Study on Operating Strategies Coupling Floor-Cooling and Cold Storage in Thermal Active System

Abstract

To explore the optimized operating strategies coupling floor cooling with cold storage for thermal active systems (TABSs), effects of operating time on indoor thermal environment, cold storage capacity, energy use and running cost were studied. Simulations were conducted based on an actual office building equipped with floor cooling. To take full advantage of the TABS and off-peak electricity, four operating cases with nighttime floor cold storage were proposed, namely C₁ (2:00–8:00), which operated only during the off-peak hours, C₂ (2:00–10:00), C₃ (2:00–12:00), and C₄ (2:00–14:00), which operated during the off-peak and flat hours. A simulation case of C₀ operating during daytime (7:00–17:00) was also proposed. Simulation results show that the C₁ and C₂ conditions with shorter operating hours result in higher indoor temperatures, which cannot ensure indoor thermal comfort. The PMV index in C₃ and C₄ conditions can be kept between −1 and 1, which meets the thermal comfort demand of Grade II. Considering that the operating duration of C₃ is the same as the occupied hours, the cold storage capacity, cooling loss, cooling supply and release process, etc., of this case are further analyzed based on data of a typical day. The floor and ceiling slabs store most of the cooling energy (72.7%) during the night; inner walls also store part of the cooling energy (23.3%) and cooling loss during cold storage accounts for approximately 3.1%. During working hours, the cooling energy released is lower than the cooling load, which makes indoor temperatures increase continuously. Compared with case C₀, case C₃ has same power use while saving 2.8% of running costs. Case C₄ provides a higher level of thermal comfort, while saving 0.9% of costs with a 1.5% increment in electricity use. This study provides detailed data about cold storage strategies coupling with floor cooling in TABS, which can be used to save running cost.

1. Introduction

Radiant cooling systems typically utilize radiant floors, ceilings, side walls, and radiant cooling panels as terminal units, offering advantages such as energy efficiency and high comfort levels [1,2]. In the cooling structure of a radiant floor, ceiling, and sidewall, heat exchange coils are usually embedded within. When chilled water circulates through these coils, the internal and surface temperatures of the structures decrease, thereby providing cooling to the interior air [3]. Such systems are also known as Thermo-active Building Systems (TABSs), a term that refers to how the heat exchange coils “activate” the thermal mass of the building structure, enabling it to dynamically store and release a certain amount of cooling or heating, thereby reflecting the thermal response characteristics of the structure [4,5,6]. For buildings such as offices that operate intermittently, the thermal capacity of the building envelope can be used to store part of the cooling energy during off-peak hours at night. This reduces or delays the operating hours of chillers during daytime, which usually corresponds to peak hours. Such changes in operating hours help to reduce operating costs [7,8].

Currently, studies on TABSs have been extensively conducted. Most studies focus on terminal unit configurations, operating control strategies, impacts on the thermal environment, and integrated applications with natural cooling sources [9,10,11,12,13,14,15,16,17,18,19,20,21], which have been listed in

Table 1. Comparison of main focus in previous studies and this study.

Sources		Primary Focus
References	[9,10,11]	Configuration of terminal units in TABS
	[12,13]	Possibility of thermal storage by slabs in TABS and primary analysis
	[14,15]	Different water temperature control strategies applied to TABS
	[16,17]	Cooling capacity in the design and operation of cooling surfaces in TABS
	[18,19]	Different cooling sources applicable in TABS
	[20,21]	Impacts on indoor temperature and energy consumption of TABS
This study		Floor cold storage operating strategies considering different operating time spans based on local peak and off-peak electricity pricing policy
		Effects of strategies on indoor thermal environment, thermal comfort, energy use, and running cost
		Quantitative analysis of cold storage capacity and cooling loss during cold storage

. In reference [11], simulations of the TABS under the thermal storage mode were analyzed. The results indicated that this approach can maintain relatively stable indoor temperatures during the day, and the total cooling capacity of the system does not differ significantly from the normal cooling operations. Meanwhile, the daytime peak cooling load was found to be reduced by approximately 25%. Floor slab cooling/heating storage was also investigated in reference [1], which confirmed the possibility of thermal storage by floor structures. Recent studies proposed the use of phase change materials (PCMs) in building structures to increase the thermal capacity in TABSs, which was beneficial for storing more cooling energy in these structures [12,13]. The above studies indicate that the use of TABSs for cooling storage is feasible for engineering applications. According to references [22,23,24], the idea of cold storage in TABSs has been applied in a few low energy consumption buildings (LECBs). In these office buildings, underground water, ground heat exchangers (GHEs), night natural ventilation, etc., were used as the cooling source. The measured data showed that combining the application of cold storage in concrete slabs during the night decreased the energy cost further. These studies showed that utilizing the cooling slabs to store cooling energy possessed a few advantages, and can be achieved by simply adjusting the operating hours of the cooling system.

