Evaluation of an intermittent mist spray cooling system for improving greenhouse eggplant cultivation

. Cultivation of crops in greenhouses faces the issue of overheating, which can harm yields. A low-cost solution is evaporative cooling, which also reduces the vapor pressure deficit. Fine mist sprays can significantly reduce temperature and vapor deficit. However, the fast evaporation rate could cause air saturation and wetting of plants, causing damage. A simple method to control the effect is intermittent spraying. In this research, air temperature was continuously measured at 20 locations for a misted 98m 2 greenhouse during eggplant cultivation, as well as a control greenhouse without mist, while also evaluating the resulting crops. An array of 28 high pressure hydraulic nozzles, spraying 2.6kg/h each, intermittently sprayed 20 seconds on, then 10 seconds off throughout the day. The temperature in the misted greenhouse averaged 2.2K to 4.8K cooler than the control greenhouse. In the misted greenhouse, the total yield (+26%) and marketable fruits (+143%) were higher than the control. The cooling effect of the misted greenhouse compared to the control case tended to correlate with outdoor wind speed. The comparison of the misted greenhouse to its own initial condition did not. Evaluations of evaporation cooling may yield better understanding if compared to a control greenhouse at the same site.


Introduction
Cultivation of crops in greenhouses faces the issue of overheating, which can reduce crop yields and product quality. In "plant factories" (controlled indoor environments optimized for cultivation) air temperature and humidly control is one growth-promoting technique. However, plant factories are costly to create and operate.
[1] When natural ventilation nor mechanical ventilation is sufficient to prevent overheating of a traditional outdoor greenhouse, a potential low-cost solution is evaporative cooling. The primary goal of an evaporative cooling system for heat-stress relief is to reduce air temperature at a relatively low system cost. Evaporative cooling yields a decrease in temperature with a proportional increase in humidity. In the case of greenhouse cooling, primary goals are to reduce temperature and to reduce evaporative water loss from the crops, by reducing the vapor pressure deficit (VPD). Evaporative cooling promotes both of these goals and has been shown to also reduce irrigation water consumption. [2]

Types of evaporation cooling
Evaporative cooling is often accomplished with wetted media and strong mechanical ventilation to promote evaporation. [3] Though due to the strong ventilation, much of the cooling effect may be lost in the exhaust air, while establishing uniformly cooled conditions throughout the space may be difficult. [4] * Corresponding author: farnham@omu.ac.jp Fine mist sprays from high pressure nozzles can evaporate completely without this strong ventilation, and may be sufficient with natural ventilation to provide significant cooling, as well as prevent air saturation which would block further mist evaporation, likely causing wetting of the plants and fruit which may be undesirable. [5] The reduced ventilation also makes it more practical to use insect-proof screens and natural ventilation. [4]. In this experiment, a mist spray evaporative cooling system is deployed for a summer growing season in a greenhouse cultivating eggplant. The temperature, humidity and VPD changes are compared to a practically identical control greenhouse without the mist spray system at the same location with the same layout growing the same crop. To the best of the authors' knowledge, studies of greenhouse mist cooling systems rarely, if ever, compare to a control case. The cooling effect is typically judged by comparison to the pre-cooled initial condition and/or conditions on a similar day without mist cooling. Here, judging the measured cooling effect as the difference between air conditions in the misted and control greenhouses is more likely to avoid the confounding factors of fluctuations in the outdoor environment conditions. The effect of changes in outdoor air temperature and humidity, insolation, wind speed and direction (thus ventilation rate) on the indoor environment is represented by the control greenhouse and taken as the basis for determining the cooling effect of the mist in the misted greenhouse.

Theory
Evaporative cooling is an exchange of sensible heat from air into latent heat of water evaporation. The result being the air temperature drops while water evaporates. The limitation is that the exchange is a net-zero exchange of enthalpy, with a maximum possible cooling effect yielding saturated air at about the wet bulb temperature. Though this condition is often not desired in many applications, including cultivation, due the potential for undesirable wetting of surfaces. A mist evaporation cooling system typically creates a spray of small droplets to maximize the air-water surface area and thus evaporation speed. Evaporation speed is also directly proportional to the VPD. As an evaporative cooling system decreases the VPD, it will reduce the speed of evaporation, possibly leading to a runaway effect toward saturated air. Thus, care in balancing the evaporation rate and spray rate is essential.
A single droplet of about 20 micron diameter at air temperature of 35˚C, relative humidity 45% and VPD of 3.1kPa (typical summer conditions in Osaka, Japan) will evaporate in about 0.5 seconds, while a 40 micron droplet takes about 2s. [6] The mist nozzles used here have a Sauter mean diameter of 20 microns, and a D50 of about 38 microns.

