Indoor thermal comfort research on the hybrid system of radiant cooling and dedicated outdoor air system

Abstract

The radiant cooling system generally operates with the dedicated outdoor air system (DOAS). Air supply modes and the corresponding setting parameters of the hybrid system may substantially influence the indoor thermal comfort. With target indexes of air diffusion performance index (ADPI) and predicted mean vote (PMV), the Taguchi method was used to choose the optimal air supply mode and to analyze the significance of different factors on the thermal comfort. The results are expected for conducting the future design and regulation of the hybrid system. Computation fluid dynamics (CFD) simulation as well as verified experiments was performed during the research. Based on the ADPI studies, it is found that the air supply mode of ceiling delivery with ceiling exhaust is an optimized option to apply in DOAS of the hybrid system. Variance analysis results show that influence fact of air supply temperature is the most dominant one to impact the indoor thermal comfort index of PMV.

Keywords

radiant cooling system / dedicated outdoor air system (DOAS) / thermal comfort / Taguchi method / variance analysis

Introduction

The radiant cooling system can solely assume the sensible heat load. The dedicated outdoor air system (DOAS) is required to remove the latent heat load, decrease the condensing risk and provide adequate fresh air [1]. These two systems are generally combined into a hybrid one in practice. Tian et al. have compared the indoor thermal comfort of the radiant systems combined with and without the air supply system under an experimental heating condition [2]. The fresh air system is found to improve the general thermal comfort as well as to reduce the risk of local discomfort. The thermal comfort study by questionnaire survey on the chilled ceiling system with displacement ventilation [3] has shown that, the vertical asymmetry, which occurs when the ceiling temperature varies, has an insignificant effect on the overall thermal comfort of the seated occupants. Slight movement of the air in the radiant cooling environment increases the comfort sensation [4]. Catalina et al. [5] have investigated a radiant cooling ceiling room by full scale model experiment with computation fluid dynamics (CFD) simulation. The vertical temperature gradient is less than 1°C/m, thermal comfort is achieved and the fresh air uniformly distributed within the test room. Another CFD simulation [6] has been conducted on the radiant ceiling combined with a mechanical ventilation system. A more comfortable condition can be obtained by adjusting the location of the vents. Studies [7] have indicated that, the introduction of radiant cooling panel system can reduce the draft risk and raise the air diffusion performance index (ADPI) value in the all air mixing ventilation system when the air jet is supplied with low Archimedes numbers. The Archimedes number of a specific vent is determined by the outlet velocity and the supply air temperature difference.

With regard to the thermal comfort index, the ADPI mainly describes the indoor thermal comfort level caused by the mixing uniformity of the supply air and the indoor air [8]. The ADPI will be influenced by the vent location and supply temperature and the indoor air velocity [9,10]. While the predicted mean vote (PMV) can assess the thermal comfort more comprehensively with the addition of mean radiant temperature (MRT) and relative humidity. Loveday et al. [11] have verified the applicability of the PMV index on the radiant cooling and displacement ventilation system by subjective questionnaire survey.

Two indexes of ADPI and PMV mentioned above were adopted to evaluate the indoor thermal comfort condition. This paper included three parts. First, the optimized air supply mode is chosen according to the ADPI index. Then, one of the above simulation conditions was conducted in the test chamber to verify the accuracy of the building model from two different aspects. With the optimized air supply mode, PMV is chosen as a target index to design an orthogonal test (namely the Taguchi method), for the investigation of the impact significances of the radiant panel temperature, the coverage rate, and the air supply temperature and volume. The main aim is to find out an optimal air supply mode for the DOAS in the hybrid system and the most remarkable influencing factor on the hybrid system for the future design and engineering application.

Methods, model and theory

From the brief literature review, the thermal comfort of the hybrid system of radiant cooling and DOAS is generally influenced by the ceiling radiant cooling panel temperature, the radiant panel coverage rate, and the air supply temperature and volume (for a specific vent with a constant area, the effect of air supply volume on the thermal comfort is reflected in the outlet velocity). During practical operation, the study on the above factors is important to maintain the stability of the thermal comfort. Besides, the foregoing researches focus on a certain air supply mode while the radiant cooling system may combine with different modes in practice. Therefore, it is necessary to compare the indoor environment differences caused by the air supply mode when designing a hybrid system.

