COUPLING STRATEGY OF HVAC SYSTEM SIMULATION AND CFD PART 2: STUDY ON MIXING ENERGY LOSS IN AN AIR-CONDITIONED ROOM

Transcription

1 12th Conference of International Building Performance Simulation Association, Sdne, November. COUPLING STRATEGY OF HVAC SYSTEM SIMULATION AND CFD PART 2: STUDY ON MIXING ENERGY LOSS IN AN AIR-CONDITIONED ROOM Satoru Iiuka 1, Mina Sasaki 1, Guoung Yoon 2, Masaa Okumia 1, Juna Kondo 3, and Yuka Sakai 4 1 Nagoa Universit, Nagoa, Japan 2 Nagoa Cit Universit, Nagoa, Japan 3 Kajima Corporation, Toko, Japan 4 Toama Cit Hall, Toama, Japan ABSTRACT A coupled analsis of heating, ventilation and airconditioning (HVAC) sstem simulation tool and computational fluid dnamics (CFD) model was carried out to assess the miing energ loss in an airconditioned room where heating and cooling operate in the perimeter and interior ones simultaneousl. To evaluate the miing energ loss, we conducted two simulations;; one was the case with airflow miing between the perimeter and interior ones and the other was the case without airflow miing between the ones. B comparing the required coil loads between the two cases, the miing energ loss was estimated. The accurac of the coupled analsis was assessed b comparing its results with those from an eperiment conducted b Ito (1988). INTRODUCTION In recent ears, effective energ management and energ saving strategies are essential in the contet of the issue of global warming. In office buildings, for instance, further improvement in the efficienc of air-conditioning sstem is required in order to achieve high energ saving. To eamine the performance of air-conditioning sstem, various simulation tools (e.g., Life Ccle Energ Management (LCEM) tool, HVACSIM+, TRNSYS, DeST, EnergPlus) have been developed. The sstem simulations performed with those tools are usuall carried out under the assumption of complete miing in the target room. However, without the consideration of spatial and temporal distributions of indoor environment, it is difficult to perform more sophisticated energ saving control. Coupled analsis of HVAC sstem simulation and CFD is a promising strateg to overcome the above problem. CFD is considered to be a powerful tool for analing airflow and thermal environments, therefore detailed information of flow field and air temperature can be obtained. As reviewed in another paper for this conference (Yoon et al., 2011), several intensive studies on coupled analsis of HVAC sstem simulation and CFD have been conducted (e.g., Lam et al., 2001;; Zhai et al., 2003, 2005, 2006;; Iida et al., 2008). In this stud, coupled simulations of a HVAC sstem simulation tool and a CFD model were carried out to assess the miing energ loss in an air-conditioned room where heating and cooling operate in the perimeter and interior ones simultaneousl. The miing energ loss is defined as the difference between the sum of the net heating/cooling load and that of the actual heating/cooling energ supplied to the room. To evaluate the miing energ loss, we conducted two simulations;; one was the case with airflow miing between the perimeter and interior ones and the other was the case without airflow miing between the ones. B comparing the required coil loads between the two cases, the miing energ loss was estimated. Furthermore, the effects of the reference points of temperature in the perimeter and interior ones on the miing energ loss were investigated. The accurac of the coupled analsis was assessed b comparing its results with those from an eperiment conducted b Ito (1988). OUTLINE OF THE COUPLED SIMULATIONS Room model and simulated cases The room model used during the eperiment conducted b Ito (1988) was the target of this stud. Figure 1 shows an illustration of the facilit with its sie being 6.0 m (Width;; direction) 3.9 m (Depth;; direction) 2.6 m (Height;; direction). The room model was divided equall in two partitions: perimeter and interior ones. Figure 1 Room model

