Large-scale multi-span greenhouses are common in

Transcription

1 AIRFLOW PATTERNS AND ASSOCIATED VENTILATION FUNCTION IN LARGE-SCALE MULTI-SPAN GREENHOUSES S. Wang, J. Deltour ABSTRACT. Understanding airflow patterns in greenhouses is essential for developing optimal natural ventilation systems. Experiments of air velocity distributions in the vertical and horizontal planes were conducted in a large multi-span greenhouse using a customized multi-point, two-dimensional sonic anemometer system. The experimental results showed that the vertical and horizontal profiles of normalized interior air velocities were well described by simple linear functions (R 2 = ). Based on the knowledge of airflow patterns within the greenhouse, a widely used nondimensional ventilation function was revised and then compared with those of similar studies in the literature. Keywords. Greenhouse, Air velocity, Natural ventilation, Sonic anemometer. Large-scale multi-span greenhouses are common in Belgium, The Netherlands, and most Northern countries of Europe. Natural ventilation, provided by roof vents, is a major tool for regulating the air temperature, humidity, and CO 2 concentration in these greenhouses. However, optimal ventilation for interior climate requirements is seldom achieved because of the contradictions between heat loss reduction and sufficient removal of water vapor in winter conditions, and between sufficient removal of excess heat and reduction of CO 2 losses in summer conditions. Therefore, detailed analyses of the natural ventilation process are needed to improve indoor climate control. Bot (1983) developed a ventilation function from a series of measurements by means of tracer gas technique in small compartments isolated inside a large greenhouse. Since then, several quantitative ventilation studies have been described for both glasshouses (De Jong, 1990; Fernandez and Bailey, 1992; Wang and Deltour, 1996) and plastic houses and tunnels (Boulard and Baille, 1995; Boulard and Draoui, 1995; Kittas et al., 1995; Papadakis et al., 1996). However, tracer techniques and energy balance methods do not allow to clearly identify the airflow in both greenhouses and vent openings. Sonic anemometry is a new measurement technique, available commercially, which can be used to map flow fields induced by roof vents in double-span plastic houses (Boulard et al., 1997) and multi-span glasshouses (Wang and Deltour, 1997). Ventilation-induced airflow patterns in greenhouses were first simulated by Okushima et al. (1989) using Computational Fluid Dynamics (CFD), whose results did Article was submitted for publication in November 1998; reviewed and approved for publication by the Structures & Environment Division of ASAE in May The authors are Shaojin Wang, Unité de Bioclimatologie, INRA, Site Agroparc, Avignon Cedex 9, France, and Jules Deltour, ASAE Member Engineer, Professor, Unité de Physique et Chimie Physique, Faculté Universitaire des Sciences, Agronomiques, Avenue de la Faculté, Gembloux, Belgium. Corresponding author: Dr. Shaojin Wang, Unité de Bioclimatologie, INRA, Site Agroparc, Domaine Saint Paul, F84914 Avignon Cedex 9, France; voice: ; fax: ; e- mail: wang@avignon.inra.fr. not agree well with the experiments (owing in part to the limited computational power). Recent progress in flow modeling by CFD has been made for a small closed greenhouse (Boulard et al., 1998), a two-span ventilated greenhouse (Mistriotis et al., 1997), and a multi-span sawtooth greenhouse (Kacira et al., 1998). Up to now, accurate investigations of the interior airflow patterns in large, multi-span greenhouses have not been presented. Objectives of this study were to investigate the lee-side ventilation-induced airflow patterns in both vertical and horizontal planes of large, multi-span greenhouses. Once the ventilation pattern was characterized, the corresponding ventilation flux was calculated and a new ventilation function is presented to assist greenhouse designs. This study concentrates on the lee-side ventilation due to its ability to provide a more homogeneous air exchange and thus a more uniform growth of the crop as compared to the windward side ventilation. LITERATURE REVIEW VENTILATION FUNCTION Historically, the relationship between ventilation rate for a specific vent type and opening angle for the particular greenhouse has been obtained through regression analysis. Experimental studies of De Jong (1990) and Fernandez and Bailey (1992) showed that ventilation flux was independent of wind direction, but proportional to wind speed and vent area. Consequently, this article proposes the expression of the dependency on vent opening angle by a widely used non-dimensional function: G α = φ v (1) U e A 0 where A 0 = vent area (m 2 ) G(α) = non-dimensional ventilation function (dimensionless) U e = external wind speed (m/s) α = vent opening angle ( ) φ v = ventilation flux (m 3 /s) Transactions of the ASAE VOL. 42(5): American Society of Agricultural Engineers 1409

