) is the central nervous system for peak shaving. It monitors real-time power consumption, controls HVAC setpoints and schedules, and manages battery dispatch — all to keep the building's peak demand below a target setpoint.
Core BEMS functions for peak shaving:
- Real-time demand monitoring — pulse counter or CT-clamp on main breaker; rolling 15-min interval tracking; alarm when approaching target
- Load shedding sequences — pre-programmed priority order for curtailing non-critical loads (parking lighting, supplemental cooling, plug load circuits)
- HVAC setpoint adjustment — raise cooling setpoint by 1–2°F during demand events; pre-condition building 1–2 hours prior
- Battery dispatch control — charge battery during off-peak; discharge when demand approaches limit; integrate weather forecasts for proactive dispatch
- Demand response integration — auto-respond to utility OpenADR signals; report response performance
Energy Management Technology Stack
| Layer | Technology | Function |
|---|---|---|
| Sensing | Smart sub-meters; CT clamps; IoT energy loggers; revenue-grade interval meters | Real-time power at circuit, panel, and equipment level; 1–15 sec resolution |
| Control | BACnet/Modbus DDC controllers; programmable logic controllers (PLCs) | Direct equipment control — chiller staging, AHU setpoints, VFD speeds |
| Supervisory | BEMS/BAS (Siemens Desigo, Honeywell Niagara, Johnson Controls Metasys) | Demand management algorithms; scheduling; setpoint optimisation across all HVAC zones |
| Analytics | Energy analytics platforms (Energy Toolbase, AutoGrid, Stem, Leap) | ML demand forecasting; battery dispatch optimisation; utility bill analysis; DR programme management |
| Grid Integration | OpenADR 2.0b virtual top node (VTN) / virtual end node (VEN) | Utility demand response signals; automated curtailment; performance verification |
| EV Charging | Smart EVSE (ChargePoint, EVgo, EV Connect) | Managed charging to prevent EV charging peaks; V2G for peak discharge |
Demand Response Programmes — Revenue from Flexibility
Beyond reducing their own bills, buildings with smart load control and storage can earn revenue from utilities and grid operators by participating in demand response (DR) programmes:
| Programme Type | Who Pays | Typical Payment | Frequency | Best Asset |
|---|---|---|---|---|
| Emergency DR (EDR) | Utility / ISO | $50–$400/MWh curtailed + capacity payment | 5–20 events/year; highest-need grid days only | HVAC with pre-conditioning; any dispatchable load |
| Economic DR | ISO (day-ahead market) | Real-time energy price (LMP) for curtailment | Daily bid into day-ahead and real-time markets | Battery storage; industrial flexible load |
| Capacity market participation | ISO (PJM, ISO-NE, CAISO) | $50–$200/kW-year for committed capacity reduction | Annual contract; called during capacity scarcity events | Large commercial/industrial with reliable response |
| Frequency Regulation (FFR) | ISO (ERCOT, PJM) | $5–$30/MW/hour plus performance score | Continuous; fast response in seconds | Battery storage only (HVAC too slow) |
| Utility direct load control (AC cycling) | Utility | $5–$15/month bill credit per unit enrolled | 10–25 events/year in summer | Residential/small commercial HVAC thermostats |
Aggregated virtual power plants: Individual buildings with small batteries or controllable HVAC can be aggregated by demand response aggregators (AutoGrid, Enel X, CPower, Voltus) into virtual power plants large enough to participate in wholesale electricity markets. A 1 MW aggregate of 50 buildings each contributing 20 kW of controlled HVAC and battery capacity can earn $50,000–$200,000/year in grid services revenue, shared among participants.
Commercial Peak Shaving Case Studies
| Building / Project | Location | System | Result | Payback |
|---|---|---|---|---|
| Walmart 100-store solar + BESS rollout | US (multiple states) | 500–1,000 kW solar + 250 kWh Tesla Powerpack per store; peak demand setpoint control | 15–25% demand charge reduction; $30,000–$60,000/year savings per store | ~6 years |
| Arizona State University, Tempe Campus | Phoenix, AZ | 7 MW rooftop solar + 2.5 MWh battery; central plant thermal storage; BAS demand management | 30% reduction in peak demand; $1.2M/year in demand charge avoidance | 5.5 years |
| LAX Airport Microgrid | Los Angeles, CA | 8 MW solar + 4 MWh BESS + demand response; HVAC pre-cooling on weather forecast signals | Peak demand reduction of 12 MW during summer events; resilience island capability | 7 years (includes resilience value) |
| Whole Foods Market (flagship stores) | Multiple US | 200–400 kW solar + ice storage for display case cooling + demand-controlled refrigeration | 20–35% demand charge reduction; 15% total energy cost savings | 4–6 years |
| Kohl's department stores (100+ locations) | US national | Rooftop solar PV (averaging 300 kW per store); smart thermostat controls; demand response enrollment | $5M+/year utility savings across portfolio; 40% of electricity from solar at enrolled stores | 6 years portfolio average |
| Singapore Changi Airport | Singapore | 1.2 MW floating solar + chilled water thermal storage + AI-optimised chiller dispatch | 22% reduction in cooling energy; 18% demand reduction during peak hours | 8 years |
| Brooklyn Microgrid (ConEd + Swell Energy VPP) | New York, NY | Aggregated 500+ residential and small commercial solar + battery systems; DERMS orchestration | 2 MW aggregate demand response capacity; $400/kW-year capacity payment to participants | Varies by participant; ~5–7 years for solar + battery |
10-Year Economics — 500 kW Peak Building, Solar + Battery
NREL PVWatts + SAM financial model; EIA commercial electricity tariff data; BloombergNEF BESS cost curves; DOE ITC/IRA incentive schedules (30% ITC through 2032). Assumptions: $15/kW/month demand charge; $0.12/kWh energy; 300 kW solar + 200 kWh BESS; Phoenix, AZ climate.
Solar + BESS Cost Trends ($/kWh installed, commercial)
BloombergNEF BESS Price Survey 2024; NREL "U.S. Solar Photovoltaic System and Energy Storage Cost Benchmarks" 2024; Wood Mackenzie storage market outlook 2025. Commercial system costs include hardware, inverters, BOS, installation, and soft costs but exclude ITC incentives.