Abstract
M.Ing.
Dry air cooled heat exchangers form a vital part of industrial heat transfer systems, especially in countries where the supply and availability of clean cooling water is limited. Headerboxes are rectangular pressure vessels that act as the inlet distribution and outlet collection devices. As rectangular pressure vessels, headerboxes are subject to design codes such as ASME (ASME Section VIII, Division 1, 2007).
Unfortunately ASME (ASME Section VIII, Division 1, 2007) offers no guidance on how to allow for the effect of external loads applied to the headerbox through the vessel’s nozzles. This creates a difficult situation, since vessel designers are mandated by ASME to consider the effects of nozzle loads by American Petroleum Institute standard 661 (API, 2006).
The aim of this project was therefore to develop a closed form design methodology that accurately predicts the stresses inside a headerbox that is subject to external loadings as well as internal pressure. After extensive research it was decided that the only viable approach would be to extend ASME’s rigid frame theory. This was done, and a new set of equations describing the stress distribution inside a headerbox were derived.
These equations were then tested using 2D Finite Element Analysis (FEA) to determine whether they represented the reality of the 2D model they described. It was found that the equations were accurate enough in 2D and the next step was to test the model experimentally and using full 3D FEA.
A local manufacturer of air cooled heat exchangers was approached and they helped design an experimental specimen and agreed to fund its construction. Unfortunately, due to time constraints, it was not possible to build and test the specimen experimentally. The specimen geometry was then analysed using the Abaqus (Dassault Systѐmes Simulia Corp., 2010) FEA package.
The 3D FEA analysis considered several different load cases. After carefully analysing the results it was seen that the rigid frame model could make useful qualitative statements about the effects of the nozzle loads, but it performed poorly as a quantitative prediction method. However, since the effects of the nozzle loads are generally quite small it is possible that, with appropriate safety factors, the rigid frame model could be used as a conservative design methodology. The usefulness of a commonly used empirical guideline was also examined.
This project is far from conclusive and much more work is required to fully examine the usefulness of rigid frame theory. That being said, this project has made important steps towards a more complete understanding of rectangular pressure vessels and has shown possible ways forward.