Heating energy savings potential from retrofitting old apartments with an advanced double-skin façade system in cold climate

Abstract

Apartments account for over 60% of total residential buildings and consume a significant portion of primary energy in South Korea. Various energy efficiency measures have been implemented for both new apartment constructions and existing apartment retrofits. Old apartment structures have poor thermal performances, resulting in a high energy consumption. The South Korean government initiated retrofitting projects to improve the energy efficiency in old apartments. Apartment owners typically replace old windows with high-performance windows; however, there is still a demand for better and more innovative retrofit methods for a highly improved energy efficiency. This paper proposes an advanced double-skin façade (DSF) system to replace existing balcony windows in old apartments. Considering the cold climate conditions of Seoul, South Korea, it mainly discusses heating energy savings. Three case models were developed: Base-Case with existing apartment, Case-1 with typical retrofitting, and Case-2 with the proposed DSF system. The EnergyPlus simulation program was used to develop simulation models for a floor radiant heating system. A typical gas boiler was selected for low-temperature radiant system modeling. The air flow network method was used to model the proposed DSF system. Five heating months, i.e., November to March, and one representative day, i.e., January 24, were selected for detailed analysis. The main heat loss areas consist of windows, walls, and infiltration. The results reveal that the apartment with the DSF retrofit saves 38.8% on the annual heating energy compared to the Base-Case and 35.2% compared to Case-1.

Keywords

double-skin façade / retrofitting / high-rise apartment / heating energy / building simulation

Introduction

Buildings generate about 23% of the total CO₂ generation in South Korea. Embedded energy used for producing and transporting subsidiary building materials and construction is estimated to reach approximately 38% of the entire CO₂ generation [¹]. Many studies have been conducted to conserve energy in buildings. New policies have also been developed to improve building energy efficiency, e.g., improved insulation, starting with the South Korean energy conservation policy in the 1980s [²]. These policies have been enforced on both new and existing buildings. For existing buildings, performance improvement is mostly conducted through retrofitting processes.

This paper presents the results of simulation modeling studies for a double-skin façade (DSF) system applied in apartments to improve energy performance and focuses on 20- to 30-year-old apartments which comprise more than 65% of the total retrofitting rate of residential buildings. Figure 1 shows the pre- and post-installation of the DSF system in an apartment complex in South Korea. Figure 1(a) is the façade of an existing apartment building where air conditioning (A/C) condensing units are installed arbitrarily. Figure 1(b) is a rendering of the future façade of the apartment after the DSF installation.

This paper uses a new and advanced DSF system as a balcony window replacement to improve the thermal performance of apartments, aiming to determine the applicability of the DSF system in apartments by analyzing energy performance. The main subject of this paper, therefore, is to study the potential of the system for heating energy savings. Since the city of Seoul is a heating-dominant area, heat loss parameters are analyzed, and energy savings achieved from installing the DSF system are discussed. Three cases are developed for comparative analyses: The Base-Case is an apartment unit built 20 to 30 years ago with typical construction at that time. Case-1 is the same apartment unit as Base-Case with a typical retrofit of windows. Case-2 is an apartment unit with the DSF system installed. All three cases have the same traditional floor radiant heating system, which uses hot water supplied from an individual boiler.

The scope of this paper is classified into two aspects.

First, the energy consumption of boilers by the installation of the DSF system is compared to analyze the annual heating energy consumption according to the thermal performance of the DSF system and characteristics of boiler part load ratio (PLR).

Second, this paper also calculates simply payback period considering manufacturing, transportation, and installation cost. The detailed cost analysis will be performed in the future with suggestions such as reduction of manufacturing, transportation, and installation cost and suggest subsidies to reduce the payback period, which is one of the disadvantages of the DSF system.

One of the novelties of this paper is to know how the installation of a DSF system can reduce building energy consumption not only for the system itself but also for the radiant heating system. This paper also analyzes floor temperature, boiler gas energy consumption, and PLR of the boiler to understand how the boiler works after the installation of the DSF system and to propose the optimal boiler capacity to further reduce heating energy consumption in the future.

Literature review

Apartment retrofitting

The South Korean government is strengthening its policies to reduce energy consumption through retrofitting in existing apartments. Windows need to have double Low-E glass or triple glazing. The heat transmission coefficient (U-value) should be 3.0 W/(m²∙K) instead of 3.84 W/(m²∙K) [³]. The majority of studies focus on the technical viewpoint, such as improvement of the insulation of exterior walls and exterior windows, improvement of infiltration, and replacement with high-efficiency equipment [³]. However, residents do not prefer replacement with high-efficiency equipment, and reinforcement of exterior walls as these entail lengthy remodeling time and discomfort or inconvenience during the replacement period. The retrofitting rate of residential buildings in South Korea is 12.7% for buildings not more than 10 years old, 12.5% for those 10 to 19 years old, 20.6% for those 20 to 29 years old, and 44.8% for those over 30 years old [⁴].

