Closed Solar House with Radiation Filtering Roof for Transplant Production in Arid Regions: Energy Consumption

Abstract

Under harsh weather conditions, closed transplant production systems (CTPS) are currently used to produce high quality transplants under artificial lighting. More than 70% of the electric energy consumed in the CTPS is for lighting. This article presents a simulation study to examine the possibility of using an alternative closed solar house, with radiation filtering roof, for transplant production in hot sunny regions to replace the artificial lighting in the CTPS with sunlight. The sidewalls of the house were insulated as in the CTPS and the roof was transparent, and made from polycarbonate hollow-channeled structure. There was a liquid radiation filter (LRF) (1.5% CuSO₄–water solution) flowing in a closed loop through the roof channels to absorb the solar heat load (i.e., the near infra-red radiation, NIR: 700–2500 nm) and transmit the photosynthetically active radiation (PAR: 400–700 nm) for plant growth. The LRF inlet temperature was assumed to be 25 °C to prevent vapor condensation on the inner surface of the cover. The evapo-transpired water vapor was removed immediately to maintain the relative humidity inside the house at 70%. The results proved that this technique can offer an appropriate air temperature inside the house less than outside air temperature by around 8–10 °C in hot summer days, and the integrated electric energy consumption during the production period was estimated to be around 43% of the CTPS consumption.

1. Introduction

Farmers in arid regions (hot and sunny deserts), such as in the Arabian Peninsula, are producing vegetable transplants either in open fields or in the greenhouses; i.e. the open transplant production systems (OTPS). In hot summer seasons, cooling the greenhouses is necessary to provide a suitable environment for transplant growth. The cooling systems (i.e., evaporative cooling) often used in these areas face many operational difficulties due to the high salinity of water resources. On the other hand, transplant production in open fields faces many problems, especially in summer, due to the high radiation intensity along with high wind speed as well as high air temperature, resulting in a high percent of wilting of transplant during production. Wilting of transplants delays the growing season and affects the final production and, consequently, the market prices of vegetables increase. Therefore, OTPS for transplant production in hot regions is not economically profitable. Promising technology has been developed in Japan, which has succeeded and been commercialized in large scales [1], later transferred to South Korea and China [2], for producing transplants in closed systems, called “closed transplant production systems” (CTPS), under artificial light using fluorescent lamps as a photosynthetic photon flux (PPF) source. CTPS is a thermally insulated and nearly airtight warehouse-like opaque structure. CTPS are able to produce high quality transplants with high photosynthetic ability using minimum resources and environmental pollution [3]. CTPS produce healthy transplants, able to produce high yields after planting in open fields or greenhouses [4,5]. There are many environmental and economical advantages for CTPS over OTPS, such as easier and faster production of pathogen-free transplants; high quality from adapting a more accurate controlled environment [6]; and more efficient space utilization by introducing multi-layer shelves in the CTPS [7,8]. Extensive previous studies revealed that the use efficiencies (i.e., the amount of the resource fixed or held in plants to the amount of the resource supplied to the CTPS) of CO₂, water, and light energy are considerably higher in CTPS than those in a greenhouse [9]. Accordingly, CTPS are appropriate technique that can be applied in hot and sunny regions during summer seasons for protecting and saving transplants from the prevailing adverse environmental conditions. However, most of the electric energy consumptions in the CTPS are consumed mainly for lighting. It has been estimated to be around 75% of the total energy consumption [10]. In most cases, the requirement of electric energy to produce transplants in CTPS is much higher than the transplant production in the greenhouse because the light source in the greenhouse is free (sunlight). In hot sunny regions such as in the Middle East and in the Arabian Peninsula, an intensive solar radiation is available all year round and is more than sufficient for transplant and crop production (annual average of daily global irradiance is 22–28 MJ·m⁻²·d⁻¹) [11,12]. Most new farms in these regions are in the desert where electricity networks are not available and water resources are brackish. Therefore, a closed solar house with a fluid-roof cover (FRC) is proposed, similar to CTPS. This is to provide a suitable internal environment for transplant growth and to use the free light source (sunlight) instead of electricity.

