- Email: [email protected]
Experimental study and numerical prediction of thermal and humidity conditions in the ventilated ice rink arena
Building and Environment 108 (2016) 171e182
Contents lists available at ScienceDirect
Building and Environment journal homepage: www.elsevier.com/locate/buildenv
Experimental study and numerical prediction of thermal and humidity conditions in the ventilated ice rink arena Agnieszka Palmowska*, Barbara Lipska Department of Heating, Ventilation and Dust Removal Technology, Silesian University of Technology, Konarskiego 20, 44-100, Gliwice, Poland
a r t i c l e i n f o
a b s t r a c t
Article history: Received 17 June 2016 Received in revised form 8 August 2016 Accepted 24 August 2016 Available online 26 August 2016
An ice rink arena is a place where the ventilation system has to deal with the presence of spectators and skaters (who have different requirements), and with the temperature of the ice surface below 0 C. An important problem that arises is excess moisture, leading to condensation of water vapour onto the ice surface and inner surface of the ceiling. The aim of this presented paper was to experimentally identify physical phenomena occurring in the actual ventilated ice rink arena in Gliwice (Poland) and to check whether the developed numerical model correctly reproduces such phenomena and how it should be improved for this purpose. Long-term and short-term experimental research was carried out to obtain data for boundary conditions, to identify changes in thermal and humidity conditions and for experimental validation of simulation results. Air parameters: speed, velocity, temperature and relative humidity were measured. Thermal imaging measurements were also carried out. The numerical model was prepared by means of the Ansys CFX 14.5. Improvement of the moisture flux numerical modelling was carried out. The scope of validation encompassed the comparison of indoor airflow pattern, air parameters (above the ice surface and on the outskirts of the ice rink) and temperature of the ceiling. The numerical model was able to map real conditions in the object with good agreement between measured and predicted values. The mean deviations for all studied cases did not exceed values of 0.03 m/s for speed, 1.1 C for temperature and 15% for relative humidity of the indoor air. © 2016 Elsevier Ltd. All rights reserved.
Keywords: Ice rinks Ventilation Humidity Measurement CFD prediction Experimental validation
1. Introduction An ice rink arena is a ventilated large-volume building, used for sports and recreational activities. Ice rink design and operation are totally unique and differ in many ways from standard buildings [1]. Therefore, the proper operation of a ventilation system with adequate conditions for users, along with the maintenance of good technical condition of the facility is required. It means that the conditions need to meet the ventilation requirements for people inside the building, and that the building is safe for the people within. However, there are no fixed standards and guidelines exist for the design of ventilation systems in ice rink arenas, especially in terms of air distribution systems, so therefore there is a great need for further research. An ice rink arena is a facility where ventilation system has to deal with an ice surface temperature of below 0 C and the presence
˛ ska, Katedra Ogrzewnictwa, Wentylacji i * Corresponding author. Politechnika Sla Techniki Odpylania, ul. Konarskiego 20, 44-100, Gliwice, Poland. E-mail address: [email protected] (A. Palmowska). http://dx.doi.org/10.1016/j.buildenv.2016.08.024 0360-1323/© 2016 Elsevier Ltd. All rights reserved.
of people, both spectators and skaters, whose requirements are different (from the point of view of thermal comfort). Ventilation must meet the requirements of both parties. However, thermal comfort conditions were not analyzed in this paper. An important problem that occurs in such objects is the risk of exceeding the permissible moisture level in the indoor air, leading to condensation of water vapour contained in the air onto the wall surfaces and formation of fog above the ice surface. Therefore, the ventilation system in an ice rink arena should fulfil the following functions: - Maintaining adequate thermal and humidity conditions for users of the ice rink arena, which is the function of ventilation or air heating, - Removing excess moisture above the ice surface, which is the function of dehumidification [2], A detailed description of the air parameter requirements was discussed by Palmowska and Lipska [3]. Due to geometry and complex flow phenomenon, it is difficult to
172
A. Palmowska, B. Lipska / Building and Environment 108 (2016) 171e182
predict the effects of the ventilation system by using traditional engineering methods. Therefore, a CFD technique is used for scientific research in this field. This method, based on the basic equations of fluid flow and heat, can be used for researching the flow of air, heat, moisture and contaminants in such facilities. The CFD allows us to check and test different design solutions [4], which is an advantage over a physical experiment. Furthermore, it can be used to improve thermal and humidity conditions in the ventilated facilities [5]. As a rule, the CFD does not replace the measurements completely, but the amount of experimentation and the overall cost can be significantly reduced. Moreover, it enables an assessment of ventilation work effectiveness, as early as at the design stage. Last but not least, the CFD modelling is a useful tool for the optimization of ventilation systems from the viewpoint of fulfiling the users' requirements. It should be noted that the CFD modelling can be burdened with errors, thus it must be validated experimentally before starting appropriate tests with its use [6]. Nowadays, the design of sports arenas is not complete without one or more simulations of the indoor air flow. Analyses of this type are used to determine the best air conditioning systems and ensure that the occupants are exposed to a predefined thermal comfort range most or all of the time [7]. So far, ventilation issues in ice rinks are relatively unknown and poorly supported by research. In literature only a few numerical studies underpinned by the experiment can be found: the Gjøvik Olympic Cavern Hall (Norway) [8], the ice rink arena in Greater Boston (USA) [9,10], and the ice rink arena in Montreal (Canada) [11,12]. The results of numerical calculations of these objects were validated by comparison with the results of the measurements. The research carried out for the facility in Gjøvik is the first experimental identification of conditions in a ventilated ice rink and its numerical modelling to be documented in literature. The distributions of temperature and air velocity were compared. In simulations, thermal radiation with the use of Discrete Transfer Model (DTM) was taken into account. Air temperature and velocity distributions were also compared for the ice rink arena in Greater Boston. Additionally, CFD technique was used to validate the spread of contaminants such as PFT and CO [10]. In the object arena in Montreal calculations results were compared with measured air velocity in the real object and with the use of its physical model. The impact of different emissivity of the ceiling on the air parameters and the ceiling's temperature was additionally studied [12]. It should be noted that the scope of all published studies encompassed just temperature and velocity distribution of air, and spread of contaminants in such objects. None of these studies took into account the modelling of moisture flow in the whole building, except an attempt of humidity modelling by Daoud et al. [13] in the ice rink arena in Montreal. Ventilation air was the only source of moisture considered in such a simulation. A humidity transport model, programmed using MATLAB, returned the heat transfer due to condensation on ice as a heat gain to the ice surface in the energy model. Results of calculations, in terms of humidity, were not validated due to the lack of experimental data in this field. In the indoor ice rink thermal imaging measurements allow detection of moisture on the surface of building partitions as in the cases of residential buildings or industrial facilities. The use of infrared thermography for building diagnostics was discussed by Balaras and Argiriou [14]. The main advantage of this method is possibility to perform non-destructive tests. Nevertheless, the infrared thermography is not free from errors (surface emissivity, camera calibration, etc.). The use of thermal imaging measurements in experimental studies to validate the results of numerical calculations for ice rinks has not been yet published in literature, as well as in research in the field of ice rink ventilation.
In literature only numerical results of tests (without experimental validation) for objects Hodynka Arena, Tsherepovets Arena (Russia) [7] and the designed ice rink in Wadowice (Poland) [2,15,16], as well as only experimental studies for ice rinks located in Honk Kong [17] and Taipei (China) [18], also can be found. For example, in Ref. [7] the CFD method was used to better understand the interior conditions and flow behavior for a range of usage scenarios. The goal of these simulations was to determine how well the planned ventilation systems work to meet the desired indoor conditions [7]. The aim of the presented research was (1) to experimentally identify characteristics of physical phenomena occurring in the actual ventilated ice rink arena and (2) to check whether CFD prediction by a developed numerical model supported by experiment reproduces correctly such phenomena and (3) how it should be improved for this purpose. 2. Experiment 2.1. Tested object and the phenomena taking place The facility used for the test was the indoor ice rink “Tafla” in Gliwice (Poland) with external dimensions: length 66 m, width 37 m and a maximum height of 11 m. Inside the building there is an ice rink with the dimensions 30 60 m (Fig. 1). The use of the hall as an ice rink is expected in the period of time from October to May for different kinds of sports and recreational activities, including curling. Inside the building thermal and humidity conditions are maintained by mechanical mixing ventilation with an integrated air distribution system. A detailed description of the air distribution systems used in ice rink arenas was discussed by Stobiecka et al. [15]. It is carried out by side air supply with the use of long-throw jet nozzles and air exhaust under the ceiling by rectangular exhaust grilles. Because of economical reasons, the designed indoor air temperature amounts to þ5 C and air dehumidification is not anticipated. The air handling unit, located on the roof of the technology building, is designed for 2 fan speeds: higher (second, II), when outdoor air temperature, te, is over þ5 C and lower (first, I), when outdoor air temperature is below þ5 C. In the ice rink arena a series of complex flow phenomena takes place. Firstly, there is movement of supply air jets, which largely shapes conditions in the facility. It has an influence on air speed distribution. The air speed value just above the ice surface is very important from the users' thermal comfort viewpoint. Moreover, low air speed minimizes load on the ice making system, while too high air speed could cause ice melting, adversely affecting its quality. The air speed above the ice surface should not exceed the value of 0.25 m/s [19]. Secondly, there are also phenomena in the ice rink arena connected with the heat transfer, e.g. heat gains or losses through the building structure, heat losses from the ice surface, heat gains from the lighting system and from people. Heat exchange in the ice rink occurs by radiation and convection. It also includes vapour diffusion or condensation on the ice sheet, which is an important contributor to the ice sheet heat load. The cold ice surface, directly opposite the ceiling, absorbs heat by radiation. As a result, the inner surface of the ceiling is colder than air below it [20]. It should be observed that the water vapour contained in the air will condense on the surfaces which are colder than the room's air dew point temperature. Thirdly, in the ice rink arena a flow of moisture also occurs from different sources. Indoor sources of moisture are only people and the temporary work of ice resurfacer. Moisture is supplied into the ice rink arena with the ventilation air and by infiltration. It is very important that moisture gains can also occur from the ice rink surface. However, in most cases, moisture
A. Palmowska, B. Lipska / Building and Environment 108 (2016) 171e182
173
dehumidifier. The limit level of specific air humidity, above which condensation of water vapour and risk of fogging can occur, is determined on the ice surface temperature, ti, and relative air humidity of 100%. For this value ice surface temperature is equal to air dew point temperature. 2.2. Measurements methods The experimental research was carried out for following reasons: 1. To obtain data for boundary conditions for numerical calculations, 2. To identify changes in thermal and humidity conditions, depending on external conditions and operating mode of ice rink, 3. For experimental validation of simulation results. Long-term measurements, 24 h a day in the period of time from January to April 2014, and short-term measurements, during the night, without the presence of people, in 3 measuring series: on 8/9 February (variant 1), 21/22 March 2014 (variant 2) and 12/13 April 2014 (variant 3), were performed. Measurements were limited to half of the rink, because of the symmetry of the facility. Fig. 1 (middle) shows the location of test points on the tested half of the rink, marking of long-throw jet nozzles and columns. 2.2.1. Field measurements to establish CFD model boundary condition Boundary conditions were obtained on the basis of the measurements, which included: long-term tests of temperature and relative humidity 4e of the outdoor air, temperature tN and relative humidity of supply air 4N, short-term tests of supply airflow rate, temperature of the ice surface, temperature distribution on the wall's inner surfaces and emissivity of ice εi by means of thermography.
