Continuous vs. pulsating flow boiling. Part 2: Statistical comparison using response surface methodology
flow boiling, flow pulsations, heat transfer enhancement, response surface methods, experiments
Design of experiments is a method for experimentation that enables the possibility to estimate effects and responses of certain factors and their interactions with statistical significance and low experimental effort. This approach includes several phases from sorting out the trivial from the vital factors, characterizing main effects, interactions, detecting non-linearity and establishing functional relationships (response surfaces) for optimizing factors and comparing categorical responses etc.. The goal of the current paper is to compare the competitive categorical heat transfer responses obtained in part 1 of the paper, using continuous and pulsating two-phase flow boiling. For in-tube continuous flow boiling in traditional tubing, the governing vital factors have been known for decades (mass flux, heat flux, vapor quality, temperature, diameter and fluid), however, for pulsating flow boiling the injection frequency (or cycle time) must be characterized too. The response surfaces are compared and tested for statistical significance in order to verify an improvement or a reduction in the time-averaged flow boiling heat transfer coefficient statistically when introducing flow pulsations. The test section consists of the two exchangeable expansion valves and a tube-in-tube evaporator with four subsections for measuring the heat transfer coefficient. Refrigerant R134a flows internally and water flows externally. The inner tube internal diameter is 8 mm and each subsection is 0.25 m long. Temperatures in and out of each subsection are measured as well as the water flow rate. Wall temperatures are measured outside on the internal tube in the top and the bottom of each subsection as well as refrigerant temperature, pressure and flow rate. The evaporation temperature is kept at 5Â°C during the experiments and the fluid state before the expansion valve is kept at a condensing temperature and subcooling of 32.5Â°C and 2 K, respectively. The time-averaged refrigerant mass flux varies from 41 to 167 kg/m2s, vapor qualities from 0.18-0.59 and heat fluxes from 0.8-45 kW/m2. The paper builds upon part 1 of the paper and includes a description of the response surface methodology, its usage and results including main effects, interactions, and response surface comparison. The main object is to compare the response surfaces for continuous and pulsation flow boiling and conclude whether pulsations may increase or decrease heat transfer with statistical significance.