Impact of Building Integrated Rooftop Greenhouse (BiRTG) on Heating and Cooling Energy Load: A Study Based on a Container with Rooftop Greenhouse

Abstract

Recently, there has been a growing interest in sustainable agricultural methods aimed at reducing energy consumption and mitigating environmental impacts. Among these methods, the integration of rooftop greenhouses into buildings has emerged as a promising solution for urban agriculture, offering various advantages such as energy-saving effects on both the host building and the rooftop greenhouse, improved resource use efficiency, enhanced food security, and reduced environmental impacts. However, there has been limited research on the energy-saving impact of rooftop greenhouses under different conditions and detailed heat transfer analysis based on actual buildings and rooftop greenhouses. The objective of this study is to investigate the energy benefits of building-integrated rooftop greenhouses by evaluating various operating scenarios for both the building and the greenhouse. A container and upper rooftop greenhouse structure were constructed, and building energy simulation models for the target system were developed. Four different operating scenarios for the greenhouse and three different operating scenarios for the building were compared. Depending on the operating scenarios, the total energy loads of the container and RTG were reduced by 18.4% to 24.7% and 0.7% to 6.3%, respectively. The findings of this study can be utilized for the development of optimized greenhouse control strategies.

1. Introduction

Recently, interest in sustainable agriculture methods has been growing in efforts to reduce energy consumption and mitigate environmental impacts. Among these methods, a rooftop greenhouse (RTG) is considered as a promising solution for urban agriculture, offering diverse advantages. Due to the benefits of an RTG, the scalability of RTGs has significant potential in urban agriculture and urban sustainability.

A rooftop greenhouse is one of the solutions to address the future food shortage. The global population is predicted to reach 9.7 billion by 2050, and more than two-thirds are expected to reside in urban areas [1,2]. There is a high probability that this phenomenon leads to serious urban food security vulnerabilities and food crisis. RTGs are receiving attention as a means to address such social issues. As urbanization accelerates, the distance between food production and consumption, known as food miles, is increasing. Emissions in transportation process represent a fifth of the total food system’s emissions; however, it has the potential to increase. Food production in urban areas is required in order to reduce transportation energy and distribution losses [3,4]. Moreover, an RTG utilizes the unused rooftop spaces on buildings for food production, effectively enabling zero-acreage farming [5,6]. Because rooftops account for one-fourth of all urban surfaces, an RTG is good for space use optimization in dense urban area, and emerges as a promising futuristic solution [4,7,8].

However, there also exists limitations affecting the scalability and spreading of RTGs. There exist related laws and regulations limiting the construction of a greenhouse on the roof or cultivation in buildings. Load capacity, roof surface slope, fire escape routes, and so on of existing buildings should be closely investigated for RTG construction. This process takes more time and cost compared to the free-standing greenhouse [9,10]. Moreover, RTGs are not yet socially widespread, resulting in the RTG field remaining still relatively small and often not commercially orientated [4]. RTGs may have higher initial costs because of a rigid structure having stability against external perturbations and observing the construction laws [4,11].

To address the cost and energy disadvantages of a RTG and enhance its economical viability, an integrated rooftop greenhouse (iRTG) is being considered. An iRTG can reduce energy consumption in both the agricultural and building sectors. Higher thermal energy and waste resources, such as CO₂ and wastewater, from the city can be used for cultivation by enhancing energy and resource efficiency in the agricultural sector [12,13,14]. The first fully integrated rooftop greenhouse was constructed on the ICTA-ICP building by the Autonomous University of Barcelona. This innovative design aims to facilitate the exchange of heat, CO₂, and rainwater between the building and RTG. Nadal et al. [12] developed a simulation model of an RTG above an ICTA-ICP building to evaluate the energy performance within the Mediterranean context. The validated results indicated that the thermal exchange between the RTG and the ICTA-ICP building significantly influences the indoor climate of the RTG. Simulation findings revealed that an iRTG successfully recycled 43.78 MWh of thermal energy, equivalent to 341.93 kWh/m²yr. Sanjuan-Delmás et al. [13] provided a comprehensive environmental assessment of food production in an iRTG, considering rainwater, residual heat (energy), residual air (CO₂), and food production. It was revealed that an iRTG enabled cultivation during winter without the need for additional heating, maintaining an average inside temperature 8 °C higher than that of the outdoor environment. Additionally, rainwater was utilized for crop irrigation, resulting in a reduction of water consumption by 80–90%.

