Experimental and Simulation Research on the Energy-Saving Potential of a Sunspace—Taking an Apartment in Qingdao as an Example

Abstract

This paper investigates the thermal performance and energy-saving potential of the sunspace of an old apartment building in Qingdao through experiments and software simulations. This study found that different modes of user behavior can have a substantial impact on the thermal performance of the studied sunspace in winter. The bedroom with a non-ventilated sunspace showed a higher average temperature than the bedroom without the sunspace. However, the bedroom with the sunspace had more heat loss than the bedroom without the sunspace when the sunspace was naturally ventilated, especially at night. In Delta temperature (DT)-controlled ventilation mode, the heating load of a bedroom can be reduced by 2.94 kWh/m² compared with the non-ventilated mode. Simply optimizing the roof configuration of the sunspace can significantly improve the heat gain of the non-ventilated sunspace and reduce the energy consumption for heating by 26 kWh/m². Compared with the unoptimized sunspace, the optimized ventilated sunspace can reduce the heating load of the bedroom by 38.82 kWh/m². In addition, the overheating of the room in summer can be solved by opening the exterior windows of the sunspace for ventilation during the day and leaving the door of the sunspace open at night.

1. Introduction

Since the 1970s, the energy crisis and environmental pollution have attracted significant attention worldwide. Energy consumption in buildings is an important part of the final energy consumption, accounting for 20% to 40% of direct energy consumption [1]. According to the Ministry of Energy, the problem of inefficient energy use in China is more severe compared to many developed countries such as America, Japan, and many countries in Europe. The 2018 China Building Energy Consumption Report states that the amount of energy consumed during building construction and operation in a year can occupy about 20.62% of the total annual energy consumption. The construction and operation of urban residential buildings consume significantly more energy than that of rural residential buildings. In 2020, China proposed a dual carbon target, which aims to hit peak carbon emissions by 2030 and achieve carbon neutrality by 2060.

The use of passive solar technologies in buildings can help reduce the energy consumption for heating in buildings. Common passive solar technologies include sunspaces, Trombe walls, and double-skin façades. Sunspaces consist of three main components: the exterior glazing, the intermediate cavity, and the interior wall or glazing. Among them, the cavity thickness of a sunspace, which is equivalent to the depth of the sunspace, is the largest [2]. Functionally, a sunspace can increase the floor area and be used as a place that allows people to breathe fresh air, hang clothes for drying, and grow plants. Therefore, sunspaces are often used in urban residential buildings. A sunspace can also separate the indoor and outdoor environments and act as a buffer to reduce heat loss from buildings [3]. The glazing performance of a sunspace plays an important role in reducing indoor energy consumption. It can allow useful visible light and short-wave radiation to penetrate into a room and block long-wave radiation emission. This can further increase the indoor air and wall-surface temperatures. In this way, the internal cavity of the sunspace can be transformed into an effective heat collector, reducing the thermal difference between the indoor and outdoor environments. The stored heat can be further transferred into the neighboring rooms through air-forced convection or by conduction through the interior walls [4].

Some studies have indicated that a sunspace can reduce 8% to 58% of a building’s energy consumption [5,6,7,8]. Hilliaho et al. [9] found that the maximum temperature difference between the outdoor air and the sunspace can reach 14 °C. Moreover, many parameters can influence the thermal performance of a sunspace. These parameters include glazing area [10,11], location [12,13,14,15], glazing type [16,17,18], form [19,20,21], shading devices [22], ventilation [23,24], thermal mass [25,26,27,28,29,30], and doors and windows [31,32]. Liu et al. [33] evaluated the sunspace temperature when the sunspace was set with three depths (150 cm, 120 cm, and 90 cm). They found that the highest indoor temperature was gained when the depth of the sunspace was 150 cm. Ulpiani et al. [34] evaluated the thermal performance of a sunspace when it was set to four depths (250 cm, 220 cm, 200 cm, and 150 cm). They also concluded that the thermal performance of the sunspace increased with decreasing depth. Babaee et al. [35] explored the thermal load of a room with different glazing types (single, double glazing, and double low-E glazing) and found that the sunspace with double low-E glazing had the best energy performance when the optimal window-to-ground ratio was 2.21. This is because low-E glass has a selective surface that permits short-wave radiation transmission to penetrate into the interior [36] and reduces heat loss transmission to the external environment [37]. Oliveti et al. [38] reached a similar conclusion. It was found that a ventilated sunspace has better energy-saving performance. Ulpiani et al. [34] found that the energy efficiency of a studied sunspace can be increased by more than 27% by using mechanically controlled ventilation. Zhu et al. [39] found that convective heat transfer from a sunspace to its adjacent rooms can reduce the energy consumption of the adjacent rooms by 28%. Ma et al. [40] used EnergyPlus to simulate the energy-saving performance of a ventilated sunspace. The results showed that compared with the bedroom without a sunspace, a bedroom with a ventilated sunspace reduced the energy consumption for heating by 14.7%. At the same time, compared with the bedroom with a non-ventilated sunspace, the bedroom with a ventilated sunspace can also reduce the energy consumption for heating by 4.4%. Similarly, Aelenei et al. [31] also used EnergyPlus software to simulate the energy performance of a sunspace in Portugal. The results showed that using a sunspace reduced the energy consumption for heating by 48% in winter, but it also increased the cooling energy consumption by 10% in summer. However, the increase in cooling energy consumption can be avoided by installing an adjustable shading system or a ventilation system. According to Chiesa et al. [41], shading and ventilation systems are two important techniques to reduce the cooling demand, and shading is more effective than ventilation. When these two techniques are combined, the overheating problem caused by a sunspace can be effectively solved. Batinah and Fayez [42] explored the optimal design of a sunspace in different climate zones in Spain in summer. They also concluded that an optimized sunspace should have an external shading system and proper ventilation.