To reduce the solar heat load inside the buildings in hot and sunny regions, the concept of the fluid roof cover (FRC) has been used since 1893, first by Van Der Heyden (Patent No. US-504544), to design sanitary house. He used Kali-alumen [KAL(SO₄)₂] and Ammonia-alumen [NH₄AL(SO₄)] in water to make solutions to be used between a double-walled glass roof, performing FRC [13]. Such FRCs act as the commercial shading materials (e.g., plastic nets, thermal screens, curtains, etc.); they reduce the transmission of solar radiation across the whole solar spectrum range including the photosynthetically active radiation (PAR). However, a liquid radiation filter (LRF) selectively absorbs the ultraviolet (UV) and NIR wave bands and highly transmits the PAR by selecting solutions with suitable optical properties. Therefore, LRFs have been investigated and used in several studies for shading a greenhouse roof instead of other shading methods. Cupper chloride–water solution (3% concentration) [14,15,16] and the copper sulfate (CuSO₄)–water solution (1.5% concentration) [17,18] were used as LRFs by flowing it through double-layer, rigid plastic, greenhouse covers to reduce the inside air temperature.

Because of the significant effects of the LRF on the light quality inside greenhouses (i.e., the transmission ratios of red/far-red, blue/red and blue/far-red), copper sulfate–water solutions (LRF) have been used as a method to regulate the plant morphology in photo-morphogenesis studies [19,20,21,22] and to reduce plant transpiration rate [23]. In 1991, Levi et al. developed a new LRF ($F_e^{+++}$ based-water solution) [24]; this LRF was evaluated by using it for FRC in greenhouses in a hot desert [25,26]. However, the spectral radiative properties of this LRF in solar spectrum bands were not published. Up to the last decade, the FRC with a selective LRF was the unique solution to regulate the solar radiation and air temperatures in greenhouses [25,26]. To avoid complexity and the possible hazards of the LRF, such as CuSO₄–water solution, efforts have been made to develop transparent covering materials (rigid sheets or films) to act as LRF [27,28]. However, the deterioration of mechanical and physical properties of these products was very fast under extensive solar radiation conditions.

The present simulation study is to design a closed solar house using the advantages of CTPS for producing high quality, low-cost transplants. In the proposed house, CTPS artificial lighting is replaced by natural sunlight. The cover of the proposed house is FRC with LRF to selectively absorb the solar heat load (UV + NIR). The sidewalls of the house are assumed to be similar to CTPS sidewalls (i.e., constructed from low cost, thermally insulated, and locally available materials) to prevent heat energy from entering the house, and the roof is assumed to be hollow-channeled polycarbonate structure. A LRF (1.5% CuSO₄–water solution) is assumed to flow in closed loop through the roof channels for filtering out the solar radiation by absorbing the heat load (UV: 200–400 nm + NIR: 700–2500 nm) and transmitting the photosynthetically active radiation (PAR: 400–700 nm) for transplant use. The hot LRF exiting from the cover will be cooled down to 25 °C using a water cooler. This study is mainly to examine the possibility of using a closed solar house with radiation filtering roof for producing transplants in arid regions and to compare the energy consumption with that in CTPS.

2. Energy Analysis

The suggested solar house used for transplant production is illustrated, not to scale, in Figure 1a. This figure illustrates the outline dimensions of the house, and different modes of energy exchange among its components (i.e., the cover, transplant trays, soil surface and inside air). Figure 1b illustrates the transplant tray used for the study. Energy exchanges among the house components are as follows.