Fig. 1. The indoor ice rink “Tafla” in Gliwice (Upper Silesia, Poland) (upper) [30], top view with marked test points (middle) and cross-section (bottom).
travels to the ice surface. Moisture flux between the ice surface and _ can be expressed by the relationship between the indoor air (W) mass heat transfer coefficient, the specific humidity in air Xa and in saturated ice Xi, in accordance with the ASHRAE Handbook [21]:
_ ¼ K·ðXa X Þ·A W i i
(1)
where K is the mass heat transfer coefficient and is equal to 0.00023 kg/s$m2.Ai is the ice rink area (m2). If the air humidity is higher than the humidity of ice (Xa>Xi) there is a positive moisture flux. It means moisture condensation on the ice surface. To prevent condensation, the room air dew point temperature should be lower than the ice surface temperature and the ceiling inner surface temperature. In practice, to avoid condensation, the moisture content need to be balanced and controlled by a
Outdoor air parameters were obtained from meteorological station DAVIS located in the distance of about 500 m away from the tested facility. Long-term measurement was carried out with a time step of 15 min. Supply airflow rate was measured indirectly by two methods based on measurement of air velocity, va. In the first method, the anemometer with Pitot tube VPT-100 VOLTCRAFT was used for the measurement of air velocity in long-throw jet nozzles 1e25 (Fig. 1). The air velocity was measured using a method based on dividing a circular cross section of the nozzle into rings, the number of which depends on the diameter of the nozzle in accordance with ISO 7194 Standards [22]. The measurement accuracy was ±0.25%. In the second method, the thermal anemometer TESTO was used for the measurement of air velocity in the ventilation duct (behind air ventilation unit). The air velocity was measured using a method based on dividing a rectangular cross section of the ventilation duct into squares of equal surface area in accordance with ISO 7194 Standards [22]. Measuring error of this device was 0.03 m/s þ 5% of measured value. Temperature and relative air humidity loggers APAR AR235 were used for measuring temperature and relative humidity of supply air in the first (nozzle 1) and last long-throw jet nozzle (nozzle 25) (Fig. 1). Humidity accuracy was in the range of 3 ÷ 5% and temperature accuracy was in the range of 0.5 ÷ 1.8 C. Longterm measurement was carried out with a time step of 1 min. The NTC probe for food TESTO was used to measure ice surface
174
A. Palmowska, B. Lipska / Building and Environment 108 (2016) 171e182
temperature. It was positioned in a drilled hole of a 15 mm depth. The measurement was made in the middle of the investigated half of the ice rink surface (point 9). Measurement accuracy was ±0.2 C. The measurement took 3 min. It was a time sufficient to stabilize measurement result. The measurement was repeated 2 times. Thermal measurements of temperature distribution on the wall's inner surfaces and emissivity of ice were performed by thermal imaging cameras FLIR E50 and ThermaCAM E45 with thermal sensitivity respectively 0.05 C and 0.1 C. The procedure for determining emissivity of ice in accordance with ISO 18434-1 Standards [23] was used. Measurement results were developed by FLIR Tools software. Remaining boundary conditions, i.e. heat transfer coefficients of building partitions, values of lamp power, were adopted on the basis of stocktaking of the facility. In Fig. 2 a selected thermogram of the northern partition of the ice rink arena is presented. Measurement was performed for the following indoor air parameters: temperature ta of 0.7 C and relative humidity 4a of 79%. Outdoor air temperature was (te) 2.1 C. The measurement area of the mean temperature of the north wall surface separating the hall from technical facilities is limited by a black frame. 2.2.2. Experimental identification of changes in thermal and humidity conditions in the ice rink arena An experimental identification of changes in thermal and humidity conditions in the ice rink arena included long-term measurements of temperature (ta) and relative humidity (4a) of indoor air with the use of the loggers APAR AR235, mentioned before. They were located along the ice rink on the selected columns E÷L (Fig. 1)
at a height of 1.7 m and additionally at heights: 1.1 m and 0.6 m on column I. Measurements were carried out with a time step of 5 min. The impact of occupant presence and the work of the ice resurfacer on the distribution of relative air humidity was observed. In Fig. 3 daily distribution of relative humidity and temperature of air, when the ventilation system was off, at the height of 1.7 m (column F) is shown. During ice resurfacer work, as well as during the presence of people, the relative humidity of air gradually increased. The lowest recorded value of this parameter of air, 4a, since the opening of the ice rink, was the value of 86.3%. The highest value of 4a was 88.6%. The temperature of air also gradually increased during the day. The lowest value of ta was 2.5 C and the highest value of ta was 0.2 C. In the period of the long-term test, when the outside air temperature was in the range of 15 C to þ25 C, indoor air temperature did not reach a value greater than 3 C and less than 3 C (Fig. 4). It means that designed indoor air temperature of 5 C was not achieved. Fig. 5 shows a Mollier chart with a marked area which contains the conditions of air inside the ice rink arena during long-term measurements. In most of these conditions moisture content was higher than the limit value, determined by ice surface temperature ti ¼ 4.5 C. Therefore, the air dew point temperature was higher than the temperature of the ice surface. There was condensation of water vapour contained in the air and then freezing of moisture onto the ice surface, which had to be removed by the ice resurfacer. 2.2.3. Experiment for validation Short-term measurements of speed, wa, temperature and relative air humidity above the ice surface and short-term thermographic measurements of temperature distribution on the ceiling's inner surface were necessary for experimental validation of simulation results. Visualization by smoke test to obtain airflow pattern information was also made. Measurements of the parameters of air above the ice surface were carried out at 9 measuring points evenly distributed on the surface of the ice rink (Fig. 1) at 4 specific heights: 0.1 m, 0.6 m, 1.1 m, 1.7 m in accordance with ISO 7726 Standards [24]. The measurement time was 15 min at each point. At point 9, it lasted 60 min. 3. Numerical method 3.1. Model description and boundary conditions
Fig. 2. Thermogram of northern partition of tested ice rink arena.
The numerical model of the tested object was prepared with the use of the Ansys CFX 14.5 code, integrated with Workbench 14.5 platform. It took into consideration the designed overall dimension of the facility and its geometry. Ice rink, lighting system and ventilation system were modelled (Fig. 6). The XY symmetry plane was introduced because of the symmetry of this object. This enabled the greater refinement of the discretization grid on the available server. The numerical model included the tested half of the ice rink and it did not take into account the presence of people, just like in the real object during the experimental research. The ice sheet was modelled as a surface of uniform temperature. The lighting system, consisting of 37 metal halide lamps, was modelled as rectangular heat sources located in the top part of the hall. Only those lamps, which were turned on during experimental measurements, were included as heat sources (marked in yellow colour on Fig. 6). Convective and radiative heat gains from lamps, as energy and radiation sources, were related to the surface of geometry of the lighting model. The long-throw jet nozzles were modelled as circular supply openings with a diameter of Ø160 and Ø80 mm, placed on the
175
0 -1
87 -2 -3
7: 31 8: 16 9: 01 9: 46 10 :3 11 1 :1 12 6 :0 12 1 :4 13 6 :3 14 1 :1 15 6 :0 15 1 :4 16 6 :3 17 1 :1 18 6 :0 18 1 :4 19 6 :3 20 1 :1 21 6 :0 21 1 :4 22 6 :2 23 1 :0 6
86
Air temperature ta. °C
1
ice resurfacer
ice skating
ice skating
ice resurfacer
ice resurfacer
ice skating
ice skating
ice resurfacer
ice skating
ice resurfacer
ice skating
ice resurfacer
ice resurfacer
ice skating
ice skating
88
ice resurfacer
Relative air humidity φa, %
89
opening
A. Palmowska, B. Lipska / Building and Environment 108 (2016) 171e182
Relative air humidity
Time, h
Air temperature
Fig. 3. Daily distribution of relative humidity and temperature of air in the tested ice rink arena (column F, height of 1.7 m).
lighting system
exhaust
Fig. 4. The relationship between indoor and outdoor air temperature in the tested ice rink arena (column J, height of 1.7 m).
symmetry plane ice surface
ventilation air supply tribune
Fig. 6. The numerical model of half of the rink hall with marked directions XYZ.
ventilation supply duct, which was modelled as a circular duct. Larger nozzles were inclined at an angle of 75 relative to the vertical axis of the ventilation duct, while the smaller at an angle of 45 . The exhausts were set as rectangular grilles with dimensions 800 150 mm, also located on the ventilation exhaust duct. The model included penetrating heat fluxes through the internal and external partitions. Heat transfer coefficients and outside temperature were set to internal and external walls. The exception was the part of the north wall separating the ice rink arena from technical facilities. The mean temperature of its surface, obtained from thermovision measurements, was used. The following averaged measured air parameters were set for supply openings: mass flow rate, temperature and specific humidity. Exhaust air mass flow rates were set for exhaust openings. Boundary conditions were prepared for the following calculation cases for the entire facility: 1) case 1, 1A: based on the variant 2 (II fan speed), 2) case 2: based on the variant 1 (I fan speed), 3) case 3: based on the variant 3 (II fan speed). Data required to perform the numerical calculations are listed in Table 1 and Table 2.
3.2. Numerical procedure Fig. 5. Mollier chart with area containing air conditions in the tested ice rink arena during long-term measurements.