Buildings account for 36% of the total world energy use, and as urbanization and population growth progress, it is anticipated that energy consumption in buildings will continue to rise. Therefore, reducing energy consumption in the building sector is considered an important issue in global environmental policies [15,16,17]. Due to the greenhouse effect in winter and the heat-blocking effect in summer, provided by RTGs, the heating and cooling energy consumption of buildings can be reduced [11,18,19]. Based on an iRTG above the ICTA-ICP building as discussed in [13], Munoz-Liesa et al. [18] conducted simulations revealing that the building has saved 31.9 kWh/m²yr of thermal energy due to the insulating effect of the iRTG. However, their simulations indicated a relatively modest cooling benefit from the iRTG. Consequently, they concluded that further investigations into the potential cooling impact of the iRTG via plant transpiration were needed. Zhang et al. [19] developed a dynamic energy simulation model to assess the impact of an iRTG and warehouse in Sweden, varying parameters such as glazing materials and shading devices for the RTG. The integration of an RTG and warehouse resulted in an overall energy saving of 10.7% for heating and 6.8% for cooling for the warehouse itself, and 10.4% for heating and 11.7% for cooling for the greenhouse. This integration proved beneficial in terms of enhancing overall energy efficiency. However, there has been limited research on the impact of energy saving from an RTG under various conditions and detailed heat transfer based on actual buildings and RTGs until now [4,19]. A literature summary of the energy research related to RTGs is suggested in

Table 1. A literature summary of energy research related to RTG.

Building Type	Location	RTG Area (m2)	Characteristic	Refs.
Research building	Barcelona (Spain)	75 (real)	- Exchange of heat, CO2, and rainwater between the building and RTG. - Demonstrate the cultivation during winter without the need for additional heating. - Need for further investigation into the cooling impact of an iRTG.	[12,13,18]
Warehouse	Malmo (Sweden)	15,898 (model)	- Analyze the impact of varying parameters such as glazing materials and shading devices for the RTG. - Investigate the heating and cooling saving effect of the warehouse and RTG.	[19]
Office	Yeongam (Republic of Korea)	358 (model)	- Consider the crop energy model (empirical equation for tomato). - Adapt alternating air temperature management (ATM) to reduce the heating load of a BiRTG. - Increase total annual energy load of BiRTG after adapting ATM.	[11]
Container	Daejeon (Republic of Korea)	27 (real)	- Construct a container and RTG for various experiments and data collection. - Compare diverse operating conditions depending on building type (office, hotel, and hospital)/RTG. - Investigate the heating and cooling saving effect of the container and RTG for various conditions. - Analyze the heat transfer through the rooftop surface to investigate the impact of RTG.	This Study

, and the distinction of this study is briefly suggested.

In the recent literature, integrated rooftop greenhouses have been appropriately referred to as BiRTGs [4]. The aim of this study is to investigate the energy benefits of building integrated rooftop greenhouses (BiRTGs) by evaluating different operating scenarios for the building and the greenhouse. A container and upper RTG, representing BiRTGs, were constructed, and building energy simulation (BES) models for the target system were developed. Four different operating scenarios for the greenhouse and three different operating scenarios for the building were compared. For each scenario, we examined how the building and upper RTG interacted with each other in terms of energy implications.