2.1. The Fluid-Roof Cover (FRC)

Polycarbonate hollow-channeled fluid-roof cover (Figure 2) consists of identical rectangular channels (15 mm × 35 mm) and LRF flowing inside them. Each channel is bounded by the surfaces of two identical vertical side webs, w (d_w = 2 mm thick), and the inner surfaces of the upper and lower sheets (c₁ and c₂, d_c = 2 mm thick) (see Figure 2). A model describing the solar radiation transmitted from the cover and absorbed in the cover elements (upper sheet, c₁; lower sheet, c₂; web, w; and LRF) was established and reported in [29]. In this model, the multiple reflections of solar radiation among c1, c2 and w were considered in addition to the multiple reflections inside the elements materials as well. The input data to this model are cover dimensions and the optical constants of the LRF and cover material. The LRF (1.5% CuSO₄–water solution) is a clear, semi-transparent and very light blue solution; therefore, its thermo-physical properties were taken as water properties. The mass flow rate and velocity of the LRF in each channel were estimated to be 0.00525 kg·s⁻¹ and 0.01 m·s⁻¹, respectively. The output parameters are the cover (FRC) transmittance and reflectance and the absorptance of the cover elements (c₁, c₂, w and LRF). The transmittances of this FRC to the PAR and NIR were estimated to be 0.63 and 0.08, respectively [29]. This means the cover can reject 92% of the solar heating load (NIR) before entering the house. For the present simulation study, the inner surface temperature of the FRC (T_c2) is needed to quantify the convection heat exchange with the inside air. The cover length is divided into N divisions in the LRF flow direction (N = 20 per meter length). The temperature profile was assumed to exist in the LRF flow direction only and the temperature profile in the direction of the cover width was neglected due to the uniformity of the incident radiation over the cover and the similarity of the flow pattern in the channels. One channel was chosen (Figure 2) to study the energy balance of the FRC. A control volume (the boundary through which the energy and mass exchange with the surrounding) was chosen for each division i (i = 1, 2, … N) (Figure 2). An energy balance was applied to each element of the division i (i.e., the upper sheet (c_1,i), lower sheet (c_2,i), web (w_i) and LRF_i) in the control volume. Each element was treated as lumped heat capacity system in transient conditions with equivalent properties and an instantaneous temperature. A number of 4N differential equations describing the energy balance of each element in the division i in the 4 N unknowns T_c1,i, T_c2,i, T_w,i and T_LRF,i, (i = 1, 2, … N), respectively, were developed and reported in [30]. By solving these equations simultaneously, the values of T_c2,i could be determined at any time of the day to be used in the upcoming analysis.

2.2. The Inside Air

The non-absorbing non-emitting inside air is treated as one unit with average temperature T_a. The air inside the house exchanges sensible heat (convection) with the FRC, transplant trays, and floor. The sidewalls of the house were insulated as CTPS and supposed to be constructed from low cost, low thermally-conducted local materials available in desert regions. Therefore, the different modes of energy exchange between the sidewalls and the inside air were neglected. The evapo-transpired water vapor from the transplant trays is assumed to be more or less removed immediately via a dehumidifier to maintain inside relative humidity (RH) equals to 70% (the desired value for transplant growth). Energy balance of the inside air is given by:

$$Q{C}_{\mathrm{c}2-\mathrm{a}}+Q{C}_{\mathrm{tr}-\mathrm{a}}+Q{C}_{\mathrm{f}-\mathrm{a}}+\kappa (ET-M_{\mathrm{dh}})=\frac{d}{dt}{\left[m_T{C}_{\mathrm{p}}T\right]}_{\mathrm{a}}$$

(1)

where QC_c2−a, QC_f−a and QC_tr−a are the convected heat from the inner surface of the cover, from the floor uncovered with transplant trays and from the transplant tray-substrate units to the inside air, respectively. ET is the evapo-transpiration rate from the transplant trays (Section 3.3) and M_dh is the mass of vapor condensed by the dehumidifier and is given by:

$$M_{\mathrm{dh}}=m_{\mathrm{D}}\left[{\mathsf{\omega }}_{\mathrm{a}}-0.75{\mathsf{\omega }}_{\mathrm{as}}\frac{(1-p_{vs}/p_t)}{(1-0.75p_{vs}/p_t)}\right]$$

(2)

where ω_a, ω_as, $m_T$ and $m_{\mathrm{D}}$ are the humidity ratio of the air, the humidity ratio of the saturated air, mass of the moist air and mass of the dry air in the proposed solar house, respectively. In addition, $p_{vs}$ and $p_t$ are the partial pressure of water vapor of the saturated air and the total pressure, respectively, inside the house. The convected heat from the inner surface of the cover to the inside air, QC_c2-a, depends on the cover element temperature, T_c2,i (i = 1, 2, 3, … N) and the inside air temperature, T_a. Thus, QC_c2−a is given by:

$$Q{C}_{\mathrm{c}2-\mathrm{a}}=N_c{\displaystyle \sum _{\mathrm{i}=1}^N{A}_i{h}_{\mathrm{c}2,\mathrm{i}-\mathrm{a}}(T_{\mathrm{c}2,\mathrm{i}}-T_{\mathrm{a}})}$$

(3)

where A_i and h_c2,i−a are the surface area of the cover element i and the convection heat transfer coefficient between the element i and the inside air, respectively. A natural convection mechanism exists on the different surfaces inside the closed house (i.e., the inner surface of cover, transplant leaves, soil substrate and floor lanes uncovered with trays). Thus, $h_{\mathrm{c}2,\mathrm{i}-\mathrm{a}}$ is given by [31] as:

$$\begin{array}{l}{h}_{c2,i-a}=Nu.k_a/L,\\ Nu=0.76R{a}^{1/4}{10}^4<Ra<{10}^7\\ Nu=0.15R{a}^{1/3}{10}^7<Ra<3\times {10}^{10}\end{array}$$

(4)

where Ra, Nu, k_a and L are Rayleigh number, Nusselt number, thermal conductivity of air and the characteristic length of the element, i (4 × area/perimeter), respectively.

