ESDU Engineer

CFD Prediction of Internal Flow Pressure Losses

ESDU’s Fluid Mechanics, Internal Flow group has for nearly 40 years been developing methods for calculating pressure losses in internal flow system components. These are industry-standard methods, and are based on the analysis of high-quality experimental data and analytical methods. However, there remain component geometries and regions of operation where no data exist. The use of Computational Fluid Dynamics, or CFD, to generate pseudo-experimental data to fill in gaps has long been mooted at ESDU. With recent advances in computer power and CFD code capabilities, ESDU has for the last year been attempting precisely that, with very encouraging results.

However, it is important to emphasize that the technology is at the stage where it can be used generally for such calculations only by highly-experienced practitioners with a solid understanding of fluid mechanics. It is essential for such calculations to be carefully validated against experimental data. Extremely rigorous validation procedures have been followed by ESDU, which have demonstrated that in order to be able to capture the complex details of internal flows the mesh and turbulence modelling sensitivity analyses need to be far more thorough than for most practical CFD applications.

The project to date has concentrated on the prediction of pressure losses across contractions. These components are widely used in industry, and reliable experimental data are scarce, particularly for a wide range of ratios of the two pipe diameters, for the different flow conditions and for various edge geometries (such as chamfered edges). For example, in laminar flow published experimental data of the wall pressure measurements required to derive the pressure loss coefficient are very few, and some measurements differ by more than 100%. Velocity measurements using Laser Doppler Anemometry (LDA) are generally most reliable as the flow field is undisturbed and the temporal and spatial resolution is high. This makes them particularly suitable for CFD validation.

In sudden contractions, the flow separation in the vicinity of the contraction plane causes an increase in pressure loss, which affects erosion rates and heat and mass transfer rates at the separation and reattachment regions. Even in axisymmetrical geometries, the flow is three-dimensional and unsteady.

The initial validation studies have been undertaken under the guidance of ESDU’s Internal Flow Panel. Also, ESDU is a member of the European Thematic Network QNET-CFD and has successfully submitted a validation report to the rigorous QNET-CFD review panel. Details of this work can be found on the web site (www.qnet-cfd.net) as part of the QNET-CFD Knowledge Base.

The CFD predictions were generated using an advanced commercial CFD package: CFX 5.6. The calculation results were compared with published experimental and numerical data for a range of flows and geometries. The CFD results predicted well the main flow features observed and measured experimentally using the LDA techniques, both qualitatively and quantitatively, and were in closer agreement with measured data than other published numerical results. The pressure loss coefficient for the different contraction ratios and flow regimes was well predicted within the experimental uncertainty, as shown in the example.

The use of CFD to analyse the flow through sudden contractions has helped improve our understanding of the flow featureswhere research meets designin a systematic manner, and provide information on pressure loss, flow separation regions, and redevelopment length, especially for those geometries and flow conditions where published experimental data are not available or inconsistent. The experience gained can now be used to supplement the available knowledge of pressure losses in other duct components.

As part of the update to Data Item 01016 on pressure losses in flow through a sudden contraction of duct area, a comprehensive guidance on the CFD methodology will be given, outlining the steps taken by ESDU in this project and giving recommendations for Best Practice CFD modelling.

Francesca Iudicello is a Senior Engineer working in the Thermofluids Group. She can be contacted at fiud@esdu.com