Experiment
The greenhouses used in this research project are located on the campus of Research Institute of Environment, Agriculture and Fisheries, Osaka Prefecture. The Koppen climate classification of Osaka is "Cfa" (Humid Subtropical Climate). [7] The greenhouses are the inner 2 in a cluster of 4, as shown in Fig. 1, likely having nearly identical exposure to sunlight, shadows and wind. Their length is 15.5m, width 6.3m and height of the peak of the roof at 4.3m. Ventilation is through windows of 1m height on the long walls, doors of 2.3m height and 2.4m total width at each end, and roof vents facing east and west along the length that can open up to 30cm. All of these openings have anti-insect screens which reduce air flow. Air flow is only through those openings.

Initial characterization experiments
Preliminary characterization experiments with nozzle sprays in the greenhouse were performed in April to choose suitable mist spray rate and nozzle locations which will yield cooling throughout the greenhouse space occupied by crops without significant wetting. When windows were closed, and on days of light wind, a constant spray tended to yield a fog-like state with wetting of many surfaces. Common options to control the total mist spray rate are; lower the per nozzle flow rate, reduce the number of nozzles, or use intermittent spray. Nozzle flow rate can be reduced by reducing the water pressure, but typically at the cost of the droplet size becoming larger, and thus more likely to cause wetting. Reducing the number of nozzles would make it more difficult to maintain uniform cooling throughout the space. Thus, intermittent spray was chosen. Intermittent sprays can allow sprayed mist to disperse and evaporate during the breaks, such that the following pulse of spray has a relatively dry parcel of air into which it can also evaporate. However, intermittent spray reduces the overall spray rate, and thus the cooling effect.
Here, it was decided that a 20-seconds ON, 10 seconds OFF intermittent spray prevented significant wetting and a runaway fogging effect. During the summer the mist is used while the windows, doors and roof vents are open to help prevent heat buildup. On-site measurements taken with 4 hot-wire anemometers during breezy (from the west and north) weather that is common at the site, an airflow of about 0.2m/s comes from the long west window, and 0.5m/s from the west roof vent and from the north door. This equates to about 30-40 air changes per hour on a breezy day. However, the influence of neighboring greenhouses on the air flow field, and that the greenhouse is on a hill slope with sudden steps of several meters nearby, yielded many localized differences in air speed and direction. A simple visualization of changes in direction was performed by attaching ribbons at 1m intervals along the centerlines of all openings. On a breezy day, it was common for airflow to reverse direction several times over the monitored span of 5 minutes, i.e. the airflow inlets became outlets and vice versa. Thus, an accurate evaluation of airflow except on the days of still air was too difficult to directly measure with the resources at hand. However, average changes in air conditions over time when mist is turned on or off can indirectly reveal the actual air exchange rate. An inspection with smoke generators found that the walls were quite airtight.