In the cold storage operation mode of TABSs, there are still some questions to be explored. For example, the optimal operation time of TABSs based on peak and off-peak electricity price and its effect on the thermal environment and running costs have not been studied systematically. Additionally, the actual cooling energy stored in the structure and the cooling loss during cold storage are not quite clear. To give a better understanding and promote the application of this technique, studies concentrating on these points should be conducted.

In this paper, the above issues are addressed based on an actual office building using radiant floor cooling [23]. Referring to the local peak and off-peak electricity pricing policy, floor cold storage operating strategies considering different operating time spans will be simulated. The effects of strategies on indoor thermal environment, thermal comfort, energy use, and operational economics are to be compared. Furthermore, a quantitative analysis of the floor’s cold storage capacity and cooling loss during cold storage will be conducted. The results will provide data about the cold storage characteristics and guidance for the practical application of this technology.

To clearly highlight the distinctions between this study and previous research, the main contributions of this paper are also listed in Table 1.

2. Introduction of Building Energy System

This office building is located in Jinan, Shandong Province, China. It has five floors, with a total height of 20.70 m and total floor area of 5450 m², of which 3815 m² are air-conditioned [1,23]. The building is designed as a LECB, and faces north–south. GHEs are used to provide cooling for the radiant floor system. A ground-source heat pump (GSHP) serves as the cooling source for the fresh air system during summer. In summer, a cooling tower is used to dissipate heat from the GSHP [23]. A thermal storage tank was equipped to store the full cooling load of the fresh air system during off-peak electricity hours. The fresh air system operates during occupied hours from 8:00 AM to 6:00 PM, which handles the entire latent heat load and part of the sensible heat load. The GHEs supply chilled water at approximately 18 °C. In the building, the floor system operates 24 h on workdays to provide sufficient cooling energy and assure the indoor thermal environment. For winter conditions, the GSHP is the heating source both for the fresh air and floor systems, which absorbs heat from GHEs. In this study, only the cooling season was studied.

According to early field measurements, the electricity consumption per unit area during cooling season was 5.16 kWh/m². The GSHP and floor circuit pumps accounted for the largest share of electricity use, with the GSHP accounting for approximately 61.8% and floor circuit pumps accounting for approximately 19.5%. Overall, the air conditioning system in the building achieves low-cost operation in summer by taking full advantage of natural cooling with GHEs, and water cold storage for the GSHP. The floor system operated 24 h, allowing some cooling energy to be stored in the floor during the night. However, this also increased the power used by the circulation pumps. If the cooling floor can be operated only during off-peak time or off-peak combined with flat time, the power use by pumps and the running cost can be reduced further. The possibility of such changes in operating time and the effect on the indoor thermal comfort should be verified.

Meanwhile, the peak and off-peak electricity policy has been continuously adjusted in response to the increased proportion of renewable energy generation in the power grids. The local government releases different electricity tariff polices each year. In the year of 2023, when this study was undertaken, there were five tariff levels corresponding to deep off-peak, off-peak, flat, peak, and sharp-peak time spans. The time periods and prices for the five levels are shown in

Table 2. Peak and off-peak periods in Shandong Province for 2023 [24].

At Any Time	Winter: December–January	Spring: February–May	Summer: June–August	Fall: September–November
0:00~1:00	flat	flat	flat	flat
1:00~2:00	flat	flat	flat	flat
2:00~3:00	flat	flat	off-peak	flat
3:00~4:00	flat	flat	off-peak	flat
4:00~5:00	flat	flat	off-peak	flat
5:00~6:00	flat	flat	off-peak	flat
6:00~7:00	flat	flat	off-peak	flat
7:00~8:00	flat	flat	off-peak	flat
8:00~9:00	flat	flat	flat	flat
9:00~10:00	flat	flat	flat	flat
10:00~11:00	off-peak	off-peak	flat	off-peak
11:00~12:00	off-peak	deep off-peak	flat	deep off-peak
12:00~13:00	deep off-peak	deep off-peak	flat	deep off-peak
13:00~14:00	deep off-peak	deep off-peak	flat	deep off-peak
14:00~15:00	off-peak	off-peak	flat	off-peak
15:00~16:00	off-peak	flat	flat	flat
16:00~17:00	sharp-peak	flat	peak	peak
17:00~18:00	sharp-peak	peak	peak	sharp-peak
18:00~19:00	sharp-peak	sharp-peak	sharp-peak	sharp-peak
19:00~20:00	peak	sharp-peak	sharp-peak	peak
20:00~21:00	peak	peak	sharp-peak	peak
21:00~22:00	peak	peak	sharp-peak	flat
22:00~23:00	flat	flat	flat	flat
23:00~24:00	flat	flat	flat	flat

. During summer, there were only four levels, i.e., off-peak period (2:00 to 8:00), peak period (16:00 to 18:00), sharp-peak period (18:00 to 22:00), flat period (rest of the time) [24].