2. Materials and Methods

2.1. The Target System Representing BiRTG

To analyze the energy-saving effect of BiRTGs, a container, considered as a building, and a greenhouse above it were constructed in Daejeon, South Korea as depicted in Figure 1. The dimensions of the container were measured 3 m in width, 9 m in length, and 2.6 m in height. The rooftop greenhouse covers the entire rooftop surface of the container, with a side height of 3.0 m and a maximum height of 3.8 m.

Table 2. Physical properties of a container wall and greenhouse covering material.

1. A Container Wall
Material		Thickness (mm)	U-Value (W/m2K)	Illustration
Wall	G	2	0.376
	A	20
	U	45
	S	2
Roof	G	2	0.588
	A	20
	U	20
	S	4
Floor	G	2	0.272
	A	20
	U	70
	S	4
	C	300
2. A Greenhouse Covering Material
Material		Thickness (mm)	U-Value (W/m2K)	Illustration
Frame	S	2	1.224	PLEXIGLAS® ALLTOP
	A	16
	S	2
Glass	PLEXIGLAS® ALLTOP		2.500
Floor	Equal to the roof of a container		0.588

shows the physical characteristics of the container wall and greenhouse covering materials. To facilitate real-time monitoring, temperature sensors were placed inside the container and the greenhouse. Ambient environmental data were obtained from a nearby weather station installed above the container, capable of monitoring ambient temperature, relative humidity, solar radiation, and wind speed. All data were logged at 1 min intervals. The sensor information, measurement accuracy, and installation location are represented in

Table 3. Information of installed sensors.

Sensor	Model (Manufacturer)	Measurement Accuracy	Measurement Location
Thermocouple	T-type (Omega, Irving, TX, USA)	±0.5 °C	Two internal columns, each with five points spaced equally along the vertical axis (Use of 10 average values)
Humidity sensor	KSH7310 (SEMSECUBE, Seoul, Republic of Korea)	±3% of full scale	Two internal columns, each with one point (Use of two average values)
Radiation sensor	SYE-2007PM (Shinyoung ELE, Seoul, Republic of Korea)	±5% of full range	External installation beside the greenhouse roof

.

2.2. The Building Energy Simulation Model for BiRTG

For the building energy simulation (BES) of the BiRTG, TRNSYS 18 (Transient system simulation, Wisconsin in USA) was applied. It is widely used for building energy analysis, including transient energy loads, thermal systems, and energy consumption [20,21]. For the BES of BiRTG, a two-zone (the container and the greenhouse) model was developed. By calculating the energy balance of each zone, the thermal changes of each zone and energy exchange between the container and RTG were examined. The main components used in the developed TRNSYS model were summarized in

Table 4. Main TRNSYS components used in the model [22].

Type	Name	Description
Type15-3	Weather Data Reader	- Reading an external weather data file. - The EPW weather data of Daejeon, Republic of Korea.
Type56	Multi-Zone Building	- Modeling the thermal behavior of a building having multiple thermal zones using the TRNBuild modeling result. - Two zones: a container, a rooftop greenhouse.
Type14h	Forcing Functions	- Employing a time-dependent forcing function with a repeated pattern. - Solar radiation constraints for shading screen operation during heating and other seasons (presented in Table 6). - Setting the temperature of a greenhouse during heating and cooling seasons (presented in Table 6).
Type516	Multiple Day Scheduler	- Generating the input schedule for a weekday, Saturday, and Sunday. - Heating, cooling, and ventilation schedule for a day (presented in Figure 3).
Type2b	Control Strategy	- Generating ON/OFF differential control signal. - ON/OFF signal for window open and ventilation fan depending on the greenhouse inside temperature.
Type25f	Printer	- Printing selected variables at specified intervals of time.
Type46a	Printegrator	- Printing the integrated values of selected variables. - Generating monthly integrated energy load data.

.