Airpak simulation tool was used to conduct the research. The office model was constructed according to an existing test chamber with a size of 4.5 m × 4.5 m × 3 m. Every opening and vent had a full scale, and the outlet velocity of the vents of the model adopted the measured value under different air supply volume. The exhaust vent was set as the boundary condition of the pressure outlet. There were four heat sources, each of which released a sensible heat of 180 W. The meshes were refined at each surface, opening and vent with prospective larger temperature gradient as shown in Figs. 1 and 2. When calculating the PMV index, human parameters were chosen according to the activity characteristics in the summer office. The metabolic rate is 58.15 W/m², the mechanical efficiency is 0 W, and the clothing thermal resistance is 0.08 m²·K/W.

The turbulence model of indoor zero-equation [12], which was exploited according to the direct numerical simulation (DNS) method, was employed in the current research. The turbulent viscosity was expressed by a single algebraic function as

(1){\mu }_t=\mathrm{0.03874}\rho vl,

(2)v={\displaystyle \frac{{\mu }_{\mathrm{eff}}}{\rho }},

(3){\mu }_{\mathrm{eff}}={\mu }_l+{\mu }_t.

where l is the length scale, ρ the air density, v the local average air velocity, μ_eff the effective viscosity, μ_l the laminar viscosity, and μ_t the turbulent viscosity. With the above model, the simulation results coincide well with the experimental data [13] according to the isothermal and non-isothermal study of indoor air flow simulation, and the calculation is ten times faster than the two-equation model [12], which is suitable in this study. The Reynolds-averaged Navier-Stokes equations (RANS) based on the indoor zero-equation can be obtained when the expression of turbulent viscosity is applied in the momentum conservation equations. With the consideration of thermal buoyancy of air, the DO model (discrete ordinates radiation model) is adopted.

As a square diffuser is applied in the ceiling, the N-point opening velocity model based on the measured data are presented [14]. The diffuser is divided into many small vents, the velocities of which are measured under different air supply volumes. Then the guide vane angle α of the diffuser is surveyed, and the above velocity can be resolved into a horizon v_h and vertical v_v.

(4)v_h=v\cos \alpha ,

(5)v_v=v\sin \alpha .

Thus, the complicated diffuser with different outlet directions can be represented by those simple small vents in the model.

Optimization of the air supply mode with ADPI

Excellent airflow uniformity is the premise of the human thermal comfort, which can be obtained by the optimization of air supply mode. To weaken the influence of the randomness in the selection of the factors, four different levels were chosen within the practical range for the air supply temperature and volume as presented in Table 1. Each air supply mode had four different combinations with air supply temperature and volume to get an average ADPI, and the combinations were designed based on the orthogonal table with the principle of uniformity and neatness. The calculated average ADPIs for different air supply modes are attached on the right side of Table 1, where down-up means that the fresh air is supplied from the floor and exhausted from the ceiling, and the rest may be deduced by analogy.

It is observed from Table 1 that the ADPI values are small. The reason for this is that the test chamber, the prototype of the simulation model, is located in a large laboratory. Thus the cooling load of the model is relatively small along with a relatively small indoor temperature. However, it does not influence the comparison as the four modes are calculated by using the same model. From the comparing results, the best indoor air uniformity can be obtained when the air is supplied from the ceiling and exhausted from the ceiling (namely the up-up mode with an ADPI of 38.3%).

Verification experiment of the model

The model used in the above simulation is verified experimentally by directly comparing the temperature field of the simulation and measurement conditions, and by calculating the indoor non-uniformity coefficient and verifying the model indirectly through the conclusion of the optimal air supply mode.

