Cooling load and environmental measurements in a Canadian indoor ice rink

Cooling load and environmental measurements in a Canadian indoor ice rink

M. Ouzzane


This paper presents the preliminary measurements performed in an indoor ice rink. The results include the cooling load, the air temperature variation at several locations, the ice temperature, and the brine temperature. The measured data will be used for the verification and calibration of a new numerical model that was developed for ASHRAE Research Project RP-1289.


Thousands of indoor ice rinks operating in North America are used for spectator events such as ice sports (hockey, curling, figure skating, speed skating) and ice shows. Although they have different sizes and configurations (ASHRAE 2002), all ice rinks have some common features: large dimensions (the ice surface is about 25 x 60 m (82 x 197 ft) with few or no dividing walls); significant refrigeration load; simultaneous need for heating, refrigeration, and ventilation; high energy consumption; and, finally, considerable emissions of greenhouse gases (GHG) due to energy consumption and synthetic refrigerant losses. The indoor air conditions (air temperature and humidity) and the motion of the ventilation air through these large buildings often lead to uncomfortable thermal conditions for the spectators, fog and condensation formation, as well as an increase of the refrigeration load and deterioration of the ice condition.

Lavoie et al. (2000) revealed that ice rinks in Quebec offer the potential for an annual energy savings of 1300 GWh (4436 x [10.sup.6] Btu) and for an annual reduction of GHG emissions of about 0.5 Mtons of equivalent C[O.sub.2]. According to that study, the current performance of ice rinks is poor due to the lack of new and improved design and operation guidelines and by not taking into consideration the interaction between the ice-making, ventilating, and heating systems.

The literature review revealed very limited information on ice rinks where both heating and refrigerating loads exist simultaneously. Numerical studies were undertaken in two-dimensional and three-dimensional configurations by Nielsen et al. (1979), Chen et al. (1989), Jones and Whittle (1992), and Yang et al. (2000). However, these studies do not generally take into account the interaction between convection, radiation, vapor diffusion, and mass transfer between the ice surface and the air. Zmeureanu et al. (2002) developed routines for the DOE-2 energy analysis software to simulate the different energy exchanges that occur within ice rinks. Experimental data were used to validate the simulation results. Recently, Bellache et al. (2005a) have studied the airflow, heat transfer, and mass diffusion within an indoor ice rink with a forced air heating system by using computational fluid dynamics (CFD) software. The study has shown the influence of supplied warm air on both spectators’ thermal comfort and refrigeration load is significant.

Chapter 34 of the 2002 ASHRAE Handbook–Refrigeration (ASHRAE 2002) contains limited information on the thermal loads of the ice sheet, which can be used to estimate the cooling load and, therefore, the required cooling capacity of the refrigeration system. Information is mostly based on practitioners’ experience, rules-of-thumb, approximations, and data collected in the 1970s.

ASHRAE TC 10.2, Automatic Icemaking Plants and Skating Rinks, issued a request-for-proposal to “develop and verify methods for determining ice sheet cooling loads.” The research project RP-1289 was awarded to CANMET Energy Technology Centre, Varennes, Quebec, Canada. The objectives of this research can be summarized as follows: (a) the development of a two-dimensional transient numerical model capable of simulating ice sheet heat transfer with transient operating conditions, (b) the long-term measurement of appropriate parameters to be used for the verification and calibration of the numerical model, and (c) the assessment of sensitivity of the numerical model to several design and operating parameters.

This paper presents the preliminary results of measurements performed in a Canadian indoor ice rink, while the paper by Bellache et al. (2005b) presents the numerical model.