Based on the energy conservation equation, the energy exchange equation for an arbitrary zone $i$ considering convection, radiation, and other heat transfer is suggested in Equation (1) [22].

$${\dot{Q}}_i={\dot{Q}}_{surf,i}+{\dot{Q}}_{inf,i}+{\dot{Q}}_{vent,i}+{\dot{Q}}_{g,c,i}+{\dot{Q}}_{cplg,i}+{\dot{Q}}_{solar, i}+{\dot{Q}}_{ISHCCI,i}$$

(1)

where ${\dot{Q}}_{surf,i}$ is the convective gain from surfaces, ${\dot{Q}}_{inf,i}$ is the infiltration gains. The infiltration rate of the container and the greenhouse described in Figure 1 was measured at 0.2 ACH and 1 ACH, respectively. ${\dot{Q}}_{vent,i}$ is the ventilation gains. Ventilation rates of a container and greenhouse were controlled according to the operating schedule suggested in Section 2.4. ${\dot{Q}}_{g,c,i}$ is the internal convective gains by people, equipment, illumination, etc., ${\dot{Q}}_{cplg,i}$ is the gain due to connective air flow from the boundary condition, ${\dot{Q}}_{solar, i}$ is the fraction of solar radiation entering an airnode through external windows which is immediately transferred as a convective gain to the internal air. For the energy balance of a container, the container has no transparent surface, unlike a greenhouse; therefore, the effect of direct heat from solar radiation was omitted. ${\dot{Q}}_{ISHCCI,i}$ is the absorbed solar radiation on all internal shading devices of the zone and directly transferred as a convective gain to the internal air.

2.3. Model Validation

To establish the reliability of the developed BES model, model verification should be conducted. For validation, the change in inside temperature of the facility, depicted in Figure 1, was acquired for 44 days. Since the seasonal change of Korea is distinct, data were collected for all seasons including hot, changing, and cold seasons, and the results are presented in Figure 2a–c. Consequently, the root mean square errors for the container and RTG are 1.6 °C and 3.7 °C, respectively. Therefore, it was determined that a reliable BES model had been developed, and further analysis was conducted.

2.4. Simulation Condition

To examine the impact of the BiRTG in terms of thermal energy, the simulation results of the container under the RTG (container B in Figure 1) were compared to that of a stand-alone container, (container A in Figure 1) under the same simulation conditions. Similarly, the RTG was analyzed in comparison to a freestanding greenhouse on the ground, which has the same structure to the RTG. For the total energy load analysis of the BiRTG, the integrated model was compared with the separated model. The integrated model includes the container and the RTG above it, while the separated model represents the simple sum of the results of the stand-alone container and the freestanding greenhouse. Furthermore, the thermal impact of operating conditions, including four different operating scenarios for the greenhouse and three different operating scenarios for the building, were investigated.

2.4.1. Operating Scenarios for a Greenhouse

Depending on the cultivated crop in a greenhouse, the required environmental conditions become different. In this paper, a crop suitable for moderate temperature conditions is selected as the target, and the optimal growth temperature and limiting temperature for representative crops are proposed in

Table 5. Optimal growth temperature and limiting temperature for crops suitable for medium temperature conditions [23,24].

Name	Minimum Limiting Temperature (°C)	Night-time Temperature (°C)	Optimal Growth Temperature (°C)	Maximum Limiting Temperature (°C)
Tomato	5	8–13	20–25	35
Watermelon	10	13–18	23–28	35
Pumpkin	8	10–15	20–25	35
Cucumber	8	10–15	23–28	35

. Based on the temperature suggested in Table 5, the set temperature of daytime and night-time for the heating season are 22 °C and 14 °C, respectively. Cooling is provided only during daytime with a set temperature of 27 °C [23,24].

The developed BES model of a greenhouse includes a shading screen (with a shading coefficient of 55%) on the skylight, thermal screen on the skylight and side windows, natural ventilation (simple window opening), forced ventilation (using ventilation fans), and an auxiliary electric heat pump (EHP). For the natural and forced ventilation, the air change rates were fixed at 10 ACH and 60 ACH, respectively, regardless of ambient conditions. The components are controlled based on the operating conditions and they are summarized in

Table 6. Operating strategy for a greenhouse.