2 Bot (1983) used a series of identical compartments in a large scale multi-span greenhouse to measure the ventilation rate by applying CO 2 as a tracer. Each airtight compartment floor area was quite small (9.6 6 m 2 ) compared to the whole greenhouse (1382 m 2 ) in his experiment. Using the whole range of window apertures (0-44 ), the experimental data were fitted to the following ventilation function: G(α) = α exp( α/50) (2) De Jong (1990) carried out his measurements in nearly the same way as Bot (1983). He used 24 identical standard compartments (9.6 6 m 2 ) located in a large glasshouse block (70 33 m 2 ). He applied the dynamic tracer gas method (N 2 O) to measure the ventilation rate and obtained a slightly different function: G(α) = [1 exp( α/21.1)] (3) Fernandez and Bailey (1992) measured the ventilation rate in a small, four-span Venlo-type glasshouse ( m 2 ) with a mature tomato crop. They used the dynamic tracer gas method (N 2 O) and fitted a linear regression equation to the measured values: G(α) = α (4) Kittas et al. (1995) performed their measurements of air exchange rate in a two-span greenhouse ( m 2 ) equipped with a continuous roof opening. They used the decay rate method with N 2 O as the tracer gas. Experimental data were fitted to the following ventilation function: G(α) = sin(α/2) (5) Wang and Deltour (1996) obtained an experimental ventilation function for a large greenhouse (2150 m 2 ) based on a dynamic energy balance model: G(α) = α (6) PROCEDURES SITE AND GREENHOUSE DESCRIPTIONS Measurements were conducted in an empty and unheated 1728 m 2 multi-span greenhouse near Roeselare, a horticultural region in Belgium (lat 50.5 N). The greenhouse is orientated east-west and bordered at the north and east sides by other greenhouses. The house is composed of 12 spans with 2.85 m eave height and 3.75 m ridge height (see fig. 1). Eighty-four vents, 1.6 m long and 0.73 m wide, are distributed along the roof, alternately on the north and south slopes of each span. Since the vents on the north slopes were closed throughout the experiments, they are not shown in figure 1. Each vent opening was controlled manually. CLIMATIC PARAMETERS MEASUREMENT Internal climatic parameters were measured at a height of 1.5 m at the center of the greenhouse. Air temperature and humidity were measured using a ventilated platinum thermometer. The roof vent opening was detected by an electric resistance potentiometer. External climatic parameters were measured at a reference height of 6 m and located 10 m away from the greenhouse. External wind speed and direction were measured by a cup anemometer and a wind vane with a start threshold of 0.4 m/s. Wind Figure 1 Schematic plan of the greenhouse with the airflow measurement locations (x: 9 vertical airflow profile measurement positions; +: 12 horizontal airflow measurement positions) TRANSACTIONS OF THE ASAE

3 direction was defined as 0 from west to east along the greenhouse ridge and increased in a counter-clockwise direction. The measurements were recorded using a programmable multimeter (Model: 8840A, Fluke) controlled by a data logger constructed in the department. All the climatic parameters were measured each second and averaged over 5 min. The final results were transferred to a personal computer (PC) through its serial port RS232. SONIC ANEMOMETER SYSTEM Sound speed is increased if the air is moving in the same direction as the sound. The basic principle of a sonic anemometer is the measurement of the flight time of an ultrasound pulse between two transducers leading to the determination of the air velocity. A two-dimensional sonic anemometer system, described by Wang et al. (1997), was used to determine the airflow pattern in the greenhouse at 3 Hz frequency. The path length for both arms of each sonic anemometer was 60 cm. Calibration and validation results