Question 7 of 8: Comparing heating and cooling options for a downtown hospital

Question 7: Comparing heating and cooling options for a downtown hospital (20 marks)

Question text not reproduced: the examination questions are © Engineers and Geoscientists BC. Open the official past paper (linked at the top of this page) to read the question, then follow the worked solution below.

A hospital is the hardest building type this list can be applied to, and the reason should frame every comment that follows. It runs continuously, it cannot be shut down for a plant changeover, its cooling load never falls to zero because of the operating theatres and imaging suites, and it is a post-disaster building under the National Building Code of Canada, so resilience and fuel diversity carry weight that they would not carry in an office. A downtown site adds two more constraints: there is no room for a large plant or for a cooling-tower field, and the neighbours will notice noise and plume. The comments below are ordered as the question lists them.

Vapour compression using R22. This is not an option for new equipment, and the comment is a regulatory one before it is a technical one. R22 is an HCFC with an ozone depletion potential of 0.055 and a global warming potential of about 1810; under the Montreal Protocol and Canada's Ozone-depleting Substances and Halocarbon Alternatives Regulations (SOR/2016-137) it can no longer be imported or manufactured for new systems, and servicing depends on a shrinking stock of recovered gas whose price has risen by an order of magnitude. An existing R22 chiller may be run out to the end of its life with careful leak management, but specifying one in a revamp would strand the asset. Technically it is a good refrigerant — high volumetric capacity, modest discharge temperature — which is exactly why its withdrawal was so disruptive; the lesson for the hospital is that refrigerant selection is now a regulatory risk decision, not only a thermodynamic one.

Vapour compression using R134a. Zero ozone depletion potential and a mature technology base, and in large centrifugal machines it delivers the best electrical efficiency available — a modern variable-speed centrifugal reaches a full-load COP near 6.1 and integrated part-load values higher still, which no thermally driven machine approaches. Capital cost is moderate, the plant is compact, and it suits a downtown site. The reservation is carbon: R134a has a global warming potential of 1430, and it is being phased down under the Kigali Amendment, which Canada implements through a declining HFC consumption cap. A hospital chiller specified today will still be running when the refrigerant is scarce, so the correct specification is a machine designed for, or convertible to, a low-GWP alternative — R1234ze (GWP 7) or R513A — with a leak-detection and record-keeping regime to match. Electricity supply is the other issue: the whole load lands on the electrical service and on the emergency generators, which is a real capital cost in a post-disaster building.

Absorption chiller on district steam. This is the option that best suits a downtown hospital that already has a steam service. A single-effect machine driven by low-pressure steam has a COP near 0.7; a double-effect machine on medium-pressure steam reaches about 1.2, a 71 % improvement that is well worth the higher first cost when the machine runs continuously. The attractions for a hospital are specific and strong: the electrical demand for cooling almost vanishes, which relieves the service and the generators; there are no large rotating machines, so noise and vibration are low; the refrigerant is water and the absorbent lithium bromide, so refrigerant regulation is irrelevant; and the plant makes productive summer use of a steam service that would otherwise sit idle, improving the district system's own load factor. The costs are a larger machine for the same duty, roughly double the heat rejection per ton, and hence a bigger cooling tower — awkward on a constrained site — plus a machine that dislikes rapid load swings and can crystallise if the cooling water runs too cold without controls.

Absorption chiller on natural gas. Thermodynamically the same machine with its own direct-fired generator, COP around 1 for a double-effect unit. It buys fuel independence from the district network and can be an excellent peak-shaving partner to an electric chiller when the electricity tariff has a high summer demand charge. But burning gas to make chilled water at a COP near 1 is poor use of a high-quality fuel when the same gas in a power station and a good centrifugal would deliver several times the cooling; the carbon intensity per ton-hour is correspondingly higher than the steam-driven version, which is recovering heat that has already done work. In a hospital it also adds a combustion appliance, a flue and a fuel train to a building where every additional hazard has to be justified. Reasonable as a supplement, weak as the primary plant.

District heating and cooling. For a downtown hospital this is often the strongest single option on total cost of ownership. It removes the boiler plant, the chillers and the cooling towers from the building, freeing valuable floor area and the roof, and it transfers maintenance, redundancy and fuel risk to an operator who carries them across many buildings and can therefore afford diversity, thermal storage and large efficient equipment. Diversity between connected buildings means the installed capacity per building is lower than the sum of the peaks. The counterweights are equally clear: the hospital surrenders control of its supply, is exposed to a single provider's tariff and to whatever fuel the central station burns, and is tied to a long-term contract; a single distribution failure affects every connected building at once. For a post-disaster building, connecting to the district system should therefore be paired with retained on-site standby capacity — typically boilers and one chiller — sized for the critical departments rather than the whole building.

Cogeneration or trigeneration. A hospital is close to the ideal cogeneration host, because the economics turn on the number of hours the recovered heat can be used and a hospital has a year-round demand for domestic hot water, sterilisation steam, humidification and reheat. A reciprocating or turbine prime mover generating on site delivers electricity at perhaps 0.36 efficiency, but recovering the jacket and exhaust heat lifts the overall fuel utilisation to about 0.8 — more than double what a condensing central station achieves, and the source of both the cost saving and the carbon saving. Trigeneration adds an absorption chiller driven by that recovered heat, which converts the summer heat surplus into cooling and so keeps the machine running at high utilisation all year; that is what makes the capital pay back. It also confers genuine resilience, since the prime mover can be arranged to carry critical load in an outage, which is worth a great deal in a post-disaster building. The reservations are that the capital cost and complexity are the highest on this list, that the economics depend on the spark spread between gas and electricity prices and can be inverted by a tariff change, that a downtown site must accommodate emissions and noise permitting, and that the plant only makes sense if it is sized to the heat demand and allowed to run at high load factor, not sized to the electrical peak.

Recommendation. On the evidence above, a downtown Canadian hospital revamp is best served by connecting to district heating and cooling where a network exists, retaining on-site standby plant for the critical departments, and evaluating trigeneration against that baseline if the site can host it and the spark spread supports it. Where new vapour compression is needed, specify a low-GWP refrigerant rather than R134a and never R22. Whatever is chosen, the decision should be made on life-cycle cost and life-cycle carbon over a 25-year horizon, with the phase-down schedules for the refrigerant written into the assumptions, because in this field the regulatory trajectory is more predictable than the fuel price.