The following variables are proposed to be measured in an existing ice rink for the verification and calibration of the numerical model:

* Total daily ice sheet cooling load

* Total ice sheet cooling load versus time, for a 24-hour schedule

* Radiant heat flow at the ice surface due to lighting fixtures

* Ice sheet load due to resurfacing versus time

* Vertical variation of indoor air temperature above the ice up to the roof at two selected locations: near the west board and near the players’ exit

* Vertical variation of air velocity and air temperature above the ice surface

* Vertical variation of dew-point temperature near the ice at a select location near the south board

* Temperature of the interior surface of the roof above lighting fixtures

* Vertical variation of interior surface temperature of the west wall

* Ice temperature at several locations across and along the ice surface, including the measurement at the blue line made with an infrared sensor

* Air temperature and relative humidity at both supply and exhaust air diffusers

* Air temperature and relative humidity supplied by dehumidifiers

Ice Rink Facility

Eight ice rinks were considered as potential candidates for the experimental program, and their features were compared according to pre-established criteria (e.g., proximity, availability of building drawings and specifications, the type of ventilation and heating system, location of supply and exhaust grilles that allows the assumption of two-dimensional air movement and heat transfer phenomena, historical data about the energy use and cost).

The selected facility, the Camillien Houde ice rink, built in Montreal in the early 1980s, is 42 m (138 ft) wide, 64 m (210 ft) long, and 9.36 m (31 ft) high, with an ice sheet of 26 x 61 m (88 x 200 ft). The ice rink is used 11 months per year. The design temperature for Montreal is -23[degrees]C (-9[degrees]F) for winter and 28[degrees]C (82.4[degrees]F) dry bulb with 21[degrees]C (69.8[degrees]F) wet bulb for summer. Six rows of stands run the whole length of the building on one side and a narrow corridor encircles the ice surface. There are about 200 seats for spectators.

The primary ventilation system is a 100% exhaust/makeup system, with the capacity of 4,000 L/s (1900 cfm) that runs continuously and contains a polymer-based heat-recovery wheel. A gas-fired duct furnace maintains the minimum temperature of makeup air at 21[degrees]C (70[degrees]F), which is primarily distributed to the player dressing room. Eight natural gas high-intensity infrared heaters of a total capacity of 140 kW (480 MBtu/h) provide local heating to the players’ and scorekeepers’ benches. The heaters installed above the spectators’ area, with a total capacity of 176 kW (600 MBtu/h), are activated by motion sensors to keep the air temperature at 15[degrees]C (59[degrees]F). The combustion products are expelled from the building through four roof-mounted exhaust fans interlocked with the burner controls. Two dehumidifiers are diagonally located in the corners of ice rink.

The power density of lights installed above the ice sheet is about 29 kW (99 MBtu/h) or 18.3 W/[m.sup.2] (5.8 MBtu/h x [ft.sup.2]) of ice sheet. They are turned on between 6:00 a.m. and 12:00 p.m. The ice is resurfaced normally at one-hour intervals, between 12:00 a.m. and 12:00 p.m. on weekdays and between 8:00 a.m. and 12:00 p.m. on the weekend.

The refrigeration system consists of two parallel brine chillers, each containing three open-drive four-cylinder compressors, using refrigerant R-22, and air-cooled condensers. One 1200 rpm split-case pump, at a nominal flow rate of about 56.7 L/s (900 GPM) with an electric motor of 30 hp, circulates the calcium chloride brine solution (with a concentration of about 20% by mass) through the 32 mm (1.25 in.) polyethylene piping embedded in the concrete rink slab. The pipes are spaced 89 mm (3.5 in.) apart and are connected to a reverse return manifold in a two-pass configuration. The chillers and the circulating pump work between 3:00 a.m. and 12:00 p.m. The return brine temperature is kept at -9[degrees]C (15.8[degrees]F) in order to maintain the ice temperature at -6[degrees]C to -5[degrees]C (21.2[degrees]F to 23[degrees]F).