showed that the given combination of path-length and low sampling frequency was accurate to about 1.3% for air velocity ranging from 0 to 1.9 m/s. The measurement system for the vertical profile of horizontal air velocities was composed of 4 aligned, twodimensional sonic anemometers placed along a vertical support at 0.40, 0.80, 1.20 and 1.60 m above the ground (fig. 2). The two components of each sonic anemometer were installed parallel to the ridge and gables, to determine the airflow direction in different horizontal planes. These four instruments were moved from one place to another at nine positions (see fig. 1). The measurement process for 5- min data collection at each position took nearly 2 h on 16 July The influence of external wind conditions on air velocity measurement was reduced by normalization. During this period, the vent opening angle at the lee-side (south) was only opened at 16. The mean internal and external air temperatures were 12.7 and 10.5 C, respectively. The mean values of internal and external relative humidity were 89.4 and 88.5%. Eighty observations at each position, obtained over 5 min, resulted each from averaging over 10 readings from the sonic anemometer. The horizontal air velocity distributions under variable vent opening angle and external wind conditions were characterized at different locations at the same time. The 12, two-dimensional sonic anemometers were distributed equally over the greenhouse at 0.40 m height (fig. 1). The measurements were performed on 29 August 1996 when it was completely overcast to minimize the buoyancy effect. In sampling phase, each sonic anemometer was waiting for a synchronizing pulse from the PC. Each reading of two velocity components was stored in their own random access memory (RAM) during a 5-min period. When the required period was completed, the PC sent an interrupt to each sonic anemometer and then the stored data were transferred one by one in turn from their RAM to the PC. Before the next cycle began, the PC sent a signal to clean the buffer of each sonic anemometer. The measurement period of the sonic anemometer system always started in phase with that of the climatic parameters measuring system. During the experiments in the horizontal plane, the mean internal and external air temperatures were 14.9 and 14.1 C, respectively. The mean values of internal and external relative humidity were 93.8 and 90.1%. Fortyeight observations were obtained by modifying the vent opening angle between 5 and 36 and the variable external wind conditions. Figure 2 Schematic view of the four, two-dimensional sonic anemometers installed for characterizing the vertical airflow profile. All dimensions are in meters. RESULTS AND DISCUSSION VERTICAL PROFILE An example of vertical airflow profile at position 2 is given in figure 3. The horizontal air speed appeared to be a predominant function of height h, as observed by Boulard et al. (1998). To reduce the influence of fluctuations in external wind conditions and vent opening angle, the speed at height of 0.40 m was chosen as a base value for normalizing the data. In this way, the airflow distribution in the horizontal plane could be determined at this height, and the relation between the vertical profile and the horizontal airflow distribution could be established. The vertical profiles were described mathematically in a satisfactory way by a linear regression function. The vertical profile equations with their R 2 values at nine positions are shown in table 1. Corresponding external wind speed and direction values, averaged over a 5-min period, are also presented. The negative values of the regression slopes show that the horizontal air speed in the vertical plane decreased with height under lee-side ventilation. At positions p2 and p3, VOL. 42(5):