System Component	Applied Period		Operation Constraint
			Parameter	Value
Shading screen	Heating	9 h–17 h	Solar radiation (W/m2)	On @ $\ge$ 1000
	Other	9 h–11 h		On @ $\ge$ 1200
		11 h–14 h		On @ $\ge$ 800
		14 h–17 h		On @ $\ge$ 1000
Thermal screen	Heating	After sunset	-	Always On
	Other	After sunset	Inside temperature (°C)	On @ $\le$ 16 Off @ $\ge$ 18
Window (N.V **)	Heating	Daytime (8 h–18 h)	Inside temperature (°C)	Open @ $\ge$ 27 Close @ $\le$ 24
	Cooling	All day	-	Open when EHP is OFF
	Other	Daytime (7 h–19 h)	Inside temperature (°C)	Open @ $\ge$ 25 Close @ $\le$ 21
		Night-time (19 h–7 h)		Open @ $\ge$ 14 Close @ $\le$ 12
Ventilation fan (F.V **)	Heating	-	-	Always Off
	Other	Daytime (7 h–19 h)	Inside temperature (°C)	Open @ $\ge$ 26 Close @ $\le$ 22
		Night-time (19 h–7 h)		Open @ $\ge$ 15 Close @ $\le$ 13
EHP ***	Heating *	Daytime (8 h–18 h)	Setting temperature (°C)	22
		Night-time (18 h–8 h)		14
	Cooling *	Daytime (7 h–19 h)		27
		Night-time (19 h–7 h)		Always Off

* Heating season: from May to September/Cooling season: from November to March. ** N.V (Natural ventilation rate): 10 ACH/F.V (Forced ventilation rate): 60 ACH. *** EHP is used when operating scenario is 2-1 or 2-2.

.

In this paper, four different operating scenarios (SCs) of a greenhouse are compared. For SC 1-1 and 1-2, the inside temperature of a greenhouse is not actively controlled using an EHP. Instead, temperature control relies on passive methods such as screens and ventilation. For SC 2-1 and 2-2, the inside temperature is controlled using an EHP, enabling the calculation of heating and cooling loads to reach the target temperature. Because of the extremely high energy costs in the harsh ambient conditions of South Korea, we designated July and August (hot summer) as the non-cultivation period for SC 1-2 and 2-2. During this period, as solar energy for photosynthesis is not required, a shading screen with a shading coefficient of 100% is used and side windows are open (natural ventilation condition) for 24 h.

2.4.2. Operating Scenarios for a Container

To investigate the effect of the operating conditions of a building on the energy load change of a BiRTG, three types of buildings are selected for the simulation. A small office, a small hotel, and a hospital are selected because of their distinct operating schedules. The small office requires daytime heating and cooling, and the small hotel represents the building which demands night-time heating and cooling. Lastly, the hospital needs heating and cooling for 24 h. Figure 3 shows the operating schedule of each building type, along with the set temperature for heating and cooling, and the required ventilation rate. The schedules are partially modified based on the commercial reference building model schedule provided by the DOE in the USA [25]. Other internal gains are not considered in this paper. For the further BES analysis, the basic simulation schedule of a container follows the operating schedule of the small office.

3. Results and Discussion

3.1. Thermal Influence of RTG on a Container

To precisely figure out the effect of the RTG, heat transfer through each surface of a container was compared under the extreme weather conditions. The heat transfer amount through each surface during a day are depicted in Figure 4. A positive value indicates heat transfer from the surface to the inside air zone. Figure 4a and Figure 4b depict the results for the peak cold day (the day with the lowest ambient temperature, 24 January) and the peak hot day (the day with the highest ambient temperature, 7 August), respectively. To check the impact of the RTG on the inside temperature change of a container, heating and cooling schedules for the containers were not set and the RTG was controlled according to SC 1-1. Temperature change of the containers is also illustrated for a day in Figure 4.