Experimental Equipment and Instrumentation

Through a combination of short- and long-term measurements, using permanently installed instruments, the following variables were measured:

* Surface temperature at several locations on the inside surface of walls and ceiling

* Temperature of the ice surface near the center of blue line

* Dew-point temperature at the dasher board at several heights above the ice surface

* Brine temperature entering and leaving both chillers

* Vertical profile of air temperature and velocity above the center of the ice surface and above the stands

* Cooling capacity of the refrigeration system

In addition, the municipality has provided operational data of the HVAC & R equipment from their own building monitoring system: (1) the operation mode (on/off) of compressors, circulating pumps, dehumidifiers, lighting systems, heating systems under the slab, and infrared heaters installed over stands; (2) the temperature of the concrete slab; (3) the ice temperature (measured with an infrared sensor); and (4) the brine temperature (on the return).

The permanently installed instruments consist mostly of thermocouples connected to distributed I/O modules whose data are logged in a remotely accessible computer through a high-speed Internet connection (Table 1). The short-term measurements were performed using four-wire platinum resistance temperature detector (RTD) probes connected to a portable data recorder and a non-intrusive transit-time ultrasonic flow meter (Table 2). The location of the sensors is presented in Figures 1 and 2.

Experimental Method and Sensor Positions

It is a difficult task to perform accurate field measurements of the capacity of refrigeration systems installed in ice skating rinks. The high flow rate of brine results in a small differential brine temperature between the inlet and outlet at the chillers. Other uncertainties in measurements are due to the possible flow rate imbalance between the two chillers, the flow rate measurement, and brine composition.

The measurement of the cooling capacity would require the installation of sensors for the suction and discharge pressure, the temperature of the refrigerant leaving the evaporator, the temperature of the refrigerant before the expansion valve, and the refrigerant flow rate on each of the six compressors. This solution is expensive and intrusive and it requires the interruption of the normal operation for installation purposes.

In this study, three different methods were used for assessing the cooling capacity of chillers:

1. The total cooling load of chillers was calculated using the manufacturer’s selection software, along with measurements of the refrigerant pressure and temperature (e.g., suction, discharge, and liquid temperatures) as well as the knowledge of the number of compressors in operation.

2. The second calculation method used the heat balance equation on the refrigerant side along the measured pressure of refrigerant entering and leaving compressor 3, the measured temperature of refrigerant entering the compressor, and the measured temperature of refrigerant liquid before the expansion valve (Figure 3). In addition, the refrigerant volumetric liquid flow rate was measured with a transit-time ultrasonic flowmeter based on the fluid sonic velocity and known pipe dimensions. This equipment can accurately measure the flow rate, resulting in data of high accuracy without intruding into the flow stream (Scott 2003). The meter was configured for the particular fluid during the flow measurement setup process. Nevertheless, while the measurement of water or water-based fluids is a common use of this technology, the measurement of refrigerant flow is more challenging. The mass flow rate was derived from the calculated liquid density and the measured refrigerant volumetric liquid flow rate, using the ultrasonic flowmeter. The enthalpy differential between the refrigerant liquid entering the expansion valve and superheated vapor leaving the evaporator (that is, the refrigeration effect), multiplied by the mass flow rate, provided the total estimated cooling load of chillers. This method was also used to estimate the average cooling capacity of one compressor by assuming that all compressors were operated at effectively the same conditions during the test period.




3. Using the heat balance equation on the brine side along with the measured brine parameters.


Short-Term Measurements of Cooling Capacity of Compressors

Short-term measurements, needed for the evaluation of the cooling capacity of compressors, were performed with five compressors in operation, outside the regular hours of use of the ice. However, during normal operation, the maximum number of compressors in operation is limited to four in order to limit the peak electrical demand.

A sample of short-term measurements used for the estimation of cooling capacity is presented in Tables 3 and 4, while the estimated cooling capacity is given in Table 5. The average values over a one-hour period of temperature and flow rate data are used with the first two methods, while the average values over a two-hour period are used for the third method.

The refrigerant volumetric liquid flow rate was measured as 0.287 L/s (0.6081 [ft.sup.3]/min) and the density at 32[degrees]C (89.6[degrees]F) (the temperature of liquid entering the expansion valve) was calculated as 1.164 kg/L (72.71 lb/[ft.sup.3]). The refrigerant mass flow rate was finally calculated as 0.3348 kg/s (44.24 lb/min).