4 Figure 4 Polar graphs of airflow distribution at 12 positions in the horizontal plane under external wind speed (U e = 4.37 m/s) and direction (φ Ue = 326 ) and vent opening angle (α = 36 ). Figure 3 Normalized vertical profile [V(h)/V(0.40)] of the horizontal air speed at position 2 as a function of height (h). The straight line was obtained by linear regression. the slopes of the linear regression were much smaller than that at other positions because the external wind speed (7.62 m/s) was much larger than at other positions ( m/s) based on the previous work (Wang and Deltour, 1999) shown in table 1. For greenhouses with roof vents, the air movement induced by natural ventilation can be considered as a dominant effect of the wind over thermal buoyancy when the external wind speed is higher than 1.5 m/s (Bot, 1983; Boulard and Baille, 1995; Papadakis et al., 1996). the external wind direction on the general inside airflow was nearly opposite, indicating a counter flow. _ The non-dimensional ratio of the mean interior air speed V 0.40 (α) at 0.40 m height to the external wind speed U e versus the vent opening angle α was linear, as shown in figure 5. An example for measurement position 5 (near the center) is given, and the directions of the internal airflow on the graph are illustrated by the different symbols. The results for the measurement positions are given in table 2 except for position 1 where the data were missing. When the vent opening angle tended to zero, the intercept values of the regression functions were explained by the combined effect of air leakage and internal circulation. NON-DIMENSIONAL VENTILATION FUNCTION It was observed that the airflow moved across the greenhouse in a highly parallel direction because the internal airflow went almost in (fig. 5). On the basis of the continuity equation, the ventilation flux might be obtained by the integral of the airflow over a cross section in the center of the greenhouse perpendicular to the average flow direction: HORIZONTAL DISTRIBUTION The polar graph shown in figure 4 gives a frequency distribution of airflow direction as per Heber et al. (1996). It provides airflow patterns in the horizontal plane for an external southward (326 ) wind with a speed (U e ) of 4.37 m/s for a vent opening angle α = 36. Evident transverse airflow occurred from the north-east corner to the south-west corner of the greenhouse. The projection of Table 1. Linear vertical profiles of the horizontal air speed (V) as a function of height (h) at 9 positions from 80 observations External Wind Coefficient of Speed Direction Vertical Profiles Determina- Positions (m/s) ( ) V(h)/V(0.4) tion (R 2 ) P h 0.64 P h 0.78 P h 0.83 P h 0.66 P h 0.70 P h 0.66 P h 0.74 P h 0.55 P h 0.75 Figure 5 Normalized interior air velocity [ _ V 0.40 (α)/u e ] in the horizontal plane versus the opening angle (α), at a height of 0.40 m, at position 5. The symbols represent the internal airflow directions with reference to north (N) for 48 observations TRANSACTIONS OF THE ASAE

5 Table 2. The linear regression equations of normalized mean horizontal velocity [ _ V 0.40 (α) /U e ] versus the opening angle (α) in the horizontal plane, at a height of 0.40 m, at 12 locations from 48 observations Coefficient of Positions V 0.40(α)/U e * Determination (R 2 ) 1 Data missing α α α α α α α α α α α * U e = External wind speed. φ v = L c hc 0 V h V 0.40 V 0.40 α dh where h = height (m) h c = mean height of the greenhouse cross-section (m) L c = length of the greenhouse cross-section (m) V(0.40) = air speed at height of 0.40 m (m/s) V(h) = air speed at height h (m/s) V 0.40 (α) = mean air speed in the horizontal plane at 0.40 m height (m/s) With respect to the limited measurement positions, the vertical and horizontal profiles were interpolated over the cross section. In this way, the following expression for the ventilation function was obtained: G(α) = α (8) A comparison of the six different non-dimensional ventilation functions is presented in figure 6. Bot (1983) and De Jong (1990) appeared to underestimate the ventilation rate because the small compartments disturbed the general air circulation in the large greenhouse. The function of Fernandez and Bailey (1992) was a little below Figure 6 Six different non-dimensional ventilation functions as a function of the opening angle. (7) that of this study because the function was only validated by the tracer gas technique below 16 of the opening angle. The overestimation of Kittas et al. s function might be due to the fact that the external wind speed was measured at 4 m height from the ground which was 2 m lower than other studies. The small difference between Wang and Deltour (1996) and this study appeared probably due to the lower air tightness and unheated situation in this study. CONCLUSIONS Direct measurement of air velocity by means of a customized two-dimensional sonic anemometer system provided a spatial description of airflow in a large multispan greenhouse. Normalized vertical profiles of the horizontal air velocities were