The replacement of windows is a method frequently applied to residential buildings, especially apartments because the construction period is relatively short. Reports suggested that buildings facing south-west with clear glazing windows had the highest energy savings of 12.3% in the cooling season [⁵]. Raeissi and Taheri investigated the optimum dimensions of overhangs to increase energy savings in A/C systems in Iran [⁶]. By using the appropriate size, cooling loads was reported to have been reduced by about 12.7% during the summer season. Meanwhile, there was an insignificant increase of 0.6% in winter heating demand.

DSF

Kim et al. analyzed the thermal and daylighting effects of a DSF system with interior and exterior blinds. The results demonstrated that the DSF model could save up to 40% for heating, 2% for cooling, and 5% in total loads compared to the base model that had no blinds or controls [⁷]. Qahtan investigated the effectiveness of a DSF system in controlling heat gain under the direct solar radiation of the west orientation in the tropics in Malaysia. The results confirmed that the DSF system effectively controlled the heat gain as reflected in the difference between outdoor, indoor, and surface temperatures. However, the DSF system increased cooling requirements because of the penetration of tropical direct solar radiation [⁸]. Chan et al. discussed the energy performance of a DSF system applied to a conventional office building in Hong Kong, China. The validated DSF model was then utilized to evaluate energy performance with various configurations, such as window type, position, and layers. The results indicated that a DSF system with a single clear window as an inner side and a double reflective window as an outer side could generate annual cooling energy savings of about 26% compared to a single-skin façade system with a single absorptive window [⁹].

Meanwhile, a variety of DSF models were selected for an office building in Tehran, using dynamic simulation programs. The results showed that the total energy consumption of the box-type DSF model with an external louver shading was reduced by 14.8% compared to the single-layer window model [¹⁰]. Gelesz and Reith evaluated the energy and comfort performance of the DSF system and compared it to double- and triple-glazed, single-skin façades for an office in Central Europe. Results revealed that the DSF system saved 7% in cooling energy compared to the double-glazed model. The DSF system also performed better than the triple-glazed model while maintaining the same level of thermal comfort [¹¹]. Alberto et al. analyzed the DSF system using Design Builder and EnergyPlus, taking advantage of the computational fluid dynamics (CFD) module, to gain better insight into building performance, air flow path, air cavity depth, openings area, and type of glazing. They found that the most critical factor that determines the efficiency of the DSF system was air flow path, while the most efficient geometry was the multi-story DSF system that reduced about 30% of heating, ventilating, and air-conditioning (HVAC) energy demands [¹²].

Simulation software

EnergyPlus simulation program is selected because it is one of the suitable simulation tools to make the DSF system and radiant floor heating system considered in this paper. The EnergyPlus simulation program uses the pressure balance equation to compute the air flow network (AFN) model. Besides, it uses conduction transfer functions (CTFs) to calculate multiple layer conduction heat transfer in radiant floor heating systems.

AFN

An AFN model in the EnergyPlus simulation program was developed using the pressure balance equation, as expressed in Eq. (1) [¹³,¹⁴]. The inlet and outlet of the components connecting the two zones are at relatively different heights and in relatively different positions from the nodes representing the volume of the zones. The analysis of air flow using the components of each zone was based on Bernoulli’s equation [¹⁴].

(1)\mathrm{\Delta }p=\left({{\displaystyle p}}_n+{\displaystyle \frac{\rho {{\displaystyle v}}_n^{\mathrm{2}}}{\mathrm{2}}}\right)-\left({{\displaystyle p}}_m+{\displaystyle \frac{\rho {{\displaystyle v}}_m^{\mathrm{2}}}{\mathrm{2}}}\right)+\rho g({{\displaystyle z}}_n-{{\displaystyle z}}_m)+{{\displaystyle p}}_{\mathrm{w}},

where Dp is the total pressure difference between nodes n and m (Pa); p_n and p_m are the inlet and outlet static pressures (Pa); v_n and v_m are the inlet and outlet air flow velocities (m/s); r is the air density (kg/m³); g is the acceleration due to gravity (9.81 m/s²); z_n and z_m are the inlet and outlet elevation (m); and p_w is the wind surface pressure relative to static pressure in the undisturbed flow (Pa). In Eq. (2), p_s represents the stacking effect. If a wind pressure term is added to the two consecutive zones, i.e., node 1 and node 2, the pressure difference of all successive nodes can be expressed as Eq. (2) and follows traditional flow direction indications. With a positive value, the flow is from node 1 to node 2; in case of a negative value, it is from node 2 to node1 [¹⁴].

(2)\mathrm{\Delta }p={{\displaystyle p}}_{\mathrm{1}}-{{\displaystyle p}}_{\mathrm{2}}-{{\displaystyle p}}_{\mathrm{s}}+{{\displaystyle p}}_{\mathrm{w}},

where p_s is the pressure difference due to density and height differences (Pa).

By using the AFN model which can calculate air flow for multi-zones, the natural ventilation of the DSF system can be analyzed. In previous studies, the EnergyPlus simulation program was utilized to analyze the energy performance of DSF systems, stacking effect, and natural ventilation characteristics of DSF [¹⁵,¹⁶].