The brine temperature (Table 2) was measured with 4W RTDs inserted into existing brine piping thermowells using a multi-channel thermometer. The same ultrasonic meter that was used for measurements on the refrigerant side (see Table 1) was also used for measuring the brine flow rate (52.62 L/s) (111.5 [ft.sup.3]/min). The density and specific heat of brine were estimated using the CoolPack program (CoolPack 2000) for a 21% CaCl m/m brine.

Table 5 reveals a difference of 2.8% between the total cooling capacity of five compressors as estimated by method 1 (277.7 kW [948 MBtu/h]), based on the manufacturer’s calculation) compared with the second method (270 kW [922 MBtu/h]), based on the heat balance on the refrigerant side), which is exceptional for field measurements. The average cooling capacity of one compressor is evaluated at 54.8 kW (15.6 ton) by calculating the average of results from methods 1 and 2.

Method 3 gives a cooling load of 218.8 kW (747 MBtu/h) that is about 20% smaller that the average value from methods 1 and 2. This result was expected due to the small differential brine temperature between the inlet and outlet at the chillers. Under these conditions, the average capacity of one compressor is estimated at 43.8 kW (12.4 ton).

Therefore, by using the average cooling capacity of one compressor of 54.8 kW (15.6 ton) and considering that four compressors are used during normal operating conditions, all of them having the same load and performance, the total cooling load of the refrigeration system is equal to about 219 kW (62 tons) or 7.2 [m.sup.2] of ice sheet /kW (274 [ft.sup.2]/ton).

Short-Term Measurements of the Air Temperature Profile above the Ice

The vertical air temperature profile, measured above the ice center up to the height of 5 m (Figure 4) indicates two notable trends: (1) the temperature increase is almost linear over the first 0.2 m (0.656 ft) from the ice where the conductive heat transfer phenomenon is dominant and (2) the temperature increase follows a parabolic trend, from 0.2 to 5 m (0.656 to 1.64 ft).

Long-Term Measurements

Data presented in Figures 5 through 10 were collected on February 19, 2005, when the outdoor air temperature varied between -18[degrees]C around 6:00 a.m. and -8[degrees]C around 2:00 p.m. Figure 5 presents the variation of the ice surface and brine temperatures (left axis) with time and the number of compressors in operation (right axis). At the beginning of that particular day, the lights were on until midnight, and from that point until about 6:00 a.m. the lights were turned off. This explains the variation of ice surface temperature that decreased from -6[degrees]C to -7[degrees]C between midnight and 3:00 a.m. However, when the compressors were off, from 3:00 a.m. to 4:30 a.m., the ice surface temperature increased slightly above -6[degrees]C. Although the compressors were turned on after 4:30 a.m., the ice temperature still increased until about 6:00 a.m. It can be seen that both brine and ice surface temperatures have the same profile of variation during the unoccupied period between 3:00 a.m. and 9:00 a.m. During that day, nine resurfacing operations were performed between 9:00 a.m. and 10:00 p.m. The ice resurfacing is indicated in the graph by temperature peaks of very short duration.

Figure 6 shows the ice temperature between 2:00 p.m. and 8:00 p.m., when all four compressors were continuously in operation. The response of ice temperature to each of the three resurfacings during this interval was fast–a sudden increase and decrease of the ice surface temperature by 3[degrees]C to 5[degrees]C. A continuous increase of the ice temperature was also noticed between resurfacings, from about -6[degrees]C at 2:00 a.m. to about -5[degrees]C at 8:00 p.m. The brine temperature also slightly increased during the same interval.

Figure 7 shows the variation with time of the air temperature at three different measuring points near the roof deck over the center of the ice sheet. During unoccupied periods, the lights were off, resulting in the reduction of air temperature between midnight and 6:00 a.m. and also after 10:00 p.m. The variation of air temperature at these selected points follows the variation of the outdoor air temperature.