Comparison of the temperature field (direct verification)

One of the above simulating conditions is chosen to conduct the experiment in the test chamber. The condition includes the up-up air supply mode, the air supply temperature of 22°C and the air supply volume of 500 m³/h. The test chamber is depicted in Fig. 3. The air supply mode can be transformed by switching different air duct valves, the air supply temperature is regulated by the valve opening of chilled water pipe and the heat from the re-heater in the air flue, and the air supply volume is changed by the fan inverter. In the test chamber, there are 4 × 180 W dummies acting as human beings. The probes of temperature and velocity are located at 6 horizontal positions and 4 vertical positions (namely 0.1 m, 0.6 m, 1.1 m and 1.7 m) to get a relatively integrated description of the indoor condition of the personnel activity area. Thus, the 24 temperatures measured can be compared with the simulating ones at the same positions, as demonstrated in Fig. 4. From Fig. 4, it is seen that the simulation data coincide well with the experimental ones, which proves that the model constructed basing on the existed test chamber has an acceptable accuracy.

Comparison of simulation and experimental conclusions (indirect verification)

In the test chamber, temperatures and velocities at those 24 positions can be measured under four different air supply modes to calculate the non-uniformity coefficients. The best uniformity of the air supply modes can be derived from the experimental data. The comparison of the simulation and experimental conclusions can indirectly verify the model. The non-uniformity coefficients of the temperature and velocity are calculated as follows.

First, the arithmetic mean values of the temperature and velocity are obtained with the measured data of 24 positions, respectively.

(6)\overline{t}={\displaystyle \frac{{\displaystyle \sum _{i=\mathrm{1}}^{\mathrm{24}}{t}_i}}{\mathrm{24}}},

(7)\overline{u}={\displaystyle \frac{{\displaystyle \sum _{i=\mathrm{1}}^{\mathrm{24}}{u}_i}}{\mathrm{24}}}.

Next, the root-mean-square errors are calculated.

(8){\sigma }_t=\sqrt{{\displaystyle \frac{{\displaystyle \sum _{i=\mathrm{1}}^{\mathrm{24}}{\left(t_i\mathrm{-}\overline{t}\right)}^{\mathrm{2}}}}{\mathrm{24}}}},

(9){\sigma }_u=\sqrt{{\displaystyle \frac{{\displaystyle \sum _{i=\mathrm{1}}^{\mathrm{24}}{\left(u_i\mathrm{-}\overline{u}\right)}^{\mathrm{2}}}}{\mathrm{24}}}}.

Finally, the non-uniformity coefficients of temperature and velocity are obtained.

(10)k_t={\displaystyle \frac{{\sigma }_t}{\overline{t}}},

(11)k_u={\displaystyle \frac{{\sigma }_u}{\overline{u}}}.

The calculation results are listed in Table 2.

It can be seen from Table 2 that the temperature can achieve a relatively ideal indoor distribution than the velocity. This is mainly caused the distribution uniformity of the indoor heat sources and the cooling terminals. According to the calculation results, the up-up air supply mode has the best air uniformity when considering either from the air temperature or from the velocity field. This result corresponds with the simulation conclusion.

Results

Based on the optimal air supply mode and the verification of the model, the thermal comfort influencing factors including the ceiling radiant cooling panel temperature, the radiant panel coverage rate, and the air supply temperature and volume are investigated by designing an orthogonal test with the target index of PMV in the up-up air supply mode. An orthogonal table L₁₆ (45) (16 combinations with 5 factors and 4 levels) as well as the variance analysis method is utilized to study the impact significance of the above factors. The factors, levels and the corresponding results of each combination are tabulated in Table 3, in which the levels are chosen based on the engineering practice.

The PMV values of each combination are between -0.5 and 0.5, which meets the standard of ISO7730. Thus, the levels of factors are relatively reasonable. The arithmetic procedure of the variance analysis and F test may be referred to in Ref. [15]. And the calculation results are given in Table 4. For a certain factor A, there is a possibility of 1-α to affirm the influence on the target index when F_A>F_α (f_A, f_error), where f_A and f_error represent the degree of freedom of factor A and error column. In this study, α = 0.05 and 0.01 are respectively chosen to define the impact significance of “notability” (marked as “**”) and “special notability” (marked as “***”). It is seen from the calculation results that the impact significance of the factors has an order of air supply temperature>radiant panel coverage rate>radiant panel temperature>air supply volume, of which, the air supply temperature has an impact of “special notability,” and the rest are “notability.”