Radiant floor heating system

When the development of a radiant floor heating system model is perceived as completely separate from an energy analysis program, it diminishes the capabilities of the simulation itself; hence, the program is unable to enhance or extend adequately [¹⁷]. Heat transfer mechanisms through multilayered slabs and heat sources play a key role in the radiant floor heating system in EnergyPlus. To calculate the surface and zone temperature in given time intervals, EnergyPlus uses the heat balance method of each surface within the conditioned zone by considering the internal heat gain and loss caused by the radiation and convection between interior surfaces [¹⁸]. With the updated zone temperature, the heating zone load is determined on a time horizon, which can be used to determine the heating system response to meet the zone heating requirement. EnergyPlus provides several types of radiant heating systems, such as low and high-temperature radiant systems with two types of flows, i.e., variable flow and constant flow. Typical residential buildings in South Korea are equipped with the low-temperature radiant floor heating system that uses hot fluid circulated through tubes within the floor slab in variable volumes. This paper uses the low-temperature radiant system with the variable flow option. Figure 2 demonstrates the radiant floor heating system applied to the conditioned zone in this paper. The system capacity of the corresponding zone floor is chosen based on the load calculation sizing with the design day. The low-temperature radiant system with the flow option is a combination of mixing valves, a pump, and the radiant heating panel that includes heat source as shown in Fig. 2.

The CTFs is employed to compute multiple layer conduction heat transfer in EnergyPlus. There are two typical methods for calculating CTFs: the Laplace transform method and the state space method. Although both methods are suitable for the CTFs calculation in an accurate manner, the current version of EnergyPlus uses the state space method to compute the CTFs [¹⁴]. With the heat transfer solver, the interior floor surface heat balance can be calculated as

(3)q^{″}[\mathrm{surface}\hspace{0.17em}\mathrm{heat}\hspace{0.17em}\mathrm{balance}]={\displaystyle {\sum }_{r=\mathrm{1}}^M{{\displaystyle X}}_r}{{\displaystyle T}}_{\mathrm{i},t-r+\mathrm{1}}-{\displaystyle {\sum }_{r=\mathrm{1}}^M{{\displaystyle Y}}_r}{{\displaystyle T}}_{\mathrm{o},t-r+\mathrm{1}}+{\displaystyle {\sum }_{r=\mathrm{1}}^s{{\displaystyle F}}_r}{q}_{\mathrm{i},t-r}^{″}+{\displaystyle {\sum }_{r=\mathrm{1}}^M{{\displaystyle W}}_r}{{\displaystyle q}}_{\mathrm{source},t-r+\mathrm{1}},

where q^{″} [surface heat balance] is the interior floor surface heat balance (W/m²), s is the number of order of CTFs; Both M and r are finite numbers defined by order of CTFs; X, Y, F, and W are the conduction heat transfer functions which are constant; T is the temperature as a function of position and time (°C); o is the outside of the building element; i is the inside of the building element; t is the current time step (h); q_{(i,t-r)}^{″} is the heat flux at a certain layer and time (W/m²), which equals -k{\displaystyle \frac{\partial T\left(x\right)}{\partial x}}, x is the position, k is the thermal conductivity (W/(m∙K)); and q_{source,t–r+1} is the internal heat source of each calculation time, which is provided by the radiation system to keep the surface heat balance (W).

The surface heat balance on the left-hand side of Eq. (3) considers several zone energy balance terms, including incident solar energy throughout exterior windows, radiation heat transfer from internal heat sources, such as lighting and electrical equipment, and radiation and convection between the floor surface and the surrounding surfaces and the air within the conditioned zone. In terms of temperature and heat flux calculation in heat transfer analysis, transient one-dimensional heat conduction through a homogeneous layer with constant thermal properties is considered in this calculation. The transient one-dimensional equation is typically coupled with Fourier’s law of conduction at any location and time to temperature.

A heat source term on the right-hand side of Eq. (3) considers heat exchange analysis for a hydronic system. In this calculation, the effectiveness-number of transfer unit (effectiveness-NTU) method is used with several assumptions, such as the constant water tubing length. Using the effectiveness-NTU heat exchanger algorithm, the energy balance rate relationship between the heat source and the water temperature is applied to each calculation step in order to estimate end energy use in the heating system. Hot water circulating in the closed loop in the radiant floor heating system indicates the amount of heat used for heating. Based on the heat balance method which the EnergyPlus simulation program used, heat is supplied to the floor piping system by the amount of heat calculated [¹⁴]. Radiant floor heating systems, unlike the conventional HVAC systems, reach the room set temperature by the surface temperature rather than by the air supplied. Certain floor surface temperatures are required to reach room set temperatures [¹⁴]. More details on load and energy calculation of the radiant heating system can be referred to in [¹⁴].

Experimental validation of the low-temperature radiant heat transfer accuracy in EnergyPlus was conducted in a model residence that has a radiant cooling and heating system by comparing measured data with predicted ones [¹⁹]. The other validation of the radiant system in EnergyPlus was made by comparing experimental measurements and test results using the Building Energy Simulation Test (BESTEST) suite [¹⁸,²⁰].