Experiment procedure
Hydraulic mist nozzles were arranged in 1 row at 2.5m height at 1m spacing between pairs of nozzles angled 45˚upwards to west and east (see Fig. 2). The water supply is pressurized to 6MPa by one pump unit. Each nozzle sprays 2.6L/hr of droplets with Sauter mean diameter of about 20μm. Two fixed circulation fans, which blow 40m 3 /min are mounted at opposing corners of the greenhouse with the aim of preventing local buildup of mist and to spread the cooling effect. The eggplants are arranged in 3 rows, with 17 plants in each row. The mist spray system is typically used throughout the day at the discretion of the staff based on weather conditions (i.e. not used on cloudy or rainy days). For the measurement trials, the mist was activated for periods of about 1 hour, then stopped for about 1 hour, so that the mist cooling effect could be compared to the control greenhouse over the simultaneous period, as well as the misted greenhouse itself by taking the average values 5 minutes before spraying as the initial condition. Measurement trials were done in 2021 on June 22 (2 periods of mist), August 5 (2 periods of mist), and August 22 (1 period of mist). Each greenhouse monitored temperature and humidity with a 20-channel electronic data logger at 2-second intervals. Most sensors were T-type thermocouples with soldered 1mm beads, placed inside horizontallyoriented open cardboard cylinders with diameter of 6cm and length of 12cm, covered with aluminium tape to both shield from sunlight and give some protection from wetting by the mist while allowing air flow past the sensor. The thermocouples as connected to the data logger were accurate to +/-0.5˚C. Relative humidity sensors were electric capacitance type accurate to +/-1.5% when less than 90%RH. In each greenhouse, sensors were placed amongst the plant leaves at heights of 80cm, 130cm and 180cm above the soil at 5 locations (as Shown in Fig. 3); in the center row of plants at the center, north and south ends, as well as the center of the east and west rows. However only 1 sensor was placed in the west row at 130cm. A sensor was placed at the north-side circulation fan inlet. Another sensor was placed near the east-west centerline at 4m height just inside the roof inlet. Outdoor temperature was recorded with a sensor 3m to the north of each greenhouse, and 1 sensor in the middle of the aisle between the 2 greenhouses. Relative humidity sensors were only placed at the center location at the 3 heights. Plant leaves eventually reach up to 2m height by the end of the growing season.

Results
The temperature and humidity at the sensor locations in the misted and control greenhouses were compared. Figure 4 shows an example of the temperature difference at the center location (130cm and 180cm), as well as the difference between the 2 outdoor sensors on the north side. During the periods of mist, the misted air temperature drops to about 5K difference within 20 minutes and continues to about 6K over 90 minutes. The curve shape reveals a time constant of about 5 minutes, thus air exchange rate of about 12 and a quasi-steady state can be assumed after 15 minutes. The curve shows signs of dropping further at about 12:45, which might be a sign of runaway saturation, but did coincide with a drop in average outdoor wind speed from about 1.0m/s to 0.4m/s, thus a drop in the natural ventilation rate.  Table 1 for all 5 mist experiments. These values are from 20 minutes after spraying started. The average of all sensors at the north, south, west, east and center (NSWEC) locations is shown at the bottom of the table. The average outdoor wind speed is also included. In the cooler month of June, the average temperature drop is over 4K, while in the hotter August experiments, the average drop was smaller, at around 3K. The average temperature drop seems to be inversely correlated with the outdoor wind speed, which influences natural ventilation rate. The humidity sensors are only located at the center location so VPD could only be evaluated there. The average temperature and relative humidity of the 3 sensors is then used to calculate the VPD, taking the values from the 5 minutes before spraying as the initial condition, and then 20 minutes after spray start as the misted condition (shown in Table 2). The VPD in June was relatively low, and dropped by about 0.9-1.2kPa. In August, the initial VPD was larger, and the drops ranged from about 0.9kPa to 1.8kPa. The effect over the entire growing season on the harvested eggplants is shown in Table 3. The total harvested eggplants and total marketable eggplants increased in all months. The relative increase was highest in August, though that is based on a relatively lower typical harvest during that period. The overall increase in marketable eggplants was increased by 143% over those from the control greenhouse.

Discussion
A mist spray evaporative cooling system was deployed and evalauted by comparing physical measurements and harvest results with a control greenhouse. The misted greenhouse yielded far better harvest in terms of number of fruits and quality. On average, for all sensors in the space, the temperature decreased by about 2.2K to 4.3K compared to the same locations in the control greenhouse. This seems to inversely correlate with outdoor wind speed, and thus the natural ventilation rate, as can be expected. The deployment of the mist nozzles along the centerline of the greenhouse, combined with the 2 circulation fans at opposite corners, seems to have successfuly spread the cooling effect to much of the crop area. When comparing changes during misting to the intial condition in the misted grenehouse, rather than the control greenhouse, a clear correlation with environment factors is more difficult to see. The average temperature drops have no clear correlation to the outdoor environment wind condition. It is possible that evaluations of greenhouse evaporation cooling would yield clearer understanding if done with a neighboring non-misted control case greenhouse, rather than comparing to intial conditions in a single greenhouse, as the change in outdoor environemnt conditions over time becomes a confounding factor. Current research is being done to model the greenhouse in a CFD simulation to better evaluate the system and the effect of wind conditions on ventilation and cooling.