Comparision of high fidelity and real-time CFD methods for simulating thermal comfort

Background

Numerical flow simulation has been successfully used for many years to calculate temperature and velocity fields in residential areas. The challenges that arise lie to a large extent in the correct choice of boundary conditions, whereby the energy input via solar radiation in particular, but also the radiation exchange between surfaces in general, should be emphasized. In most cases, a stationary, i.e. time-independent solution of the temperature and flow velocity field can be obtained, assuming the correct wall properties and temperatures on the outer surfaces. In some cases, however, a transient consideration must be carried out. The correct modelling of air inlets is also of great importance: turbulence degrees of the velocities at air inlets are normally not known and must be estimated, a correct modelling of the close-up range of the nozzles is essential in order to be able to obtain the correct temperature-dependent flow field.
The steady-state solution of the Navier Stokes equations requires considerable computing time, depending on the size of the room and consideration of additional physics (radiation exchange, turbulence, humidity), which - depending on the number of processors used - is several hours. However, a considerable part of every simulation project lies in the geometry and computational mesh creation. In this area, the manufacturers of commercial CFD tools have invested a lot in recent years, so that computationally powerful, highly automated methods are available.
While the use of GPUs has been steadily expanding in recent years, real-time GPU-based flow solvers (e.g. ANSYS Discovery Live ) have only recently become available. Neglecting or simplifying computationally intensive partial aspects, it is possible to calculate the influence of changes in geometry or boundary conditions in real time. This possibility is to be used for the calculation and visualisation of temperature and flow and ultimately of thermal comfort. The extent to which the simplifications required for this influence the accuracy and informative value of the simulations is subject of the thesis, in the context of which the real-time results will be compared with high-fidelity (conventional) CFD methods.

Thesis/Learning objectives

After the thesis has been performed the student should be able to:
- Identify and describe a gap in knowledge
- Perform a comprehensive literature review
- Describe the essential indoor thermodynamics including the influence of radiation
- Utilize real time CFD tools and high fidelity CFD methods
- Make informed conclusions supported by documented observations

Method of attack

Following a literature review, several simplified geometries of rooms in buildings will be selected. After generating computational meshes, the indoor comfort will be assessed by the use of high fidelity CFD methods as well as utilizing real time CFD approaches. The results will be compared and the necessary simplifications described and reviewed.

Proposed time schedule, including milestones and intermediate reports

The thesis is expected to start under January 2022 (wk. 3) and be completed in June 2022 (wk. 23). Intermediate reports will be due in early March (wk. 10) and late April (wk. 17).

Contact persons

Supervisor at KTH, Department of Energy Technology

Supervisor at AIT, Center for Energy
Christoph Reichl , Senior Scientist christoph.reichl@ait.ac.at