Passive heating

Windows are controlled open or close to bringing warm air from the DSF system module to the indoor area to achieve passive thermal heating. It is essential to optimally control the window to save as much heating energy as possible. Figure 3 depicts a window opening algorithm for the heating season. Figure 4(a) shows the window opening conditions while Fig. 4(b) the air nodes in EnergyPlus. In the EnergyPlus program, the vertical and horizontal window openings in AFN are used for airflow connected zones in multiple linkages. Horizontal openings in the DSF system are used for surfaces between individual DSF zones to consider the air pressure and temperature differences even though the actual DSF system is one zone. Vertical openings in the DSF system are used for the interior window between the zone and the middle of the DSF system for passive heating. Table 1 lists the input values for the AFN used in this paper. Since the values in Table 1 cannot be obtained from actual experiments, the values presented in the previous studies and the EnergyPlus simulation input and output reference were used [⁷,¹³,¹⁶].

To consider the ventilation between the DSF system and the interior zone through internal windows of the DSF, it was assumed that the interior window could be controlled. If the air temperature of the DSF system is lower than the indoor air temperature, both the outer and inner windows are closed. Discharge coefficient values C_d for the vertical and horizontal opening are set to 0.65 and 0.2, respectively, directly obtained from Ref. [⁷]. The infrared transparent (IRT) surface function in EnergyPlus is used to consider the impacts of radiation, conduction, and convection in the DSF system. A “Full interior and exterior method” is used to consider direct and diffuse solar radiation factors through exterior and interior windows of the DSF system [⁷,¹³].

Thermal characteristics of DSF system

Optical properties, thermal properties, the effect of buffer area, and the convection on each glass surface are important factors considered in the theoretical analysis of the DSF system with multi-layers. Depending on the condition of the external environment, the optical properties that occur in the DSF system include absorption, reflection, and transmission of direct/diffuse/reflective solar radiations [²¹,²²]. The thermal properties of each surface are affected by the interfacial long-wave radiation of the buffer area in the DSF system and the convective heat transfer of the inner and outer surfaces. The effect on air flow can be caused by upper and lower ventilation damper conditions and the temperature in the buffer area. The characteristics of the DSF system can be represented in a two-dimensional form. Figure 5 illustrates the air flow and heat transfer in the DSF system. The optical properties of each surface and the buffer area reflect the passage of direct and diffuse solar radiations through Glass #1 from the outside. The solar radiation that passes through Glass #1 is the direct solar radiation. The process of reflection and absorption is repeated and finally flows into the room through Glass #2. It is necessary to adopt a theoretical analysis method that can mathematically verify the environment of the above physical phenomena. Although simulation modeling technologies can provide valuable results to analyze various physical phenomena, it is necessary to analyze unsteady-state heat conduction, radiation, and convection heat transfer. Moreover, the high operation speed is required to cope with various parameter configurations. Accordingly, complex heat transfer analyses can be conducted by using EnergyPlus on the complex physical phenomena of the DSF system, such as solar radiation and temperatures of indoor, outdoor, and buffer areas [²³].

Simulation model

Proposed DSF system

Figure 6 illustrates the outer and inner layer compositions of the proposed DSF system. The system has operable windows and insulated walls in the outer layer which includes a damper for the outdoor unit, photovoltaic (PV) systems, fixed/operable windows, and the overhang. The inner layer includes the outdoor unit, fixed and operable windows, and shelves. The depth of the DSF system is 700 mm. The DSF system can be used for retrofitting apartment units to improve energy efficiency. It would be interesting to see how much heating energy could be saved from retrofitting the existing balcony with the DSF system.

The radiant heating system which is mainly used in apartment buildings in South Korea consumes gas energy. As previously mentioned, this paper aims to reduce the heating energy of residential buildings through the installation of a DSF system. Accordingly, PV systems that generate electrical energy are excluded. Future studies on annual energy savings will be conducted to explore electricity energy generated by the PV system.

Simulation cases

The construction specifications of a typical existing apartment, i.e., 20 years old, were used to create the details of the case study models. Figure 7 provides a drawing of an old apartment built in December 1999 in Seoul, South Korea. The 2016 Population and Housing Census, an annual report published by Statistics Korea, reported that the average size of high-rise apartment units is 80.4 m² for units built before 1979, 65.6 m² for those built between 1980 and 1989, and 70.0 m² for those constructed between 1990 and 1999 [²⁴]. The floor plan used in this paper is a common high-rise residential building unit with a size of 75 m². The apartment balcony space was initially used as an auxiliary living space, i.e., unconditioned zone, for residents and for other functions, such as ventilation and lighting. Nonetheless, it retained the form of a semi-outdoor space typical of the 1960s when the first apartments were built [²⁵]. In December 2005, balcony expansion was officially permitted. Apartment owners removed the inner balcony windows and extended the area as part of the living room space, thus increasing both the total conditioned space and the apartment resale value [²⁵]. As a result, the previous buffer area of the balcony disappeared, which meant an increase in energy consumption both in summer and in winter. The proposed DSF system is intended to be a replacement for the remaining outer wall/window of the balcony to eventually bring back the thermal buffer area and increase energy savings.