In general, current studies indicate that there are numerous studies that focus on the thermal performance of non-ventilated sunspaces. In comparison, studies about the optimization of a ventilated sunspace are rather limited. Most studies investigated the energy-saving potential of sunspaces in detached houses, while very few studies analyzed the energy-saving potential of sunspaces in multi-story residential apartments. In this study, an old six-story apartment in Qingdao, which is located in the cold climate zone of China, was selected for a case study. The novelty of this paper is to propose an approach for evaluating the sunspace performance that combines parametric design and performance evaluation. A simplified interzone airflow model ‘ZoneCrossMixing’ was used to simulate the air exchange between the bedroom and the sunspace. The impact of different modes of user behavior and physical configuration on the thermal performance of a sunspace was analyzed and investigated through actual measurements and software simulations. The purpose of this study is not only to propose an optimal sunspace solution but also to provide various feasible optimization options through a large number of simulated cases. It is expected to make some suggestions for residents to use the sunspace rationally in different seasons and to provide various possible optimization solutions for designers to improve residential thermal comfort and energy-saving efficiency. Figure 1 shows the flow chart of the paper structure.

2. Methodology

2.1. Experimental Study

In this paper, an old apartment in Qingdao (elevation: 40 m, longitude: 120.4° E, latitude: 36.1° N), which is located in the cold climate zone of China, was selected for a case study. Qingdao is cold and windy in winter and hot and humid in summer. It is in the transition zone between the humid subtropical climate and the humid continental climate. The appearance of the sunspace is shown in Figure 2. The apartment has two bedrooms, one of which is a bedroom without a sunspace, denoted as bedroom 0. The other one is a bedroom with a sunspace, denoted as Bedroom 1 (Figure 3). The sunspace has four large, south-facing windows and an opaque, flat roof. The apartment is heated by radiators through a district heating system (running 24 h) in winter and cooled by an air conditioner in summer.

The Vantage Pro2 weather station (Davis, CA, USA) was used to measure solar radiation and air temperature. The TC96-V DT controller (Rockson, Shanghai, China) was used to control the air supply between the sunspace and bedroom. The testo-174H thermometers (Testo, Berlin, Germany) were used to measure the temperature of the sunspace and bedroom. The testo-175T3 surface thermometers (Testo, Germany) were used to measure the surface temperature of the radiators. The room plan and the position of the thermometers are shown in Figure 4. Among them, thermometers No. 5 and No. 6 were placed in the position of the air inlet and outlet, respectively.

Table 1. Uncertainties of experimental devices.

Device	Parameter	Resolution	Range	Accuracy
Testo-174H [43]	Temperature	0.1 °C	−20 to +70 °C	±0.5 °C
Testo-175T3 [44]	Temperature	0.1 °C	−40 to +400 °C	±0.5 °C
TC96-V DT controller [45]	Delta temperature	0.1 °C	0 to +60 °C	±1 °C
Vantage Pro2 weather station [46]	Solar radiation	1 W/m²	0 to 1800 W/m²	±5%
	Temperature	1 °C	−68 to +64 °C	±2 °C

shows the uncertainties of these devices in this experiment.