3. Simulation Model

3.1. Simulation Model Development

The building model was established with SketchUp, which is shown in Figure 1. TRNSYS 18 software was used to simulate the building energy systems. The model consisted of building modules and system modules. The thermal settings for the building modules were based on the actual structures [25]. The exterior walls are composed, from the inside out, of mixed mortar (20 mm), aerated concrete (300 mm), cement mortar (20 mm), and extruded polystyrene insulation (25 mm). the interior walls primarily consist of brick walls (200 mm) and cement mortar (20 mm on each side). The roof comprises reinforced concrete (200 mm), cement mortar (20 mm), extruded polystyrene boards (80 mm), cement mortar (20 mm), and cement tiles (20 mm). The floor is based on a reinforced concrete slab (100 mm), with an insulation layer (25 mm), heat exchange piping, a cement mortar leveling layer (30 mm), and a decorating marble layer (20 mm) laid on top; exterior windows use a Low-E double-glazed structure. The thermal parameters of the envelope materials are shown in

Table 3. Thermal properties of envelope materials [23].

Materials	Thickness (mm)	Thermal Conductivity (W/(m·K))	Heat Capacity (kJ/(kg·K))	Density (kg/m3)
Extruded polystyrene board	25	0.03	1.38	35
Cement mortar	20	0.93	1.05	1800
Aerated concrete	300	0.93	1.05	700
Mixed mortar	38	0.93	1.05	1700
Brick wall	200	1.10	1.05	1900
Decorating marble layer	20	3.49	0.92	2800
Reinforced concrete slab	100	1.74	0.92	2500
Cement tile	20	0.73	1.05	1400
Low-E double-glazed window	5 + 12 + 5	1.60	2.45	2000

.

The cooling pipes embedded in the floors were laid on top of the insulation layer. High-temperature resistant polyethylene pipes were used, with a thermal conductivity of 0.4 W/(m·K). The inside pipe diameter is 20 mm with a wall thickness of 2 mm. The pipe spacing in the floor is 250 mm.

During cooling, the indoor air temperature is set to 26 °C with a relative humidity of 60%, and the enthalpy of indoor air is 59 kJ/kg. In the simulation, meteorological parameters of typical years were used. Under the design condition, the summer outdoor dry-bulb temperature is 34.8 °C, the outdoor wet-bulb temperature is 26.7 °C, and the outdoor air enthalpy is 84.4 kJ/kg. Indoor heat sources were set according to reference [23], including a lighting power of 11 W/m², equipment power of 20 W/m², and heat generation by occupants of 150 W/person. The occupancy density is 4 m²/person and the supplied fresh air volume is 30 m³/h per person. The total fresh air volume is 6800 m³/h. Based on the latent load, the supply air moisture is determined to be 9.5 g/kg, corresponding to a supply temperature of 20 °C and an enthalpy of 44.6 kJ/kg. The fresh air also shares part of the sensible heat load. The cooling season usually ranges from June 15 to September 15 in the local city.

Based on field investigation reported in Reference [23], GHEs provided cooling to the floor system in the cooling season, and the fresh air system operated according to occupied hours. Though the GHEs also supply all the heating energy during winter, this paper focuses on the summer condition. Therefore, in the model, system facilities and operating time were set according to summer conditions. The GSHP provides cooling only for the fresh air system, which operates during off-peak hours (2:00–8:00) to store cooling energy for the fresh air system. The GSHP dissipates heat via the cooling tower. The rated cooling capacity of the GSHP is 119 kW with a rated input power of 23.4 kW. The volume of the cold storage water tank is 163 m³.

In the TRNSYS model, the floor circuit includes GHEs, circulating pumps, and radiant floor terminals. For the floor terminal, an active Layer is defined in TRNBuild and placed within the floor concrete. The fresh air system comprises the GSHP, cooling tower, water tank, air handling unit, pumps and fans. The energy model is shown in Figure 2.

The rated power of equipment in the system was set in accordance with Reference [23], which included the GSHP, pump for each circulating loop, fresh air fan, and cooling tower fan. The flow rate of pumps and air volume of fans were set as constant.

3.2. Model Validation

To verify the building and energy model, a simulation case with the same operating condition as the field investigation was set [23]. The measured data from a typical room located on the third floor of the building with an area of 32 m² were selected from the literature. The maximum sensible cooling load for this typical room is 1390 W, and the latent cooling load is 452 W. The fresh air system handles the entire latent load and 355.5 W of sensible cooling load, while the radiant floor handles the remaining cooling load. The simulated data of indoor air temperatures, and supply and return water temperatures of the GHEs in the week of 1–8 July were compared with the measured data in Reference [23], as shown in Figure 3.

The measured indoor air temperature ranged from 22 to 26.5 °C, and the simulated temperature was in the range of 22.5 to 26.5 °C. The variation in measured and simulated indoor temperature was generally consistent, with an average deviation of 0.5 °C. The root mean square error (RMSE) was 0.39 °C. The deviation can be caused by a few factors, such as the fact that the number of occupants fluctuated during the field test, whereas in the simulation a fixed value was assumed. Furthermore, the simulation utilized data from a typical meteorological year, which differs somewhat from the actual parameters during the field test. Overall, the difference between the simulated and measured values is quite small, indicating good predictive accuracy of the model.