Conclusions

With the target thermal comfort indexes of ADPI and PMV, the Airpak simulation tool as well as required verified experiment is utilized for the current study to analyze the influencing factors of the indoor thermal comfort. Besides, the indoor environments of radiant cooling system combined with DOAS are compared. Fresh air is supplied in different air supply modes. By means of the Taguchi method and intuitive comparisons, the main conclusions of this paper can be drawn as follows:

1) According to the study on the ADPI and the non-uniformity coefficient, the indoor air owns a comparably perfect uniformity when fresh air is supplied from the ceiling and exhausted from the ceiling (namely the up-up air supply mode).

2) The influencing factors of the indoor thermal comfort have significance order of air supply temperature>radiant panel coverage rate>radiant panel temperature>air supply volume, of which, the air supply temperature has an impact of “special notability,” and the rest are “notability.” This finding may provide some reference for the operation and regulation of the hybrid system.

3) The up-up supply mode is recommended for the DOAS, when the DOAS is coupled with the radiant cooling system.

References

[1] Feustel H E, Stetiu C. Hydronic radiant cooling-preliminary assessment. Energy and Building, 1995, 22(3): 193–205

[2] Tian Z, Ding Y, Wang S, Yin X L, Wang M L. Influence of the ventilation system on thermal comfort of the chilled panel system in heating mode. Energy and Building, 2010, 42(12): 2360–2364

[3] Hodder S G, Loveday D L, Parsons K C, Taki A H. Thermal comfort in chilled ceiling and displacement ventilation environments: Vertical radiant temperature asymmetry effects. Energy and Building, 1998, 27(2): 167–173

[4] Kitagawa K, Komoda N, Hayano H, Tanabe S. Effect of humidity and small air movement on thermal comfort under a radiant cooling ceiling by subjective experiments. Energy and Building, 1999, 30(2): 185–193

[5] Catalina T, Virgone J, Kuznik F. Evaluation of thermal comfort using combined CFD and experimentation study in a test room equipped with a cooling ceiling. Building and Environment, 2009, 44(8): 1740–1750

[6] Chiang W H, Wang C Y, Huang J S. Evaluation of cooling ceiling and mechanical ventilation systems on thermal comfort using CFD study in an office for subtropical region. Building and Environment, 2011, 48(1): 113–127

[7] Corgnati S P, Perino M, Fracastoro G V, Nielsen P V. Experimental and numerical analysis of air and radiant cooling systems in offices. Building and Environment, 2009, 44(4): 801–806

[8] Zang X G, Zhao S X, Li W J. On comparison and analysis of thermal comfort with application of ADPI and PMV evaluations. Shanxi Architecture, 2010, 36(33): 174–175

[9] Ng K C, Kadirgama K, Ng E Y K. Response surface models for CFD predictions of air diffusion performance index in a displacement ventilated office. Energy and Building, 2008, 40(5): 774–781

[10] Abu-El-Hassan M B, Hosni M H, Miller P L. Evaluation of turbulence effect on air distribution performance index (ADPI). ASHRAE Transactions, 1996, 102: 322–331

[11] Loveday D L, Parsons K C, Taki A H, Hodder S G. Displacement ventilation environments with chilled ceiling: Thermal comfort design within the context of the BS EN ISO7730 versus adaptive debate. Energy and Building, 2002, 34(6): 573–579

[12] Chen Q Y, Xu W R. A zero-equation turbulence model for indoor airflow simulation. Energy and buildings, 1998, 28(2): 137–144

[13] Zhao B, Li X T, Yan Q S. Simulation of indoor air flow in ventilated room by zero-equation turbulence model. Journal of Tsinghua University (Science & Technology), 2001, 41(10): 109–113

[14] Zhao B, Li X T, Yan Q S. Simplified method for numerical simulation of indoor airflow. HV&AC, 2003, 33(3): 102–104

[15] Turgut E, Cakmak G, Yıldız C. Optimization of the concentric heat exchanger with injector turbulators by Taguchi method. Energy Conversion and Management, 2012, 53(1): 268–275

RIGHTS & PERMISSIONS

Higher Education Press and Springer-Verlag Berlin Heidelberg