In order to study the thermal performance of a sunspace under different modes of user behavior in winter, the thermal performance of the sunspace in DT-controlled ventilation mode, ‘Door-closed’ mode (non-ventilated sunspace), and ‘Door-open’ mode (naturally ventilated sunspace) were measured in December (heating) and April (HVAC system was shut down). All measurements were obtained over a 24 h period. The Delta-temperature-controlled ventilation system was composed of an intelligent DT controller and two fans. The DT controller was externally connected to two air-temperature sensors, which were used to detect the temperature difference between the sunspace and Bedroom 1. During the actual measurement, the EPS board was used as the door. The air supply fan and the return air fan were installed at a distance of 15 cm from the upper and lower edges of the EPS board. When the temperature difference reached the target temperature, the air in the sunspace was sent to the bedroom by the supply fan, and the air in the bedroom was sent to the sunspace by the return fan. Figure 4 shows the indoor condition of Bedroom 1, which was controlled by the Delta temperature-controlled ventilation system. DT (Delta temperature) is equal to the supply (inlet) air temperature minus the return (outlet) air temperature. The initial temperature difference was set to 3 °C, the final temperature difference was set to 1 °C, and the ventilation volume of the fan was set to 78 m³/h. When the DT was greater than 3 °C, the ventilation equipment started working. When DT was less than or equal to 1 °C, the ventilation equipment stopped working. When the ‘Door-closed’ mode was enabled, the door between the sunspace and Bedroom 1 was closed. When the ‘Door-open’ mode was enabled, the door between the sunspace and Bedroom 1 was kept open allowing the air to ventilate between the sunspace and Bedroom 1. In addition, actual measurements were conducted to address the potential overheating of the sunspace in summer (the HVAC system was shut down). Three modes of user behavior were recorded: ‘Door-closed-window-closed (DCWC)’, ‘Door-open-window-closed (DOWC)’, and ‘Door-closed-window-open (DCWO)’. DCWC means that the door between the sunspace and bedroom is closed, and the window of the sunspace is closed. DOWC means that the door between the sunspace and bedroom is opened, and the window of the sunspace is closed. DCWO means that the door between the sunspace and bedroom is closed, and the window of the sunspace is opened.

Table 2. The monitoring schedule.

Time		Mode of User Behavior	Details
			Door	Window	Ventilation
Heating	2021/12/18~2021/12/21	DT-controlled ventilation	Closed	Closed	The air supply starts when the DT is 3 °C and stops when the DT is 1 °C.
	2021/12/23~2021/12/26	Door-closed	Closed	Closed	No ventilation.
	2021/12/28~2021/12/30	Door-open	Open	Closed	No ventilation.
HVAC system shut down	2022/04/17~2022/04/18	DT-controlled ventilation	Closed	Closed	The air supply starts when the DT is 3 °C and stops when the DT is 1 °C.
	2022/04/14~2022/04/16	Door-closed	Closed	Closed	No ventilation.
	2022/04/10~2022/04/13	Door-open	Open	Closed	No ventilation.
HVAC system shut down	2022/07/05~2022/07/07	DCWC	Closed	Closed	No ventilation.
	2022/07/11~2022/07/14	DCWO	Closed	Open	No ventilation.
	2022/07/16~2022/07/19	DOWC	Open	Closed	No ventilation.

shows the monitoring schedule.

2.2. Simulation Study

The energy performance of the studied sunspace was evaluated by using the proposed method: the parametric performance design. This method combines parametric design and performance evaluation. It is an ‘all-in-one’ toolbox for non-professionals to assess and optimize the energy-saving performance of a sunspace. Developed from the Ladybug plugin, the simulation method is operated on the EnergyPlus engine in the grasshopper platform of Rhino. A parametric model of the study room was built and simulated by using Ladybug Tools. A notable feature of EnergyPlus is that heat and mass transfer models can be integrated to measure the air movement between different zones. Compared to the Airflow network model, ‘ZoneCrossMixing’ is a simplified interzone airflow model in the software. It is ideal to simulate the airflow exchange between two zones. Therefore, the ‘ZoneCrossMixing’ model was applied to simulate the air exchange between the bedroom and the sunspace instead of the airflow network model. In the ‘ZoneCrossMixing’ model, Delta temperature was the key parameter that was used to control the airflow from the sunspace to Bedroom 1. Because Delta temperature cannot be a negative value, there would be two conditions. If the Delta temperature was positive, it means that the sunspace temperature was higher than the bedroom temperature, and the warm air was being drawn from the sunspace to the bedroom. This can lead to a mixing of the warm and cold air in the bedroom. If the Delta temperature was zero, the warm and cold air was still mixed regardless of the relative temperatures of the two rooms.

During the heating period, the heating system (running 24 h) was used to maintain the bedroom temperature at 18 °C.

Table 3. Parameters of the structural layers of the wall.

Layer	Thickness (m)	λ (W/m K)	ρ (kg/m³)	cp (J/kg K)
Stucco	0.025	0.69	1858	837
Brick	0.24	0.81	1920	790
Stucco	0.025	0.69	1858	837

,

Table 4. Parameters of the structural layers of the floor.

Layer	Thickness (m)	λ (W/m K)	ρ (kg/m³)	cp (J/kg K)
Tile	0.019	0.06	368	590
Concrete	0.101	0.53	1280	840
Gypsum	0.015	0.16	784.9	830

,

Table 5. Parameters of the structural layers of the window.