2 12th Conference of International Building Performance Simulation Association, Sdne, November. A fan coil unit (FCU) was installed on the floor of the perimeter one for heating, while a suppl inlet from an air handling unit (AHU) was mounted on the ceiling of the interior one for cooling. The suppl air volumes of FCU and AHU were constant (FCU: 320 m 3 /h, AHU: 325 m 3 /h). The lighting in the interior one was considered as the onl internal heat source (320 W). All walls were adiabatic ecept the window and its surrounding wall in the perimeter one, making it the onl place where heat transfer occurs. The air temperature outside the window and its surrounding wall was set at 5 C. The setting temperature in both perimeter and interior ones was 19 C. The reference points of temperature, as shown in Figure 1, were different for the two test cases (Cases 1 and 2). In Case 1, the reference points in the perimeter and interior ones were located near the suppl inlet of FCU and at the center of the wall with an ehaust outlet, respectivel, while in Case 2, the reference points in the perimeter and interior ones were set at the suction opening of FCU and at the center of the interior one, respectivel. In the eperiment conducted b Ito (1988), for both Cases 1 and 2, additional tests with a vinl-sheet partition at the boundar between the perimeter and interior ones were carried out. The miing energ loss in the eperiment was estimated b comparing the required coil loads between the cases with and without the vinl-sheet partition. In the coupled simulations conducted in this stud, we performed additional simulations with a virtual adiabatic wall at the boundar between the perimeter and interior ones. Like the estimation in the eperiment, we assessed the miing energ loss in the room model b comparing the required coil loads between the cases with and without the virtual adiabatic wall. HVAC sstem simulation The HVAC sstem simulations were performed using LCEM tool ver.3.02 developed under the supervision of the Ministr of Land, Infrastructure, Transport and Tourism (MLIT) of Japan. Table 1 shows the specifications of FCU (perimeter one) and AHU (interior one) determined based on the results of heat load calculations. Table 1 Specifications of FCU and AHU FCU (PERIMETER ZONE) Suppl air volume : 320 m 3 /h Hot water volume : 5.0 l/min Hot water temperature : C Heating capacit : 3.2 kw AHU (INTERIOR ZONE) Suppl air volume : 325 m 3 /h Chilled water volume : 3.6 l/min Chilled water temperature : 7-12 C Cooling capacit : 1.3 kw CFD The airflow and thermal simulations in the room model (cf. Figure 1) were conducted using a commercial CFD software, STREAM ver.8. The analsis conditions are shown in Table 2. We used the standard k- model as the turbulence model. The total number of the grid points was 70,848. In both Cases 1 and 2, the value of the wall coordinate at the first grid point adjacent to the wall was around Table 2 CFD analsis conditions Grid points 64 () 41 () 27 () = 70,848 Scheme for convection 1st-order upwind scheme for all governing equations terms Turbulence Standard k- model model Inlet boundar FCU (PERIMETER ZONE) condition Velocit : 320 m 3 /h Temp. : PID control (226 C) Humidit : Results from HVAC sstem simulation k : m 2 /s 2 : m 2 /s 3 AHU (INTERIOR ZONE) Velocit : 325 m 3 /h Temp. : PID control (14-18 C) Humidit : Results from HVAC sstem simulation k : m 2 /s 2 : m 2 /s 3 Outlet boundar condition Zero-gradient conditions for all variables Wall boundar condition Heat generation Velocit Temp. : Logarithmic law : Overall heat transfer coefficients for window and its surrounding wall were 3.33 W/m 2 K and 0.50 W/m 2 K, respectivel. For other walls, adiabatic conditions were used. Humidit : Impermeable conditions without condensation Lighting (interior one): 320 W Coupling of HVAC sstem simulation and CFD The detailed information on the coupling method of HVAC sstem simulation (LCEM tool) and CFD (STREAM) is described in another paper for this conference (Yoon et al., 2011). In the coupled simulations conducted here, for both FCU and AHU, the air temperature and humidit at the ehaust outlet calculated b the CFD simulation were imposed to the HVAC sstem simulation. The humidit at the suppl inlet obtained with the HVAC sstem simulation was given to the CFD simulation as boundar conditions. The temperature at the