The simulated supply and return water temperatures of the GHEs were compared with the measured data, which is shown in Figure 4. The measured water temperature returning to GHEs ranged from 19.0 °C to 20.5 °C, and the supplied water temperature from the GHEs ranged from 18.0 °C to 19.0 °C. In the simulation, the corresponding water temperature ranged from 19.0 °C to 21.0 °C, and 18.0 °C to 19.0 °C, respectively. The mean difference in water temperature returning to GHEs between the measured and simulated values was 0.2 °C, and the mean difference was 0.1 °C for the water temperature supplied from the GHEs. The RMSE was 0.24 °C and 0.15 °C, respectively. Therefore, the model can predict the water temperatures of GHEs accurately.

The measured total power use by the air conditioning system in the cooling season was 19,703.04 kWh, and the power use per unit of air-conditioned area was 5.16 kWh/m². In the simulation, the total power use was 21,073 kWh for the cooling season, and the unit power use was 5.52 kWh/m². The simulated electricity use is slightly higher than the measured value. In the simulation, the power of pumps and fans was set according to their rated values, while the actual power cannot be exactly the same. The overall deviation of power use is approximately 6.5%, which indicates that the prediction of energy use is quite accurate.

In summary, the simulation model is considered to be accurate in predicting the indoor thermal environment and energy use of the systems.

3.3. Setting Up Simulation Conditions

To save more operating fees, it is necessary to avoid operating during peak and sharp-peak periods and make full use of the off-peak and flat time. Meanwhile, thermal comfort during office hours should be assured. Based on the current peak and off-peak electricity policy shown in Table 2, four operating strategies for the floor-cooling system are proposed. These strategies include one that operates exclusively during off-peak hours (2:00–8:00), and three strategies combining off-peak and flat hours during the morning (2:00–10:00, 2:00–12:00, and 2:00–14:00). Considering that normal office hours are 8:00–18:00, a comparative case based on occupied hours is also compared, in which the start and stop time are both set an hour earlier (7:00–17:00) due to the thermal inertia of the floor. The operation of the fresh air system and the GSHP is identical across all simulated cases. The air conditioning system operates from Monday to Friday during occupied hours, and is shut down on weekends.

The five simulated cases are listed in

Table 4. Description of simulation cases.

Operating Conditions	Actual Operating Conditions [22]	Simulation Cases
		Operating During Occupied Hours	Operating Only During Off-Peak Hours	Operating during Off-Peak and 2 h of Flat Time	Operating During Off-Peak and 4 h of Flat Time	Operating During Off-Peak and 6 h of Flat Time
Case names	T0	C0	C1	C2	C3	C4
Operating time	0:00~24:00	7:00~17:00	2:00~8:00	2:00~10:00	2:00~12:00	2:00~14:00
Operating duration	24 h	10 h	6 h	8 h	10 h	12 h
Analyzed data	Energy use, operating costs	Thermal environment parameters, energy consumption, operating costs

, which are named as C₀, C₁, C₂, C₃ and C₄. The actual operating condition is also shown in the table, which is named as T₀. The simulated cases were set in the TRNSYS model separately, and the simulated data of room temperatures, floor-cooling capacity, cooling loss, operating energy use, and costs were analyzed.

4. Simulation Results and Analysis

The simulation period covered the entire cooling season. The week with the highest outdoor temperatures (6–12 August) was selected as a representative week for data analysis, and the midweek day (Wednesday, August 8) was chosen as a representative day. The outdoor temperature and humidity parameters for the typical week are shown in Figure 5. The outdoor air temperature ranged from 23 to 36.5 °C during the week, staying between 23 and 27 °C at night. The outdoor relative humidity ranged from 44.0% to 98%.

Preliminary analysis showed that the impact of different control strategies on indoor thermal environmental parameters across different zones is quite similar. In later analysis, data from a representative room were used for comparison regarding indoor thermal environment parameters, cold storage and cooling energy supply, etc. The selected room is the same one employed earlier in the model validation section, which is located on the third floor.

4.1. Typical Weekly Indoor Temperature Parameters

The indoor temperatures under the five simulated cases during the typical week are shown in Figure 6. In all the cases, the floor temperature is the lowest, followed by the ceiling temperature, while the temperatures of the interior walls, operative temperature, and air were relatively similar; the inner surface temperature of the exterior walls was the highest. The ceiling is directly on top of the floor, and the radiation angle factor is quite large between the two surfaces, which increases the radiant heat exchange between the floor and the ceiling, resulting in a corresponding decrease in the ceiling temperature when the floor temperature drops. All temperatures fluctuated daily, with the floor and ceiling exhibiting the smallest fluctuations, whereas the inner surface temperature of the exterior walls varied significantly due to the influence of outdoor air temperature. Under case C₀, floor and ceiling temperatures declined between 7:00 and 17:00. Under the other cases, temperatures dropped quickly during the nighttime cooling storage period from 2:00 to 8:00, while increasing during working hours, and the magnitude of the increase was influenced by the duration of the cooling supply during the flat-rate period. The fluctuation ranges and differences in each temperature parameter during office hours from Monday to Friday under the five simulated cases are summarized in

Table 5. Statistics of indoor temperatures in typical week.

