In thermal models, the ventilation airflow rates normally arc input parameters, to be defined by the user or to be calculated by the program on the basis of a nominal air exchange (or flow rate) and some control parameters (demand — controlled ventilation, variable air volume flow ventilation systems). In airflow models, on the other hand, room air temperatures must be defined in the input (see Fig. 11.49).

Stepwise procedure

Thermal simulation

Natural ventilation

Input: Air change rates as boundary condition

Input: Temperatures as boundary condition

Output: Temperatures Thermal loads

Output: Air change rates

Integrated approach

Airflow model integrated into thermal model

Input: Thermal and aerodynamic properties Loads, climatic conditions, etc.

Output: Temperatures and airflow rates

| FIGURE 11.49 In thermal models, the ventilation airflow rates are input parameters. In airflow models, on the other hand, room air temperatures must be defined in the output. Since natural ventila­tion airflows and room air temperatures are interdependent, both parameters must be integrally consid­ered In the solution process. This is possible only by an integration of the natural airflow model into the thermal model.

However, in thermally driven naturally ventilated enclosures, airflow rates and room air temperatures are interdependent and can be determined only si­multaneously by using an integrated airflow and thermal mode!.} 4

For the data exchange between the airflow and the thermal model, several levels can be distinguished:

1. Sequential coupling. Here the link between one model and the other has to be established by the user, e. g., by defining specific input and output files for the airflow and the temperature results respectively and by proceeding in an iterative way toward a solution.

2. “Ping-pong” coupling. Here the results of one model for a certain time step are used in the other model for the calculation in the next time step. In the iterative solution process at a certain time step, however, there is no data exchange between the two models. This means that the calculated airflows and the room air temperatures physically do not fully comply. The ping-pong cou­pling may be established by the user or may be realized in an interoperable environment where the data exchange between the airflow and thermal models is automated.

3. Integration. Here, at a certain time step in the simulation, the airflow model is repeatedly called within the iterative solution process in the thermal model, and the airflow results are considered within this iterative solution process. Thus, the resulting airflows and room air temperatures fully comply in terms of the underlying physics.

Also, for this option, several levels can be observed concerning the integra­tion of the airflow model. The lowest level of integration still requires separate input for the thermal and the airflow model respectively with a high degree of redundancy in the input parameters (e. g., zones must be input for both models), and the connectivity of the airflows and the zones in the thermal model must be established manually by the user. More sophisticated levels have reduced redun­dancy and automatic establishment of the link connectivity.

A higher level of integration is achieved when the airflow model is fully in­tegrated in the thermal model and the input for the airflow part is just an addi­tion to the input required for the thermal model.

The highest level of integration would be to establish one large set of equa­tions and to apply one solution process to both thermal and airflow-related variables. Nevertheless, a very sparse matrix must be solved, and one cannot use the reliable and well-proven solvers of the present codes anymore. Therefore, a separate solution process for thermal and airflow parameters respectively re­mains the most promising approach. This seems to be appropriate also for the coupling of computational fluid dynamics (CFD) with a thermal model.

For the models described, the usual assumption for air nodes in regard to the room air distribution is still valid. This means that each air node represents a vol­ume of perfectly mixed air. Thus, the same limitations as for thermal and airflow models apply: Local air temperatures and air velocities as well as local contami­nant concentrations can be neither considered nor determined. This also means that thermal comfort evaluations in terms of draft risk cannot be performed.

II time-dependent spatial airflow and temperature distributions in a room are requested, the coupling of CFD with a thermal model has to be consid­ered.1