Question 6 of 8: Refrigerant selection and plant strategy for a 1970s hospital chiller

Question 6: Refrigerant selection and plant strategy for a 1970s hospital chiller (20 points)

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) Refrigerants available for a new system. R-12 has not been manufactured in Canada since 1 January 1996 under the Montreal Protocol, and the hydrochlorofluorocarbons that first replaced it — R-22 and R-123 — are themselves through their phase-out, with R-22 import and manufacture ended in 2020 and R-123 permitted only for servicing existing equipment. The field for a new hospital chiller is therefore divided between the synthetic and the natural refrigerants. Among the synthetics, R-134a remains widely installed in centrifugal and screw machines and has zero ozone depletion potential, but its global warming potential of about 1,430 puts it squarely inside the Kigali Amendment hydrofluorocarbon phase-down that Canada implements through the Ozone-depleting Substances and Halocarbon Alternatives Regulations; it is a defensible choice for a fifteen-year horizon, not a thirty-year one. The current generation is the unsaturated hydrofluoroolefins — R-1234ze(E) and R-1234yf — and the blends built from them, such as R-513A and R-514A, all with GWP below 10 in the pure fluids and around 600 in the blends, and all with zero ODP. Among the natural refrigerants, ammonia (R-717) has zero ODP and zero GWP and the best thermodynamic performance of any candidate; carbon dioxide (R-744) is attractive where heat recovery at high temperature is wanted; and water (R-718) is used in large centrifugal machines where the low pressure can be accommodated. Hydrocarbons are excluded here by the occupancy: a hospital is an institutional building where CSA B52 restricts flammable refrigerants sharply.

(b) Recommendation for a thirty-year horizon. The system that will still be legal and economic in 2045 is one whose refrigerant is not on any phase-down schedule. For a hospital of this size the recommendation is a central water-cooled plant using either ammonia in a packaged indirect arrangement or an HFO such as R-1234ze(E) in a centrifugal machine, with the choice turning on the plant-room situation. Ammonia is the stronger long-term answer thermodynamically and environmentally, but it is a B2L refrigerant — toxic and mildly flammable — so CSA B52 requires a dedicated machinery room with gas detection, emergency ventilation and a discharge to a safe location, and the ammonia must be confined to that room with a secondary glycol or chilled-water loop distributing cooling to the building. In a hospital, where the plant room is often deep in the building and adjacent to occupied space, that requirement frequently decides against ammonia, and an HFO centrifugal chiller with a magnetic-bearing oil-free compressor becomes the practical recommendation. Either way the plant should be specified with multiple chillers for redundancy and turndown — a hospital cannot lose cooling to its operating suites, imaging equipment or isolation rooms — with a water-side economiser for the long Canadian shoulder season, and with condenser heat recovery feeding domestic hot water preheat and reheat coils. Given the expansion driving the project, the plant should also be arranged so a further module can be added without shutting the existing machines down.

(c) Coefficients of performance. Comparing systems on COP is only meaningful at stated conditions, but the ranking is stable. A modern water-cooled centrifugal chiller achieves a full-load COP of about 6.0 to 6.5 (roughly 0.55 kW per ton) and, because centrifugal machines unload well against a falling condenser water temperature, an integrated part-load value substantially higher still; a water-cooled screw or scroll machine sits at 4.5 to 5.5; an air-cooled chiller, penalised by the higher condensing temperature, falls to 2.8 to 3.5. Absorption plant is in a different class because it consumes heat rather than work: a single-effect lithium-bromide machine driven by low-pressure steam returns a COP of only 0.6 to 0.7 and a double-effect machine 1.1 to 1.3, so absorption is chosen for its ability to use waste or district heat, never for its efficiency. The comparison must be made on primary energy: an absorption chiller at COP 1.2 driven by district heat that is itself a by-product of electricity generation can beat an electric chiller at COP 6.0 supplied from a thermal grid, while on the largely hydroelectric grids of British Columbia, Manitoba and Quebec the electric chiller wins decisively on both cost and emissions.

