Buildings on Ice: How Thermal Energy Storage Cuts Peak Cooling Costs

Some of the most valuable efficiency features in a green building hide below grade. One Bryant Park, the 2.1 million square foot (195,000 square meter) office tower in New York City that pursued LEED Platinum, drew attention for its rainwater harvesting, onsite wastewater treatment, optimized daylighting, and combined heat and power plant. The feature that pulls engineers down to the sub-basement is quieter: 44 insulated tanks, each holding more than 1,600 gallons (6,100 liters) of water, where ice is made every night and melted every day to help cool the building. The tanks provide about a quarter of the tower’s total cooling.

The same ice-powered cooling approach is spreading far beyond flagship projects. Thermal energy storage (TES) freezes water during off-peak hours and releases the stored cooling when demand spikes, and the technology keeps appearing in utility programs, design guides, and efficiency standards because the economics and the engineering both hold up.

How Thermal Energy Storage Works

Thermal energy storage banks cooling capacity in a tank of water during one period and spends it later. The physics is simple: freezing one pound of water absorbs about 144 Btu of heat, the latent heat of fusion, and melting releases that energy back. A chiller running at night makes ice inside the tanks; during the day, warm water or glycol from the building loop returns to the tanks, melts the ice, and drops back to supply temperature without the chillers carrying the full load.

The tanks at One Bryant Park show the geometry in practice. Each 8 foot diameter, 8.5 foot tall (2.4 by 2.6 meter) insulated tank holds more than 1,600 gallons (6,100 liters) of water and three miles (4.8 km) of plastic tubing. About 150 gallons (570 liters) of glycol solution circulate through that tubing, picking up cold from the ice and delivering it to the cooling loop.

The Daily Charge and Discharge Cycle

  1. Night charge. Chillers run on off-peak electricity and freeze water in the tanks, building an ice bank over six to eight hours.
  2. Morning transition. Controls switch from charging to cooling mode as occupancy and internal loads ramp up.
  3. Day discharge. Warm glycol from the building loop passes through the tank tubing, melts ice, and returns chilled to the cooling coils.
  4. Trim mode. If the ice bank runs low on a hot afternoon, the chillers top up the remaining load.
  5. Recharge. The cycle repeats the next night, with ice thickness sensors preventing overcharging.

Ice Versus Chilled Water Storage

Water can store cooling as sensible heat, a temperature drop, or as latent heat, a phase change. Chilled water storage uses the first approach and needs roughly seven times the tank volume of ice storage for the same capacity, because melting a pound of ice absorbs about 144 Btu, more than one hundred times the heat needed to warm a pound of water by one degree Fahrenheit. That density advantage explains why ice tanks fit into existing basements and tight urban sites.

Ice is only one storage medium. The same load-shifting logic appears in a thermal storage guide that explains how to store solar energy under a concrete slab for winter heating, proof that thermal mass can carry energy across seasons, not just across a day.

AttributeIce storageChilled water storageConventional cooling
Storage mechanismLatent heat of fusion at 32°F (0°C)Sensible heat in a 39 to 44°F (4 to 7°C) tankNone
Volume for 1,000 ton-hoursAbout 10,000 gallons60,000 to 70,000 gallonsNot applicable
Chiller sizing30 to 50 percent below peakModerately reducedFull peak capacity
Best fitRetrofits, tight sites, demand responseNew plants with tank spaceSimple, low-cost plants
Supply temperature34 to 38°F (1 to 3°C)39 to 44°F (4 to 7°C)42 to 45°F (6 to 7°C)

Why Buildings Shift Cooling Off the Peak

Cooling is usually the single largest driver of a commercial building’s peak electric demand. Utilities bill demand charges on the highest 15 minute draw of the month, so a few hot afternoons can dominate the bill. Ice storage flattens that curve by moving chiller operation to the night hours, when rates are lower and demand is scarce.

The grid benefits as well. Night charging runs when wind and baseload generation are plentiful, and discharging during the afternoon reduces the need for expensive peaker plants. Utilities in many markets now pay for this behavior through demand response programs, and rating systems recognize it: LEED awards points for demand response and for shifting load off peak.