Figure 8 presents the geometry models of the three cases. The Base-Case model is an apartment with an enclosed balcony. Two cases were then developed to determine the effect of different retrofitting methods. The simulation model has three floors and three units per floor. The central unit located in the middle of the second floor was used for analysis. The Case-1 model, i.e., typical remodeling, includes a balcony window replaced with new glazing type, low-emissivity window (low-e window), which is a typical retrofitting method for energy savings. The Case-2 model includes the DSF system, as well as the window opening control, which is the proposed retrofitting method in this paper. Window opening control is not optional but required for the DSF system to save the heating energy consumption of the residential building, and it is already confirmed [²⁶]. That is the reason for not considering the DSF system without window opening control in this paper.

Most apartments have separate heating and cooling systems in South Korea. Space heating is provided by a floor radiant heating system while space cooling is provided by a stand-alone A/C system. Both heating and cooling systems have individual on/off control. As previously mentioned, this paper focuses only on the heating system, considering the heating-dominant climate of Seoul, South Korea. A gas boiler is installed in residential buildings for space heating, in general. The air temperature of the DSF system is compared with the indoor air temperature to control windows for free heating in winter.

Simulation model

A weather file for Seoul, South Korea was used for the simulation. The latitude and longitude of the Seoul are 37.33°N and 126°58′E. The city of Seoul has four seasons with hot/humid summer and cold/dry winter. Outdoor air temperature ranges from ‒11.8°C to 32.7°C. The heating energy analysis focused on the heating season from November to March during which the outdoor air temperature ranged from - 11.8°C 16.6°C.

Figure 9 presents a simulation model of Case-2, which is the proposed retrofitting method. This simulation model was validated in Ref. [²⁷].

Tables 2 and 3 outline the properties of building materials and construction sets of the building and the DSF system [²⁷,²⁸,²⁹]. Case apartment constructions and materials were all based on the drawing of the apartment building in Fig. 7.

Table 4 shows the properties of the windows such as U-value, solar heat gain coefficient (SHGC), and visible transmittance (VT). For Base-Case, all windows installed in the apartment building were double glazed with air. Double low-e windows with argon gas were used for Case-1 (i.e., the typical retrofitting method that entails replacement of the balcony window). Case-2 had exterior windows replaced with the DSF system that consisted of double low-e windows with argon gas. Interior windows for the DSF system were double glazed with air. The specifications of the windows were obtained from a window manufacturing company. The 24 mm double-glazed window consisted of 6 mm clear glass, 12 mm air gap, and 6 mm clear glass. The 22 mm double low-e glazed window consisted of 5 mm clear glass, 12 mm argon gas gap, and 5 mm low-e glass.

Internal loads

Table 5 shows internal heat gain and heating set points. These values were based on the Study on Heating and Cooling Load Standard per Area for Apartments in 2017, published by the Korea Research Institute of Mechanical Facilities Industry (KRIMFI) [³⁰]. Heating and internal heat gain schedules were based on the report Energy Technology Transfer and Diffusion 2007 [³¹].

Yoon et al. calculated the infiltration value of an apartment building installed with a DSF system by using the infiltration value of apartment buildings and effective leakage area values in the ASHRAE Handbook of Fundamentals [²⁷,³²]. The air change per hour (ACH) of Base-Case and Case-1 was 2.82, while the ACH of Case-2 was 2.15 [²⁷].

Gas boiler heating system

For the heating system, the low temperature radiant: Variable flow function was used in EnergyPlus. This function is a water-based radiant system. Energy is either supplied or removed through the surface, such as walls or floor [¹⁶]. A gas boiler which had a thermal efficiency of 80% and produced hot water with a temperature of 60°C was used as a radiant floor heating system. The inside diameter of hydronic tubing for the hot water coil of the radiant floor was 16 mm. The spacing between coils was 200mm [³³]. The length of each heating coil located inside the floor was calculated by considering each zone size and the spacing of coils. As tabulated in Table 5, the heating setpoint temperature was 21°C during the occupied hours and 13°C during the unoccupied hours. The set temperature of hot water was fixed at 60°C [³⁴]. Figure 10 illustrates the typical efficiency and coefficients for non-electric and non-condensing boilers used in this paper [¹⁶]. Equation (4) is the boiler efficiency equation [¹⁴]. Boiler efficiency ranged from 68.4% at a PLR of 10% to 80.9% at a PLR of 100%. The a, b, and c are the PLR coefficients.

(4)\mathrm{B}\mathrm{oiler\; Efficiency\; =}a+b(\mathrm{P}\mathrm{L}\mathrm{R})+c{(\mathrm{P}\mathrm{L}\mathrm{R})}^{\mathrm{2}}+d{(\mathrm{P}\mathrm{L}\mathrm{R})}^{\mathrm{3}}.