Simulated Cases		Indoor Air Temperature (°C)	Indoor Operative Temperature (°C)	Surface Temperature (°C)
				Floor	Ceiling	Exterior Wall	Interior Walls
C0	Highest	26.3	26.1	25.0	25.1	27.2	26.3
	Lowest	24.6	24.6	23.9	24.3	24.8	24.8
	Difference	1.7	1.5	1.1	0.9	2.3	1.6
C1	Highest	28.2	28.1	27.1	27.6	29.0	28.2
	Lowest	25.8	25.8	25.0	25.3	26.1	26.2
	Difference	2.4	2.3	2.0	2.3	2.9	2.1
C2	Highest	27.3	27.2	26.0	26.4	28.0	27.3
	Lowest	25.0	25.1	24.3	24.5	25.4	25.4
	Difference	2.2	2.1	1.7	1.9	2.7	1.9
C3	Highest	26.5	26.3	25.3	25.5	27.2	26.5
	Lowest	24.4	24.4	23.6	23.8	24.7	24.8
	Difference	2.1	1.9	1.8	1.7	2.5	1.7
C4	Highest	25.8	25.6	24.8	24.9	26.5	25.8
	Lowest	23.8	23.8	23.0	23.2	24.1	24.2
	Difference	2.0	1.7	1.8	1.7	2.3	1.6

. The lowest floor temperature is 23 °C (C₄) in all the simulated cases, which is higher than the room dew temperature (18–19 °C). Therefore, there is no condensation risk in any of the simulated cases.

Under case C₀, during the typical weekday office hours of 8:00–18:00, the indoor air temperature generally remained within the range of 24.6–26.3 °C, with the operative temperature ranging from 24.6 to 26.1 °C, thereby meeting indoor comfort requirements. Under case C₁, the indoor air temperature remained within the range of 25.8–28.2 °C, with the operative temperature ranging from 25.8 to 28.1 °C. Temperatures are higher in the afternoon, making it difficult to meet the thermal comfort requirements. In case C₂, the indoor air temperature and operative temperature are 25.0–27.3 °C and 25.1–27.2 °C, respectively. The overall temperatures of case C₂ are slightly lower than in case C₁. For cases C₃ and C₄, the indoor temperatures further decreased, with air temperatures ranging from 24.4 to 26.5 °C and 24.4 to 26.3 °C, respectively, while operative temperatures ranged from 23.8 to 25.8 °C and 23.8 to 25.6 °C. During the typical week, indoor temperatures remained within the target range for case C₃ and C₄.

For the four cold storage cases, as the total cooling duration increased, both the maximum and minimum indoor air temperatures tended to decrease, and the fluctuations of indoor temperatures also declined. It can also be noticed that the indoor temperature is lowest in the morning due to night cooling, rising gradually during occupied hours due to factors such as internal heat gains, rising outdoor temperatures, and the cessation of floor cooling. The indoor temperature reaches the peak value by 6:00 PM when the workday ends. After this period, the indoor temperature decreases gradually as occupants have left and both lighting and equipment have been switched off.

Fluctuations in indoor temperature can affect thermal comfort. Kolarik et al. [25] suggested that when the indoor temperature fluctuations were less than 2 °C, thermal comfort would not be significantly influenced. Considering that the four cold storage cases operate only during nighttime and the morning off-peak periods, a comparison of the typical daily operative temperature under the four cases is presented in Figure 7. The changes in operative temperatures in the four cases are quite similar, which were at their lowest at 8:00 and kept increasing till 18:00; after that, the temperatures decrease gradually. When the night cold storage starts, the temperatures drop quickly. For each cold storage case, the operative temperature fluctuation remained within 2 °C on a typical day, meeting the basic requirements for temperature drifts. Condition C₁, which stores cold only at night, is insufficient to provide the required cooling energy, resulting in indoor air temperatures exceeding 27 °C for more than half of the working hours. With prolonged cooling duration in the four cases, the operative temperatures decrease obviously. Under case C₃ and C₄, the maximum room temperature values are both below 26.5 °C, which meets the thermal comfort requirements of lower than 27 °C.

The above analysis shows that relying solely on nighttime floor-cooling storage cannot fully satisfy the daytime cooling demand. However, combining nighttime cooling with supplementary cooling during daytime off-peak hours can effectively maintain indoor temperature at a comfortable level. Further comparison indicates that the distribution of indoor thermal environment parameters under case C₀ is comparable to that under C3, as both cases share the same total daily operation duration of 10 h. The difference lies in the off-peak-hour operation time: case C₀ only allocates 1 h to off-peak operation, while case C₃ allocates 6 h. Therefore, case C₃ achieves a more economically efficient operation while maintaining equivalent indoor thermal environment quality.