	U-Factor (W/m² K)	SHGC	VT
Window	2.953	0.39	0.27

and

Table 6. Parameters of the structural layers of the roof.

Layer	Thickness (m)	λ (W/m K)	ρ (kg/m³)	cp (J/kg K)
Concrete	0.2	0.52	1280	5000
Air gap	0.18	-	-	-
Tile	0.02	0.06	368	590

show the structural parameters of the sunspace as a parametric simulation model. The typical meteorological year (TMY) data of Qingdao (Qingdao_TMYx.2007–2021) was used for the simulation. The basic climatic parameters including global horizontal radiation, dry bulb temperature, dew point temperature, relative humidity, wind direction, and wind speed are shown in

Table 7. Weather data summary.

Monthly Means	January	February	March	April	May	June	July	August	September	October	November	December
Global Horiz Radiation (Avg Hourly) (Wh/sq.m)	297	351	416	443	457	414	365	373	364	341	266	257
Global Horiz Radiation (Avg Daily Total) (Wh/sq.m)	2918	3719	4910	5738	6384	5973	5181	4967	4466	3756	2682	2465
Dry Bulb Temperature (Avg Monthly) (°C)	0	2	7	13	19	22	26	26	22	16	8	1
Dew Point Temperature (Avg Monthly) (°C)	−6	−5	0	4	9	17	22	21	15	8	2	−5
Relative Humidity (Avg Monthly) (%)	62	56	64	60	57	76	80	74	68	62	65	59
Wind Direction (Monthly Mode) (degrees)	330	340	180	180	180	160	160	160	160	340	0	330
Wind Speed (Avg Monthly) (m/s)	3	3	4	4	3	4	3	2	2	3	3	2

. The simplified dynamic parametric simulation model is shown in Figure 5. Using this model, heat collection and energy savings were simulated to assess the different configurations in various parameter settings. The optimized model was also coupled with the DT-controlled ventilation device to compare and evaluate the energy-saving potential of the optimized ventilated sunspace.

Table 8. Input parameters of the sunspace in the simulation program.

Parameters	Range	Increment	Amount
θ (slope angle of the roof)	0~60°	10°	7
WRR (window-to-roof ratio)	0~90%	10%	10
RGT (roof glazing type)	Single glazing/double low-E glazing	-	2
FGT (façade glazing type)	Single glazing/double low-E glazing	-	2
WWR (window-to-wall ratio)	10~90%	10%	9
Depth	0.6~1.5 m	0.1 m	10
FR (flow rate)	0.00~0.06 m3/s	0.01 m3/s	6
DT (delta temperature)	0~8 °C	1 °C	9

shows the parameters used in the simulation programs. There are eight parameters, including θ (slope angle of the roof), WRR (window to roof ratio), RGT (roof glazing type), FGT (façade glazing type), WWR (window to wall ratio), depth (width of the sunspace), FR (flow rate), and DT (delta temperature). The input range of θ is between 0° to 60°, with a total of seven values at intervals of 10°. The range of WRR is from 0% to 90% with a total of nine values, at an interval of 10%. RGT and FGT both have two types of settings: single glazing and double low-E glazing. The input range of the WWR is from 10% to 90% with a total of nine values at intervals of 10%. The input range of depth is from 0.6 m to 1.5 m, with a total of ten values at intervals of 0.1 m. The input range of FR is from 0.00 m³/s to 0.06 m³/s, with a total of seven values at intervals of 0.01 m³/s. The input range of DT is between 1 and 6 °C, with a total of seven values at intervals of 1 °C. A total of 1,058,400 configurations were calculated to explore and evaluate the energy-saving potential of the sunspace in different physical forms.

Table 9. Characteristics of the glazing materials of the sunspace.

Glazing Type	Thermal Conductivity (W/m²·K)	SHGC	VT
Single glazing	4.6	0.28	0.6
Double low-E glazing	1.48	0.67	0.78

shows the characteristics of the glazing materials of the sunspace.

3. Results

3.1. Results of the Experimental Study

Figure 6 shows the relationship between room temperature and solar radiation variations in the ‘Door-closed’ mode during the heating period. It can be seen that the overall temperature of Bedroom 1 (with the sunspace) was higher than that of Bedroom 0 (without the sunspace). The maximum temperature difference reached 1.9 °C. The average temperature of Bedroom 1 was 23.8 °C, which was 1.6 °C higher than the average temperature of Bedroom 0. On 26 December, there was a significant increase in solar radiation. The room temperature of Bedroom 0 reached its peak value in the late afternoon with a maximum temperature of 30.4 °C, creating an uncomfortable indoor environment. In contrast, Bedroom 1 had fewer temperature fluctuations and no overheating in the afternoon.