The numerical calculations were carried out using ANSYS CFX
176
A. Palmowska, B. Lipska / Building and Environment 108 (2016) 171e182
Table 1 The boundary conditions for the ice rink arena, obtained from measurements. Boundary condition
Unit
Outdoor air temperature Temperature of supply air Specific humidity of supply air ¼ Specific humidity of outdoor air Mass flow rate of supply air ¼ Mass flow rate of exhaust air Ice surface temperature Mean temperature of the north wall surface separating hall from technical facilities Mean ice surface emissivity
Table 2 The boundary conditions for the ice rink arena, obtained from facility stocktaking. Boundary condition Heat transfer coefficient Heat transfer coefficient Heat transfer coefficient Heat transfer coefficient Lighting power Ground temperature
Unit of of of of
external walls (N, W, E) wall in the ground (E) floor ceiling
Value 2
W/m K W/m2K W/m2K W/m2K W/m2 C
0.187 0.140 0.203 0.301 1206 2.0
14.5 software, in steady-state, three-dimensional and nonisothermal conditions. Discretization of model equations was solved by Finite Volume Method. The high resolution discretization scheme and the Rhie Chow redistribution algorithm for coupling pressure-velocity were used. The CFD model used the Shear Stress Transport (SST) turbulence model from the EVM models group. The choice of the turbulence model was based on study results [25]. No-slip condition was adopted on the wall surface. A minimum wall distance Yþ was equal to 34. Thermal radiation between walls, ice surface and objects located inside the ice rink arena was performed by the Discrete Transfer Model, selected as a result of our own research [26]. This model is based on tracing the domain by multiple rays leaving from the bounding surfaces. The technique was developed by Shah in 1979 and depends upon the discretisation of the transfer equation along rays [27]. The applied unstructured discretization grid was composed of 12574808 cells, built mainly with tetrahedral elements, and 3161134 nodes (Fig. 7). It also included local refinement around outlet and inlet openings, and just above the ice rink surface (Fig. 7). The basic mesh element size was 35 cm, with a refinement over the ice surface up to 7 cm. 5 layers consisting of prismatic
Calculation cases 1
2
3
C C kg(H2O)/kg(dry air)
11.2 12.5 0.0040
6.3 19.8 0,0044
4.8 14.2 0.0048
kg/s C C e
4.8 4.45 5.9 0.96
2.2 4.70 3.6 0.96
4.8 4.50 3.5 0.96
elements, with a total maximum thickness of 17 cm in the boundary layer at the surface of the partitions were included in the numerical model. A grid independence test was performed for different grids in order to select the grid, which most accurately reproduces the conditions in the facility. The calculations were carried out using iteration method as long as the convergent solutions were obtained. Achieved convergence after approximately 6000 iterations was in the range of 1.0E-6 (for turbulence, momentum and mass, wallscale) to 1.0E-4 (for heat transfer and mass fractions).
3.3. Improvement of the moisture flux numerical modelling The moisture flux between the ice surface and indoor air in preliminary calculations for the case 1 was modelled by treating the ice sheet as a source of moisture with default mass flow boundary condition. Constant value of this stream was determined based on Formula (1). It did not take into account local changes in flow direction of moisture: from the ice into the air or inversely, depending on the difference between the specific humidity in air and saturated ice. Therefore, the distribution of specific air humidity above the ice surface could be burdened with errors. To prevent this, our own modification of the method of numerical modelling of such emission for case 1A was proposed. It was based on the introduction in Ansys CFX user function, developed using the Formula (1). This procedure allowed calculation of the mass flow of moisture separately for each node of the discretization grid with an unknown value of specific humidity of air above the ice surface Xa. The unknown value Xa was replaced in the subsequent iteration steps of numerical calculations by the calculated value of specific humidity in the first mesh grid above the ice surface. Fig. 8a shows contour maps of the specific air humidity on the
Fig. 7. Cross-section through the mesh discretization (left) and local refinement of the discretization grid around the long-throw jet nozzles (right).
A. Palmowska, B. Lipska / Building and Environment 108 (2016) 171e182
177
Fig. 8. Maps of the specific air humidity in the tested ice rink arena a) in the horizontal plane ZX, Y ¼ 0.1 m, b) at the height of 0.1 m as a result of the experiment, c) in the vertical plane XY, Z ¼ 24 m.
height of 0.1 m above the ice surface from CFD simulations by means of two methods of moisture emission modelling. The calculation results were related to the distribution shown in Fig. 8b. It was made by using the values of specific humidity, determined on the basis of measurements of relative humidity and temperature. For case 1 the distribution of specific air humidity was uniform over the entire ice surface, which was not compatible with the measurements results (b). An average relative deviation between the calculation results and measurements was 11.5%. In case 1A the distribution was irregular and an improvement of modelling accuracy with an average deviation of 8.5% has been obtained. A method of moisture emission modelling from the ice surface had an impact on the distribution of specific humidity not only directly above the ice surface, but also in the upper regions of the facility. Fig. 8c shows the comparison of this parameter distribution in the vertical plane XY (Z ¼ 24 m). In the case 1A specific air humidity values were lower than in the case 1. The lowest values occurred directly over the ice surface due to air dehumidification as a result of higher air humidity than in the saturated ice. This caused a humidity decrease in the area over the ice surface (the area limited by the red frame), which could not be predicted in case 1. Further numerical calculations were carried out taking into account our own function for the moisture flux modelling from the ice surface.
4. Validation of the CFD results and discussion 4.1. Indoor airflow pattern For case 3, the predicted airflow pattern as an isosurface of air velocity va ¼ 0.7 m/s was compared with the visualization by smoke test (Fig. 9). Both in the experiment and in the CFD modelling the initial falling and then, at the height of 3 m above the floor, the deflection of supply air jet were observed. The numerical model of the object well reproduced the airflow pattern. 4.2. Indoor air parameters Predicted values of the indoor air parameters such as speed, temperature and relative humidity, were validated using the measurement data for different variants. The air velocity values, calculated directly by means of CFD code, were converted to speed values on the basis of the method proposed in Ref. [28]. Relative air humidity was determined by direct simulation results of specific humidity using our own user function based on humidity conversion formulas. For the needs of validation it enabled the comparison of the predicted parameters with the values of speed and relative air humidity obtained during the measurements. For case 3 a detailed comparison of CFD results with the measurements results of the air parameters in the facility at 9 test
Fig. 9. The comparison of airflow pattern: predicted as an isosurface of air velocity (left) and visualized by smoke test (middle, right) (case 3).
178
A. Palmowska, B. Lipska / Building and Environment 108 (2016) 171e182
Fig. 10. The comparison of the results of numerical prediction and measurements for the indoor air parameters at the height of 0.6 m above the ice surface a) speed, b) temperature, c) relative humidity (case 3).
points above the ice surface was carried out. This is shown in Fig. 10 for the height of 0.6 m. The mean deviation between the measured and calculated values at each point was: 0.1 m/s for the speed, 0.5 C for the temperature and 9% for the relative air humidity. It was noted that the calculated values of relative air humidity in part were not in the range of the measured values, which included the higher values of this parameter. In the case of calculated temperature and speed values they were not always exactly at the point where the data values were measured, but in another part of the ice surface. However, in most cases they did not exceed the range of values from the measurements, except some small local areas of
b)
Y = 0.6 m
Y = 0.1 m, 0.6 m, 1.1 m, 1.7 m
100
100
80
80
Probability, %
Probability, %
a)
speed above 0.2 m/s. The evaluation of relative deviation dispersion between the results of calculations and measurements (Fig. 11) for the analyzed plane (a) and in total for all heights (b) for each air parameter, with marked measurement error bars, was carried out. The relative deviation was calculated as the difference between calculated and measured value in relation to the measured value. At the height of 0.6 m (a) the probability of relative deviation below 20% between the results of measurements and calculations for relative air humidity was 100%. For the temperature it was 33% probability and 11%, for the speed. In contrast, the chance of
60 40 20
60 40 20 0
0
Relative deviation, % relative air humidity
air temperature
Relative deviation, % air speed
Fig. 11. The probability of relative deviation between the results of numerical prediction and measurements for the temperature, speed and relative air humidity a) at a height of 0.6 m b) in total for all test heights above the ice surface (case 3).
A. Palmowska, B. Lipska / Building and Environment 108 (2016) 171e182 Table 3 The mean values of predicted and measured indoor air parameters at the height of 0.6 m above the ice surface. Unit
Mean measured value
Mean predicted value
Speed Temperature Relative humidity
m/s C %
0.06 ± 0.02 2.3 ± 0.2 96 ± 5
0.11 2.6 87
deviation below 50% was 89% and 11% respectively. Small probability values of the speed resulted from large differences in relative deviation values, and they in turn were caused by difference values in comparison with small values of this parameter. Similar results were obtained in total for all test heights (b). Therefore it was found that the main source of deviation was not the air parameter value, but its other distribution on the plane. However, from the viewpoint of reproducing the thermal and humidity conditions in the occupied zone and their assessment, the point values are not as important as their mean values at a given height. Table 3 shows the comparison of mean values of measurements and calculations for a selected plane at the height of 0.6 m above the ice surface. The deviations between these values, without taking into account the measurement error, were as follows: 0.05 m/s for the speed, 0.3 C for the temperature and 9% for the relative humidity. Taking this into consideration, for other cases (1A, 2) the results of numerical prediction and measurements of the indoor air parameters at the test heights of 0.1 m, 0.6 m, 1.1 m and 1.7 m above the ice surface were compared, as shown in Fig. 12 and Fig. 13. The deviations between the measured and calculated values of
25
50
75
100
Mean relative air humidity, % measurements CFD Fig. 13. The comparison of the results of numerical prediction and measurements for the mean relative air humidity at different test heights for case 1A.
mean air speed were higher in the upper part of the facility than directly over the ice surface. For case 2 (a) they were in the range of 0.01 ÷ 0.03 m/s and 0.01÷ 0.04 m/s for case 1A (b). The mean value of the deviation for both cases was 0.02 m/s. The deviations of mean air temperature for case 2 (a) were within the range of 0.4 ÷ 3.1 C and 0.4 ÷ 0.9 C for case 1A (b). The mean value of the deviation for both cases was 1.1 C. The comparison of the relative air humidity results was carried out only for case 1A, because the experiment for 1 variant (case 2)
1.7
1.7
1.1 0.6
1.1 0.6 0.1 0.0
-0.05
0.00
0.05
-3.0-2.0-1.0 0.0 1.0 2.0 3.0 4.0
0.10
Mean air speed, m/s
Mean air temperature, °C
1.7
1.7
Height, m
Height, m
0.6
0
0.1 0.0
b)
1.1
0.1 0.0
Height, m
Height, m
a)
1.7
Height, m
Indoor air parameter
179
1.1
1.1
0.6
0.6
0.1 0.0
0.1 0.0
0.00
0.06
0.12
Mean air speed, m/s
0.18
-1.0 0.0 1.0 2.0 3.0 4.0
Mean air temperature, °C measurements CFD
Fig. 12. The comparison of the results of numerical prediction and measurements for the mean air speed (left) and mean air temperature (right) at different test heights a) case 2, b) case 1A.
