Team course project (6 people)

Heat pump or adiabatic cooler? Building cooling across three climates

Université de Technologie de Compiègne · BA01 Building Technical Equipment (spring 2025)

A 44 × 23.5 × 6 m office building is placed in Compiègne, Marseille and Fort-de-France to compare air-to-air heat pumps, air-to-water heat pumps and a direct adiabatic cooler on energy, water, life-cycle impact and annualised cost. In dry heat the adiabatic cooler uses about a third of the electricity, but it fails above 75% humidity and has the highest water use.

Background

This was the BA01 “Building Technical Equipment” course project at UTC, completed by a team of six in spring 2025. The brief started from a concrete tension: cooling demand rises with heatwaves, while air conditioning and heat pumps themselves push up electricity load and carbon emissions. The question was not simply “which device saves more electricity”, but under which climate and water conditions each option is actually reasonable.

The case study is a two-storey office building with RT2012-level insulation, a 44 × 23.5 m footprint and a height of 6 m, giving a volume of 6,204 m³. The same building is placed at three virtual sites: Compiègne (temperate oceanic climate, summer relative humidity around 70%), Marseille (Mediterranean, around 59%) and Fort-de-France (tropical, around 77%). Operating hours are the weekdays when the outdoor temperature exceeds 25 °C, from 11:00 to 17:00.

What is compared

Three systems are compared: an air-to-air heat pump (EER 4.03), an air-to-water heat pump (EER 4.56) and a direct adiabatic cooler (EER 12.5). The EER values come from manufacturer datasheets.

The cooling load is estimated with a simplified formula:

P_cooling = V · C · ΔT

where V = 6,204 m³, C = 0.75 (recent building envelope) and ΔT is the temperature drop the adiabatic cooler can actually achieve in each climate (4 °C in Marseille, 3 °C in Compiègne, 4 °C in Fort-de-France). Annual cooling energy is the load times the number of operating days and 6 hours per day; electricity is cooling energy divided by EER. Primary energy uses the French coefficient CEP = 2.3.

Water has two components: the adiabatic cooler consumes 1.52 m³ per MWh of cooling directly, while every system also carries the water used to generate electricity, weighted to the French 2024 mix at 2.2 L per kWh of primary energy. The life-cycle assessment uses the default FDES declarations from the Inies database, normalised to a 22-year functional unit. Cost uses the EDF blue tariff of €0.2026/kWh and an average water price of €4.34/m³.

Energy: first ask whether the system can deliver

SiteOperating daysAnnual coolingAir-to-air HPAir-to-water HPAdiabatic
Marseille (ΔT = 4 °C)106 days11,826 kWh2,935 kWh2,594 kWh946 kWh
Compiègne (ΔT = 3 °C)41 days3,434 kWh852 kWh753 kWh275 kWh
Fort-de-France (ΔT = 4 °C)45 days5,022 kWh1,246 kWh1,101 kWh—

The ranking does not change when converted to primary energy: 6,751 / 5,966 / 2,176 kWh in Marseille, and 1,960 / 1,732 / 633 kWh in Compiègne.

The adiabatic cooler uses about one third of the electricity of the air-to-air heat pump and 36% of the air-to-water unit — but only if it can reach the required temperature drop. Direct adiabatic cooling depends on outdoor humidity: at 59% relative humidity it can deliver at most 7 °C, at 70% only 3 °C, and at 77% just 1 °C. Fort-de-France is therefore excluded: the system does not fail because it is inefficient, but because it cannot cool enough. The air-to-air heat pump, by contrast, can deliver 10 °C, at the cost of the highest electricity use of the three.

Water: the price of the adiabatic advantage

SiteAir-to-air HPAir-to-water HPAdiabatic (direct + indirect)
Compiègne10.26 m³9.07 m³0.96 + 6.27 = 6.87 m³
Marseille2.97 m³2.63 m³3.31 + 21.57 = 24.88 m³
Fort-de-France4.36 m³3.85 m³—

Two situations emerge. In Compiègne the adiabatic cooler actually uses less total water than either heat pump, because its very low electricity use offsets the direct evaporation. In Marseille the opposite happens: 21.57 m³ of direct water pushes the total to 9.5 times the air-to-water heat pump. Placed against the building’s total water use (about 5,110 m³/year, assuming 140 people at 100 L/day), the maximum additional demand is around 0.5%. The share is small, but it falls precisely in the summer period when water is already scarce.

Life cycle: manufacturing and end of life

Normalised to a 22-year functional unit, the adiabatic cooler’s manufacturing, transport and disposal impacts are far below those of the heat pumps:

IndicatorAir-to-air HPAir-to-water HPAdiabatic
Climate change (kg CO₂eq)20,90014,700267
Abiotic depletion, minerals and metals (kg Sb eq)2.372.160.00199
Non-renewable primary energy (MJ)96,900115,0003,420
Net freshwater use (m³)1261531.15
Non-hazardous waste (kg)15,60015,70091.2

The adiabatic cooler’s carbon footprint is about 1/78 of the air-to-air heat pump, and the air-to-water unit is the highest on most waste and resource indicators. Two caveats matter: the figures come from “default FDES” data, which tends to be conservative, and the adiabatic FDES records zero water use during operation, which contradicts its actual consumption. The table is therefore useful for orders of magnitude, not for a precise comparison.

Cost: the same order of magnitude

Amortising the purchase price over the service life and adding electricity and water:

SystemAnnualised total cost
Air-to-air heat pump€3,214 – 5,591 / year
Air-to-water heat pump€4,579 – 7,312 / year
Adiabatic cooler€6,755 – 7,132 / year

The air-to-air heat pump is the cheapest, and the adiabatic cooler sits at the upper end of the air-to-water range, but all three remain within the same order of magnitude. Cost is not the main obstacle to deploying adiabatic cooling; climate and water resources are.

Conclusion: no global optimum

  • Compiègne (temperate, moderate humidity, low water stress): the adiabatic cooler is acceptable on energy, life cycle and cost, and is the option to prioritise.
  • Marseille (hot and dry, best cooling performance, but summer water scarcity): the adiabatic cooler is technically the best, yet it would worsen local water stress — the Rhône-Mediterranean-Corsica water agency reports a summer deficit of about 35 million m³, so large-scale deployment needs caution.
  • Fort-de-France (hot and humid, fragile supply network): humidity makes adiabatic cooling ineffective, and the air-to-water heat pump is the safer choice.

Climate projections move the conclusion further: mainland France is expected to warm by about 4 °C by 2100, with twice as many summer droughts and about 20% less rainfall. Places that suit adiabatic cooling today will become rarer, while places that need cooling with even less water will become more common. The report’s recommendation is therefore hybrid systems (adiabatic + heat pump) and passive measures — shading, green roofs, ground cooling — rather than betting on a single device.

Limitations

  • Only direct adiabatic cooling is compared; indirect and hybrid systems lack standardised operating data, so an equivalent energy and LCA comparison was not possible.
  • The FDES data are default values, and the adiabatic declaration records zero operational water use; life-cycle results should be read as orders of magnitude.
  • The building is simplified to a regular rectangular block, ignoring thermal inertia, occupancy schedules, internal gains and part-load behaviour.
  • The indirect water factor is based on the French electricity mix; another grid would change the results significantly.
  • Operating days come from typical Météo France records, not from metered operation.

Related skills

  • HVAC
  • Energy analysis
  • Life-cycle assessment
  • Excel
  • Academic writing