Figure 7 shows the relationship between temperature and solar radiation variations in the ‘Door-closed’ mode during the non-heating period. The overall temperature of Bedroom 1 remained generally higher than that of Bedroom 0, and the average temperature of Bedroom 1 was 0.8 °C higher than that of Bedroom 0. Unlike the heating period, the temperature difference between Bedroom 1 and Bedroom 0 was smaller during the afternoon hours of the day, and the temperature difference was even more pronounced at night. This indicates that the ‘Door-closed’ mode can reduce the bedroom’s heat loss during the night. In addition, compared with the average bedroom temperature during the heating period, the average temperature of the two bedrooms during the non-heating period decreased. The overall temperature fluctuation of the two bedrooms was small, and the overheating of the two bedrooms disappeared in the afternoon.

In summary, the thermal performance of the studied sunspace was very efficient in winter when the ‘Door-closed’ mode was enabled. The sunspace can reduce the bedroom’s heat loss at night and improve the room’s overall thermal stability. In addition, Bedroom 1 can be overheated because it received a lot of direct solar heat in the afternoon. Using radiators for heating can further lead to overheating in winter. This means that Bedroom 1 can better withstand the effects of the outdoor environment and increase the indoor thermal stability.

Figure 8 shows the relationship between the temperature and solar radiation variations in the ‘Door-open’ mode during the heating period. Compared with the bedroom under the ‘Door-closed’ mode, the temperature of Bedroom 1 in the ‘Door-open’ mode fluctuated more during the day. However, the overall thermal stability of Bedroom 1 was still better than that of Bedroom 0. However, the average temperature of Bedroom 1 was 1 °C lower than that of Bedroom 0. This is especially true at night, as Bedroom 1 saw more heat loss than Bedroom 0.

Figure 9 shows the relationship between the temperature and solar radiation variations in the ‘Door-open’ mode during the non-heating period. The average temperature of Bedroom 1 was 0.98 °C higher than that of Bedroom 0, and the maximum temperature difference between the two rooms reached 1.38 °C. However, because the outdoor temperature and solar radiation dropped dramatically from 12 April to 13 April, the additional heat gain created by the sunspace also decreased and eventually disappeared. This phenomenon was observed as the temperature of Bedroom 1 became the same as that of Bedroom 0. This indicates that the ‘Door-open’ mode of the sunspace can reduce the thermal stability of the adjacent bedroom. When the outdoor temperature becomes warmer, the ‘Door-open’ mode can preserve the heat. However, the bedroom suffered significant heat loss in the ‘Door-open’ mode at night when the outdoor temperature dropped drastically.

Figure 10 shows the relationship between the temperature and solar radiation variations in the DT-controlled ventilation mode during the heating period. The DT-controlled ventilation mode was applied from 18 to 21 December, and the ‘Door-closed’ mode was enabled from 23 to 26 December. Due to the lack of solar radiation and cold weather outside, the sunspace temperature was lower than Bedroom 1′s temperature throughout the day (Figure 6), and, therefore, DT-controlled ventilation was stopped. Furthermore, the average temperature of Bedroom 1 was above 20 °C during the heating period. It was not suitable to use the DT-controlled ventilation mode in the sunspace under the current heating conditions because of the lack of proper temperature difference conditions and thermal comfort requirements.

Figure 11 shows a schematic diagram of the effective working zones in DT-controlled ventilation mode during the non-heating period. Due to the significant temperature difference between Bedroom 1 and the sunspace during the non-heating period, the DT-controlled ventilation mode produced an effective working zone. In the effective working zone, the temperature of Bedroom 1 was slightly increased, proving that the DT-controlled ventilation had some positive heating effect on Bedroom 1.

In order to understand and avoid the indoor overheating that may be caused by the sunspace in summer, three modes of use of the sunspace were tested: the ‘Door-closed-window-closed’ mode, ‘Door-closed-window-open’ mode, and ‘Door-open-window-closed’ mode. Among them, Figure 12 shows the relationship between the indoor and outdoor temperatures and the solar radiation variations in the summer when the sunspace was in the ‘Door-closed-window-closed’ mode. As shown in the figure, the overall temperature of Bedroom 1 was significantly higher than that of Bedroom 0. The average daily temperature difference between Bedroom 1 and Bedroom 0 was 0.4 °C, and the maximum temperature difference reached 0.6 °C. Therefore, Bedroom 1 was overheated in ‘Door-closed-window-closed’ mode in summer, but the degree of overheating was mild.

Figure 13 shows the relationship between the temperature and solar radiation variations of the sunspace in the ‘Door-closed-window-open’ mode in summer. Compared with the two rooms in the ‘Door-closed-window-closed’ mode, the average temperature difference between Bedroom 1 and Bedroom 0 decreased and was only 0.27 °C. At noon and in the afternoon, the temperature of Bedroom 1 was similar to Bedroom 0. Therefore, in ‘Door-closed-window-open’ mode, opening the external windows of the sunspace for ventilation during the daytime can minimize the heat storage effect and further mitigate the overheating problem of the adjacent bedroom.