180
A. Palmowska, B. Lipska / Building and Environment 108 (2016) 171e182
all cases. They were in the range of 0.6 ÷ 2.3 C and their mean value was 1.4 C. However, all values obtained from the simulation fit the range of measuring error of temperature register, except the column I at a height of 1.7 m for the case 1A (b). The calculated values of mean relative air humidity were lower than the measured values. The deviations between them were similar for all cases. They were within the range of 11 ÷ 23% and their mean value was 17%. Nevertheless, all values obtained from the numerical calculations did not fit the range of measuring error for each case.
due to technical reasons did not include measurement of this parameter above the ice surface. The deviation values were similar at all test heights and were in the range of 13 ÷ 16%. Their mean value was equal to 15%. Comparing both measured and calculated values of the relative humidity and temperature of air in the area above the ice surface to the Mollier chart in Fig. 5, it was found that in the facility an excess of air moisture occurred, causing the condensation of water vapour on the ice surface and the risk of fog formation. The scope of validation also encompassed temperature and relative humidity of air at the outskirts of the ice rink. The comparison of their measured and calculated values is shown in Fig. 14 for the selected columns G, K and I. The calculated values of mean air temperature were higher than the measured values. The deviations between them were similar for
4.3. The temperature of the ceiling The results of numerical calculations of the inner surface temperature of the ceiling for cases 2 and 3 were also validated. For this
I 1.7 m
I 1.7 m
Column
K 1.7 m
Column
a) K 1.7 m
I 1.1 m
I 1.1 m
I 0.6 m
I 0.6 m
G 1.7 m
G 1.7 m
-1 0 1 2 3 4 5 Mean air temperature, °C
0 25 50 75 100 Mean relative air humidity, %
I 1.7 m
I 1.7 m
Column
K 1.7 m
Column
b) K 1.7 m
I 1.1 m
I 1.1 m
I 0.6 m
I 0.6 m
G 1.7 m
G 1.7 m 0
-1 0 1 2 3 4 5 6 Mean air temperature, °C K 1.7 m
K 1.7 m
I 1.7 m
I 1.7 m
I 1.1 m
I 0.6 m
G 1.7 m
G 1.7 m 1
2
3
4
5
Mean air temperature, °C
75 100
I 1.1 m
I 0.6 m
0
50
Mean relative air humidity, %
Column
Column
c)
25
0
25
50
75 100
Mean relative air humidity, % measurements CFD
Fig. 14. The comparison of the results of numerical prediction and measurements for the mean air temperature (left) and mean relative air humidity (right) at different heights at the outskirts of the ice rink a) case 2, b) case 1A, c) case 3.
A. Palmowska, B. Lipska / Building and Environment 108 (2016) 171e182
181
Fig. 15. The comparison of the thermal imaging research (left) and numerical prediction (right) for the temperature distribution on the ceiling a) case 2, b) case 3.
purpose they were compared with the results of thermal imaging research as shown in Fig. 15. Comparing the simulations and measurements results for case 2 (a) the similar temperature distribution on the ceiling surface can be clearly observed. The ceiling surface temperature oscillated around the value of 1.5 C. The higher values of temperature occurred only in the region of the lighting system. A similar situation was observed in case 3 (b), where ceiling temperature was in the range of 0 ÷ 2.5 C. In case 2 (a) warmer areas on the ceiling over the supply air duct, shown in the measurement results, but not found in the numerical prediction, were the result of heating the ceiling surface from the hot surface of the ventilation duct, which in turn was heated from warm supply air flowing inside. In the simulations supply air was set directly in the supply opening, thus the CFD model did not take into account the flow of air through the modelled duct. Whereas, in case 3 (b) the warmer edge of the ceiling (so-called thermal bridge) was clearly visible on the thermogram, which also did not occur in simulation results due to the geometry of the numerical model which assumed no leaking between walls. The thermal imaging research is useful for the validation as well as to check whether the risk of condensation of water vapour contained in the air onto ceiling surface occurs, when its temperature is below the air dew point temperature. In studied cases the condensation appeared locally, both in the CFD and during the experiment. A detailed way to verify the risk of condensation along with the other results for the tested object was presented in Ref. [29]. 5. Conclusions 1. The experimental research, carried out in the investigated ice rink arena, allowed us to identify thermal and humidity conditions and validate numerical calculation results, performed for the developed model of this facility, which was the aim of this paper. 2. As a result of the experimental identification of thermal and humidity conditions in the ice rink arena it was found that regardless of external conditions there was low indoor air temperature, which did not exceed the value of 3 C. There was
3.
4.
5.
6.
7.
8.
also an excess of moisture in the air, leading to the condensation of water vapour onto the ice surface and inner surface of the ceiling. By means of CFD it was possible to take into account most of the phenomena associated with the flow of air, heat and moisture in a ventilated ice rink arena. The results of measurements were used to obtain data for boundary conditions for numerical prediction as well as for experimental validation of its results. They were also useful to identify changes in indoor thermal and humidity conditions. During the numerical research our own method of moisture flux modelling based on dependence (1) was tested and an improvement of moisture flow between ice surface and indoor air was obtained. As a result of the experimental validation of the numerical model of the ice rink arena, it was proven that it well reproduced the thermal and humidity conditions in the actual facility. The exception was only the area directly over the ice surface, where deviations were greatest, in spite of refinements carried out in the CFD model. The maximum deviations between measured and calculated values for all cases and in total for all test heights (both over ice surface and on the outskirts) were following: 0.05 m/s for the speed, 3.1 C for the temperature and 23% for the relative humidity. In comparison, the mean deviations were 0.03 m/s, 1.1 C and 15%, respectively. The relative humidity had the greatest deviation, but it should be noted that its predicted value depends on the temperature value, which was also burdened with some error during simulation. It might be worth comparing indirectly real moisture content in the future. The temperature distribution on the ceiling, predicted numerically, was very close to the thermogram obtained from thermal imaging research. The validated numerical model is useful in solving the problems relating to the effectiveness of the ventilation system operation under adequate thermal and humidity conditions for users, and maintaining a good technical condition of the object. In particular this includes the choice of an efficient way of air desiccation to prevent the condensation of moisture onto the ice surface and the ceiling.
182
A. Palmowska, B. Lipska / Building and Environment 108 (2016) 171e182
References [1] International Ice Hockey Federation, Technical Guidelines of an Ice Rink. Chapter 3, Switzerland, 2003. [2] B. Lipska, P. Koper, K. Jopert, Z. Trzeciakiewicz, Numerical study of air distribution in indoor ice rink arena, Ciepłownictwo Ogrzew. i Went. 42 (10) (2011) 431e437 (in Polish). [3] A. Palmowska, B. Lipska, The experimental validation of numerical modelling of the air distribution in the indoor ice rink arena, in: Proceedings of the 2nd International Conference on Heat Transfer and Fluid Flow, Spain, 2015. [4] P. Rohdin, B. Mosfegh, Numerical modelling of industrial indoor environment: a comparison between different turbulence models and supply systems supported by field measurements, Build. Environ. 46 (2011) 2365e2374. [5] B. Lipska, Z. Trzeciakiewicz, J. Ferdyn-Grygierek, Z. Popiołek, The improvement of thermal comfort and air quality in the historic assemble hall of a university, Indoor Built Environ. 21 (3) (2012) 332e337. [6] W. Liu, J. Wen, C.-H. Lin, J. Liu, Z. Long, Q. Chen, Evaluation of various categories of turbulence models for predicting air distribution in an airliner cabin, Build. Environ. 65 (2013) 118e131. [7] S. Lestinen, T. Laine, T. Sundman, Scoring an HVAC Goal for Hockey Spectators. CFD Is Used to Design Ventilation Systems for Sports Arenas. ANSYS Advantage I/1, 2007, pp. 11e12. [8] Report IEA Annex 26. Ventilation of Large Spaces in Buildings Part 3: Analysis and Prediction Techniques. Chapter 4.1: Olympic Mountain Hall, GjØvik. Norvay, 1998, pp. 191e203. [9] C. Yang, P. Demokritou, Q. Chen, J. Spengler, Experimental validation of a computational fluid dynamics model for IAQ applications in ice rink arenas, Indoor Air 11 (2) (2001) 120e126. [10] P. Demokritou, C. Yang, Q. Chen, J. Spengler, An experimental method for contaminant dispersal characterization in large industrial buildings for indoor air quality (IAQ) applications, Build. Environ. 37 (3) (2002) 305e312. [11] O. Bellache, M. Ouzzane, N. Galanis, Numerical prediction of ventilation patterns and thermal processes in ice rinks, Build. Environ. 40 (3) (2005) 417e426. [12] O. Bellache, M. Ouzzane, N. Galanis, Coupled conduction, convection, radiation, heat transfer with simultaneous mass transfer in ice rinks, Numer. Heat Transf. Part A Appl. An Int. J. Comput. Methodol. 48 (3) (2005) 219e238. [13] A. Daoud, N. Galanis, O. Bellache, Calculation of refrigeration loads by convection, radiation and condensation in ice rinks using a transient 3D zonal
model, Appl. Therm. Eng. 28 (14e15) (2008) 1782e1790. [14] C.A. Balaras, A.A. Argiriou, Infrared thermography for building diagnostics, Energy Build. 34 (2) (2002) 171e183. [15] A. Stobiecka (Palmowska), P. Koper, B. Lipska, The comparison of air distribution systems in ice rink arena ventilation, Sci. Future Lithuania 5 (4) (2013) 429e434. [16] K. Jopert, Modeling of Airflow in the Ice Rink Arena, Msc thesis, Silesian University of Technology, Poland, 2010 (in Polish). [17] H. Guo, S.C. Lee, L.Y. Chan, Indoor air quality in ice skating rinks in Hong Kong, Environ. Res. 94 (3) (2004) 327e335. [18] J-T. Lin, Y.K. Chuah, Prediction of infiltration rate and the effect on energy use for ice rinks in hot and humid climates, Build. Environ. 45 (1) (2010) 189e196. [19] P. Sormunen, T.L. Sundman, S. Lestinen, The design challenges of multipurpose Arenas, in: Proceedings of Clima WellBeing Indoors, 2007 (Finland). [20] Desert Aire, Indoors Ice Rink Dehumidification. Application Note 13 Rev.10, 2007. [21] ASHRAE Handbook: Refrigeration. 1791 Tullie Circle, Atlanta, 2010. GA 30329 2010. [22] ISO 7194: Measurement of Fluid Flow in Closed Conduits. Velocity-area Methods of Flow Measurement in Swirling or Asymmetric Flow Conditions in Circular Ducts by Means of Current-meters or Pitot Static Tubes. [23] ISO 18434e1: Condition Monitoring and Diagnostics of Machines. Thermography. Part 1: General Procedures. [24] ISO 7726: Ergonomics of the Thermal Environment - Instruments for Measuring Physical Quantities. [25] P. Ciuman, Influence of the Turbulence Modeling on the Prediction of Numerical Supply Air Jets, Msc thesis, Silesian University of Technology, Poland, 2013 (in Polish). [26] A. Palmowska, Modeling of Ventilation Air Distribution in the Indoor Ice Rink Arena, PhD thesis, Silesian University of Technology, Poland, 2016 (in Polish). [27] Ansys CFX Help: ANSYS CFX- Solver, Release 10.0: Radiation Modelling, 2005. [28] M. Hurnik, M. Blaszczok, Z. Popiołek, Air distribution measurement in a room with a sidewall jet: a 3D benchmark test for CFD validation, Build. Environ. 93 (2015) 319e330. [29] A. Palmowska, G. Miczka, The usage of a thermal imaging camera to study the thermal and humidity conditions in indoor ice rink arena, INSTAL 3 (2015) 44e49 (in Polish). [30] https://gliwice.eu/miasto/sport-i-rekreacja/obiekty-sportowe/lodowiska/ lodowisko-tafla (last accessed 17.06.16).
