mollierSolver - A CFD Solver for Phase Changes
mollierSolver is an unsteady, density based RANS-solver which models the Mollier-T-H-X-diagram. It was developed to be used with OpenFOAM. The first implementation was designed for the system ice/water/vapour from -50 to 100°C. All possible phase changes in the carrier fluid (air) are modelled, i.e. evaporation, condensation, melting, crystallisation, sublimation and re-sublimation. The solver is extremely versatile and applicable for any condensation and evaporation simulation (not only those of ice/water/vapour). In a development of the solver we adapted it for the use with arbitrary substances.
The use cases for mollierSolver are manifold. In principle it can be used for all problems where phase change is relevant for the solution. This could be problems in AC-Systems (humidity in clean rooms, rooms with electronic equipment, ...) or in industrial cooling processes. Unwanted condensation (e.g. recuperation of off-heat in dryers, cooling in film-extrusion plants, condensation on walls and the ceiling in production facilities etc.) can be simulated and the problem further optimized.
We designed our code with flexibility in mind so that we can tune a number of model parameters to deal with issues like adsorption / desorption problems and conjugate heat transfer from/to solids. This allows us to define specific evaporation and condensation rates for arbitrary parts of geometries including the full effect of heat transfer from fluid to solid phases. Effectively this gives us a model for adsorption (chemisorption and physisorption) for CFD simulations. Want to know how the fluid mechanics influence the activity of a catalyzer, optimize a fixed adsorption bed? There is a plethora of industrial applications from protein adsorption on implants and other biomaterial to adsorption chillers where these mechanisms play a crucial role.
A practical example that everybody has been confronted with is de-fogging of a windshield. Following video shows defogging of the windshield and the driver side window under heavy winter conditions.
The images of the simulation results show the distribution of relative humidity in the range of 35 to 100% in the interior and on the fogged windows of a sports car 10 seconds after turning on the front fan and heating. When the relative humidity reaches 100% (red areas) water condenses as small droplets and hinders free view through the windshield. Warm air reduces relative humidity (blue areas) causing water droplets to evaporate and allowing for a free, unobstructed view.
If you ask yourself, why the blue areas are de-fogged after 10 seconds already: our simulation does not take into account the warm-up phase of the engine. In the simulation warm air is immediately available. Of course it would be possible to account for this effect.
From the driver's perspective:
A model was developed to simulate realistic influence of persons in the car that breathe and sweat. Other sources of humidity are implemented in the code.
Validation of CFD-Code
An experimental set-up for fogging and de-fogging was designed to test the performance of mollierSolver under real world conditions. The experiment was done in a sealed glass chamber containing a polished metal mirror. The mirror was well insulated on its backside and kept at constant temperature (+/- 0.01 K) by means of a very sensitive thermostat. From the top of the glass chamber we then blew in an impinging free-jet at a certain temperature with constant humidity just a bit below its saturation loading. When the humid air jet hits the cool mirror it cools down and water precipitates out of the jet onto the mirror. The condensation and evaporation rates to and from the metal surface are measured (by recording the energy that is needed to keep the polished metal surface at constant temperature) and compared with simulation results.
In essence the experiment is very similar to what you experience when you switch on the hot water in a cold bathroom: warm, humid air rises and the mirror above the tap fogs over within a couple of seconds.
The video below shows the experiment in action: the left panel shows the polished metal mirror (at the beginning you can even see the camera stand, if you look closely), the middle panel shows the condensation in the simulation (blue means dry, the more red the surface becomes, the more humidity condensed) and the graph on the right side compares the total condensed mass on the metal surface of the experiment (light blue) and the simulation (dark blue).