4.2. Analysis of Typical Daily Cold Storage and Cooling Supply Performance

For case C₁ to C₄, the floor-cooling loop is operated for cold storage from 2:00 to 8:00. A portion of the cooling energy is stored within the floor and ceiling slabs, while the indoor air temperature decreases overnight and further cools the building internal envelopes, including interior walls. Consequently, the cold energy is ultimately stored in the slab structure, internal walls and indoor air. Meanwhile, a certain amount of cooling loss occurs due to heat transfer through the external building envelopes.

Among the four cold storage cases, case C₃ is selected for further detailed analysis of the cold charging and discharging processes on a typical day, as it has the same operating duration as the referenced condition C₀ and provides satisfactory indoor thermal comfort.

The nighttime cold storage distribution and cooling losses under case C₃ on a typical day are presented in Figure 8. Between 2:00 and 8:00, the total cooling energy supplied by the water-side floor circulation system is 27,014 kJ, according to the circulating water flow rate and supply–return water temperature differences. During this period, the temperatures of the floor and ceiling slabs, internal walls and indoor air decrease continuously. The cooling energy stored in the floor and floor slabs is 19,651 kJ, accounting for 72.7% of the total supplied cooling; the internal walls store 6297 kJ, accounting for 23.3%; and the indoor air stores 242 kJ, accounting for 0.9%. The total cooling loss through external walls and windows during the cold storage period is 824 kJ, which constitutes approximately 3.1% of the total cooling supply. The hourly cooling energy stored in the floor and ceiling keeps decreasing during the nighttime. This is primarily because the indoor air temperature drops gradually as the cold storage process proceeds, which reduces the heat exchange temperature difference between the slab surfaces and the indoor space. Generally, owing to the low outdoor temperature and the small indoor–outdoor temperature difference during the night, the heat loss through the external envelopes remains relatively low.

Figure 9 shows the hourly heat flux of indoor surfaces in the typical day including night cooling storage and office hours. Positive heat flux means that the surface absorbs cooling energy, while the negative one shows that the surface releases cooling energy.

The net heat flux on the surfaces of the floor and the ceiling is negative, indicating that these two surfaces are continuously absorbing heat. The ceiling is affected by the radiation from the floor, and its temperature is lower than the air and the other non-cooled surfaces. The two surfaces are cooled by night cooling, while during the day, the continuous increase in air temperature makes the two surfaces absorb heat from the room continuously. Since the air temperature rises significantly during the day, the heat absorption of these two surfaces also increases during the period 8:00 to 12:00. The heat flux on the floor surface is approximately between −5 and −17 W/m², and that on the ceiling is approximately between 0 and −15 W/m². The heat flux on the surface of the window is negative from 2:00 to 6:00 at night and positive at other times, mainly because the outdoor temperature is slightly lower than the indoor temperature at night and much higher during the day. The specific value is between −2 and 55 W/m². The surface heat flux of the inner walls is positive throughout the day, and approximately between 2 and 5 W/m². The heat flux of the exterior wall is similar to that of the inner wall at night, while increasing significantly during the day. The heat flux of exterior wall ranges from 2 to 10 W/m².

Figure 10 shows the hourly cooling load, cooling energy stored, and cooling supplied to the room on the typical day. Between 8:00 and 18:00, the cooling load first increases and then decreases, ranging between 45 and 65 W/m². During the same period, the fresh air bears part of the cooling load, which is basically stable in the range of 6–8 W/m².

From 2:00 to 8:00, the total cooling energy supplied from the water side is 27,014 kJ. During this period, a major portion of the cooling energy is stored, while the rest is released into the room through the floor and ceiling surfaces, which includes those stored in the internal walls, air, and cooling losses. When the cooling system is turned on at 2:00, the cooling energy stored in the slabs is the highest of 42 W/m². During cold storage, the floor temperature continuously drops, resulting in a decrease in the heat exchange temperature difference between the water side and the floor, leading to a gradual reduction in the cold charging rate and a corresponding increment of the cold discharging rate from the surfaces of the floor and ceiling.

By 8:00, the cooling energy discharged from the floor and ceiling surfaces is approximately 8 W/m² and 6 W/m², respectively. Subsequently, during occupied hours, internal heat gains cause the room temperature to rise, thereby increasing the heat transfer temperature difference. Consequently, the cooling discharged from the slabs continues increasing, and reaches 28 W/m² by 12:00. Two hours after the cooling system is turned off (i.e., at 14:00), the discharging rate reaches the peak of 29 W/m². After that, it gradually declines to 22 W/m² by 18:00. It is noteworthy that after 18:00, the slabs’ surface temperatures remain lower than the indoor air temperature. A residual amount of stored cooling continues to be discharged, which lowers the indoor temperature before the next cold storage cycle. Furthermore, during the 8:00–12:00 period, the cooling energy supplied to the slabs exceeds the cooling released from the surfaces; thus, there is still a small amount of cooling energy being stored. This stored cooling energy decreases quickly.

After 6:00 PM, the office room is closed, while the floor and ceiling surface temperatures are still lower than the indoor air temperature. Before the nighttime cold storage begins operating, a portion of the cooling energy is still released continuously causing the indoor air temperature to drop gradually.

