J. Cent. South Univ. (2012) 19: 633-638
DOI: 10.1007/s11771-012-1049-7
Coupled simulation of BES-CFD and performance assessment of
energy recovery ventilation system for office model
Yunqing FAN, T. Hayashi, K. Ito
Interdisciplinary Graduate School of Engineering Science (IGSES), Kyushu University, Fukuoka, Japan
? Central South University Press and Springer-Verlag Berlin Heidelberg 2012
Abstract: Thermal comfort and indoor air quality as well as the energy efficiency have been recognized as essential parts of sustainable building assessment. This work aims to analyze the energy conservation of the heat recovery ventilator and to investigate the effect of the air supply arrangement. Three types of mixing ventilation are chosen for the analysis of coupling ANSYS/FLUENT (a computational fluid dynamics (CFD) program) with TRNSYS (a building energy simulation (BES) software). The adoption of mutual complementary boundary conditions for CFD and BES provides more accurate and complete information of indoor air distribution and thermal performance in buildings. A typical office-space situated in a middle storey is chosen for the analysis. The office-space is equipped with air-conditioners on the ceiling. A heat recovery ventilation system directly supplies fresh air to the office space. Its thermal performance and indoor air distribution predicted by the coupled method are compared under three types of ventilation system. When the supply and return openings for ventilation are arranged on the ceiling, there is no critical difference between the predictions of the coupled method and BES on the energy consumption of HVAC because PID control is adopted for the supply air temperature of the occupied zone. On the other hand, approximately 21% discrepancy for the heat recovery estimation in the maximum between the simulated results of coupled method and BES-only can be obviously found in the floor air supply ventilation case. The discrepancy emphasizes the necessity of coupling CFD with BES when vertical air temperature gradient exists. Our future target is to estimate the optimum design of heat recovery ventilation system to control CO2 concentration by adjusting flow rate of fresh air.
Key words: building energy simulation; computational fluid dynamics (CFD); FLUENT; TRNSYS; energy saving
1 Introduction
Building energy simulation (BES), a calculation method of energy consumption for heating and cooling of buildings, has been adopted to the design of sustainable low energy building in recent years. The BES is based on the assumption that room air is uniform hybrid. However, this approximation is not satisfactory in stratified indoor air environment. On the contrary, computational fluid dynamics (CFD) simulation can provide more detailed information such as temperature, pressure, velocity distribution of airflow, heat transfer and contaminant transportation indoor and outdoor building environment. Therefore, the integration of CFD and BES is important to analyze non-uniform air distribution inside enclosure zone and provide more accurate prediction of building thermal performance. NEGRAO [1] described coupling between BES and CFD, in which the value of convective heat transfer coefficient (CHTC) used in BES was calculated in CFD. ZHAI and CHEN [2-5] proved the data coupling method which transferred inside surface temperatures from BES to CFD and returned convective heat transfer coefficient as well as indoor air temperature gradients from CFD to BES. They proposed several staged coupling strategies such as static, dynamic and bin coupling and defined implemental condition for each coupling process. DJUNAEDY et al [6] proposed a new concept of external coupling between CFD and BES to reduce computing time. They used existing software packages in difference domains such as thermal domain for energy simulation and the airflow domain for CFD, and provided the mechanism for communication with each other. According to the investigation and comparison among 20 building energy simulation softwares, CRAWLEY [7] explained that only few of these tools allowed for a simulation of thermal load or air flow models with CFD packages. TRNSYS (Transient system simulation program) was advised to be the most flexible simulation tool in terms of calculation and control of building energy consumption evaluation with component-based black box method. Transporting the mutual complementary boundary conditions in the ANSYS/FLUENT and TRNSYS, DIEGO [8] presented the implementation of an automatic fully-dynamic coupling in simple flow network as an example, and predicted the future application of air flow performance estimation and pollutant dispersion under complex boundary condition by creating new component.
The aim of this work is to analyze the energy conservation of the heat recovery ventilator and to investigate the effect of the air supply location on the design and performance of the mixing ventilation by coupling two sophisticated commercial softwares. ANSYS/FLUENT [9] provides detailed three- dimensional information about indoor air flow and TRNSYS indicates integrated characteristic. To reduce the computational cost, standard k-ε turbulent model has been used for CFD simulating and PID control because HVAC has been introduced in building energy simulation in this work.