As shown in Figure 4a, a stand-alone container (container A) absorbs heat from the ground surface and south surface on the peak cold day, and loses heat through other surfaces, including the roof surface. Conversely, a container under the rooftop greenhouse (container B) mainly gains heat through the roof surface during a day, since the RTG functions as a heat source for the container. The inside temperature of container B is on average 8.6 °C higher than that of container A. For the peak hot day scenario depicted in Figure 4b, both containers obtain heat from the roof surface, but the amount for container B is 64.2% larger than that for container A. It leads to an average temperature inside container B that is 1.4 °C higher. It is important to note that the effect of the RTG on the container may vary depending on simulation conditions.

3.2. Influence of RTG Operating Conditions on the Energy Benefits of BiRTG

Depending on the operating conditions, the RTG could have a different effect on energy loads. Thus, the effect of the four different operating scenarios suggested in Section 2.4.1 were compared. During the simulation, the operating schedule of a container followed the schedule of the small office suggested in Section 2.4.2. Monthly heating and cooling loads of the container for the cases of SC 1-1 and 1-2 are presented in Figure 5. The heating load of container B is consistently lower than that of container A. The most significant reduction occurs in January, with a decrease in the heating load of 6.7 kWh/m² over the month. However, the total cooling load of container B is 36.8% higher than that of container A for the case of SC 1-1. The result shows that heat from RTG has a positive effect on building heating; however, it also increases the cooling load of a building. For SC 1-2, RTG is covered by an external shading screen, and the side windows are open for natural ventilation for 24 h in July and August. Thus, only the results of the cooling load in July and August differ from the results of SC 1-1. As shown in Figure 5, the cooling load of container B for that period is 6.4% lower than that of container A for SC 1-2. Because the shading screen with a shading coefficient of 100% blocks a portion of solar energy reaching a container, the cooling load decreases.

The temperature of the greenhouse is not actively controlled for SC 1-1 and 1-2, so energy loads for heating and cooling cannot be compared. Instead, the effect of a container on the greenhouse can be inferred by analyzing the inside temperature of the greenhouses. A freestanding greenhouse on the ground and a rooftop greenhouse above a container were compared. Temperature conditions inside a greenhouse are intuitively depicted by 1 h steps in Figure 6 throughout the year. The vertical axis represents the hour of the day, and the horizontal axis indicates the day of the year. In this paper, the crop suitable for moderate temperatures, as suggested in Table 5, was considered. When the temperature exceeds 35 °C or falls below 9 °C, there is a possibility of crop damage [23,24]. The green area indicates that the greenhouse temperature is appropriate for crop growth, while the black area indicates that the crop can be damaged because of extremely high or low temperatures. In Figure 6a, it turns out that a freestanding greenhouse becomes difficult to grow crops for 1333 h a year. On the other hand, this is reduced to 846 h for the RTG in Figure 6b. The inappropriate period for the cultivation is primarily diminished during the winter nights, but it is slightly increased in the summer for the RTG compared to the freestanding greenhouse. This result demonstrates that waste heat through the roof surface of a container can be effectively used to provide the necessary thermal energy to the greenhouse, especially in winter.

Figure 7 presents monthly energy loads for SC 2-1 and 2-2, where a greenhouse is actively heated and cooled according to the conditions in Table 6. The impact of the RTG on the load change of a container is illustrated in Figure 7a, with results similar to those in Figure 5. While the heating load of container B is decreased by 40.3% due to transferred heat from the RTG, the cooling load of container B is increased by 32.6% for SC 2-1. The heating and cooling loads for container A are 81.3 kWh/m² and 25.9 kWh/m², respectively, indicating that the heating load is 3.1-times larger than the cooling load. Therefore, the total energy load for container B is reduced by 22.7%, although the cooling load is increased compared to container A. By comparing SC 2-1 and 2-2, the cooling load for container B in July and August has decreased by 14.2% due to the external shading effect. For SC 2-1, the heating and cooling energy loads of a greenhouse can be analyzed. Monthly energy loads of the RTG compared to that of a freestanding greenhouse (GH) are presented in Figure 7b. The heating load is reduced by 21.2%; however, the cooling load is increased by 6.4% in the RTG. The total annual energy load of the RTG is 0.7% lower than that of the GH.