(d) District heating. District heating is the centralised production of heat at one plant and its distribution as hot water or steam through a buried piping network to a group of buildings, each of which takes heat through a substation heat exchanger rather than operating its own boilers. The central plant may burn natural gas or biomass, recover heat from a waste-to-energy facility or an industrial process, run large heat pumps on sewage or lake water, or operate as a combined heat and power plant generating electricity with the rejected heat sold into the network. Canadian examples include the Southeast False Creek Neighbourhood Energy Utility in Vancouver, which recovers heat from raw sewage, and the Enwave systems in Toronto and Windsor. The economic case rests on the diversity of the connected load, on the higher efficiency and better emissions control achievable in one large plant than in many small ones, and on the ability to change the heat source once without touching any of the connected buildings.

(e) Winter heating if district heating is adopted. Connecting the hospital to a district system replaces the boiler plant with a substation: a plate heat exchanger, control valves, an energy meter and the isolation required by the utility. The building's own hydronic distribution is largely unaffected, though the design return temperature becomes critical, because the utility charges on energy and its network economics depend on a large temperature difference across the service. Terminal equipment sized for a small ΔT — typical of older converted plant — will need larger coils or lower flow to return water cold enough. The advantages for a hospital are considerable: the boiler room, its flues, its fuel supply and its combustion-safety obligations disappear, which frees floor area and removes a maintenance and licensing burden, and the reliability of the network is usually higher than that of two building boilers. The counterweights are the loss of control over the heat price, dependence on a single supplier, and the need for a standby source, since a hospital is a post-disaster building under the National Building Code and cannot be left without heat if the network fails. In practice the existing boilers are retained as standby, and the substation is sized for the full winter peak with the summer domestic-hot-water load also taken from the network so the connection is used year-round.

(f) Maintaining the existing R-12 plant if it is not changed. Continuing to run the machine is legal in Canada but tightly constrained. No new R-12 may be manufactured or imported, so the only supply is reclaimed refrigerant recovered from decommissioned equipment, and its price and availability both make leakage the dominant operating risk. The maintenance programme therefore centres on containment. The Federal Halocarbon Regulations require that any leak be repaired or the system shut down and drained, that systems above the threshold charge be leak-tested at prescribed intervals, that service be performed only by certified technicians, that recovered refrigerant never be vented, and that a service log recording every charge and recovery be maintained for inspection. Practically, this means an annual leak test with an electronic detector, attention to the shaft seal and to flanged joints, a purge unit whose run time is trended as a leak indicator, and vacuum and moisture control since R-12 systems with mineral oil are intolerant of water. Beyond containment, the plant should be kept efficient: clean condenser and evaporator tubes annually, monitor the approach temperatures as a fouling indicator, verify the purge and the oil analysis, and keep the chilled and condenser water treatment programme current. Retrofitting to a service blend such as R-401A, R-409A or R-413A is possible and buys time, but it requires attention to oil compatibility — most blends need an alkylbenzene or polyol-ester oil in place of the original mineral oil — and to elastomer swelling in the seals, and it typically costs 3 to 8 % of capacity. Given that the plant is already more than forty years old and the hospital is expanding, retrofit is a bridging measure, not an alternative to replacement.

(g) Ozone depletion potentials. The values below are the Montreal Protocol ODPs, referenced to R-11 = 1.0. Every refrigerant recommended in parts (a) and (b) has an ODP of exactly zero, because none of them contains chlorine or bromine; what distinguishes them now is global warming potential, which is why both are tabulated.

Taken together, the seven parts describe a single decision. The plant is at the end of a technological era rather than merely at the end of its service life: its refrigerant is banned, its replacements are themselves being phased down, and the expansion that prompted the discussion is the natural moment to move to a fluid and a plant configuration that will survive the regulatory horizon. The engineering recommendation is to replace rather than retrofit, to choose a zero-ODP, low-GWP refrigerant, to build in redundancy appropriate to a hospital, and to evaluate a district-energy connection for the heating side on the same thirty-year basis.