2 BES-CFD integration
FLUENT is adopted as CFD software for predicting the non-uniform airflow stratification and TRNSYS [10] is employed to estimate the energy performance. Figure 1 illustrates the framework of coupling simulation. The automatic fully-dynamic coupling simulations of TRNSYS and FLUENT are carried out based on the mutual complementary boundary conditions exchanging at each BES time step. At the time series domain, TRNSYS is a main timing continuity process in thermal performance calculation. FLUENT exports non-uniform information of temperature and velocity distribution data to TRNSYS when it achieves converged solution at fixed intervals. The coupling between the two programs takes place by running the FLUENT through a “script file” written by TRNSYS component. The FLUENT issues the “Results file” which contains CFD information such as non-uniform temperature distribution, average temperature in occupied zone, and air temperature of HRV. The file is returned to the TRNSYS component module as the boundary condition at every time step (1 h in this analysis).
Fig. 1 Coupling computational algorithm framework
At each time step, CFD conducts steady state calculation and secures at least 500 iterations to update the data input to the corresponding component of TRNSYS. During the initial time step manipulations, at least one day running-up period is needed for supporting this fully-dynamic coupling. TRNSYS finishes each time-step when its components reach convergence by reading hourly updating boundary condition provided by FLUENT.
3 Project description
Numerical simulation of air distribution under mixing ventilation is carried out for the office located in Gifu Prefecture, Japan. Figure 2 illustrates the geometry and configuration of the office.
Fig. 2 Schematic of office model and layout of AC, SA and RA opening
The office is 28.225 m (x)×15.45 m (y)×2.4 m (z) and consists of main working place and two fitting rooms. Two representative double glazing windows are set on the north and east walls of the main office space. The window ratio is 0.297 for the north wall and 0.251 for east wall. With supply jet diffuser and return slit, the air-conditioner (terminal units of heat pump air condition system) is installed on the ceiling, which provides supply air more directly to the occupied zone instead of attempting to condition the whole space. The air volume of each air conditioner is 1 620 m3/h. The ventilation system operates under a 13.94 ACH (air changes per hour) air change rate. Heat recovery ventilator (HRV) is equipped independently from AC system. HRV originally can recover sensible and latent heat. However, temperature exchange efficiency for sensible heat is only taken into account and the effect of latent heat is ignored. The supply and return air volumes of HRV are all 350 m3/h. Therefore, the air change rate of the space is 1.43 ACH. Three types of ventilation systems have been considered in this case study. Type 1 is setting up ventilation supply openings on the ceiling at interior zone and return openings are also on the ceiling at perimeter zone. These arrangements are as same as the real situation. Type 2 is reverse arrangement of openings against Type 1. Type 2 is designed to compare the effect of openings arrangement on energy consumption. For these ceiling cases, both the supply and return openings are installed as the same height as openings for air-conditions. Therefore, the fresh air enters the room from the ceiling and the warm air exhausts from the upper part of the office. In Type 3, ventilation supply and return opening are located at floor level. The supply air velocity is as same as the ceiling cases.
3.1 BES model
TRNSYS software can describe the transient thermal behavior both of the individual components and the complete ventilation system. Multi-zone building module (Type 56) is used to analyze thermal behavior of multi space. According to real geographical environment features, meteorological data of Matsumoto City has been chosen as TRNSYS boundary conditions. Table 1 gives the information of wall structure of the office model. Outer wall is a lightweight, well adiabatic and noise abatement structure. Non-humidity is considered in this work. Therefore, analysis is focused only on the sensible heat and the effect of latent heat is disregarded.
A detailed description of boundary condition parameter setting is given in Table 2.
PID (proportional–integral–derivative) controller (Type 23) is applied to minimizing the supply temperature fluctuation of air-conditioner by adjusting process control inputs. Several governments in East Asia issue recommendation to keep relatively higher room average temperature in summer to reduce energy consumption in buildings. The Ministry of Environment (MoE) of Japan has been encouraging people to set the temperature of air conditioners at office to 28 °C during summer. In our research, target zone temperature (Tave_B: perfect mixing temperature for BES; Tave_BC: occupied zone temperature for BES with CFD) is kept at 28 °C. The HRV (heat recovery ventilator) model (Type 91) with 60% exchanging co-efficiency validated in accordance with the experimental data are installed. As to the duct system, the heat loss or heat gain of the duct (Type 31) is not been taken into account.
Table 1 Wall structures
3.2 CFD model
CFD solves governing transportation equations to discretize the equation with finite-volume method. RANS (Reynolds Averaged Navier-Stokes) eddy- viscosity model (EVM) is applied in this coupling model. Based on Boussinesq approximation, buoyancy-driven force is treated as a source term in the Navier-Stokes equation. As typical EVM of RANS, standard k-ε turbulent model is selected with log low type wall functions to predict airflow distribution and thermal behavior. According to Ref. [10], it is well anticipated that the airflow and turbulence in enclosed environment can be calculated within acceptable computing cost and accuracy.