Contents lists available at ScienceDirect
Building and Environment journal homepage: www.elsevier.com/locate/buildenv
Experimental study and numerical prediction of thermal and humidity conditions in the ventilated ice rink arena Agnieszka Palmowska*, Barbara Lipska Department of Heating, Ventilation and Dust Removal Technology, Silesian University of Technology, Konarskiego 20, 44-100, Gliwice, Poland
a r t i c l e i n f o
a b s t r a c t
Article history: Received 17 June 2016 Received in revised form 8 August 2016 Accepted 24 August 2016 Available online 26 August 2016
An ice rink arena is a place where the ventilation system has to deal with the presence of spectators and skaters (who have different requirements), and with the temperature of the ice surface below 0 C. An important problem that arises is excess moisture, leading to condensation of water vapour onto the ice surface and inner surface of the ceiling. The aim of this presented paper was to experimentally identify physical phenomena occurring in the actual ventilated ice rink arena in Gliwice (Poland) and to check whether the developed numerical model correctly reproduces such phenomena and how it should be improved for this purpose. Long-term and short-term experimental research was carried out to obtain data for boundary conditions, to identify changes in thermal and humidity conditions and for experimental validation of simulation results. Air parameters: speed, velocity, temperature and relative humidity were measured. Thermal imaging measurements were also carried out. The numerical model was prepared by means of the Ansys CFX 14.5. Improvement of the moisture flux numerical modelling was carried out. The scope of validation encompassed the comparison of indoor airflow pattern, air parameters (above the ice surface and on the outskirts of the ice rink) and temperature of the ceiling. The numerical model was able to map real conditions in the object with good agreement between measured and predicted values. The mean deviations for all studied cases did not exceed values of 0.03 m/s for speed, 1.1 C for temperature and 15% for relative humidity of the indoor air. © 2016 Elsevier Ltd. All rights reserved.
Keywords: Ice rinks Ventilation Humidity Measurement CFD prediction Experimental validation
1. Introduction An ice rink arena is a ventilated large-volume building, used for sports and recreational activities. Ice rink design and operation are totally unique and differ in many ways from standard buildings [1]. Therefore, the proper operation of a ventilation system with adequate conditions for users, along with the maintenance of good technical condition of the facility is required. It means that the conditions need to meet the ventilation requirements for people inside the building, and that the building is safe for the people within. However, there are no fixed standards and guidelines exist for the design of ventilation systems in ice rink arenas, especially in terms of air distribution systems, so therefore there is a great need for further research. An ice rink arena is a facility where ventilation system has to deal with an ice surface temperature of below 0 C and the presence
˛ ska, Katedra Ogrzewnictwa, Wentylacji i * Corresponding author. Politechnika Sla Techniki Odpylania, ul. Konarskiego 20, 44-100, Gliwice, Poland. E-mail address: [email protected] (A. Palmowska). http://dx.doi.org/10.1016/j.buildenv.2016.08.024 0360-1323/© 2016 Elsevier Ltd. All rights reserved.
of people, both spectators and skaters, whose requirements are different (from the point of view of thermal comfort). Ventilation must meet the requirements of both parties. However, thermal comfort conditions were not analyzed in this paper. An important problem that occurs in such objects is the risk of exceeding the permissible moisture level in the indoor air, leading to condensation of water vapour contained in the air onto the wall surfaces and formation of fog above the ice surface. Therefore, the ventilation system in an ice rink arena should fulfil the following functions: - Maintaining adequate thermal and humidity conditions for users of the ice rink arena, which is the function of ventilation or air heating, - Removing excess moisture above the ice surface, which is the function of dehumidification [2], A detailed description of the air parameter requirements was discussed by Palmowska and Lipska [3]. Due to geometry and complex flow phenomenon, it is difficult to
172
A. Palmowska, B. Lipska / Building and Environment 108 (2016) 171e182
predict the effects of the ventilation system by using traditional engineering methods. Therefore, a CFD technique is used for scientific research in this field. This method, based on the basic equations of fluid flow and heat, can be used for researching the flow of air, heat, moisture and contaminants in such facilities. The CFD allows us to check and test different design solutions [4], which is an advantage over a physical experiment. Furthermore, it can be used to improve thermal and humidity conditions in the ventilated facilities [5]. As a rule, the CFD does not replace the measurements completely, but the amount of experimentation and the overall cost can be significantly reduced. Moreover, it enables an assessment of ventilation work effectiveness, as early as at the design stage. Last but not least, the CFD modelling is a useful tool for the optimization of ventilation systems from the viewpoint of fulfiling the users' requirements. It should be noted that the CFD modelling can be burdened with errors, thus it must be validated experimentally before starting appropriate tests with its use [6]. Nowadays, the design of sports arenas is not complete without one or more simulations of the indoor air flow. Analyses of this type are used to determine the best air conditioning systems and ensure that the occupants are exposed to a predefined thermal comfort range most or all of the time [7]. So far, ventilation issues in ice rinks are relatively unknown and poorly supported by research. In literature only a few numerical studies underpinned by the experiment can be found: the Gjøvik Olympic Cavern Hall (Norway) [8], the ice rink arena in Greater Boston (USA) [9,10], and the ice rink arena in Montreal (Canada) [11,12]. The results of numerical calculations of these objects were validated by comparison with the results of the measurements. The research carried out for the facility in Gjøvik is the first experimental identification of conditions in a ventilated ice rink and its numerical modelling to be documented in literature. The distributions of temperature and air velocity were compared. In simulations, thermal radiation with the use of Discrete Transfer Model (DTM) was taken into account. Air temperature and velocity distributions were also compared for the ice rink arena in Greater Boston. Additionally, CFD technique was used to validate the spread of contaminants such as PFT and CO [10]. In the object arena in Montreal calculations results were compared with measured air velocity in the real object and with the use of its physical model. The impact of different emissivity of the ceiling on the air parameters and the ceiling's temperature was additionally studied [12]. It should be noted that the scope of all published studies encompassed just temperature and velocity distribution of air, and spread of contaminants in such objects. None of these studies took into account the modelling of moisture flow in the whole building, except an attempt of humidity modelling by Daoud et al. [13] in the ice rink arena in Montreal. Ventilation air was the only source of moisture considered in such a simulation. A humidity transport model, programmed using MATLAB, returned the heat transfer due to condensation on ice as a heat gain to the ice surface in the energy model. Results of calculations, in terms of humidity, were not validated due to the lack of experimental data in this field. In the indoor ice rink thermal imaging measurements allow detection of moisture on the surface of building partitions as in the cases of residential buildings or industrial facilities. The use of infrared thermography for building diagnostics was discussed by Balaras and Argiriou [14]. The main advantage of this method is possibility to perform non-destructive tests. Nevertheless, the infrared thermography is not free from errors (surface emissivity, camera calibration, etc.). The use of thermal imaging measurements in experimental studies to validate the results of numerical calculations for ice rinks has not been yet published in literature, as well as in research in the field of ice rink ventilation.
In literature only numerical results of tests (without experimental validation) for objects Hodynka Arena, Tsherepovets Arena (Russia) [7] and the designed ice rink in Wadowice (Poland) [2,15,16], as well as only experimental studies for ice rinks located in Honk Kong [17] and Taipei (China) [18], also can be found. For example, in Ref. [7] the CFD method was used to better understand the interior conditions and flow behavior for a range of usage scenarios. The goal of these simulations was to determine how well the planned ventilation systems work to meet the desired indoor conditions [7]. The aim of the presented research was (1) to experimentally identify characteristics of physical phenomena occurring in the actual ventilated ice rink arena and (2) to check whether CFD prediction by a developed numerical model supported by experiment reproduces correctly such phenomena and (3) how it should be improved for this purpose. 2. Experiment 2.1. Tested object and the phenomena taking place The facility used for the test was the indoor ice rink “Tafla” in Gliwice (Poland) with external dimensions: length 66 m, width 37 m and a maximum height of 11 m. Inside the building there is an ice rink with the dimensions 30 60 m (Fig. 1). The use of the hall as an ice rink is expected in the period of time from October to May for different kinds of sports and recreational activities, including curling. Inside the building thermal and humidity conditions are maintained by mechanical mixing ventilation with an integrated air distribution system. A detailed description of the air distribution systems used in ice rink arenas was discussed by Stobiecka et al. [15]. It is carried out by side air supply with the use of long-throw jet nozzles and air exhaust under the ceiling by rectangular exhaust grilles. Because of economical reasons, the designed indoor air temperature amounts to þ5 C and air dehumidification is not anticipated. The air handling unit, located on the roof of the technology building, is designed for 2 fan speeds: higher (second, II), when outdoor air temperature, te, is over þ5 C and lower (first, I), when outdoor air temperature is below þ5 C. In the ice rink arena a series of complex flow phenomena takes place. Firstly, there is movement of supply air jets, which largely shapes conditions in the facility. It has an influence on air speed distribution. The air speed value just above the ice surface is very important from the users' thermal comfort viewpoint. Moreover, low air speed minimizes load on the ice making system, while too high air speed could cause ice melting, adversely affecting its quality. The air speed above the ice surface should not exceed the value of 0.25 m/s [19]. Secondly, there are also phenomena in the ice rink arena connected with the heat transfer, e.g. heat gains or losses through the building structure, heat losses from the ice surface, heat gains from the lighting system and from people. Heat exchange in the ice rink occurs by radiation and convection. It also includes vapour diffusion or condensation on the ice sheet, which is an important contributor to the ice sheet heat load. The cold ice surface, directly opposite the ceiling, absorbs heat by radiation. As a result, the inner surface of the ceiling is colder than air below it [20]. It should be observed that the water vapour contained in the air will condense on the surfaces which are colder than the room's air dew point temperature. Thirdly, in the ice rink arena a flow of moisture also occurs from different sources. Indoor sources of moisture are only people and the temporary work of ice resurfacer. Moisture is supplied into the ice rink arena with the ventilation air and by infiltration. It is very important that moisture gains can also occur from the ice rink surface. However, in most cases, moisture
A. Palmowska, B. Lipska / Building and Environment 108 (2016) 171e182
173
dehumidifier. The limit level of specific air humidity, above which condensation of water vapour and risk of fogging can occur, is determined on the ice surface temperature, ti, and relative air humidity of 100%. For this value ice surface temperature is equal to air dew point temperature. 2.2. Measurements methods The experimental research was carried out for following reasons: 1. To obtain data for boundary conditions for numerical calculations, 2. To identify changes in thermal and humidity conditions, depending on external conditions and operating mode of ice rink, 3. For experimental validation of simulation results. Long-term measurements, 24 h a day in the period of time from January to April 2014, and short-term measurements, during the night, without the presence of people, in 3 measuring series: on 8/9 February (variant 1), 21/22 March 2014 (variant 2) and 12/13 April 2014 (variant 3), were performed. Measurements were limited to half of the rink, because of the symmetry of the facility. Fig. 1 (middle) shows the location of test points on the tested half of the rink, marking of long-throw jet nozzles and columns. 2.2.1. Field measurements to establish CFD model boundary condition Boundary conditions were obtained on the basis of the measurements, which included: long-term tests of temperature and relative humidity 4e of the outdoor air, temperature tN and relative humidity of supply air 4N, short-term tests of supply airflow rate, temperature of the ice surface, temperature distribution on the wall's inner surfaces and emissivity of ice εi by means of thermography.