Analysis

Gas consumption on January 24

January 24 is the date selected as the winter representative day to analyze hourly boiler gas consumption for space heating. January 24 has the typical heating loads during the winter, because of this, January 24 also was used for the winter representative day in Ref. [³⁵]. The boiler size was 18 kW for all cases. Figure 11 presents the PLR of the boiler for each case. The distribution of PLR during boiler operation was 0%–86% in Base-Case, 0%–84% for Case-1, and 0%–65% for Case-2. This distribution meant that Case-2 had less heating load than the other cases. Notably, in all three cases, the boiler was repeatedly turned on and off during the boiler operation time, in which the zone radiant HVAC heating rate changed accordingly (see Fig. 12). The radiant HVAC heating rate is the heating rate of the low-temperature radiant system. The heating rate is calculated by considering the zone conditions and the control of the low-temperature radiant system. The radiant floor heating system heats the floor via radiant heat instead of heating the indoor air directly. The room temperature was set at 21°C. The boiler was turned on when room temperature reached 20.5°C and turned off when it reached 21.5°C.

Figure 13 displays the temperature changes in indoor air and radiant floor surface during the day. The floor was heated by the hot water coming from the boiler. The floor surface temperature was increased so that the indoor air temperature reached the heating setpoint temperatures. Accordingly, the floor surface was heated up to 34°C by the boiler supply water that had a temperature of 60°C. This paper compared floor surface temperature with a previous experimental study to confirm the range of the floor surface temperature. Cho et al., conducted an experimental study on the application of low-temperature radiant floor heating system which this paper used [³⁶]. They used hot water of 50°C for radiant floor heating system, and maximum floor surface temperature was 37.8°C which meant the surface floor temperature in this paper was in a reasonable temperature range. In the unoccupied hours between 09:00 and 19:00, floor surface temperature gradually decreased since there was no hot water supply. Meanwhile, indoor air temperature gradually increased after 14:00 due to solar radiation coming into space through the glazing during the daytime.

Figure 14 compares boiler gas consumption in three cases. The gas consumption rate reached up to 24 kWh for Base-Case, 23 kWh for Case-1, and 18 kWh for Case-2. Case-2 had the least gas consumption rate. Between19:00 and 20:00, no heating energy was used in Case-2 due to the heat generated in the DSF area during the day.

All three cases had the same gas consumption pattern as the PLR pattern. The boiler efficiency used in this paper can explain why gas consumption pattern and PLR pattern are similar. There was a difference of about 3% in boiler efficiency from 40% to 100% for each PLR sections described in Fig. 10, which is not a big difference. Besides, in Fig. 11, it can be seen that the boilers in the three cases operate over 40% of the PLR except for a few hours. As a result, the PLR pattern is similar to the gas consumption pattern.

As shown in Fig. 7, 90% of the south wall is covered with windows. A large amount of heat loss is expected through the window area during the wintertime, thus increasing the heating load. Figure 15 illustrates the window heat loss of each case. The respective maximum and minimum heat losses through the window area were 1.11 and 0.62 kWh for Base-Case, 0.64 and 0.24 for Case-1, and 0.40 and 0.07 kWh for Case-2. Notably, window heat loss decreased significantly in Case-2 due to the improved U-value of the window, the air cavity of the DSF system, and reduced infiltration.

Figure 16 shows that the outdoor air temperature ranged from ‒6.5°C to ‒3.2°C while the DSF inside air temperature ranged from 15.0°C to 24.4°C on January 24. Case-2 had a thermal buffer from the DSF system, which significantly reduced delta T in the heat loss calculation, ranging from 21.5°C to a maximum of 37.6°C. This thermal buffer resulted in substantial heating energy savings in winter.

Total heating energy consumption and savings

Figure 17 presents the daily window heat loss for each case in the winter months, i.e., November to March. The daily heat loss value excluded heat loss occurred from 09:00 until 19:00 since the boiler did not typically operate during the daytime when there were no occupants. The window heat loss ranged from 2.79 to 18.30 kWh/d for Base-Case, 0.46 to 10.23 kWh/d for Case-1, and 0.66 to 6.24 kWh/d for Case-2. As expected, window heat loss of Case-1 was 45% less than that of Base-Case due to high-performance windows replacement. The window heat loss of Case-2 was more substantial than that of Base-Case: 66% less compared to Base-Case, which was due to improved window performance, reduced infiltration, and the thermal buffer area that the DSF system created. This considerable heat loss reduction decreased boiler gas consumption in Case-2.

Figure 18 provides the annual heat loss summary for the three main heat loss areas, i.e., window, wall, and infiltration, of the three cases during the heating season. Case-2 had the least heat loss of 2.04 MWh/a in the window area, which is about 39% less than the heat loss in Base-Case. There was no significant difference between the heat loss through the wall in the three cases, since minor changes in the wall areas. Notably, 90% of the south wall is covered by windows, resulting in minimal heat loss differences. Heat loss through infiltration had a significant impact on the heating energy in winter (see Fig. 18). The infiltration heat loss was nearly four times higher than the window heat loss and nine times higher than the wall heat loss. Case-2 had a significant heat loss reduction in infiltration, i.e., 24% reduction, compared to Base-Case. The ACHs were about 2.6–2.8 in both Base-Case and Case-1 and about 1.9–2.1 in Case-2 [²⁷]. Case-2 had a DSF system that significantly prevented infiltration due to its double-layer construction and improved air tightness.