The convective heat transfers from the transplant tray-substrate units to the inside air QC_tr−a at different growth stages is the summation of the convective heat from the transplant leaf and from the substrate soil surface to air, and is given by:

$$Q{C}_{tr-a} = A_s\left[h_{s-a}+LAI\times h_{p-a}\right]\left(T_{tr}-T_a\right)$$

(5)

where T_tr is the equivalent temperature of the tray-substrate unit, A_s is the substrate soil surface area (i.e., the trays area), LAI is the leaf area index (leaves surface area/tray surface area) and h_s−a, h_p−a are the convective heat transfer coefficients between the substrate surface and air and between the transplant leaves and air, respectively. Correlation to estimate the convective coefficient, h_p−a is given by [32,33] as:

$$h_{p-a}=(k_a/L_p)\left[0.37G{r}^{1/4}\right]$$

(6)

where Gr and L_p are the Grashof number and the leaf characteristic length (2 cm), respectively, and correlation to estimate h_s−a is given by [34] as:

$$\begin{array}{l}{h}_{s-a}=Nu.k_a/L_s,\\ Nu=\{\begin{array}{l}0.27R{a}^{1/4},T_{\mathrm{tr}}<T_a\\ 0.54R{a}^{1/4}{T}_{\mathrm{tr}}>T_a\end{array}\end{array}$$

(7)

where L_s is the characteristic length of the substrate soil surface (i.e., the tray length). The convective heat from the floor, uncovered with trays, Q_f−a in Equation (1) was estimated as Q_f−a = A_fh_f−a(T_f − T_a). In the proposed closed solar house with FRC, h_f−a was assumed to be h_s-a (Equation (7)) and T_f equals to T_tr. Such assumptions did not jeopardize the accuracy of energy analysis in a closed solar house with FRC [30]. A_f was taken to be 0.35 of the total floor area of the house.

2.3. Transplant Tray-Substrate Unit

The transplants and the substrate (i.e., soil matrix, water in the soil, transplant roots and air in the soil) are treated as one unit characterized by an average temperature T_tr and average thermo-physical properties. This is because the transplant’s transpiration and the evaporation from the substrate were measured together as evapo-transpiration (ET) and the latent heat associated with the evapo-transpiration cannot be divided for plant and soil. The tray energy balance is given by:

$$Q{S}_{\mathrm{tr}}+Q{T}_{\mathrm{tr}}-Q{E}_{tr}-Q{C}_{\mathrm{tr}-\mathrm{a}}-\kappa ET=\frac{d}{dt}{\left[\mathsf{\rho }V{C}_{\mathrm{p}}T\right]}_{\mathrm{tr}}$$

(8)

The solar radiation absorbed by the trays at specified solar incidence (QS_tr) and the thermal radiation absorbed by the trays (QT_tr) are defined, respectively, by:

$$Q{S}_{\mathrm{tr}}=A_c(1-{\mathsf{\rho }}_{\mathrm{se}}){\overline{\mathsf{\tau }}}_{\mathrm{c}}G$$

(9)

$$Q{T}_{\mathrm{tr}}=(1-{\mathsf{\rho }}_{\ell \mathrm{e}})\left\{Q{E}_{\mathrm{c}2}+{\mathsf{\rho }}_{\mathrm{c}2}Q{E}_{\mathrm{tr}}\right\}$$

(10)

where ${\mathsf{\rho }}_{\mathrm{se}}$ is the total effective short wave reflectance (albedo) of the trays at specified incident direction and ${\mathsf{\rho }}_{\ell \mathrm{e}}$ is the total effective long wave reflectance of the tray. In addition, ${\overline{\mathsf{\tau }}}_{\mathrm{c}}$ and G are the short wave cover transmittance and the solar radiation flux incident over the fluid-roof cover at the specified incident direction, respectively. The emissive power from the inner surface of the cover QE_c2 is given by:

$$Q{E}_{\mathrm{c}2}=N_c{\displaystyle \sum _{i=1}^N{A}_i{E}_{\mathrm{c}2,\mathrm{i}}}$$