Figure 8a and Figure 8b represent the annual energy demands for the container and the greenhouse, respectively, and Figure 8c shows the total annual demand. In Figure 8a, the container is operated according to the schedule of a small office. According to the operating scenarios of the RTG, the energy demand of the container under the RTG (container B) becomes different. Container B requires a lower total energy load for all scenarios compared to container A. The heating load of container B is reduced by 36.0–40.3%, while the cooling load increases by 24.4–36.8% depending on the scenarios. The case of SC 2-2 exhibits the most significant annual total energy-saving effect, with the total energy demand of container B decreased by 24.7%.

Similar results are suggested in Figure 8b in terms of the greenhouse. The total annual energy demands of the RTG are decreased by 0.7% and 4.0% for SC 2-1 and 2-2, respectively, compared to those of the freestanding greenhouse. As the cooling load of the greenhouse constitutes a large portion of the total energy load (the cooling load is 2.9 times higher than that of the heating load for SC 2-1), the energy-saving effect in the RTG is rather insignificant relative to that of the container. In Figure 8c, energy-saving effects of the total system are depicted. The separated model is the sum of an independent container and greenhouse on the ground, while the integrated model indicates the container and the RTG. For the integrated model, the total annual energy is saved up to 6.2% and 11.4% for SC 2-1 and SC 2-2, respectively. As a result, it is revealed that the BiRTG has an advantage in energy saving for both the container and the greenhouse.

3.3. Influence of Building Operating Conditions on the Energy Benefits of BiRTG

Depending on the building’s operating conditions, the heat exchange pattern between a building and RTG varies. Three types of building such as a small office, a small hotel, and a hospital for the container were selected, and compared. The annual energy load and energy reduction rate for each type are presented in

Table 7. Annual energy load and energy reduction rates depending on the building operating types for scenario 2-2 (Sep.: separated model, Int.: integrated model and Cont.: container).

Annual Energy Load (kWh/m2yr)
		Small Office			Small Hotel			Hospital
		Heating	Cooling	Total	Heating	Cooling	Total	Heating	Cooling	Total
Sep.	Cont.	81.3	25.9	107.3	106.0	13.4	119.4	279.1	93.0	372.1
	GH	81.1	112.7	193.8	81.1	112.7	193.8	81.1	112.7	193.8
Int.	Cont.	48.5	32.3	80.8	78.7	15.9	94.6	230.0	102.3	332.3
	RTG	63.8	122.2	186.1	59.9	123.4	183.3	60.7	120.8	181.5
Energy Reduction Rate (%)
		Small Office			Small Hotel			Hospital
		Heating	Cooling	Total	Heating	Cooling	Total	Heating	Cooling	Total
Cont.		40.3	−24.4	24.7	25.7	−18.2	20.8	17.6	−10.0	10.7
RTG		21.2	−8.4	4.0	26.1	−9.5	5.4	25.1	−7.2	6.3
Total		30.8	−11.4	11.4	25.9	−10.4	11.3	19.3	−8.5	9.2

, while the RTG is controlled according to the SC 2-2. For all cases, the integrated model has an advantage of reducing the energy load of the container and the greenhouse. When examining the results by the building type, the most significant decrease in total energy load of the building was observed in the small office at 24.7%, followed by the small hotel at 20.8%, and the hospital at 10.7%. However, regarding the greenhouse, the RTG above the hospital shows the best energy reduction, amounting to 6.3%. The highest energy reduction rate for the total system is 11.4% for the small office case. As the hospital is a high-energy-consumption facility, the amount of the total reduced energy load becomes higher than others at 52.1 kWh/m²yr. Since the heat exchange pattern between a container and an RTG changes according to their operating schedule, the amount of energy reduction becomes different depending on the types.

