Numerical upgrading and experimental verification of MatriHX, a matrix-based solver for plate-and-tube heat exchangers

This thesis project is part of a wider research, within the collaboration between Deca S.r.l and Politecnico di Milano, about new commercial technologies of heat transfer enhancement. The selected solution envisages the application of finned inserts inside tubes. The insert promotes the heat exchange, with the objective of diminish the size of the exchanger guaranteeing the same performances. Within this work, an upgrading of the simulation code, the MatriHX, is performed. New circuitry is implemented, in order to simulate aftercoolers performances. Four configurations are investigated, sizing the exchanger in order to match the performances of the conventional product. Two of the four simulations underline what the most delicate point of this technology is, obtaining a surface reduction up to 64%; but an increase of pressure drops of over 170%. On the other hand, the other two simulations demonstrate how insert can be promising, diminishing the encumbrance of ≈ 40% with, in the best case, a decrease of pressure drops of 20%. Moreover, an improvement in pressure drops calculation is added. Concentrated pressure drops and drops in curves across passages are now considered in the exchanger simulations. The new version of MatriHX is compared successfully against Aspen EDR, in order to validate the modifications performed. The second goal of this work is the design and assembly of an experimental apparatus, whose final aim is to test the performances of the finned insert. The project of the test bench is finalised, the heat exchange section is commissioned and manufactured, the instrumentation is selected and supplied. A first experimental campaign is executed, in order to validate the behaviour of the apparatus in standard conditions, i. e. internal smooth tube. Two main cases are effectuated, varying the inlet air mass flow rate. On the single test, a comparison between air and water thermal powers is performed, in order to validate the test itself. Experimental results highlight a mismatch between thermal powers up to 26% for the low Reynolds case, and 22% for the high Reynolds case. A comparison against MatriHX is effected, obtaining a relative deviation between air thermal power and the expected one < 2%.