(11)

where N is the number of cover division in the channel and N_c is the number of channels in the cover. In fact, E_c2,i is a function of the spectral directional emissivity of the cover ${\mathsf{\epsilon }}_{\mathrm{c}}(\mathsf{\lambda },\mathsf{\theta })$ and the black body distribution function $I_{\mathrm{c}}(\mathsf{\lambda },T_{\mathrm{c}2,\mathrm{i}})$. For simplicity, the emissive power from the cover element i (E_c2,i), per unit area, was correlated as a function of the element temperature T_c2,i and reported in [30] as:

E_c2,i = 1179.6 − 10614(T_c2,i) + 0.027(T_c2,i)², R² = 0.96

(12)

The transplant tray was assumed as gray surface and the emission from the transplant trays QE_tr is defined as:

$$Q{E}_{\mathrm{tr}}=A_{\mathrm{tr}}{\mathsf{\epsilon }}_{\mathrm{tr}}\mathsf{\sigma }{T}_{\mathrm{tr}}^4$$

(13)

The emittance of the transplant trays ${\mathsf{\epsilon }}_{\mathrm{tr}}$ was assumed to be constant value and equals to 0.90 and $\mathsf{\sigma }$ is the Stefan–Boltzmann constant [35,36].

2.4. Simulation Procedure

For the components of the fluid-roof cover system (Figure 2), the temperature profiles of the upper sheet (T_c1), lower sheet (T_c2), web (T_w) and LRF (T_f) were obtained by solving the 4-N differential equations describing the energy balance of these components; detailed analysis and method of solution were reported in [30]. The unknowns T_a and T_tr were obtained by solving Equations (1)–(8) simultaneously using Predictor Corrector Method [37]. The meteorological data used to solve these equations, such as the ambient temperature, wind speed and solar radiation flux, were recorded during 16 consecutive clear sunny days (the production period). In addition, required parameters used in the simulation, such as the leaf area index (LAI), transplant volume and density (V_p, ρ_p), tray evapo-transpiration (ET), water content in the substrate (W_sub) and physical and radiative properties of the transplant tray-substrate unit were measured in the conventional evaporation-cooled greenhouse and adapted to be useful in the proposed fluid-roof solar house. For example, the water content in the substrate soil, W_sub, measured in the greenhouse cannot be used directly as an input parameter for the simulation of the closed solar house because the environment is different; however, the ratio between W_sub and the cumulative evapo-transpiration can be used (more details are in Section 3.6).

3. Measuring and Estimating the Required Parameters

Experiments to determine the required parameters used in this study were conducted in the Agricultural Research and Experiment Station, Agriculture Engineering Department, King Saud University (Riyadh, Saudi Arabia, 46°47′ E, longitude and 24°39′ N, latitude). The proposed solar house (32 m² floor area) dimensions, orientation (N–S) and the modes of energy exchanges through the house are illustrated in Figure 1a. An evaporation-cooled greenhouse with a floor area of 32 m² was used to grow the transplants. In the greenhouse and in the proposed solar house, the trays (Figure 1b) were arranged into three rows in the N–S direction (0.75 m width for each); the surface area covered with transplant trays was estimated to be 65% of the floor area. The remaining 35% of the floor area was a dry bare soil (four lanes) used for labor movement to serve the trays. Cabbage, (Brassica campestris L.) transplants were grown in the trays for 16 sunny days (the production period 1–16 May 2015) using artificial soil as supporting material. The measured meteorological parameters inside and outside the greenhouse were: (i) the air temperatures using aspirated psychrometers; (ii) the global solar radiation fluxes using CMP3 solar meters (Kipp & Zonen B.V. Inc., Bohemia, NY, USA); and (iii) the photosynthetically active radiation fluxes (PAR) using quantum sensors LI-190SA (LI-COR Inc., Lincoln, NE, USA). The other parameters used in the simulation modeling were estimated as follows.