Figure 19 shows the monthly gas consumption of the three cases during the heating months. As expected, all cases had the highest boiler gas consumption in January and the lowest in March. The gas consumption was about 5.1 MWh/month in Base-Case, 4.9 MWh/month in Case-1, and 3.3 MWh/month in Case-2 in January. Annual heating energy savings in Case-2 were 38.8% compared to Base-Case and 35.2% compared to Case-1.

Discussion

This paper compared heating energy consumption by the installation of the DSF system using a validated EnergyPlus simulation model. Three case models were developed, Base-Case, i.e., existing apartment, Case-1, i.e., typical remodeling/retrofitting with high-performance double-glazed windows, and Case-2, i.e., proposed remodeling/retrofitting with DSF system.

After the installation of the DSF system, the boiler works in the low PLR section. The less PLR of the boiler means that the boiler efficiency will also be decreased depending on the boiler efficiency.

If boiler capacity decreases, the boiler works in the high PLR section, which means that the boiler efficiency will be higher, and the heating energy consumption would then be further reduced. It shows that there is an opportunity to reduce the more heating energy consumption by the installation of the DSF system, which means that the heating energy consumption can be reduced by not only the installation of the DSF system itself but also a smaller boiler capacity than before.

Heat losses are one of the main factors to explain the heating energy saving by the installation of the DSF system because of the internal heat gains, internal heat gain schedules, type of heating system, and heating set-point are the same in all cases. Window and wall heat losses can be reduced because the DSF system works as a thermal buffer, and south windows and walls are not adjacent outside directly but the DSF system. Due to the high-insulated wall, high-performance window, and the captured air of the DSF system, the range of the air temperature inside the DSF is from 21.5°C to 37.6°C which is higher than the outdoor temperature (‒11.8°C to 16.6°C).

Because of the higher temperature than the outside air temperature, the temperature differences between indoor and outdoor air temperatures are reduced by the installation of the DSF system. Small temperature differences have less effect on heat transfer, which accounts for the reduction in window and wall heat losses. Besides, the low U-value window also reduces window heat loss.

The old apartment building which is the target building of this paper has poor-insulated walls, low-performance windows, and high infiltration values because of poor construction code at that time it was built. By the installation of the DSF system, the south side of the apartment unit has an additional space and a double-layered construction, with 700 mm air, high-insulated exterior wall, and high-performance exterior windows, which not only decreases the window and wall heat loss but also decreases the infiltration heat loss.

By the installation of the DSF system, the south side surface of the building has an additional space and a sturdy construction, which reduced the infiltration value of the target apartment unit. Therefore, the infiltration heat loss is reduced.

These analyses indicate that 38.8% of gas energy consumption for space heating can be saved. In addition, if a DSF system is installed in a new building, not for retrofitting, the boiler capacity can be reduced by about 30% due to the low PLR of the boiler by reduced the heating load after installation of the DSF system.

Conclusions

This paper proposed a DSF system as a balcony replacement for 20- to 30-year-old apartments in South Korea. It explained how heating energy consumption can be saved by the installation of the DSF system.

The DSF system will be installed in a real building in the near future to calibrate the simulation model and confirm the energy performance of the DSF system. One of the disadvantages of the DSF system is the payback period. The estimated cost of the DSF system, including production, transportation, installation, and labor cost, is about $24500, and the payback period is more than 30 years. Notably, this estimation only considers energy saving costs. Attempts will consequently be made to find subsidies and to minimize the production and transportation costs to reduce the payback period. In the future, a cost analysis will be conducted for providing suggestions to achieve a reasonable payback period.

References

[1] Seo B M, Lee K H. Detailed analysis on part load ratio characteristics and cooling energy saving of chiller staging in an office building. Energy and Building, 2016, 119: 309–322

[2] Oliver T, Lew D, Redlinger R, Prijyanonda C. Global energy efficiency and renewable energy policy options and initiatives. Energy for Sustainable Development, 2001, 5(2): 15–25

[3] Song D S, Choi Y J. Effect of building regulation on energy consumption in residential buildings in Korea. Renewable & Sustainable Energy Reviews, 2012, 16(1): 1074–1081

[4] Lim J H, Han H G, Kang D R. Analysis on green remodeling status, satisfaction, and energy saving effect–focusing on window replacing of residential buildings. Journal of the Architectural Institute of Korea Structure & Construction, 2017, 33: 75–80

[5] Chan A L S, Chow T T. Investigation on energy performance and energy payback period of application of balcony for residential apartment in Hong Kong. Energy and Building, 2010, 42(12): 2400–2405

[6] Raeissi S, Taheri M. Optimum overhang dimensions for energy saving. Building and Environment, 1998, 33(5): 293–302

[7] Kim D S, Cox S J, Cho H, Yoon J. Comparative investigation on building energy performance of double skin façade (DSF) with interior or exterior slat blinds. Journal of Building Engineering, 2018, 20: 411–423

[8] Qahtan A M. Thermal performance of a double-skin façade exposed to direct solar radiation in the tropical climate of Malaysia: a case study. Case Studies in Thermal Engineering, 2019, 14: 100419