Figure 14 shows the relationship between temperature and solar radiation variations in the summer in the ‘Door-open-window-closed’ mode, and the data shows a 0.02 °C positive temperature difference between Bedroom 1 and Bedroom 0. The temperature of Bedroom 1 was lower or equal to Bedroom 0, especially at night. This suggests that opening the door of the sunspace at night can effectively mitigate the overheating of the adjacent bedroom. Therefore, overheating in the summer caused by the sunspace can basically be avoided by opening the external windows of the sunspace for ventilation during the daytime and leaving the door between the sunspace and the bedroom open at night. Bataineh [47] also had a similar conclusion in their study of a sunspace in Amman-Jordan.

3.2. Results of Simulation Study

3.2.1. Model Validation

This study used ASHRAE Guide 14-2014 [48] to calibrate the simulation model. Two dimensionless metrics for determining the error were applied: the MBE (mean bias error) and the CV(RMSE) (coefficient of variation of the root mean square error). Simplified MBE formulas and CV(RMSE) formulas are shown in Equations (1) and (2).

$$MBE=\frac{{\Sigma }_{\mathrm{i}=1}^n\left(M_i-S_i\right)}{{\Sigma }_{\mathrm{i}=1}^n{M}_i}\times 100\%$$

(1)

$$CV\left(RMSE\right)=\frac{1}{\overline{y}}\sqrt{\frac{{\Sigma }_{\mathrm{i}=1}^n{\left(M_i-S_i\right)}^2}{n}}\times 100\%$$

(2)

According to ASHRAE guidelines, the simulation model should be calibrated if the hourly MBE value is less than ±10% and the hourly CV (RMSE) value is less than 30%. Figure 15 shows the comparison of simulated and measured temperatures. The results indicate that the simulation data generally match the measured data. The error indicators for each room are shown in

Table 10. Simulation error index of each room.

Room	MBE	CV (RMSE)
Sunspace	−3%	6%
Bedroom 1	2%	2%
Bedroom 0	1%	1%

. This indicates that the simulation model is able to generate relatively valid data that match reality.

3.2.2. Efficiency Analysis

The sunspace roof was first optimized by using a software simulation. Figure 16 shows the results of different combinations of WRR (window-to-roof ratio) and θ (roof tilt angle) configurations. Figure 16a shows the average temperature of the sunspace without heating in winter; Figure 16b shows the amount of energy used for heating Bedroom 1.

The results show that when the roof was flat (θ = 0°) and the WRR increased from 0% to 90%, the average temperature of the sunspace decreased by 0.1 °C, and the heating load decreased by 1 kWh/m². This may be because the increasing WRR can increase the incident solar radiant heat and the heat conduction in the sunspace. While increasing the size of the skylight, more solar radiant heat entered Bedroom 1 through the inner wall glazing, and therefore the energy consumption decreased. However, in general, there was no additional heat gain if the roof of the sunspace was a skylight with single glazing. When the roof was tilted at certain angles, increasing the WRR increased the heat gain of the sunspace. When the WRR was greater than or equal to 30%, the larger the roof tilt angle, the more heat gain the sunspace can attain. Bastien [49] also obtained a similar result. Therefore, a tilted roof was suitable for a larger WRR. Because the height of the sunspace should be more than 1.8 m, the maximum roof tilt angle was 40°. When the WRR was 90% and the roof tilt angle was increased from 0° to 40°, the heating load was significantly reduced by 4 kWh/m².

Figure 17 shows the simulation results of different combinations of the WRR (window-to-roof ratio) and θ (roof tilt angle) of the roof configurations when the roof glazing type was changed to double low-E glazing. Among them, Figure 17a shows the average temperature of the sunspace without heating, and Figure 17b shows the energy consumption for heating in Bedroom 1. Better thermal performance of the sunspace can be achieved by optimizing the roof material. More heat gain can be achieved when the roof glazing type is changed to double low-E glazing. The results show that increasing the WRR of a flat roof can significantly increase the heat gain. Similar to single glazing, increasing WRR or θ can also improve the thermal performance and heat gain of the sunspace. Moreover, the overall thermal performance of the sunspace was substantially improved compared to single glazing. When the WRR was set to 90% and θ was 40° and the double low-E glazing was applied, the heating load of the bedrooms was reduced by 21.11 kWh/m² and the average temperature of the sunspace was increased by 6 °C. The simulation indicates that the best sunspace roof configuration was the double low-E glazing with a 40° tilt angle and 90% WRR.