3 Proceedings of Building Simulation 2011: 12th Conference of International Building Performance Simulation Association, Sdne, November. suppl inlet was provided b a PID control. Those processes were performed ever 30 seconds. RESULTS AND DISCUSSION The following results obtained with the coupled simulations were those at 6000 seconds from the start of the simulations. The indoor environment at that time was considered as a statisticall stead state. Air temperature distributions Figures 2 and 3 show the distributions of air temperature at the center section ( = 1.95 m) obtained with the coupled simulations for Cases 1 and 2, respectivel. The star- shaped and diamond- shaped smbols represent the reference points of the temperature. For reference, Figures 4 and 5 depict the distributions of velocit vectors at the same section ( = 1.95 m) for Cases 1 and 2, respectivel. (1) Without airflow miing (2) With airflow miing Figure 4 Distribution of velocit vectors in Case 1 (1) Without airflow miing (1) Without airflow miing (2) With airflow miing (2) With airflow miing Figure 2 Distribution of air temperature in Case 1 Figure 5 Distribution of velocit vectors in Case (1) Without airflow miing (2) With airflow miing Figure 3 Distribution of air temperature in Case 2 Figures 2(1) and 3(1) illustrate the results of the case without airflow miing (with the virtual adiabatic wall) between the perimeter and interior ones. Figures 2(2) and 3(2) show the results of the case with airflow miing (without the virtual adiabatic wall) between the perimeter and interior ones. In both Figures 2(1) and 3(1), there are large differences between the perimeter and interior ones due to the eistence of the virtual adiabatic wall. In the perimeter one, for both Cases 1 and 2, the warm suppl air from FCU rises (cf. Figures 4(1) and 5(1)), and the cold air caused b the heat transfer through the window and its surrounding non- adiabatic wall stas in the lower region. We can observe an apparent thermal stratification in both perimeter ones. On the other hand, in the interior one, unstable flow is formed b the effect of the cold suppl air from the ceiling, and thus the air in the

4 12th Conference of International Building Performance Simulation Association, Sdne, November. entire interior one is well mied for both Cases 1 and 2. The air temperature is around 19 C (the setting temperature), ecept for the vicinit of the suppl inlet on the ceiling. In both Figures 2(2) and 3(2), airflow miing arises between the perimeter and interior ones because there is no partition at the boundar. The warm suppl air from FCU rises and flows into the upper region of the interior one (cf. Figures 4(2) and 5(2)). In the lower region of the perimeter one, the cold air region epands due to not onl the heat transfer through the window and its surrounding nonadiabatic wall but also the cold flow from the interior one. A remarkable thermal stratification is formed in the entire room in both Figures 2(2) and 3(2). As for the reference temperature, onl the reference temperature in the perimeter one for Case 2 with airflow miing is different from the setting temperature (19 C) (cf. Figure 3(2)). The temperature is 1.3 C lower than the setting temperature. The reference point in the perimeter one for Case 2 is located in a lower position than that of Case 1. Therefore, the reference temperature in the perimeter one of Case 2 is often lower and thus the temperature of the suppl air from FCU is higher. On the other hand, in case with airflow miing, the temperature of the suppl air from AHU in the interior one is lower due to the effect of the warmer suppl air from FCU. As a result, the reference temperature in the perimeter one for Case 2 with airflow miing decreases further. Actuall, in Case 2 with airflow miing, the temperatures of the suppl air from FCU and AHU were the maimum value (26 C) and almost the minimum value (14.1 C) of the PID controls, respectivel. Figures 6 and 7 compare the vertical profiles of air temperature between the eperimental and simulated results ( ST means the setting temperature (19 C)). In general, the results obtained with the coupled simulations correspond well to the eperimental results of Ito (1988). However, there are some differences between the eperimental and simulated results;; 1) In Cases 1 and 2 both with and without airflow miing, the temperatures in the lower region of the perimeter one predicted b the coupled simulations are lower than those of the eperiment. Therefore, in the perimeter one, the temperature differences in the vertical direction given from the coupled simulations are larger than the eperimental results. 2) In Cases 1 and 2 without airflow miing, the temperatures in the interior one predicted b the coupled simulations are higher than those of the eperiment. Those discrepancies might be related to the problems with the eperiment such as air leakage from the vinl-sheet partition and those with the accurac of CFD such as the performance of turbulence model. (1) Eperiment ST 2 ST 2 (2) Coupled simulation Figure 6 Vertical profiles of air temperature in Case 1 (dot line: without airflow miing;; solid line: with airflow miing) ST 2 (1) Eperiment ST 2 (2) Coupled simulation Figure 7 Vertical profiles of air temperature in Case 2 (dot line: without airflow miing;; solid line: with airflow miing)