The second-order upwind scheme in which quantities at cell faces are computed using a multi-dimensional linear reconstruction is adopted for the convection term. SIMPLE (Semi-implicit method for pressure-linked equations) is used for steady-state analyses. The temperature fields are analyzed on the basis of convective and radiative heat transfers by using the coupled simulation of CFD-BES. Boussinesq approximation is adopted to reproduce the buoyancy effect by assuming a linear relationship between temperature and density. The view factors are calculated by a hemicube method, and an iterative solution of radiosity method is used to analyze mutual radiation. Unstructured grid with size function is adopted to configure desks, internal partitions and bookshelves.
Table 2 Analysis parameters of BES
Air supply from the air-condition system and HRV are assumed as constant airflow velocity and set as the values of maker’s catalogs. Workers and PC devices are simply assumed as the heat source at the floor and the desk surface, respectively. Lighting apparatus is evenly set at ceiling surface. Numerical and boundary conditions of CFD are summarized in Table 3.
4 Results and discussion
4.1 Energy consumption
Three types of BES-CFD cases with different supply opening positions of ventilation system are conducted to compare the results with BES-only simulation case (Case 0). Case 1 is to set up ventilation supply opening on the ceiling at interior zone in accordance with the existing office. Case 2 is to arrange the supply opening in perimeter zone. Case 3 is to arrange the supply opening at floor level. As shown in Fig. 3, the SA (supply air) temperature of HRV in both BES-only (Case 0) and coupled CFD and BES approach (Case 1-3) are in similar order at the peak time of the day as the result of PID control. RA (return air) temperature in Case 1 and 2 are about 1 °C higher than that of Case 0.
Figure 4 depicts that the average temperature in occupied zone (Tave_BC) estimated by CFD and perfect mixing temperature (Tave_B) calculated by BES become nearly equivalent due to PID control that sets the target temperature at 28 °C as well. Figure 5 shows time series of the supply air temperature of air conditioning estimated by BES (Tac_B) and the supply air temperature of air conditioning estimated by BES with CFD (Tac_BC). There is no obvious difference among Tac_B, Tac1_BC and Tac2_BC. Hence, the flow and temperature fields in this office model are assumed to be in approximately perfect mixing condition.
Table 3 Numerical and boundary conditions of CFD
Fig. 3 Time series of supply inlet and outlet temperature of HRV estimated by single BES and CFD_BES
Fig. 4 Time series of room averaged temperature estimated by single BES and occupied zone temperature by CFD_BES
Fig. 5 Time series of supply inlet temperature of air conditioning system estimated by single BES and CFD_BES
Under the initial condition, only sensible heat load is taken into account for the arrangement of opening for ventilation. It may cause inevitably inaccuracy results in temperature field estimation. Therefore, to neglect the effect of this uncertainty, the corresponding setting is identical for the three types of mixing ventilation.
Figure 6 illustrates the energy consumption of each case under the same daily operation schedule. The energy consumption of air-condition system in Case 1 and Case 2 analyzed by couple BES and CFD has no significant discrepancy compared with that of BES only. Case 3 shows a critical difference, about 21.27% larger energy compared with Case 0. The result indicates that relatively higher temperature airflow directly enters the target zone through SA on the floor and becomes a heat source to increase the cooling load.
Fig. 6 Energy consumption of air conditioning system estimated by single BES and CFD_BES
4.2 Airflow pattern
Figure 7 shows the sectional view of the airflow and temperature distributions. Non-uniform air flow distribution is confirmed only in the vicinity of supply opening. Stagnant and uniform
flow filed is formed in entire office space. Similar temperature distributions appear in both of ceiling supply cases. No significant difference of distribution is found between Case 1 and Case 2. In the occupied zone, the average temperature is around 28 ?C. In the upper zone close to the ceiling level, the air temperature is about 33 °C, 5 °C higher than the average room temperature. Since we assume the sensitive heats from human bodies and computers are generated from the desk surface, radiative heat exchange between those fixed desks and ceiling results in an non-uniform thermal range in the upper zone. In Case 3, a slight effect of the floor supply on the temperature gradients is observed. The temperature in the occupied zone varies from 27 to 28 °C.