Figure 8 shows the variation with time of the west wall temperature at different heights. Due to the position of the radiant heaters, the wall temperature at the 3.7 m level is greater than at 2.7 m.

Figure 9 shows the variation of air temperature above the spectator stands. This area has high-intensity radiant heaters controlled by motion sensors. When the stands are occupied, the air temperature increases significantly, for instance, between noon and 3:00 p.m.

Figure 10 shows the measurements of ice temperature taken on March 26, 2005, when the outdoor air temperature was between -3[degrees]C (around 6:00 a.m.) and 8[degrees]C (around 2:00 p.m.). The results indicate the number and intensity of resurfacings that took place between 9:00 a.m. and 6:00 p.m. had a significant impact on the ice temperature, which remained above -6.0[degrees]C.


This paper presents the preliminary measurements performed in a Canadian indoor ice rink. The results help to better understand the thermal and energy behavior of this ice rink. They will also be used for the verification and calibration of the numerical model developed by Bellache et al (2005b). During the summer of 2005, the brine circulation system was renovated. Additional instruments were installed to improve the measurement of the brine loop temperature differential and the heat flux at the ice surface.







The authors acknowledge the financial support received from ASHRAE research and technical support from Technical Committee TC 10.2, Automatic Ice Making Plants/Skating Rinks. The authors are also grateful to Mr. C. Dumas who is an engineer in the Department of Buildings, City of Montreal, Quebec, Canada, for his helpful assistance.




ASHRAE. 2002. 2002 ASHRAE Handbook–Refrigeration. Atlanta: American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc.

Bellache, O., M. Ouzzane, and N. Galanis. 2005a. CFD calculation of air flow and thermal comfort in an indoor ice rink. Building and Environment Journal 40:415-24.

Bellache, O., N. Galanis, M. Ouzzane, and R. Sunye. 2005b. Two-dimensional transient model of ice sheet heat transfer. Submitted for presentation at the ASHRAE Meeting in Quebec City, June 2006.

Chen Q., T.G. Hoornstra, and J. Vanderkoon. 1989. Energy analysis of building with different air supply and exhaust systems. ASHRAE Transactions 96(1):1041-49.

CoolPack. 2000. Version 1.46. Department of Mechanical Engineering Technical, University of Denmark.

Holman, J.P. 1984. Experimental Methods For Engineers, 4th ed. New York: McGraw-Hill Book Company.

Jones P.J., and G.E. Whittle. 1992. Computational fluid dynamics for building air flow prediction-current status and capabilities. Building and Environment 27(3)321-38.

Lavoie, M., R. Sunye, and D. Giguere. 2000. Potentiel d’economies d’energie en refrigeration dans les arenas du Quebec. Report prepared by the CANMET Energy Technology Center.

Nielsen P.V., A. Restivo, and J.H. Whitelaw. 1979. Buoyancy-affected flows in ventilated rooms. Numerical Heat Transfer 2:115-27.

Scott, J. 2003. Ultrasonic chiller validation: Liquid flow vs. refrigerant flow, Seminar 19. ASHRAE Winter Meeting, Chicago.

Yang C., P. Demokritou, and Q. Chen. 2000. Ventilation and air quality in indoor ice skating arenas. ASHRAE Transactions 160(2):339-46.

Zmeureanu, R., E.M. Zelaya, and D. Giguere. 2002. Simulation de la consommation d’energie d’un arena a l’aide du logiciel DOE-[2.1.sup.E]. ESIM 2002 Conference, September, Montreal, Quebec, Canada.

M. Ouzzane, PhD

R. Zmeureanu, PhD, PEng


J. Scott


R. Sunye, PhD

D. Giguere, PEng

O. Bellache, PhD

M. Ouzzane and O. Bellache are research scientists, R. Sunye is the manager of the present project (RP-1289) and a research scientist, J. Scott is a senior refrigeration specialist, and D. Giguere is a technological expert at CANMET Energy Technology Centre-Varennes, Quebec, Canada. R. Zmeureanu is a professor in the Department of Building, Civil, and Environmental Engineering, Concordia University, Quebec, Canada.