5 12th Conference of International Building Performance Simulation Association, Sdne, November. Table 3 Suppl air temperature, return air temperature, and required coil load in the coupled simulations PERIMETER ZONE: FCU Without airflow miing With airflow miing Suppl air Return air Suppl air Return air Case C 17.7 C kw 22.5 C 17.7 C kw Case C C kw 26.0 C 17.7 C kw INTERIOR ZONE: AHU Without airflow miing With airflow miing Suppl air Return air Suppl air Return air Case C C kw 17.2 C 18.8 C kw Case C 19.1 C kw 14.1 C C kw Table 4 Estimation of miing energ loss/gain PERIMETER ZONE: FCU INTERIOR ZONE: AHU TOTAL Eperiment Simulation Eperiment Simulation Eperiment Simulation Case kw kw kw kw kw kw Case kw kw kw kw kw kw As an eample of the investigation into the performance of turbulence model of CFD, Figure 8 shows the results of the RNG k- model. The correspondence with the eperimental results is worse, compared to the results of the standard k- model (cf. Figure 6(2)). In the net phase of this stud, we will conduct further investigations into the performance of turbulence model of CFD b introducing more sophisticated models. ST 2 Figure 8 Performance of the RNG k- model in Case 1 (dot line: without airflow miing;; solid line: with airflow miing) Estimation of miing energ loss Table 3 shows the results of suppl air temperature, return air temperature, and required coil load obtained with the coupled simulations for Cases 1 and 2 with and without airflow miing at the boundar between the perimeter and interior ones. Table 4 compares the miing energ loss estimated b the eperiment (Ito, 1988) and that b the coupled simulations. Here, minus values mean the miing energ gain. As described above, the miing energ loss (or gain) was assessed b comparing the required coil loads between the cases with and without the partition (vinl sheet or adiabatic wall) at the boundar of the perimeter and interior ones. In Case 1, the miing energ losses of the perimeter and interior ones estimated b the eperiment are kw and kw, respectivel. The total miing energ loss in the room model is kw. On the other hand, in the coupled simulations for Case 1, the miing energ gain occurs in both perimeter ( kw) and interior ( kw) ones. The total miing energ gain predicted b the coupled simulations is kw. In Case 1, therefore, the correspondence of the eperimental and simulated results is not good. This implies that it is difficult to predict accuratel small amounts of miing energ loss, although further improvement in the accurac of the coupled simulation is still required. In Case 2, the miing energ losses of the perimeter and interior ones obtained with the eperiment are kw and kw, respectivel. The total miing energ loss in the room model is kw. In the coupled simulations for Case 2, the miing energ loss also occurs in both perimeter (0.272 kw) and interior (0.218 kw) ones. The total miing energ loss predicted b the coupled simulations is kw. In Case 2, the simulated results correspond ver well to the eperimental results. CONCLUSION In this stud, the coupled analsis of HVAC sstem simulation and CFD developed b the authors was applied to estimate the miing energ loss in an airconditioned room where heating and cooling operate in the perimeter and interior ones simultaneousl. In case that the miing energ loss was relativel large, the simulated results of the miing energ loss were in good agreement with the results obtained with the eperiment conducted b Ito (1988). On the other hand, in case that the miing energ loss was

6 12th Conference of International Building Performance Simulation Association, Sdne, November. relativel small, the correspondence of the eperimental and simulated results was not good. Although further improvement in the accurac is still required, such coupled simulations are ver useful to anale indoor environment and to perform effective energ management and energ saving in airconditioned rooms where complete miing cannot be assumed. ACKNOWLEDGEMENT The present work was supported b the Environment Research and Technolog Development Fund (Principal investigator: Prof. Takauki Morikawa, Nagoa Universit). REFERENCES Iida, R., Shiraishi, Y., Sagara, N Stud on control and diagnosis of HVAC sstems for an office space, Part 3, Case studies for coupled simulation of CFD analsis and HVACSIM+(J), Summaries of technical papers of Annual Meeting Architectural Institute of Japan, D, (in Japanese). Ito, N Ph.D. Thesis, Nagoa Universit. Lam, J.C., Chan, A.L.S CFD analsis and energ simulation of a gmnasium, Building and Environment, 36, Yoon, G., Kondo, J., Sakai, Y., Watanabe, T., Iiuka, S., Okumia, M Coupling strateg of HVAC sstem simulation and CFD, Part 1, Stud on air conditioning sstem using OA floor in design phase, Submitted to Building Simulation Zhai, Z., Chen, Q Solution characters of iterative coupling between energ simulation and CFD programs, Energ and Building, 35, Zhai, Z., Chen, Q Performance of coupled building energ and CFD simulations, Energ and Buildings, 37, Zhai, Z., Chen, Q Sensitivit analsis and application guides for integrated building energ and CFD simulation, Energ and Buildings, 38,