3.1. Leaf Area Index (LAI)

Leaf area per transplant was measured at different growth stages during the production period (on Days 6, 9, 12 and 16 after sowing) using one tray. The LAI was estimated (leaf area per transplant × number of transplants per tray (i.e., 128) divided by the tray area (i.e., 0.18 m²)) and plotted in Figure 3 at different days after sowing and could be correlated as a function of the day number after sowing (d) as follows:

$$LAI=0.36-0.315(d)+0.059{(d)}^2-0.002{(d)}^3,{\mathrm{R}}^2=0.98$$

(14)

3.2. Transplants Volume (V_p)

The transplant volume, V_p was estimated at different growth stages by measuring the mass of fresh transplants (including the roots), m_p using electronic balance (EB-6200) and plotted in Figure 4. Average transplant density ρ_p was estimated to be equal to 720 kg·m⁻³ and m_p could be correlated as a function of d as follows:

$$m_{\mathrm{p}}=36.345-37.69(d)+8.32{(d)}^2-0.16{(d)}^3,{\mathrm{R}}^2=0.99$$

(15)

3.3. Evapo-Transpiration Rate (ET)

Tray weight was recorded every 5 min using an electronic balance (EB-6200) every day from Day 3 to Day 16 after sowing in the greenhouse. Thus, Evapo-transpiration rates ET, per unit area of tray, have been estimated. Hourly average values of ET have been plotted vs. the transmitted solar radiation flux into the greenhouse (${\overline{\mathsf{\tau }}}_{c}G$) every day during the production period. In CTPS, the daily average of ET was almost the same during the production period, except for the first three days (germination period) due to the steady state controlled environment in CTPS [8]. Similarly, in a closed solar house in sunny regions, the environment is expected to also be steady and the daily average of ET is expected to be nearly uniform during the production period. Therefore, correlations representing ET can be used for all the days in production period. Two correlations for the hourly average values of ET were obtained before noon and after noon (Figure 5 and Figure 6) by fitting the values of (${\overline{\mathsf{\tau }}}_{c}G$) vs. ET for the production period. These correlations are given as follows:

(Before noon, 12:00 a.m.–12:00 p.m.)

$$ET=1.024+0.322({\overline{\mathsf{\tau }}}_{\mathrm{sc}}G)+0.0001{({\overline{\mathsf{\tau }}}_{\mathrm{sc}}G)}^2,{\mathrm{R}}^2=0.86$$

(16)

(After noon, 12:00 p.m.–12:00 a.m.)

$$ET=20.59+0.43({\overline{\mathsf{\tau }}}_{sc}G)-4.45\times {10}^{-4}{({\overline{\mathsf{\tau }}}_{sc}G)}^2,{\mathrm{R}}^2=0.86$$

(17)

3.4. Transplant Body and Soil Temperature (T_tr)

The transplant trays including the soil substrate and the transplants were characterized by an average temperature, T_tr. To emphasize this assumption, the plant body temperature and the soil substrate temperature were measured on Day 10 (the substrate was fully covered with transplants) using a sheathed thermocouple (cupper constantan, T-type, 0.2 mm dia.) and recorded at 5 min intervals using a data logger (CR23X Micro-logger, Campbell Scientific, Inc., Logan, UT, USA). Figure 7 illustrates the plant body, the substrate, the greenhouse air and the outside ambient temperatures. As shown in Figure 7, the plant body and soil substrate temperatures are almost the same and the greenhouse air temperature during daytime is much higher than the outside ambient temperature in the naturally ventilated greenhouse used for measurements. However, the conventional ventilated (natural or forced) greenhouse cannot be used in hot and sunny regions for transplant production purposes due to the excessive solar radiation and extremely high air temperature.

3.5. Tray Radiative Properties

Trays short wave reflectance ${\mathsf{\rho }}_{se}$ (i.e., apparent reflectance of transplant trays due to multiple reflections of solar beams between the transplant leaf and substrate) was recorded every 1 min, averaged every 30 min, and illustrated in Figure 8 for Day 1 and Day 15 (the beginning and end of the production period). Albedo-meter CMA-11 (Kipp & Zonen B.V. Inc., Bohemia, NY, USA) was used and the measured data were recorded in a data logger (CR23 Micrologger Campbell Scientific Inc., Logan, UT, USA). Value of ${\mathsf{\rho }}_{se}$ was taken from Figure 8 at each time interval at every day after sowing and used as input parameter in the simulation (for growing transplants in the proposed solar house). Long wave effective reflectance of the trays ${\mathsf{\rho }}_{\ell e}$ was assumed to be constant value and equal to 0.1 [35].