Table 1. List of Instruments Used for Long-Term Measurements

Item Quantity Description

1 50 Type-T thermocouple, encapsulated tip, 24 AWG premium

grade polytetrafluorethylene (PTFT) wire leads, 40 ft


2 15 Type-T thermocouple, encapsulated tip, 24 AWG premium

grade PTFT wire leads, 60 ft long

3 1 Edgetech Model 200 Dewtrack #200C1/D1; 24 VDC, 4-20mA

4 1 Sample chamber (SC1)

5 1 Sensor model # TX-S-LT-SF (33:1 D:S Focus) 4-20 mA


6 8 #CB-7018–8 channel T/C input or [+ or -]20 mA module, 6

diff. and 2 single ended

7 1 #CB-7017–8 channel voltage or [+ or -]20 mA input


8 1 #CB-7520 isolated RS-232 to RS-485 converter module

9 1 #CB-7067–7 channel relay output module

10 2 Panel mounted 24VDC 3 amp. power supply

11 1 RS-CABLE M/F DBP connector 6 ft cable

12 4 ASCO 0.25 in. FPT direct acting solenoid # 8262G208 C/W

24 VDC coil

13 2 Sampling rotary vane pump

14 2 Sampling air flowmeter–Dwyer RMA-12

Table 2. List of Instruments Used for Short-Term Measurements

Instrument Quantity Accuracy

RTDs for short-term 8 [+ or -]0.1[degrees]C

measurements ([+ or -]0.18[degrees]F)

Portable ultrasonic flowmeter* 1 [+ or -]0.5-2%

for short-term measurements

Instrument Sensitivity

RTDs for short-term 0.01[degrees]C (0.018[degrees]F)


Portable ultrasonic flowmeter* 0.1% @ 40[degrees]C (104[degrees]F)

for short-term measurements

* Varies with ultrasonic transducer utilized.

Table 3. Measurements on the Refrigerant Side Used for the Estimation

of Cooling Capacity of Chillers

Parameter Compressor Suction Compressor Discharge

Pressure 263.4 kPa 1549 kPa

(38.2 psia) (225 psia)

Temperature of refrigerant -13.9[degrees]C

leaving the evaporator (7[degrees]F)

Temperature of refrigerant 32[degrees]C

entering the expansion (89.6[degrees]F)


Table 4. Measurements on the Brine Side Used for the Estimation of

Cooling Capacity

Inlet Brine Outlet Brine

Temperature Temperature Flow Rate

Chiller 1 -5.91[degrees]C -7.37 [degrees]C 50% of total

(21.36[degrees]F) (18.73[degrees]F)

Chiller 2 -5.85 [degrees]C -6.66 [degrees]C 50% of total

(21.47[degrees]F) (20.01[degrees]F)

Total 52.62 l/s (834 gpm)

Table 5. Estimated Cooling Capacity of One Compressor

Total Capacity Average Capacity

for Five of One

Compressors, Compressor,

Method Main Input Parameters kW (tons) kW (tons)

1 Compressor model, refrigerant, 277.7 55.54 (15.8)

saturated suction (79.0)


Saturated condensing


Superheated vapor leaving


Liquid temperature entering

TXV valve

2 Liquid temperature and 270.0 54.0 (15.36)

pressure before the TXV (76.8) for

Temperature and pressure compressor 3

entering the compressor

Refrigerant mass flow rate

3 Evaporator inlet and outlet 218.8 43.8 (12.4)

brine temperature at each (62.2)


Brine flow rate, density, and

specific heat

COPYRIGHT 2006 American Society of Heating, Refrigerating, and Air-Conditioning Engineers, Inc.

COPYRIGHT 2008 Gale, Cengage Learning