Fig. 1. The indoor ice rink “Tafla” in Gliwice (Upper Silesia, Poland) (upper) [30], top view with marked test points (middle) and cross-section (bottom).
travels to the ice surface. Moisture flux between the ice surface and _ can be expressed by the relationship between the indoor air (W) mass heat transfer coefficient, the specific humidity in air Xa and in saturated ice Xi, in accordance with the ASHRAE Handbook [21]:
_ ¼ K·ðXa X Þ·A W i i
(1)
where K is the mass heat transfer coefficient and is equal to 0.00023 kg/s$m2.Ai is the ice rink area (m2). If the air humidity is higher than the humidity of ice (Xa>Xi) there is a positive moisture flux. It means moisture condensation on the ice surface. To prevent condensation, the room air dew point temperature should be lower than the ice surface temperature and the ceiling inner surface temperature. In practice, to avoid condensation, the moisture content need to be balanced and controlled by a
Outdoor air parameters were obtained from meteorological station DAVIS located in the distance of about 500 m away from the tested facility. Long-term measurement was carried out with a time step of 15 min. Supply airflow rate was measured indirectly by two methods based on measurement of air velocity, va. In the first method, the anemometer with Pitot tube VPT-100 VOLTCRAFT was used for the measurement of air velocity in long-throw jet nozzles 1e25 (Fig. 1). The air velocity was measured using a method based on dividing a circular cross section of the nozzle into rings, the number of which depends on the diameter of the nozzle in accordance with ISO 7194 Standards [22]. The measurement accuracy was ±0.25%. In the second method, the thermal anemometer TESTO was used for the measurement of air velocity in the ventilation duct (behind air ventilation unit). The air velocity was measured using a method based on dividing a rectangular cross section of the ventilation duct into squares of equal surface area in accordance with ISO 7194 Standards [22]. Measuring error of this device was 0.03 m/s þ 5% of measured value. Temperature and relative air humidity loggers APAR AR235 were used for measuring temperature and relative humidity of supply air in the first (nozzle 1) and last long-throw jet nozzle (nozzle 25) (Fig. 1). Humidity accuracy was in the range of 3 ÷ 5% and temperature accuracy was in the range of 0.5 ÷ 1.8 C. Longterm measurement was carried out with a time step of 1 min. The NTC probe for food TESTO was used to measure ice surface
174
A. Palmowska, B. Lipska / Building and Environment 108 (2016) 171e182
temperature. It was positioned in a drilled hole of a 15 mm depth. The measurement was made in the middle of the investigated half of the ice rink surface (point 9). Measurement accuracy was ±0.2 C. The measurement took 3 min. It was a time sufficient to stabilize measurement result. The measurement was repeated 2 times. Thermal measurements of temperature distribution on the wall's inner surfaces and emissivity of ice were performed by thermal imaging cameras FLIR E50 and ThermaCAM E45 with thermal sensitivity respectively 0.05 C and 0.1 C. The procedure for determining emissivity of ice in accordance with ISO 18434-1 Standards [23] was used. Measurement results were developed by FLIR Tools software. Remaining boundary conditions, i.e. heat transfer coefficients of building partitions, values of lamp power, were adopted on the basis of stocktaking of the facility. In Fig. 2 a selected thermogram of the northern partition of the ice rink arena is presented. Measurement was performed for the following indoor air parameters: temperature ta of 0.7 C and relative humidity 4a of 79%. Outdoor air temperature was (te) 2.1 C. The measurement area of the mean temperature of the north wall surface separating the hall from technical facilities is limited by a black frame. 2.2.2. Experimental identification of changes in thermal and humidity conditions in the ice rink arena An experimental identification of changes in thermal and humidity conditions in the ice rink arena included long-term measurements of temperature (ta) and relative humidity (4a) of indoor air with the use of the loggers APAR AR235, mentioned before. They were located along the ice rink on the selected columns E÷L (Fig. 1)
at a height of 1.7 m and additionally at heights: 1.1 m and 0.6 m on column I. Measurements were carried out with a time step of 5 min. The impact of occupant presence and the work of the ice resurfacer on the distribution of relative air humidity was observed. In Fig. 3 daily distribution of relative humidity and temperature of air, when the ventilation system was off, at the height of 1.7 m (column F) is shown. During ice resurfacer work, as well as during the presence of people, the relative humidity of air gradually increased. The lowest recorded value of this parameter of air, 4a, since the opening of the ice rink, was the value of 86.3%. The highest value of 4a was 88.6%. The temperature of air also gradually increased during the day. The lowest value of ta was 2.5 C and the highest value of ta was 0.2 C. In the period of the long-term test, when the outside air temperature was in the range of 15 C to þ25 C, indoor air temperature did not reach a value greater than 3 C and less than 3 C (Fig. 4). It means that designed indoor air temperature of 5 C was not achieved. Fig. 5 shows a Mollier chart with a marked area which contains the conditions of air inside the ice rink arena during long-term measurements. In most of these conditions moisture content was higher than the limit value, determined by ice surface temperature ti ¼ 4.5 C. Therefore, the air dew point temperature was higher than the temperature of the ice surface. There was condensation of water vapour contained in the air and then freezing of moisture onto the ice surface, which had to be removed by the ice resurfacer. 2.2.3. Experiment for validation Short-term measurements of speed, wa, temperature and relative air humidity above the ice surface and short-term thermographic measurements of temperature distribution on the ceiling's inner surface were necessary for experimental validation of simulation results. Visualization by smoke test to obtain airflow pattern information was also made. Measurements of the parameters of air above the ice surface were carried out at 9 measuring points evenly distributed on the surface of the ice rink (Fig. 1) at 4 specific heights: 0.1 m, 0.6 m, 1.1 m, 1.7 m in accordance with ISO 7726 Standards [24]. The measurement time was 15 min at each point. At point 9, it lasted 60 min. 3. Numerical method 3.1. Model description and boundary conditions
Fig. 2. Thermogram of northern partition of tested ice rink arena.
The numerical model of the tested object was prepared with the use of the Ansys CFX 14.5 code, integrated with Workbench 14.5 platform. It took into consideration the designed overall dimension of the facility and its geometry. Ice rink, lighting system and ventilation system were modelled (Fig. 6). The XY symmetry plane was introduced because of the symmetry of this object. This enabled the greater refinement of the discretization grid on the available server. The numerical model included the tested half of the ice rink and it did not take into account the presence of people, just like in the real object during the experimental research. The ice sheet was modelled as a surface of uniform temperature. The lighting system, consisting of 37 metal halide lamps, was modelled as rectangular heat sources located in the top part of the hall. Only those lamps, which were turned on during experimental measurements, were included as heat sources (marked in yellow colour on Fig. 6). Convective and radiative heat gains from lamps, as energy and radiation sources, were related to the surface of geometry of the lighting model. The long-throw jet nozzles were modelled as circular supply openings with a diameter of Ø160 and Ø80 mm, placed on the
175
0 -1
87 -2 -3
7: 31 8: 16 9: 01 9: 46 10 :3 11 1 :1 12 6 :0 12 1 :4 13 6 :3 14 1 :1 15 6 :0 15 1 :4 16 6 :3 17 1 :1 18 6 :0 18 1 :4 19 6 :3 20 1 :1 21 6 :0 21 1 :4 22 6 :2 23 1 :0 6
86
Air temperature ta. °C
1
ice resurfacer
ice skating
ice skating
ice resurfacer
ice resurfacer
ice skating
ice skating
ice resurfacer
ice skating
ice resurfacer
ice skating
ice resurfacer
ice resurfacer
ice skating
ice skating
88
ice resurfacer
Relative air humidity φa, %
89
opening
A. Palmowska, B. Lipska / Building and Environment 108 (2016) 171e182
Relative air humidity
Time, h
Air temperature
Fig. 3. Daily distribution of relative humidity and temperature of air in the tested ice rink arena (column F, height of 1.7 m).
lighting system
exhaust
Fig. 4. The relationship between indoor and outdoor air temperature in the tested ice rink arena (column J, height of 1.7 m).
symmetry plane ice surface
ventilation air supply tribune
Fig. 6. The numerical model of half of the rink hall with marked directions XYZ.
ventilation supply duct, which was modelled as a circular duct. Larger nozzles were inclined at an angle of 75 relative to the vertical axis of the ventilation duct, while the smaller at an angle of 45 . The exhausts were set as rectangular grilles with dimensions 800 150 mm, also located on the ventilation exhaust duct. The model included penetrating heat fluxes through the internal and external partitions. Heat transfer coefficients and outside temperature were set to internal and external walls. The exception was the part of the north wall separating the ice rink arena from technical facilities. The mean temperature of its surface, obtained from thermovision measurements, was used. The following averaged measured air parameters were set for supply openings: mass flow rate, temperature and specific humidity. Exhaust air mass flow rates were set for exhaust openings. Boundary conditions were prepared for the following calculation cases for the entire facility: 1) case 1, 1A: based on the variant 2 (II fan speed), 2) case 2: based on the variant 1 (I fan speed), 3) case 3: based on the variant 3 (II fan speed). Data required to perform the numerical calculations are listed in Table 1 and Table 2.
3.2. Numerical procedure Fig. 5. Mollier chart with area containing air conditions in the tested ice rink arena during long-term measurements.