3.6. Soil Matrix

Artificial soil was used in the study (i.e., the substrate) consisting of 70% vermiculite, 20% peat and 10% sand by volume. Time courses of water content in the substrate (W_sub), dependent on ET, have been estimated during the production period in the greenhouse. Similar trends of W_sub were observed from Day 1 to Day 16 (sunny days). Therefore, Day 10 was chosen to represent time course of W_sub and the cumulative evapo-transpiration (CET) in the greenhouse (Figure 9). The W_sub time function (i.e., decreasing of water content in the substrate) in the greenhouse cannot be used directly for the simulation in the proposed fluid-roof solar house. This is because the environmental conditions such as air temperature, relative humidity and solar radiation intensity affecting ET as well as W_sub in the greenhouse are different from those in the closed solar house. ET in Equation (1) is estimated in the solar house using Equations (16) and (17) based on the transmitted solar radiation into the house and the FRC transmittance; consequently, the CET can be estimated. To estimate the diurnal variation of W_sub inside the solar house, the ratio between W_sub and CET (Figure 10) was determined and W_sub was correlated in Equation (18) to be used in the solar house.

$$W_{\mathrm{sub}}=CET{\displaystyle \sum _{L=0}^7{C}_{\mathrm{L}}{(t)}^L},{\mathrm{R}}^2=0.95$$

(18)

where t is the local time (in hour) and the coefficients C_L are: (C_o = 37,761, C₁ = −22,100, C₂ = 5468.4, C₃ = −741.14, C₄ = 59.426, C₅ = −2.8199, C₆ = 0.073359 and C₇ = 0.0008075).

The decrease in W_sub with time was assumed to be replaced by air. Physical properties of soil compositions (vermiculite, peat and sand) were obtained from [35] and those for air and water were obtained from [36]. Accordingly, average thermo-physical properties of substrate are given by:

$${\mathsf{\rho }}_{\mathrm{sub}}=\frac{{\mathsf{\rho }}_{\mathrm{v}}{V}_{\mathrm{v}}+{\mathsf{\rho }}_{\mathrm{p}}{V}_{\mathrm{p}}+{\mathsf{\rho }}_{\mathrm{s}}{V}_{\mathrm{s}}+{\mathsf{\rho }}_{\mathrm{w}}{V}_{\mathrm{w}}+{\mathsf{\rho }}_{\mathrm{a}}{V}_{\mathrm{a}}}{V_{\mathrm{sub}}}$$

(19)

$$C{p}_{\mathrm{sub}}=\frac{m_{\mathrm{v}}C{p}_{\mathrm{v}}+m_{\mathrm{p}}C{p}_{\mathrm{p}}+m_{\mathrm{s}}C{p}_{\mathrm{s}}+m_{\mathrm{w}}C{p}_{\mathrm{w}}+m_{\mathrm{a}}C{p}_{\mathrm{a}}}{{\mathsf{\rho }}_{\mathrm{sub}}{V}_{\mathrm{sub}}}$$

(20)

where the subscripts v, p, s, w and a are referring to substrate contents of vermiculite, peat, sand, water and air, respectively. Tray average properties including substrate and transplants (Cp_tr and ${\mathsf{\rho }}_{\mathrm{tr}}$) can be obtained by the same manner of Equations (19) and (20).

4. Results and Discussion

Time courses of the incident (measured) and the transmitted (estimated) photosynthetic photon flux (PPF) into the proposed fluid-roof solar house are illustrated in Figure 11 on two sunny summer days (May 1 and 15, 2015) for 12 h photo period (from 6:00 a.m. to 6:00 p.m.). Figure 11 shows that the transmitted PPF into the solar house is lower than 200 µmole·m⁻²·s⁻¹ (Wm⁻² = 4.628 µmole·m⁻²·s⁻¹) during short periods around sunrise and sunset and about 1200 µmole·m⁻²·s⁻¹ around noon. However, the optimum PPF for transplant production is generally 250–350 µmole·m⁻²·s⁻¹ [4,7]. This means that, during most of the day, the transmitted PPF is much higher than the transplant production demands. This leads to the possibility of using multi-shelves of transplant trays and radiation reflectors in the solar house to effectively distribute the excessive PPF to the trays. This in turn increases the house space utilization and increases the transplant production per unit area.