4.3. Comparison of Thermal Comfort, Energy Use, and Running Costs

4.3.1. Statistics on Thermal Comfort in Typical Room

Table 6. Statistics on indoor PMV indices for various operating cases.

Operating Conditions	Proportion of |PMV| < 0.5	Proportion of |PMV| < 1	Proportion of |PMV| > 1
C1	26%	64%	36%
C2	60%	93%	7%
C3	87%	100%	0
C4	98%	100%	0

presents the statistical analysis of the predictive mean vote (PMV) for thermal comfort in the typical room under the four cold storage strategies during the whole cooling season. For case C₁ and C_2, the proportions of time during which the PMV index exceeds 1 were 36% and 7%, respectively. These two cases cannot meet the requirements for Thermal Comfort Grade II (PMV within the range of −1 to 1) as specified in the reference [26]. case C₃ and C₄ fully meet the requirements of this grade. In Case C₄, within 98% of the time PMV can be kept in the range of −0.5 to 0.5, which meets the requirement of Thermal Comfort Grade I [26].

4.3.2. Power Use and Costs

According to the electricity pricing policy, the sharp-peak rate is 1.2 RMB/kWh, the peak rate is 0.71 RMB/kWh, the flat rate is 0.58 RMB/kWh, the off-peak rate is 0.35 RMB/kWh, and the deep off-peak rate is 0.25 RMB/kWh. The operating hours for each time period under the six operating cases are summarized in

Table 7. Statistics of operating hours in different periods during cooling season.

Cases	Sharp-Peak Hours	Peak Hours	Flat Hours	Off-Peak Hours	Deep Off-Peak Hours	Total Operating Hours
T0	238	141	802	346	33	1560
C0	11	119	465	22	33	650
C1	0	108	66	324	0	498
C2	0	0	196	324	0	520
C3	0	0	304	335	11	650
C4	0	0	412	335	33	780

. In the statistics, there are 54 working days from June to August and 11 days in September after excluding weekends and public holidays.

The total electricity use per unit of air-conditioned area during the cooling season of the building is 5.52 kWh (T₀), 5.0 kWh (C₀ and C₃), 4.85 kWh (C₁), 4.92 Wh (C₂), and 5.07 kWh (C₄). With decreased operating hours, the power use also decreases accordingly, primarily due to reduced power use by the radiant floor circulation pumps. Compared to T₀, the energy savings of the other five cases range from 8.1% to 12.2%. The operating costs are calculated based on the operating hours, electricity rates, and electricity use for each time period. The results are shown in

Table 8. Comparison of electricity use and costs during cooling season.

Simulating Cases	Total Electricity Use Per Unit Air-Conditioned Area (kWh)	Cost Per Unit Air-Conditioned Area (RMB/m2)	Compared to T0 Conditions		Compared to C0 Conditions
			Changes in Electricity Use	Changes in Operating Costs	Changes in Electricity Use	Changes in Operating Costs
T0	5.52	2.73	/	/	/	/
C0	5.00	2.38	−9.5%	−12.7%	/
C1	4.85	2.22	−12.2%	−18.6%	−3.0%	−6.8%
C2	4.92	2.27	−10.9%	−16.9%	−1.5%	−4.8%
C3	5.00	2.31	−9.5%	−15.2%	0.0%	−2.8%
C4	5.07	2.36	−8.1%	−13.4%	+1.5%	−0.9%

.

Under case T₀, the operating cost per unit area is 2.73 RMB/m². Under case C₀, with 10 h of normal operation, the cost is 2.38 RMB/m², representing a 12.7% cost savings. Case C₃ has the same operating duration as C₀, which achieves 15.2% cost savings. Case C₄, operating for 12 h, achieves 13.4% savings. Compared with case C₀, the four cold storage cases (C₁ to C₄) also yield varying degrees of cost savings. Case C₃ and C₄, which meet the thermal comfort requirements, achieve savings of 2.8% and 0.9%, respectively.

Cases C₁ and C₂ have shorter operating hours, which lead to more significant savings in power use and operating costs. However, both cases cannot assure indoor thermal comfort. Cases C₃ and T₀ have equal operating durations, and both meet thermal comfort requirements of Grade II, while case C₃ offers better economic efficiency. Case C₄ leads to the best thermal comfort situation, and the indoor thermal environment can meet the requirements of Grade I. Compared to case C₀, the electricity use of C₄ increases by 1.5%, while the operating costs decrease by 0.9%. The two cold storage strategies of cases C₃ and C₄ are more economical than of case C₀, which operates normally during office hours.

5. Limitations

The above results quantify the cold storage capacity, cooling losses, indoor temperature fluctuations, power use, and running costs associated with cold storage operating strategies by envelopes of TABSs. The study offers detailed insights for engineering applications. For radiant cooling systems, thermal comfort and cost savings can be achieved by simply rescheduling the running hours of radiant terminals based on local peak and off-peak electricity tariffs. Inevitably, there are a few limitations of the study, which are discussed in the following section.