The numerical calculations were carried out using ANSYS CFX
176
A. Palmowska, B. Lipska / Building and Environment 108 (2016) 171e182
Table 1 The boundary conditions for the ice rink arena, obtained from measurements. Boundary condition
Unit
Outdoor air temperature Temperature of supply air Specific humidity of supply air ¼ Specific humidity of outdoor air Mass flow rate of supply air ¼ Mass flow rate of exhaust air Ice surface temperature Mean temperature of the north wall surface separating hall from technical facilities Mean ice surface emissivity
Table 2 The boundary conditions for the ice rink arena, obtained from facility stocktaking. Boundary condition Heat transfer coefficient Heat transfer coefficient Heat transfer coefficient Heat transfer coefficient Lighting power Ground temperature
Unit of of of of
external walls (N, W, E) wall in the ground (E) floor ceiling
Value 2
W/m K W/m2K W/m2K W/m2K W/m2 C
0.187 0.140 0.203 0.301 1206 2.0
14.5 software, in steady-state, three-dimensional and nonisothermal conditions. Discretization of model equations was solved by Finite Volume Method. The high resolution discretization scheme and the Rhie Chow redistribution algorithm for coupling pressure-velocity were used. The CFD model used the Shear Stress Transport (SST) turbulence model from the EVM models group. The choice of the turbulence model was based on study results [25]. No-slip condition was adopted on the wall surface. A minimum wall distance Yþ was equal to 34. Thermal radiation between walls, ice surface and objects located inside the ice rink arena was performed by the Discrete Transfer Model, selected as a result of our own research [26]. This model is based on tracing the domain by multiple rays leaving from the bounding surfaces. The technique was developed by Shah in 1979 and depends upon the discretisation of the transfer equation along rays [27]. The applied unstructured discretization grid was composed of 12574808 cells, built mainly with tetrahedral elements, and 3161134 nodes (Fig. 7). It also included local refinement around outlet and inlet openings, and just above the ice rink surface (Fig. 7). The basic mesh element size was 35 cm, with a refinement over the ice surface up to 7 cm. 5 layers consisting of prismatic
Calculation cases 1
2
3
C C kg(H2O)/kg(dry air)
11.2 12.5 0.0040
6.3 19.8 0,0044
4.8 14.2 0.0048
kg/s C C e
4.8 4.45 5.9 0.96
2.2 4.70 3.6 0.96
4.8 4.50 3.5 0.96
elements, with a total maximum thickness of 17 cm in the boundary layer at the surface of the partitions were included in the numerical model. A grid independence test was performed for different grids in order to select the grid, which most accurately reproduces the conditions in the facility. The calculations were carried out using iteration method as long as the convergent solutions were obtained. Achieved convergence after approximately 6000 iterations was in the range of 1.0E-6 (for turbulence, momentum and mass, wallscale) to 1.0E-4 (for heat transfer and mass fractions).
3.3. Improvement of the moisture flux numerical modelling The moisture flux between the ice surface and indoor air in preliminary calculations for the case 1 was modelled by treating the ice sheet as a source of moisture with default mass flow boundary condition. Constant value of this stream was determined based on Formula (1). It did not take into account local changes in flow direction of moisture: from the ice into the air or inversely, depending on the difference between the specific humidity in air and saturated ice. Therefore, the distribution of specific air humidity above the ice surface could be burdened with errors. To prevent this, our own modification of the method of numerical modelling of such emission for case 1A was proposed. It was based on the introduction in Ansys CFX user function, developed using the Formula (1). This procedure allowed calculation of the mass flow of moisture separately for each node of the discretization grid with an unknown value of specific humidity of air above the ice surface Xa. The unknown value Xa was replaced in the subsequent iteration steps of numerical calculations by the calculated value of specific humidity in the first mesh grid above the ice surface. Fig. 8a shows contour maps of the specific air humidity on the
Fig. 7. Cross-section through the mesh discretization (left) and local refinement of the discretization grid around the long-throw jet nozzles (right).
A. Palmowska, B. Lipska / Building and Environment 108 (2016) 171e182
177
Fig. 8. Maps of the specific air humidity in the tested ice rink arena a) in the horizontal plane ZX, Y ¼ 0.1 m, b) at the height of 0.1 m as a result of the experiment, c) in the vertical plane XY, Z ¼ 24 m.
height of 0.1 m above the ice surface from CFD simulations by means of two methods of moisture emission modelling. The calculation results were related to the distribution shown in Fig. 8b. It was made by using the values of specific humidity, determined on the basis of measurements of relative humidity and temperature. For case 1 the distribution of specific air humidity was uniform over the entire ice surface, which was not compatible with the measurements results (b). An average relative deviation between the calculation results and measurements was 11.5%. In case 1A the distribution was irregular and an improvement of modelling accuracy with an average deviation of 8.5% has been obtained. A method of moisture emission modelling from the ice surface had an impact on the distribution of specific humidity not only directly above the ice surface, but also in the upper regions of the facility. Fig. 8c shows the comparison of this parameter distribution in the vertical plane XY (Z ¼ 24 m). In the case 1A specific air humidity values were lower than in the case 1. The lowest values occurred directly over the ice surface due to air dehumidification as a result of higher air humidity than in the saturated ice. This caused a humidity decrease in the area over the ice surface (the area limited by the red frame), which could not be predicted in case 1. Further numerical calculations were carried out taking into account our own function for the moisture flux modelling from the ice surface.
4. Validation of the CFD results and discussion 4.1. Indoor airflow pattern For case 3, the predicted airflow pattern as an isosurface of air velocity va ¼ 0.7 m/s was compared with the visualization by smoke test (Fig. 9). Both in the experiment and in the CFD modelling the initial falling and then, at the height of 3 m above the floor, the deflection of supply air jet were observed. The numerical model of the object well reproduced the airflow pattern. 4.2. Indoor air parameters Predicted values of the indoor air parameters such as speed, temperature and relative humidity, were validated using the measurement data for different variants. The air velocity values, calculated directly by means of CFD code, were converted to speed values on the basis of the method proposed in Ref. [28]. Relative air humidity was determined by direct simulation results of specific humidity using our own user function based on humidity conversion formulas. For the needs of validation it enabled the comparison of the predicted parameters with the values of speed and relative air humidity obtained during the measurements. For case 3 a detailed comparison of CFD results with the measurements results of the air parameters in the facility at 9 test
Fig. 9. The comparison of airflow pattern: predicted as an isosurface of air velocity (left) and visualized by smoke test (middle, right) (case 3).
178
A. Palmowska, B. Lipska / Building and Environment 108 (2016) 171e182
Fig. 10. The comparison of the results of numerical prediction and measurements for the indoor air parameters at the height of 0.6 m above the ice surface a) speed, b) temperature, c) relative humidity (case 3).
points above the ice surface was carried out. This is shown in Fig. 10 for the height of 0.6 m. The mean deviation between the measured and calculated values at each point was: 0.1 m/s for the speed, 0.5 C for the temperature and 9% for the relative air humidity. It was noted that the calculated values of relative air humidity in part were not in the range of the measured values, which included the higher values of this parameter. In the case of calculated temperature and speed values they were not always exactly at the point where the data values were measured, but in another part of the ice surface. However, in most cases they did not exceed the range of values from the measurements, except some small local areas of
b)
Y = 0.6 m
Y = 0.1 m, 0.6 m, 1.1 m, 1.7 m
100
100
80
80
Probability, %
Probability, %
a)
speed above 0.2 m/s. The evaluation of relative deviation dispersion between the results of calculations and measurements (Fig. 11) for the analyzed plane (a) and in total for all heights (b) for each air parameter, with marked measurement error bars, was carried out. The relative deviation was calculated as the difference between calculated and measured value in relation to the measured value. At the height of 0.6 m (a) the probability of relative deviation below 20% between the results of measurements and calculations for relative air humidity was 100%. For the temperature it was 33% probability and 11%, for the speed. In contrast, the chance of
60 40 20
60 40 20 0
0
Relative deviation, % relative air humidity
air temperature
Relative deviation, % air speed
Fig. 11. The probability of relative deviation between the results of numerical prediction and measurements for the temperature, speed and relative air humidity a) at a height of 0.6 m b) in total for all test heights above the ice surface (case 3).
A. Palmowska, B. Lipska / Building and Environment 108 (2016) 171e182 Table 3 The mean values of predicted and measured indoor air parameters at the height of 0.6 m above the ice surface. Unit
Mean measured value
Mean predicted value
Speed Temperature Relative humidity
m/s C %
0.06 ± 0.02 2.3 ± 0.2 96 ± 5
0.11 2.6 87
deviation below 50% was 89% and 11% respectively. Small probability values of the speed resulted from large differences in relative deviation values, and they in turn were caused by difference values in comparison with small values of this parameter. Similar results were obtained in total for all test heights (b). Therefore it was found that the main source of deviation was not the air parameter value, but its other distribution on the plane. However, from the viewpoint of reproducing the thermal and humidity conditions in the occupied zone and their assessment, the point values are not as important as their mean values at a given height. Table 3 shows the comparison of mean values of measurements and calculations for a selected plane at the height of 0.6 m above the ice surface. The deviations between these values, without taking into account the measurement error, were as follows: 0.05 m/s for the speed, 0.3 C for the temperature and 9% for the relative humidity. Taking this into consideration, for other cases (1A, 2) the results of numerical prediction and measurements of the indoor air parameters at the test heights of 0.1 m, 0.6 m, 1.1 m and 1.7 m above the ice surface were compared, as shown in Fig. 12 and Fig. 13. The deviations between the measured and calculated values of
25
50
75
100
Mean relative air humidity, % measurements CFD Fig. 13. The comparison of the results of numerical prediction and measurements for the mean relative air humidity at different test heights for case 1A.
mean air speed were higher in the upper part of the facility than directly over the ice surface. For case 2 (a) they were in the range of 0.01 ÷ 0.03 m/s and 0.01÷ 0.04 m/s for case 1A (b). The mean value of the deviation for both cases was 0.02 m/s. The deviations of mean air temperature for case 2 (a) were within the range of 0.4 ÷ 3.1 C and 0.4 ÷ 0.9 C for case 1A (b). The mean value of the deviation for both cases was 1.1 C. The comparison of the relative air humidity results was carried out only for case 1A, because the experiment for 1 variant (case 2)
1.7
1.7
1.1 0.6
1.1 0.6 0.1 0.0
-0.05
0.00
0.05
-3.0-2.0-1.0 0.0 1.0 2.0 3.0 4.0
0.10
Mean air speed, m/s
Mean air temperature, °C
1.7
1.7
Height, m
Height, m
0.6
0
0.1 0.0
b)
1.1
0.1 0.0
Height, m
Height, m
a)
1.7
Height, m
Indoor air parameter
179
1.1
1.1
0.6
0.6
0.1 0.0
0.1 0.0
0.00
0.06
0.12
Mean air speed, m/s
0.18
-1.0 0.0 1.0 2.0 3.0 4.0
Mean air temperature, °C measurements CFD
Fig. 12. The comparison of the results of numerical prediction and measurements for the mean air speed (left) and mean air temperature (right) at different test heights a) case 2, b) case 1A.
