A new image analysis technique is proposed to determine the macromixing time in a transparent stirred tank. It consists of capturing on video a decolorization process by using a fast acid−base indicator reaction and employing image analysis to quantify the color evolution. The color change is quantified by means of individual thresholds on the RGB color model and provides a direct measurement of the macromixing evolution as it can be seen by an operator in front of the vessel. It is shown that this technique removes the subjectivity of the estimation of macromixing time by the naked eye, has a high degree of reliability and repeatability, and can yield accurate macromixing information by considering the possible presence of segregated regions and dead zones. Moreover, applications show that the macromixing curves bring new insights to study and compare mixing efficiency of different impellers or multiple impeller mixing systems.

No abstract

In this paper,experimental and numerical studies of heat and mass transfer in a falling film absorber are presented. The investigated absorber is a plate heat exchanger used ina falling film configuration. The ammonia-water solution flows in a falling film mode along the plates. The vapour flows co-current with the falling film and the coolant fluid is in a countercurrent flow with the falling film. A prototype of ammonia-water absorption chiller isused to experimentally study the absorber behaviour in real operating conditions. Amacro study of the absorber and a local analysis deduced from local temperatures measurements along the falling film are presented. A numerical model and a simulation tool aredeveloped in order to complete the experimental investigations.The associatednumerical parametric studyaims to separatethe coolant mass flow rate impact.The model is validated with experimental dataand a maximal relative error of 15 % is observedbetween experimental and numerical results. The results of this study suggest that during the absorption process,mass transfers are controlled by the falling film mass transfer resistance and that the liquid-side heat transfer resistance is negligible.

• A commercial absorption chiller has been tested in dynamic operating conditions on a semi-virtual test bench. • The absorption chiller was modelled in a dynamic way using artificial neural networks. • The model is validated using experimental data. • The neural model predictions are very satisfactory, absolute relative errors of the transferred energy are in 0.1-6.6%.