The impact of different façade types and depths on the thermal performance of the sunspace was further investigated by using the software simulation. Figure 18 shows the results of different combinations of depth and WWR (window-to-wall ratio) of the sunspace. Figure 18a shows the average temperature of the sunspace without a heating system; Figure 18b shows the amount of energy used for heating Bedroom 1. The results show that the sunspace’s heat gain was positively correlated with WWR and negatively correlated with depth when the roof was an opaque, flat roof. Bataineh [47] also had a similar conclusion. However, as the WWR increased, the additional heat gain rate kept decreasing when the depth remained unchanged. In particular, when the WWR was greater than 60%, the increasing rate of heat gain decreased as the WWR increased. Furthermore, increasing the depth can lead to inefficient thermal performance. This might be due to the extra shadow casting created by the untransparent roof cover.

To eliminate the negative impact of the opaque roof shading on the sunspace, the roof configuration was set to the optimized roof parameters: θ was set to 40°, WRR was 90%, and RGT was double low-E glazing. Then, the thermal performance of the sunspace was simulated when different depths and WWRs (window-to-wall ratios) of the sunspace were combined. The results are shown in Figure 19. It can be seen that the heat gain of the sunspace was negatively correlated to WWR, and the heat gain of the sunspace was positively correlated to depth. This was the opposite of the previous result of the opaque roof. Increasing the depth can increase the heat gain of the sunspace. This is largely because increasing the depth allows more solar radiation to penetrate the roof and enter the sunspace. Furthermore, increasing the WWR can decrease the thermal performance of the sunspace when the depth remains unchanged. This is because the large window opening area can increase the thermal conductivity of the sunspace, resulting in weak thermal insulation. When the roof tilt angle θ was 40°, the optimal depth of the sunspace was 1.1 m. Moreover, this study assumed that the minimum WWR should be no less than 50%, as a small WWR can affect indoor daylight. Therefore, in the simulation, when the sunspace depth was reduced from 1.2 m to 1.1 m and the WWR was reduced from 70% to 50%, the heating load of the bedroom was reduced by 0.85 kWh/m² and the average temperature of the sunspace was increased by 0.5 °C.

The thermal performance of the sunspace can be further optimized by changing the façade glazing type from single glazing to double low-E glazing. In Babaee’s study [35], they also obtained a similar result. Figure 20 shows the thermal performance of the sunspace configurations with different combinations of depth and WWR (window-to-wall ratio). The data show that the range of WWR values decides whether increasing depth has a positive or negative impact on the sunspace. When the WWR was less than 50%, the heat collection capacity was positively correlated to the depth. This is because increasing depth can increase the glazing area of the façade, and this means that the sunspace can receive more solar radiation and heat. When the WWR was 50%, the effect of the depth on the sunspace was negligible. When the WWR was greater than 50%, the thermal performance was negatively correlated to depth. This result might be generated because of the excessive window-opening area of the façade, which increased the thermal conductivity of the sunspace. In addition, the high-performance façade glazing greatly increased the heat collection and insulation of the sunspace. When the WWR was set to 50% and the depth was 1.1 m, changing the façade glazing type from single glazing to double low-E glazing reduced the heating load by 11.92 kWh/m² and increased the average temperature of the sunspace by 9.64 °C. By optimizing the structural configuration of the sunspace, the thermal performance of the sunspace was significantly improved. The average temperature of the optimized sunspace reached 25.38 °C in winter when double low-E glazing was used.

Figure 21a shows the average temperature of Bedroom 1 without heating by using different combinations of FR (flow rate) and DT (delta temperature). Figure 21b shows the amount of energy consumed for heating Bedroom 1 by using different combinations of FR and DT. The results show that the larger the FR, the more energy was consumed to heat the bedroom when DT was equal to 0° and DT-controlled ventilation mode was enabled. When DT was larger than 0°, the energy consumption decreased significantly, and more specifically, it decreased as the FR increased. Moreover, the energy consumption for heating was the lowest when DT was set to 1 °C, resulting in the optimal 0.03 m³/s air flow rate. Under these parameter settings, using the DT-controlled ventilation mode reduced the heating load by 0.58 kWh/m² compared to the ‘Door-closed’ mode. Compared with the unoptimized conditions, the heating load was reduced by 38.82 kWh/m².

The optimized sunspace shows a significant improvement in energy performance and saving.

Table 11. Optimized sunspace configurations and their energy savings.

Configurations	θ (o)	WRR	RGT	FGT	WWR	Depth (m)	FR (m3/s)	DT (°C)	Heating Loads (kWh/m2)	Efficiency
Conf_0	0	0%	None	Single glazing	70%	1.2	0	-	39.15	-
Conf_1	0	90%	Single glazing	Single glazing	70%	1.2	0	-	38.24	2%
Conf_2	20	90%	Single glazing	Single glazing	70%	1.2	0	-	36.11	7%
Conf_3	40	90%	Single glazing	Single glazing	70%	1.2	0	-	34.25	13%
Conf_4	40	90%	Double low-E glazing	Single glazing	70%	1.2	0	-	13.14	66%
Conf_5	40	90%	Double low-E glazing	Single glazing	50%	1.1	0	-	12.83	67%
Conf_6	40	90%	Double low-E glazing	Double low-E glazing	50%	1.1	0	-	0.91	98%
Conf_7	40	90%	Double low-E glazing	Double low-E glazing	50%	1.1	0.03	1	0.33	99%

shows some optimized sunspace configurations and their energy savings. These results can provide data support for the renovation and design of the sunspace. Conf_0 represents the unoptimized model. By optimizing the configuration of the sunspace and ventilation mode, the energy used for heating the bedroom can be reduced by up to 99%.