Ventilation and cooling loads interact in dense buildings, and the two conversations increasingly overlap. The same design teams that are making the case for ERVs in multifamily buildings, where heat recovery trims fresh-air loads, often pair those systems with peak-shaving strategies such as ice storage to keep the whole plant small.

Reading the Rate Structure

Three numbers decide whether TES pays: the on-peak demand charge, the spread between on-peak and off-peak energy rates, and the number of peak hours per day. A rate with a $10 to $20 per kilowatt demand charge and a 3 to 5 cent per kilowatt-hour energy spread justifies storage in most climates; a flat rate does not.

  • Demand charge above $10 per kilowatt: strong incentive to shave peaks.
  • On-peak window longer than six hours: full storage starts to make sense.
  • Off-peak rate below 60 percent of the on-peak rate: energy arbitrage adds up.
  • Capacity or subscription charges: reward lower connected loads.

Sizing and Design Considerations

Storage sizing starts before the plant room. Design teams cut the cooling load first by tightening the building thermal envelope with air barriers, vapor retarders, insulation, and high-performance fenestration, because every Btu of envelope gain becomes a Btu of storage capacity that someone pays for. A well-sealed, well-insulated shell can shrink the ice plant by a third before a single tank is specified.

Partial Storage Versus Full Storage

Partial storage is the common choice. The chiller is sized to 50 to 70 percent of the design peak and the ice bank shaves the difference, which cuts first cost and still captures most of the demand savings. Full storage sizes the chiller to the average load and covers the entire peak from ice, which eliminates daytime chiller operation but requires a much larger tank room.

Chiller Selection

Ice-making chillers must produce lower evaporator temperatures, typically 22 to 26°F (-6 to -3°C), which reduces their efficiency during the charge cycle. Designers weigh that penalty against lower nighttime rates and cooler ambient temperatures that improve condenser performance. Options include dedicated ice-making chillers, dual-duty machines that switch between ice mode and direct cooling, and packaged systems with internal heat exchangers that isolate the glycol loop from the building loop.

Costs, Savings, and Payback

The business case combines first cost, demand savings, and energy arbitrage. Storage adds tanks, glycol, heat exchangers, and controls to a plant, but it also lets the chiller shrink. In favorable rate markets, the added cost pays back in three to seven years, and demand response incentives can shorten that window.

Ice storage belongs on a longer menu of energy-saving technologies for buildings, and its ranking on that menu depends more on the local rate structure than on climate. A hot, humid city with steep demand charges is a better market than a mild one with flat rates.

ParameterTypical range
Chiller plant size reduction30 to 50 percent
On-peak demand reduction25 to 40 percent
Installed cost premium over a conventional plant10 to 30 percent
Simple payback in favorable rate markets3 to 7 years
Useful life of storage tanks25 years or more

Operation and Maintenance Considerations

Ice systems change the operating rhythm of a plant. The controls team now manages a charge schedule, ice thickness targets, and a discharge strategy that responds to weather forecasts, occupancy, and utility price signals. Well-tuned controls recover most of the theoretical savings; sloppy ones leave ice in the tanks at midnight.

Thermal losses are the enemy of storage. The same principles taught in any course on thermal insulation in buildings apply to the tanks, headers, and piping, which need vapor-sealed insulation to keep the ice bank cold through the afternoon. Water treatment and glycol concentration checks round out the maintenance routine.

What the Maintenance Crew Should Check

  • Ice thickness and charge duration against the schedule.
  • Glycol concentration and freeze protection each season.
  • Tank insulation integrity, including vapor seals and penetrations.
  • Pump, valve, and heat exchanger condition on the storage loop.
  • Control sequences after rate changes or load shifts.

Questions to Ask Before Installing Ice Storage

Owners evaluating a TES project should run through a short checklist before committing to a design.

  1. What does the utility rate structure reward: demand, energy, or both?
  2. How many hours per day does the building actually peak?
  3. Is there room for tanks in a basement, parking level, or yard?
  4. How old is the chiller plant, and does it support ice-making temperatures?
  5. Do local incentives or demand response programs offset first cost?
  6. Who will operate the charge and discharge controls?

Finally, plan to measure and document the results. Buildings pursuing energy performance certificates for buildings can translate verified demand reductions into a better rating, which keeps the ice plant working for the balance sheet long after installation.