Fig. 7 Temperature (a) and velocity (b) distribution of coupled cases in specific time
5 Conclusions
1) By transporting the mutual complementary boundary conditions in CFD package FLUENT and the system simulation package TRNSYS, the estimation of dynamic and the non-uniform indoor air flow performance is examined. The supply opening temperature of air-conditioner and heat exchanger are transmitted from TRNSYS and FLUENT as the boundary condition variable. FLUENT steadily calculates 500 iteration times until convergence and feedbacks non-uniform return air temperature of the heat recovery ventilators at each TRNSYS time step. Due to the unsteady heat condition through walls caused by different thermal conductivities, at least one-day pre-running period is needed in coupling simulation to obtain the relatively stable temperature gradient between two sides of the wall surfaces. The simulation result demonstrates the detailed information of dynamic air flow such as distribution of air temperature, velocity and wall surface heat flux.
2) Owing to the well mixed airflow situation, the discrepancy between BES-only and coupled BES-CFD is not obvious. There are no critical differences of airflow distribution and energy consumption between the two ceiling cases: ventilation supply opening installed in the interior zone (Case 1) or perimeter zone (Case 2). For floor supply ventilation system case (Case 3), a slight effect of the displacement on the temperature gradients is observed. However, an approximately 21% discrepancy for the heat recovery estimation in the maximum between the simulated results of coupled method and BES-only can be obviously found in the floor air supply ventilation case.
Acknowledgement
The authors would like to express deep appreciate to Mr. Keita Hattori who provided us valuable support on this research.
References
[1] NEGRAO C O R. Integration of computational fluid dynamic with building thermal and mess flow simulation [J]. Building and Environmental, 1998, 27: 155-165.
[2] ZHAI Zhi-qiang, CHEN Qing-yan, HAVES P. On approach to couple energy simulation and computational fluid dynamic programs [J]. Building and Environmental, 2002, 37: 857-864.
[3] ZHAI Zhi-qiang, CHEN Qing-yan. Solution characters of iterative coupling between energy simulation and CFD programs [J]. Energy and Building, 2003, 35: 493-505.
[4] ZHAI Zhi-qiang, CHEN Qing-yan. Performance of coupled building energy and CFD simulations [J]. Energy and Building, 2005, 37: 333-344.
[5] ZHAI Zhi-qiang, CHEN Qing-yan. Sensitivity analysis and application guides for integrated building energy and CFD simulation [J]. Energy and Building, 2006, 38: 1060-1068.
[6] DJUNAEDY E, HENSEN J, LOOMANS M. External coupling between CFD and energy simulation: Implementation and validation [J]. ASHRAE Transactions, 2005: 612-624.
[7] CRAWLEY D B. Contrasting the capabilities of building energy performance simulation programs [J]. Building and Environment, 2008, 43: 661-673.
[8] ARIAS D A. Advance on the coupling between a commercial CFD package and a component-based simulation program [C]// Second National IBPSA-USA Conference. Cambridge MA. 2006: 231-237.
[9] ANSYS. ANSYS/FLUENT 12 [M]. ANSYS Japan Ltd. 2009.
[10] KLEIN S A, BECKMAN W A, MITCHELL J W, et al. TRNSYS 17: A Transient system simulation program, SEL [D]. Madison, USA: University of Wisconsin, 2006.
[11] ZHAI Zhi-qiang, CHEN Qing-yan. Evaluation of various turbulence models in predicting airflow and turbulence in enclosed environments by CFD: Part 1: Summary of prevalent turbulence models [J]. HVAC & R Research, 2007, 13(6): 853-870.
[12] WARGOCKI P, WYON D P, SUNDELL J, CLAUSEN G, FANGER P O. The effects of outdoor air supply rate in an office on perceived air quality, sick building syndrome (SBS) symptoms and productivity [J]. Indoor Air, 2000(10): 222-236.
[13] SEPP?NEN O, FISK W J, LEI Q H. Ventilation and performance in office work [J]. Indoor Air, 2006, 16: 28-36.
[14] LAU J, NIU J L. Measurement and CFD simulation of the temperature stratification in an atrium using a floor level air supply method [J]. Indoor and Built Environment, 2003, 12: 265-280.
[15] BARTAK M, BEAUSOLEIL-MORRISON I. CLARKE J A, DRKAL D F, LAIN M, MACDONALD I A, MELIKOV A, POPIOLEK Z, STANKOV P. Integrating CFD and building simulation [J]. Building and Environment, 2002, 37(8/9): 865-871.
[16] SORENSEN D N, NIELSEN P V. Quality control of computational fluid dynamics in indoor environments [J]. Indoor Air, 2003, 13(1): 2-17.
(Edited by YANG Bing)
Foundation item: Project supported by Grant-in-Aid for Scientific Research (JSPS KAKENHI for Young Scientists (S), 21676005)
Received date: 2011-07-26; Accepted date: 2011-11-14
Corresponding author: Yun-qing FAN, PhD Candidate, Tel: +81-92-583-7628; E-mail: fanfangoogle@sina.com