4. Discussion

This paper explores the energy-saving potential of a sunspace in an old residential apartment by using six modes of user behavior. The thermal performance of the sunspace in winter was investigated by enabling the ‘Door-closed’ mode, ‘Door-open’ mode, and DT-controlled ventilation mode, and the sunspace overheating in summer was also studied by enabling the ‘Door-closed-window-closed’ mode, ‘Door-closed-window-open’ mode, and ‘Door-open-window-closed’ mode. By using the parametric performance design method, 1,058,400 possible configuration combinations were simulated. Among them, 468 options were selected to evaluate the energy-saving effect of the sunspace in the six ventilation modes.

The thermal performance of a sunspace with different modes of user behavior in winter was different. During the heating period, the average temperature of the studied bedroom with the sunspace in ‘Door-closed’ mode was 1.6 °C higher than the bedroom without the sunspace. In the non-heating period, the temperature difference between the two conditions was only 0.8 °C. In ‘Door-open’ mode, the average temperature of the bedroom with the sunspace was 1 °C lower than the bedroom without the sunspace during the heating period, and the temperature difference between the two was 0.98 °C in the non-heating period. This means that it is not suitable to use the DT-controlled ventilation mode when district heating is applied in Qingdao. Therefore, it is necessary to integrate the HVAC system with the DT-controlled ventilated sunspace in a more flexible way. For example, in order to achieve more energy-efficient performance, the HVAC system can be set to ‘Low’ when using the DT-controlled ventilated sunspace.

From this study, two major suggestions are provided for further studies. First and foremost, a series of parametric modeling tools in modular form should be further explored from engineering and economic perspectives. These tools can be upgraded with deep reinforcement learning functions to automatically gather the best combinations of various parameters of a subject. Moreover, due to the recent research trend being more focused on the development of multifunctional sunspaces, this research recommends inventing new types of multifunctional sunspaces that can integrate sunspaces with other ‘green’ functions at the design stage, such as humidification and dehumidification, air purification, oxygen supply, carbon dioxide removal (photosynthesis), power generation, natural lighting, and sound insulation.

5. Conclusions

The differing thermal performance of a sunspace in various modes of user behavior was tested in this study, and the energy-saving potential of the ventilated sunspace was further evaluated and optimized by using a proposed parametric performance design method. By optimizing the configuration of the sunspace and ventilation mode, the energy consumed for heating in the studied Bedroom 1 can be reduced by up to 99%.

The results of DT-controlled ventilation show that the energy used for heating Bedroom 1 was the lowest when the Delta temperature was set to 1 °C. In this setting, the larger the flow rate, the more energy that can be saved. Applying the DT-controlled ventilation mode reduced the heating load by 0.58 kWh/m² compared to the ‘Door-closed’ mode. Compared with its unoptimized condition, the heating load was reduced by 38.82 kWh/m². Roof configuration has a great impact on the thermal performance of a sunspace. Simply optimizing the roof configuration of a non-ventilated sunspace can significantly improve its heat gain and reduce the energy consumed for heating the bedroom by 26 kWh/m².

Increasing the θ or WRR of the sunspace can also improve the thermal performance and heat gain of the sunspace. Glazing material also shows a great impact on the thermal performance of a sunspace. The best sunspace roof configuration was the double low-E glazing with a 40° tilt angle and 90% WRR. When using the double low-E glazing on the roof of the sunspace, the heating load of the bedroom can be reduced by 21.11 kWh/m². When the roof was an opaque, flat roof, the heat gain of the sunspace was negatively correlated to depth and positively correlated to WWR. When the roof tilt angle θ was 40°, the optimal depth of the sunspace was 1.1 m. When the sunspace depth was reduced from 1.2 m to 1.1 m and WWR was reduced from 70% to 50%, the average temperature of the sunspace was increased by 0.5 °C and the heating load of the bedroom was reduced by 0.85 kWh/m². When using double low-E glazing on the façade, the heating load of the bedroom can be reduced by 11.92 kWh/m², and the average temperature of the sunspace can be increased by 9.64 °C. In summer, the overheating problem of the bedroom in Qingdao can be basically solved by opening the exterior windows for ventilation during the day and opening the door of the sunspace at night.