Time courses of air temperature (T_a) inside the house and transplant tray temperature (T_tr) for two days (1 and 15 May 2015) during the production period compared with the outside ambient temperature (T_am) are illustrated in Figure 12a,b, showing that the fluid-roof solar house can reduce the inside air and transplant-tray temperatures (T_a, T_tr) by about 8–10 °C lower than the outside ambient temperature (T_am) through the production period in hot summer days. The energy released with evapo-transpiration reduces T_tr lower than T_a all the time. However, the value of T_a is higher than T_tr at earlier growth stages (Figure 12a, day 1) and is almost equal to T_tr at later growth stages (Figure 12b, day 15). This is due to the relatively low thermal energy absorbed by transplants at low growth stages and most of solar and thermal radiation inside the house is converted into convection heat transfer to the inside air via the house roof inner surface and transplant tray surfaces, which increases the internal energy of the air as well as its temperature T_a. The sidewalls and the FRC of the house act as thermal insulators, preventing the outside thermal energy to transmit into the house; therefore, such house can keep T_a and T_tr lower than T_am, even at nighttime.

Daily amount of the evapo-transpired water in the solar house estimated using Equations (16) and (17) compared with the measured values in the greenhouse and in the CTPS [8] are illustrated in Figure 13. Values of ET strongly depend on the relative humidity of the inside air and the vapor pressure deficit as well. Therefore, the integrated values of ET in the greenhouse fluctuate due to the variation of RH during the production period. Nearly uniform estimated values were obtained in the solar house due to the assumption of constant relative humidity (RH = 70%) during the production period. Values in the CTPS do not fluctuate much because of the steady state condition during the production period (which lasted for 11 days), except for values on first two days, which were lower due to the decrease of the controlled PPF in the germinating stage.

A schematic diagram of the daily integral of solar and thermal radiation energy distributed among the components of the solar house (MJ per unit area of tray) is illustrated in Figure 14. Sky thermal radiation together with some of solar radiation is absorbed completely in the cover system (about 14.1 MJ·m⁻²·d⁻¹). Some of the absorbed energy was removed by the LRF and the rest is converted in the cover to internal energy. Around 70% of the transmitted PAR (10.8 MJ·m⁻²·d⁻¹) is converted to latent heat during the evapo-transpiration process and 15% is converted to radiation emission and/or convection from the tray surface to the surrounding air inside the house. The remaining 15% of the transmitted PAR is reflected backward and assumed to escape from the cover to outside ambient because the cover has high PAR transmittance.

Electric energy consumption in the proposed solar house is mainly consumed for cooling the LRF, W_c (water cooler) and for dehumidifying the house air to maintain 70% RH in the house, W_d (dehumidifier). Minor electric energy consumption for auxiliaries’ equipment, W_e, such as air circulating fans and LRF pumping was estimated to be around 3.2 MJ·m⁻² during the production period. Electric energy consumption for air conditioning, W_a, (if T_a > 25 °C), water cooler and dehumidifier operate under hot condition were included as heat pump with a coefficient of performance (COP) equal to 2.5. Accordingly, cumulative power consumption, i.e., W_c, W_a, W_d and W_e, during production period compared with CTPS electric energy consumption is illustrated in Figure 15. This figure shows that the energy consumption in the fluid-roof solar house is lower than that in CTPS because the energy for lighting (75% of the total energy consumption) is free in the solar house. Total electric energy consumption during the production period for the solar house was 183 MJ·m⁻² (50.83 kWh·m⁻²) while in CTPS it was 430 MJ·m⁻² (119.4 kWh·m⁻²). In sunny regions with extensive solar irradiance, the electric energy needed for the solar house can be supplied renewably using PV solar panels to supply the required electric energy to the house.

5. Conclusions and Recommendations

According to the previous results and discussion, the conclusion can be summarized as follows:

• Fluid-roof solar house can be used economically for transplant production in arid regions (usually hot sunny desert) where electric energy resources are not prevalent.
• High values of PPF are transmitted into the house every day; thus, the house space could be better utilized using radiation reflectors and growing transplants in multi-layers of trays arranged in tray shelves.
• The electric energy consumption in the proposed solar house was around 43% of the electric energy consumed in CTPS. This proves the profitability of using closed solar houses in arid regions.
• Using CuSO₄-water solution as LRF may not be safe due to the possible leakage from the cover over the trays; however, at the moment, there is no alternative, cheap LRF available to use for absorbing the NIR effectively. Further research is needed to develop effective, safe and suitable LRFs for cooling the roofs of residential and agricultural structures.