What is it about?

Flows of carbon dioxide coolant around clusters of pipework, such as those found in the boilers of EDF Energy's civil nuclear reactor fleet in the UK, are poorly understood. This is due to the complicated geometry, the flow being in a turbulent regime, and the effects of intense heat and radiation over a prolonged period changing the surface texture of the boiler pipes, leading to variable heat transfer rates everywhere. This article covers simulations of an idealized geometry which bears some similarity to that used in the Advanced Gas-Cooled Reactors under representative conditions, intended to improve understanding of the flow and heat transfer rates in an ideal case, before introducing other complicating factors discussed in my doctoral thesis.

Featured Image

Why is it important?

EDF Energy is the sole nuclear energy operator in the UK, and as their reactors are aging with few being replaced in the near future, they require information on the flow and heat transfer characteristics of the overall system to make the safety case for extending the lifespan of the existing fleet. This is both to ensure the UK's energy needs in the short-term and provide the government enough opportunity to put together its long-term Nuclear New Build strategy to improve the UK's energy security in the future. A key finding was that flow at the representative conditions has a tendency to be deflected from navigating arrays of tubes in a straight-through manner to a diagonal trajectory, which is not accounted for in EDF Energy's in-house flow diagnostics tools.


There has been much discussion as to the optimal strategy to reliably simulating flows about bundles of tubes, of which this text forms a small part. Provided computational expense is of no object, dynamic Large Eddy Simulation (LES) appears to be the only method that consistently captures the complex features exhibited by these intricate flows. Unusually, the high-Reynolds number class of turbulence models, which are widely considered to be less descriptive of turbulence than their low-Reynold number counterparts, appear to capture the diagonalized behavior of the flow along with the LES, but at the expense of accurate wall shear stresses, whereas the low-Reynolds number models return reasonable wall shear stresses but do not predict diagonal flow. It seems that a definitive consensus has yet to be reached, but the addition of the Analytical Wall Function strategy endorsed by the paper may yet prove to overcome the shortcomings of the high-Reynolds number models.

James Blackall
University of Manchester

Read the Original

This page is a summary of: Modeling of In-Line Tube Banks Inside Advanced Gas-Cooled Reactor Boilers, Heat Transfer Engineering, August 2019, Taylor & Francis, DOI: 10.1080/01457632.2019.1640486.
You can read the full text:




The following have contributed to this page