[9] Chan A L S, Chow T T, Fong K F, Lin Z. Investigation on energy performance of double skin façade in Hong Kong. Energy and Building, 2009, 41(11): 1135–1142

[10] Zomorodian Z S, Tahsildoost M. Energy and carbon analysis of double skin façades in the hot and dry climate. Journal of Cleaner Production, 2018, 197: 85–96

[11] Gelesz A, Reith A. Climate-based performance evaluation of double skin facades by building energy modelling in Central Europe. Energy Procedia, 2015, 78: 555–560

[12] Alberto A, Ramos N M M, Almeida R M S F. Parametric study of double-skin facades performance in mild climate countries. Journal of Building Engineering, 2017, 12: 87–98

[13] Dutton S, Shao L, Riffat S. Validation and parametric analysis of EnergyPlus: air flow network model using contam. In: 3rd National Conference of International Building Performance Simulation Association (IBPSA-USA) (Simbuild 2008), Berkeley, California, USA, 2008, 124–131

[14] EnergyPlus engineering reference. The reference to Energyplus calculations, version 9.0. 2019–08–15, available at the website of energyplus

[15] Winkelmann F C. Modelling windows in EnergyPlus. In: 3rd National Conference of International Building Performance Simulation Association (IBPSA) (Building Simulation 2001), Rio de Janeiro, Brazil, 2011, 457–464

[16] EnergyPlus input output reference. The encyclopedic reference to Energyplus input and output, version 9.0. 2019–08–15, available at the website of energyplus

[17] Strand R K, Pedersen C O. Modeling radiant systems in an integrated heat balance based energy simulation program. ASHRAE Transactions, 2002, 108: 979–987

[18] Raftery P, Lee K H, Webster T, Bauman F. Performance analysis of an integrated UFAD and radiant hydronic slab system. Applied Energy, 2011, 90(1): 250–257

[19] Chantrasrisalai C, Ghatti V, Fisher D E, Scheatzle D G. Experimental validation of the EnergyPlus low-temperature radiant simulation. ASHRAE Transactions, 2003, 109(2): 614–623

[20] Henninger R H, Witte M J, Crawley D B. Analytical and comparative testing of EnergyPlus using IEA HVAC BESTEST E100–E200 test suite. Energy and Building, 2004, 36(8): 855–863

[21] Gratia E, Herde A D. Optimal operation of a south double-skin façade. Energy and Building, 2004, 36(1): 41–60

[22] Gratia E, Herde A D. Are energy consumptions decreased with the addition of a double skin? Energy and Building, 2007, 39(5): 605–619

[23] Kim D W, Park C S. Energy performance assessment of a double skin façade with different control strategies. Journal of Architectural Institute of Korea, 2010, 26: 389–398

[24] Statistics Korea. 2016 population and housing census. 2019–03–25, available at the website of kostat

[25] Kim S Y, Oh C O. A study of current status and residents’ needs of balcony extension in Korean apartment housing. Journal of the Korean Institute of Interior Design, 2012, 21: 152–162

[26] Yoon Y B, Koh B B, Cho S. An analysis of energy efficiency of a smart envelope package in residential buildings. In: Proceeding of ARCC-EAAE 2018 International Conference, Philadelphia, PA, USA, 2018, 2: 27–35

[27] Yoon Y B, Seo B M, Koh B B, Cho S. Performance analysis of a double-skin façade system installed at different floor levels of high-rise apartment building. Journal of Building Engineering, 2019, 26: 100900

[28] Ministry of Land, Infrastructure and Transport (MOLIT), Korea Energy Agency (KEA). Guidebook of Building Energy Conservation Design Standard. 2017

[29] U.S. Department of Energy (DOE). EnergyPlus version 9.0, ASHRAE 2005 HOF Materials.id. 2018

[30] Korea Research Institute of Mechanical Facilities Industry (KRIMFI). Study on heating and cooling load standard per area for apartments in 2017. 2019–08–07, available at the website of krimfi

[31] Korea Institute of Energy Research (KIER). Energy Technology Transfer and Diffusion 2007 Report. 2007

[32] American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE). ASHRAE Handbook: Fundamentals. 2009

[33] Lee E J, Lee D Y, Hong H K, Kim K Y. Measurement and simulation of heating energy for apartments with district heating. Korean Journal of Air-Conditioning and Refrigeration Engineering, 2014, 26: 572–578

[34] Yu B H, Seo B M, Moon J W, Lee K H. Analysis of the part load ratio characteristics and gas energy consumption of a hot water boiler in a residential building under Korean climatic conditions. Korean Journal of Air-Conditioning and Refrigeration Engineering, 2015, 27: 455–462

[35] Lee D Y, Seo B M, Hong S H, Choi J M, Lee K H. Part load ratio characteristics and energy saving performance of standing column well geothermal heat pump system assisted with storage tank in an apartment. Energy, 2019, 174: 1060–1078

[36] Cho J, Park B, Lim T. Experimental and numerical study on the application of low-temperature radiant floor heating system with capillary tube: thermal performance analysis. Applied Thermal Engineering, 2019, 163: 114360

RIGHTS & PERMISSIONS

Higher Education Press