Firstly, in this paper, the floor cold storage operating strategies were proposed. However, the actual operating effect in the modeled building was not verified in practice. It would be better if the strategies can be applied in the building to provide more solid proof.

Secondly, the thermal capacity of a concrete floor is quite limited, which limits the cold storage amount during off-peak hours. In recent years, building envelopes combining PCMs have been studied widely [27,28,29,30,31,32]. The use of PCMs increases the thermal capacity of building envelopes and thermal stability, and shows significant promise in enhancing energy efficiency. If PCMs can be incorporated into building structures such as floors and walls, the cold storage capacity of these building envelopes under the proposed cold storage operating modes can be significantly enhanced; thereby, operating costs can be further reduced. The optimal PCM temperature range, cold storage capacity, and operating impacts can be further investigated.

Furthermore, the model building in the study is a LECB that uses GHEs to provide cooling for the floor circuit, and cold storage for the fresh air system. Consequently, the running cost savings under the proposed cold storage strategies are not quite generalizable. In practice, the cooling sources can also be high-temperature chillers. For such systems, the running cost savings can be more significant, which should be further investigated.

Additionally, the water temperature from the GHEs is relatively high, which reduces the cold storage capacity during cold storage. Further research could be conducted to explore the impact of reducing the chilled water temperature on both cold storage capacity and operating efficiency while avoiding floor surface condensation.

In addition, the simulation analysis in this study is primarily based on a specific physical building. Changes to the building characteristics or the climate zone would inevitably alter parameters such as thermal performance and load profiles, thereby affecting the effect of cold storage in this research. Meanwhile, the impact of operating time was primarily considered, while other parameters such as supply water temperature were not taken into account. Further analysis adding the impact of these parameters on operating effect would be useful to verify the results and address this limitation.

In the simulation, the operating start time, occupancy density, and supply water temperature of fresh air system were fixed. The primary purpose of the fixed-parameter setting for the parameters was to simplifying the analysis, and compare the specific impacts of different cold storage strategies under identical conditions. Nevertheless, the impact of these fixed settings should be considered in a future study, in which parametric or sensitivity analysis can be conducted to identify the most influential variables and evaluate the robustness of the proposed operating strategies.

Finally, it should be noted that the four cold storage strategies proposed are based on the peak and off-peak tariff of summer time (from June to August). It can be seen from Table 2 that the tariff policies in September are quite different. During 8:00 to 18:00, most of the time falls into the off-peak, deep off-peak and flat periods, except the peak and sharp-peak periods between 16:00 and 18:00. Meanwhile, nighttime falls entirely within the flat-rate period. Consequently, throughout the operational period of 11 working days in September, the system can be operated more economically during normal occupied hours. In the simulation, this factor was neglected, which exerts a negligible impact on the overall results of the control strategies. However, this reminds us that the operating strategies must be evaluated according to the changes in local electricity pricing to achieve the best economical savings.

6. Conclusions

This paper proposes four operating cases for cold storage during nighttime off-peak hours in floor radiant cooling systems. The impacts of operating duration on thermal environmental parameters, thermal comfort, cooling supply, total power use, and operating costs are analyzed. The cold storage process and cooling loss are quantified based on a typical case. The study provides detailed information about the cold storage thermal process for TABSs. The main conclusions are as follows:

(1)

During the nighttime cold storage, the indoor temperatures keep decreasing, while increasing during the occupied hours, resulting in fluctuations in indoor temperature throughout the day. Both the air and operative temperature drifts are less than 2 °C within the typical day. The cold energy stored during the six off-peak hours alone cannot meet the total cooling load. With extended cooling duration during flat hours, the indoor temperatures can be reduced effectively. In case C₃, with the same operating duration as occupied hours, the PMV index can be maintained within the ±1 range. Under case C₄, operating for 12 h, indoor thermal comfort is further enhanced, and the PMV is maintained within the ±0.5 range for 98% of the time.

(2)

Analysis of the cooling energy stored and supplied is conducted for case C₃, which meets the thermal comfort requirement and has the same operating time as occupied hours. The total cooling supply during the nighttime cold storage is 27,014 kJ, among which 72.7% is stored in the floor and ceiling slabs, 23.3% in the interior walls, and 0.9% in the air. The cooling energy losses through the building envelopes account for 3.1%, indicating a very low level of cooling capacity loss during nighttime cold storage.

(3)

During office hours, the hourly cooling energy supplied to the room is lower than the required cooling load, leading to increasing indoor temperatures. With GHEs as the cooling source, the supply water temperature in the floor circuit is relatively high, which limits the cooling capacity of the floor slabs. The thermal capacity of slabs allows the cooling energy to be released slowly during periods when cooling systems are turned off, which decreases the room temperature after the office is closed.

(4)

The total power use of case C₃ is the same as case C₀, while the running costs of C₃ are 2.8% less than those of case C₀. Case C₄ uses more power than case C₀, while the running costs are still 0.9% lower. With the price difference between peak and off-peak times, the operating strategies of using the building envelopes of the TABS for cold storage during the night not only ensure thermal comfort but also reduce the running cost.