180
A. Palmowska, B. Lipska / Building and Environment 108 (2016) 171e182
all cases. They were in the range of 0.6 ÷ 2.3 C and their mean value was 1.4 C. However, all values obtained from the simulation fit the range of measuring error of temperature register, except the column I at a height of 1.7 m for the case 1A (b). The calculated values of mean relative air humidity were lower than the measured values. The deviations between them were similar for all cases. They were within the range of 11 ÷ 23% and their mean value was 17%. Nevertheless, all values obtained from the numerical calculations did not fit the range of measuring error for each case.
due to technical reasons did not include measurement of this parameter above the ice surface. The deviation values were similar at all test heights and were in the range of 13 ÷ 16%. Their mean value was equal to 15%. Comparing both measured and calculated values of the relative humidity and temperature of air in the area above the ice surface to the Mollier chart in Fig. 5, it was found that in the facility an excess of air moisture occurred, causing the condensation of water vapour on the ice surface and the risk of fog formation. The scope of validation also encompassed temperature and relative humidity of air at the outskirts of the ice rink. The comparison of their measured and calculated values is shown in Fig. 14 for the selected columns G, K and I. The calculated values of mean air temperature were higher than the measured values. The deviations between them were similar for
4.3. The temperature of the ceiling The results of numerical calculations of the inner surface temperature of the ceiling for cases 2 and 3 were also validated. For this
I 1.7 m
I 1.7 m
Column
K 1.7 m
Column
a) K 1.7 m
I 1.1 m
I 1.1 m
I 0.6 m
I 0.6 m
G 1.7 m
G 1.7 m
-1 0 1 2 3 4 5 Mean air temperature, °C
0 25 50 75 100 Mean relative air humidity, %
I 1.7 m
I 1.7 m
Column
K 1.7 m
Column
b) K 1.7 m
I 1.1 m
I 1.1 m
I 0.6 m
I 0.6 m
G 1.7 m
G 1.7 m 0
-1 0 1 2 3 4 5 6 Mean air temperature, °C K 1.7 m
K 1.7 m
I 1.7 m
I 1.7 m
I 1.1 m
I 0.6 m
G 1.7 m
G 1.7 m 1
2
3
4
5
Mean air temperature, °C
75 100
I 1.1 m
I 0.6 m
0
50
Mean relative air humidity, %
Column
Column
c)
25
0
25
50
75 100
Mean relative air humidity, % measurements CFD
Fig. 14. The comparison of the results of numerical prediction and measurements for the mean air temperature (left) and mean relative air humidity (right) at different heights at the outskirts of the ice rink a) case 2, b) case 1A, c) case 3.
A. Palmowska, B. Lipska / Building and Environment 108 (2016) 171e182
181
Fig. 15. The comparison of the thermal imaging research (left) and numerical prediction (right) for the temperature distribution on the ceiling a) case 2, b) case 3.
purpose they were compared with the results of thermal imaging research as shown in Fig. 15. Comparing the simulations and measurements results for case 2 (a) the similar temperature distribution on the ceiling surface can be clearly observed. The ceiling surface temperature oscillated around the value of 1.5 C. The higher values of temperature occurred only in the region of the lighting system. A similar situation was observed in case 3 (b), where ceiling temperature was in the range of 0 ÷ 2.5 C. In case 2 (a) warmer areas on the ceiling over the supply air duct, shown in the measurement results, but not found in the numerical prediction, were the result of heating the ceiling surface from the hot surface of the ventilation duct, which in turn was heated from warm supply air flowing inside. In the simulations supply air was set directly in the supply opening, thus the CFD model did not take into account the flow of air through the modelled duct. Whereas, in case 3 (b) the warmer edge of the ceiling (so-called thermal bridge) was clearly visible on the thermogram, which also did not occur in simulation results due to the geometry of the numerical model which assumed no leaking between walls. The thermal imaging research is useful for the validation as well as to check whether the risk of condensation of water vapour contained in the air onto ceiling surface occurs, when its temperature is below the air dew point temperature. In studied cases the condensation appeared locally, both in the CFD and during the experiment. A detailed way to verify the risk of condensation along with the other results for the tested object was presented in Ref. [29]. 5. Conclusions 1. The experimental research, carried out in the investigated ice rink arena, allowed us to identify thermal and humidity conditions and validate numerical calculation results, performed for the developed model of this facility, which was the aim of this paper. 2. As a result of the experimental identification of thermal and humidity conditions in the ice rink arena it was found that regardless of external conditions there was low indoor air temperature, which did not exceed the value of 3 C. There was
3.
4.
5.
6.
7.
8.
also an excess of moisture in the air, leading to the condensation of water vapour onto the ice surface and inner surface of the ceiling. By means of CFD it was possible to take into account most of the phenomena associated with the flow of air, heat and moisture in a ventilated ice rink arena. The results of measurements were used to obtain data for boundary conditions for numerical prediction as well as for experimental validation of its results. They were also useful to identify changes in indoor thermal and humidity conditions. During the numerical research our own method of moisture flux modelling based on dependence (1) was tested and an improvement of moisture flow between ice surface and indoor air was obtained. As a result of the experimental validation of the numerical model of the ice rink arena, it was proven that it well reproduced the thermal and humidity conditions in the actual facility. The exception was only the area directly over the ice surface, where deviations were greatest, in spite of refinements carried out in the CFD model. The maximum deviations between measured and calculated values for all cases and in total for all test heights (both over ice surface and on the outskirts) were following: 0.05 m/s for the speed, 3.1 C for the temperature and 23% for the relative humidity. In comparison, the mean deviations were 0.03 m/s, 1.1 C and 15%, respectively. The relative humidity had the greatest deviation, but it should be noted that its predicted value depends on the temperature value, which was also burdened with some error during simulation. It might be worth comparing indirectly real moisture content in the future. The temperature distribution on the ceiling, predicted numerically, was very close to the thermogram obtained from thermal imaging research. The validated numerical model is useful in solving the problems relating to the effectiveness of the ventilation system operation under adequate thermal and humidity conditions for users, and maintaining a good technical condition of the object. In particular this includes the choice of an efficient way of air desiccation to prevent the condensation of moisture onto the ice surface and the ceiling.
182
A. Palmowska, B. Lipska / Building and Environment 108 (2016) 171e182
References [1] International Ice Hockey Federation, Technical Guidelines of an Ice Rink. Chapter 3, Switzerland, 2003. [2] B. Lipska, P. Koper, K. Jopert, Z. Trzeciakiewicz, Numerical study of air distribution in indoor ice rink arena, Ciepłownictwo Ogrzew. i Went. 42 (10) (2011) 431e437 (in Polish). [3] A. Palmowska, B. Lipska, The experimental validation of numerical modelling of the air distribution in the indoor ice rink arena, in: Proceedings of the 2nd International Conference on Heat Transfer and Fluid Flow, Spain, 2015. [4] P. Rohdin, B. Mosfegh, Numerical modelling of industrial indoor environment: a comparison between different turbulence models and supply systems supported by field measurements, Build. Environ. 46 (2011) 2365e2374. [5] B. Lipska, Z. Trzeciakiewicz, J. Ferdyn-Grygierek, Z. Popiołek, The improvement of thermal comfort and air quality in the historic assemble hall of a university, Indoor Built Environ. 21 (3) (2012) 332e337. [6] W. Liu, J. Wen, C.-H. Lin, J. Liu, Z. Long, Q. Chen, Evaluation of various categories of turbulence models for predicting air distribution in an airliner cabin, Build. Environ. 65 (2013) 118e131. [7] S. Lestinen, T. Laine, T. Sundman, Scoring an HVAC Goal for Hockey Spectators. CFD Is Used to Design Ventilation Systems for Sports Arenas. ANSYS Advantage I/1, 2007, pp. 11e12. [8] Report IEA Annex 26. Ventilation of Large Spaces in Buildings Part 3: Analysis and Prediction Techniques. Chapter 4.1: Olympic Mountain Hall, GjØvik. Norvay, 1998, pp. 191e203. [9] C. Yang, P. Demokritou, Q. Chen, J. Spengler, Experimental validation of a computational fluid dynamics model for IAQ applications in ice rink arenas, Indoor Air 11 (2) (2001) 120e126. [10] P. Demokritou, C. Yang, Q. Chen, J. Spengler, An experimental method for contaminant dispersal characterization in large industrial buildings for indoor air quality (IAQ) applications, Build. Environ. 37 (3) (2002) 305e312. [11] O. Bellache, M. Ouzzane, N. Galanis, Numerical prediction of ventilation patterns and thermal processes in ice rinks, Build. Environ. 40 (3) (2005) 417e426. [12] O. Bellache, M. Ouzzane, N. Galanis, Coupled conduction, convection, radiation, heat transfer with simultaneous mass transfer in ice rinks, Numer. Heat Transf. Part A Appl. An Int. J. Comput. Methodol. 48 (3) (2005) 219e238. [13] A. Daoud, N. Galanis, O. Bellache, Calculation of refrigeration loads by convection, radiation and condensation in ice rinks using a transient 3D zonal
model, Appl. Therm. Eng. 28 (14e15) (2008) 1782e1790. [14] C.A. Balaras, A.A. Argiriou, Infrared thermography for building diagnostics, Energy Build. 34 (2) (2002) 171e183. [15] A. Stobiecka (Palmowska), P. Koper, B. Lipska, The comparison of air distribution systems in ice rink arena ventilation, Sci. Future Lithuania 5 (4) (2013) 429e434. [16] K. Jopert, Modeling of Airflow in the Ice Rink Arena, Msc thesis, Silesian University of Technology, Poland, 2010 (in Polish). [17] H. Guo, S.C. Lee, L.Y. Chan, Indoor air quality in ice skating rinks in Hong Kong, Environ. Res. 94 (3) (2004) 327e335. [18] J-T. Lin, Y.K. Chuah, Prediction of infiltration rate and the effect on energy use for ice rinks in hot and humid climates, Build. Environ. 45 (1) (2010) 189e196. [19] P. Sormunen, T.L. Sundman, S. Lestinen, The design challenges of multipurpose Arenas, in: Proceedings of Clima WellBeing Indoors, 2007 (Finland). [20] Desert Aire, Indoors Ice Rink Dehumidification. Application Note 13 Rev.10, 2007. [21] ASHRAE Handbook: Refrigeration. 1791 Tullie Circle, Atlanta, 2010. GA 30329 2010. [22] ISO 7194: Measurement of Fluid Flow in Closed Conduits. Velocity-area Methods of Flow Measurement in Swirling or Asymmetric Flow Conditions in Circular Ducts by Means of Current-meters or Pitot Static Tubes. [23] ISO 18434e1: Condition Monitoring and Diagnostics of Machines. Thermography. Part 1: General Procedures. [24] ISO 7726: Ergonomics of the Thermal Environment - Instruments for Measuring Physical Quantities. [25] P. Ciuman, Influence of the Turbulence Modeling on the Prediction of Numerical Supply Air Jets, Msc thesis, Silesian University of Technology, Poland, 2013 (in Polish). [26] A. Palmowska, Modeling of Ventilation Air Distribution in the Indoor Ice Rink Arena, PhD thesis, Silesian University of Technology, Poland, 2016 (in Polish). [27] Ansys CFX Help: ANSYS CFX- Solver, Release 10.0: Radiation Modelling, 2005. [28] M. Hurnik, M. Blaszczok, Z. Popiołek, Air distribution measurement in a room with a sidewall jet: a 3D benchmark test for CFD validation, Build. Environ. 93 (2015) 319e330. [29] A. Palmowska, G. Miczka, The usage of a thermal imaging camera to study the thermal and humidity conditions in indoor ice rink arena, INSTAL 3 (2015) 44e49 (in Polish). [30] https://gliwice.eu/miasto/sport-i-rekreacja/obiekty-sportowe/lodowiska/ lodowisko-tafla (last accessed 17.06.16).