To analyze the heat exchange between the container and the RTG, the simulation data were examined for each case. We presented how heat transfer occurs through the roof surface during a day from the perspective of the container in Figure 9 and Figure 10. The temperatures of container B and the RTG are also provided for reference. By comparing the heat flow of container A and B, the effect of the RTG on the heat exchange of the roof surface can be analyzed. A positive value of heat flow indicates that heat flows from the roof surface to the inside of the container zone. The results on the peak cold day are presented in Figure 9a–c. Due to the heat storage effect of the RTG, the heat absorption from the roof during daytime of container B is higher than that of container A for all cases. Consequently, the absorbed heat from the RTG is utilized for daytime heating of the small office and the hospital, as depicted in Figure 9a,c. For cases requiring heating during night-time, such as the small hotel and the hospital, the RTG also has an advantage by diminishing heat loss through the roof surface, as proved in Figure 9b,c.

The same analysis was conducted for the peak hot day in Figure 10a–c. Similar to the results of Figure 9, the RTG transfers more heat to the container during daytime for all cases and even during night-time for the small hotel and the hospital. Considering the cooling, the RTG can cause the increase in the cooling loads. However, when comparing Figure 9 and Figure 10, the difference in heat absorption between container A and B is more pronounced on the peak cold day. This is because the thermal energy inflow into the greenhouse on the peak hot day is regulated using screens and ventilation. Thus, the RTG influences on the container more on the peak cold day than on the peak hot day, and the RTG is especially effective for buildings with larger heating demands during daytime.

4. Conclusions

The objective of this investigation is to assess the energy benefits of building integrated rooftop greenhouses by analyzing different operating scenarios for both the building and the greenhouse. Four different operating scenarios for the greenhouse and three different operating scenarios for the building are compared. A container and upper RTG are constructed, as well as building energy simulation models for the target system. Each scenario is examined to understand the energy dynamics between the building and upper RTG. For the simulation, the result of the integrated model (container and RTG above it) was compared with that of the separated model (simple sum of the stand-alone container and freestanding greenhouse on the ground).

Analysis of heat transfer across the container surfaces revealed that thermal energy from the RTG mainly flows into the container via the roof surface. This heat transfer led to a reduction in container heating load by 36.0–40.3% depending on the operating scenarios of the RTG. Although the cooling load was also increased, it was verified that the cooling load can be reduced by controlling the components of the greenhouse in scenarios 1-2 and 2-2. Due to the large heating load reduction, the annual energy saving of the container was 18.4-24.7%. From the perspective of the greenhouse, heat released through the roof surface of the container can be used for RTG heating, particularly during winter nights, even though this resulted in increased cooling loads by 6.4–8.4%. The total energy loads of the RTG compared to the freestanding greenhouse were reduced by 0.7–4.0%, indicating that the BiRTG has an advantage in effectively utilizing unused thermal energy in both the container and the greenhouse.

Furthermore, the impact of building operational conditions, such as a small office, a small hotel, and a hospital, was discussed. The heat exchange pattern between the building and RTG varies depending on the building’s operational conditions. As a result, the integrated model consistently has an advantage of reducing the total energy load of the container and the greenhouse across all scenarios. For scenario 2-2, the most significant decrease in total energy load is observed in the small office, with a reduction rate of 11.4%, resulting in energy reductions of 24.7% for the container and 4.0% for the RTG.

This study analytically investigated how buildings and greenhouses can have benefits from the energy perspective due to the structural effects of installing the greenhouse on the rooftop of buildings. Additionally, it is demonstrated that the effect can be different depending on the operation strategies. As a future research direction, the optimization of greenhouse control strategies is considered. The greenhouse operating scenarios investigated in this study were defined by establishing constraints for each component, and the effect of each scenario was compared. The energy benefits of a BiRTG can be maximized by identifying the optimal operational strategy for each component, and this will be discussed in succeeding research.