successfully represented by means of linear regression. The air speed under lee-side ventilation decreased with measurement height, and the decreasing slope was dependent on the external wind speed. Horizontal air velocities at 12 locations were proportional to the external wind speed and the vent opening angle. The significant transverse airflow in the horizontal plane of the greenhouse was observed by means of polar graph. The ventilation flux can be expressed as a nondimensional ventilation function which allows comparison of the ventilation rate for different types of greenhouses. A new ventilation function was based on the continuity equation and the clear airflow patterns in both vertical and horizontal planes. The final estimated linear ventilation function was in conformity with that obtained by the dynamic energy balance method in the same type greenhouse. ACKNOWLEDGMENT. The authors wish to express their sincere thanks to Dr. J. Pieters for useful discussions and help in finding the experimental site, to Prof. David B. Hannaway and Mr. Brooks Saucier for their English reviews, to Mr. M. Yernaux for his technical help with the measurement system, and to Mrs. and Mr. J. Luyckx for providing us with the measurement site in their greenhouse in Roeselare. REFERENCES Bot, G. P. A Greenhouse climate: From physical process to a dynamic model. Ph.D. thesis. The Netherlands: Agric. Univ. Wageningen. Boulard, T., and A. Baille Modeling of air exchange rate in a greenhouse equipped with continuous roof vents. J. Agric. Engng. Res. 61(1): Boulard, T., and B. Draoui Natural ventilation of a greenhouse with continuous roof vents: Measurements and data analysis. J. Agric. Engng. Res. 61(1): Boulard, T., G. Papadakis, C. Kittas, and M. Mermier Air flow and associated sensible heat exchanges in a naturally ventilated greenhouse. Agric. Forest. Meteorol. 88: Boulard, T., M. A. Lamrani, J. C. Roy, A. Jaffrin, and L. Bouirden Natural ventilation by thermal effect in a one-half scale model mono-span greenhouse. Transactions of the ASAE 41(3): De Jong, T Natural ventilation of large multi-span greenhouses. Ph.D. thesis. The Netherlands: Agric. Univ. Wageningen. VOL. 42(5):

6 Fernandez, J. E., and B. J. Bailey Measurement and prediction of greenhouse ventilation rates. Agric. Forest. Meteorol. 58: Heber, A. J., C. R. Boon, and M. W. Peugh Air patterns and turbulence in an experimental livestock building. J. Agric. Engng. Res. 64: Kacira, M., T. H. Short, and R. R. Stowell A CFD evaluation of naturally ventilated, multi-span, sawtooth greenhouses. Transactions of the ASAE 41(3): Kittas, C., B. Draoui, and T. Boulard Quantification of the ventilation of a greenhouse with a roof opening. Agric. For. Meteorol. 77: Mistriotis, A., G. P. A. Bot, P. Picuno, and G. Scarascia Analysis of the efficiency of greenhouse ventilation using computational fluid dynamics. Agric. Forest. Meteorol. 85: Okushima, L., S. Sase, and M. Nara A support system for natural ventilation design of greenhouses based on computational aerodynamics. Acta Hortic. 248: Papadakis, G., M.Mermier, J. F. Meneses, and T. Boulard Measurement and analysis of air exchange rates in a greenhouse with continuous roof and side openings. J. Agric. Engng. Res. 63: Wang, S., and J. Deltour An experimental ventilation function for large greenhouses based on a dynamic energy balance model. Agric. Eng. J. 5(3&4): Wang, S., and J. Deltour Natural ventilation induced airflow patterns measured by an ultrasonic anemometer in Venlo-type greenhouse openings. Agric. Eng. J. 6(3&4): Wang, S., and J. Deltour Lee-side ventilation induced air movement in a large scale multi-span greenhouse. J. Agric. Engng. Res. (Accepted for publication). Wang, S., M. Yernaux, and J. Deltour A multi-point ultrasonic anemometer system for measurement of airflow distribution. In Proc. 3rd Int. Conference on Fluid Dynamic Measurement and Its Applications. Beijing, PR China: Int. Acad. Publ SYMBOLS A 0 vent area (m 2 ) G(α) non-dimensional ventilation function (dimensionless) h measurement height (m) h c mean height of the central cross-section (m) L c length of the central cross-section (m) U e wind speed (m/s) V(0.40) air speed at height of 0.40 m (m/s) V(h) _ air speed at height h (m/s) V 0.40 (α) mean air speed in the horizontal plane at 0.40 m height (m/s) α vent opening angle ( ) φ Ue wind direction ( ) φ v ventilation flux (m 3 /s) 1414 TRANSACTIONS OF THE ASAE