This thesis aims to increase the knowledge of simulation of thermal indoor climate for nearly zero energy buildings in a sub-arctic climate. Air heating systems in cold climate generate temperature gradients, which negatively affects the thermal indoor climate. Stand-ard multi-zone modeling has problemswithpredicting these gradients.

In this work, Computerized Fluid Dynamics (CFD) simulations are used to model the tem-perature gradients. The consequences of reducing the cell sizes for the simulation volume are estimated and case studies of different building and heating systems are presented. The CFD method is validated for a traditional underfloor heating system and also for an air heating system.

Furthermore, the effects of snow on heat losses for common building foundations are in-vestigated, and snow is shown to be an important boundary for CFD simulations of a build-ing. The snow and ground freezing areshown to reduce the annual heat losses between 7-10%.

The CFD method is shown to be a suitable method for predicting thermal indoor climate. The method can determine the temperature variations inside a building, for different rooms, floors and heating systems. The CFD method is most appropriate for local distributed sys-tems. For traditional hydronic systems the method may be overambitious,since a good indoor climate is usually achieved anyway.

Heat transfer coefficients are inaccurate when calculated using standard wall functions used in many turbulence models (like the k-ε model) for surfaces with a high heat transfer rate and natural convection. Automatic wall functions have shown better accuracy for this type of problem, but they require more cells. In order to still use the k-εmodel, a user defined wall function is investigated. This method gave good results and a significant re-duction in the number of necessary cells in the simulation volume. The validation of the indoor climate shows that the wall boundary conditions are important for predicting the indoor temperaturevariations for steady state simulations.

New buildings have a higher thermal inertia, which affects the heat losses. It is important to include this effect in the boundary condition calculations for a CFD model.

The CFD simulations show that air heatingand local distributed heating systems have dif-ficulties infulfillinga good thermal comfort inside all rooms. This is especially true for rooms with exhaustair and closed doors and multi-storybuildings. Results from a CFD simulation can be used to improve the thermal comfort in these cases.