Flow Boiling Heat Transfer and Pressure Drop. Experiments and Model Development for Complex Geometries

Doctoral thesis, 2004

Two very different evaporator geometries, a dimple plate and a herringbone microfin tube, have been studied. Heat transfer and pressure drop for both pure refrigerants and refrigerant mixtures have been determined experimentally. For the purpose of modelling heat transfer, the general flow boiling heat transfer correlation by Steiner and Taborek has been modified to take the different geometries into account. The dimple plate (3.0 m x 0.5 m) was taken from the evaporator of a large district heating heat pump. This sub-project was part of work dealing with the identification of possible substitutes for R22. Hence, a test facility was built which enabled global measurements of heat transfer and pressure drop. R22, R134a and three mixtures were investigated. Measured overall heat transfer coefficients were generally higher for pure fluids than for mixtures. Operational aspects, such as ice formation, are discussed. The model developed overestimates heat transfer by 12% for pure fluids and by 15% for mixtures. Pressure drops are calculated to within 5% of measured values. The herringbone microfin tube is a relatively new type of microfin-enhanced heat exchanger tube, typically used in direct expansion refrigeration evaporators and condensers. The tube is 3.8 m long and has an outer diameter of 9.53 mm. The main objective here was to increase knowledge about herringbone microfin behaviour in terms of heat transfer and pressure drop. For this purpose a water-heated test section was constructed that enabled local measurements of heat transfer and pressure drops. Experimental results for R134a, R407C and R410A show that heat transfer coefficients for all fluids improve with increasing mass velocities, and peak generally at vapour qualities around 60%. A clear heat flux dependence was not found. Heat transfer coefficients for R410A and R407C, at equal mass flow rates and temperatures, are in general lower than coefficients for R134a. Possible causes of this are discussed. Also a heat transfer prediction method is proposed, which is based on a hypothesis that explains measured heat transfer peaks. The method predicts heat transfer coefficients for R134a, R410A and R407C with average residuals of -5%, -2% and 17% respectively.