Diffuser model

Abstract

Numerical Modelling of Supply Openings for Room Air Distribution

Degree: Master of Science (Building Science)
Dept of Building, National University of Singapore

The inlet boundary condition of an air supply diffuser can be complex and modelling of airflow in indoor environment can become difficult. Computational Fluid Dynamic (CFD) method requires these complex quantities to be simplified into a certain mathematical format and, with the use of graphical interface, they were further simplified into a graphical representation. Such simplification can cause the lost of details of flow parameters around the supply. This dissertation elaborates the development of methods of modelling for the inlet supply without sacrificing the accuracy. The performance of the diffuser models was assessed based on the prediction of airflow pattern in the room as compared to the experiment result. The result shows that it is possible to develop an inlet model with limited data on the flow parameters while at the same time capturing the details of flow.

Keywords: square diffuser, diffuser modelling, CFD, boundary condition, numerical modelling, validation.

Summary

Since the introduction of numerical method in indoor air environment application more than twenty years ago, the method has gained a lot of attention as a promising method to complement the experimental and theoretical methods. However, it is not until recently that Computational Fluid Dynamic (CFD) could be made affordable for engineering consulting firm. CFD was known to be a privileged reserved for the researchers (and manufacturers) who have research and development funds.

The rapid advancement in computer technology reduces the cost of computer hardware substantially. The cost of CFD software package has also significantly reduced (more than half over the last ten years). In addition, the computing resource requirement for operating the CFD software has changed from a mainframe computer system to a desktop Pentium PC. This advancement makes the widespread application of CFD, including in indoor air environment problems, possible.

CFD packages have become more user-friendly. This means the user can just select some options provided in the package, without having to write any computer code, and it will solve the flow problem. With the advancement of computer graphics technology, the CFD packages become even more user-friendly. Specification of modelling parameters could be done graphically, instead of numerically.

However, this development can cause oversimplification of the flow parameters. Such oversimplification can result in the flow simulation to become unrealistic.

The inlet boundary condition is usually very complex for indoor air environment problem. CFD requires these complex quantities to be simplified in a certain mathematical format and, with the use of graphical interface, they were further simplified into a graphical representation.

Air diffusion in a room depends on the diffuser type and the air supply parameters of the diffuser. Reliability of room airflow simulation depends on the correct description of airflow around diffuser. There is a critical need to develop a suitable modelling technique for the air diffusers.

This dissertation elaborates the development of methods of modelling for the inlet supply by using the graphical method without the need of writing a computer program. The standard procedure to specify an inlet supply in a CFD package is to define a plane with a certain shape and a certain area with a velocity vector assigned to the plane. A diffuser could be divided into one or more such planes. This study uses this standard procedure and elaborates some ways to capture as much detail as possible with minimum data requirement on the airflow around diffuser.

Five diffuser models are developed and used to simulate airflow pattern in a Seminar Room. Measurements of air velocity were also carried out in this room and airflow patterns in some critical locations were also determined by smoke visualization tests. The performance of the diffuser models was assessed based on the prediction of airflow pattern in the room.

One of the diffuser models shows very good similarity of airflow pattern between the predicted and experimental result. Airflow rate is the only data needed to develop the model. This is an important feature since the other parameters, e.g. air velocity below the diffuser, are very difficult to obtain.

This study includes the measurement for validation of the simulation result. Indoor air velocity measurement was found to be very difficult to perform. With the limitation on the instrument capabilities, the accuracy of the numerical predicted results are not very good as compared to the measurement results. The predicted results always underestimate the air velocity in the room. However, qualitatively the numerical prediction still gives a consistent result with the visualized airflow pattern and measured data.

Further Study

The above models has not been properly validated in my MSc research as the measurement data resulted from the experiment was not qualified to be the validation data due to the instrumentation limitation. However, after submission of the dissertation, I found a published measurement data that can be used for validating the models. This finding encouraged me to do another study to validate the models.

I conducted a series of simulation to repeat the whole experimental protocol described in the published data, and compare the measurement result with the simulation result. The result of the validation shows that one of the model shows a very good agreement with the experimental result.

The main contribution of this study (and also my MSc research) is not the model but the method to develop the model. The simplified method requires only the design airflow rate of the diffuser to develop the diffuser model. This is certainly an advantage as we can avoid the air velocity which is very difficult data to obtain.

This study has been published in a refereed journal.