Numerical Investigation of Steady-State Thermo-Fluid Performance of a Reducer for a High-Temperature Dynamic Pressure Sensor Calibration Rig : Minimizing Heat Loss and Pressure Loss in a Combustion-Dynamics Calibration Rig

Sammanfattning

High-temperature laboratory test rigs are essential for evaluating components and measurement systems intended for gas-turbine-related applications. In Siemens Energy’s Pulsatorn dynamic pressure calibration rig, the reducer connecting the air heater to the downstream sensor section was identified as a likely source of both aerodynamic losses and unwanted heat dissipation. The legacy reducer is a commercially available hydraulic fitting that produces two abrupt contractions (35.05 mm → 11.25 mm → 6.0 mm) and was selected primarily for mechanical compatibility rather than thermo-fluid performance. A redesigned reducer was therefore developed using a quintic C²-continuous contraction profile over a solid length of 150 mm, providing a smooth transition from the 35.05 mm heater bore to the 6.0 mm outlet. The project was initially intended to include experimental high-temperature testing on the Pulsatorn rig. However, heater hardware failure prevented sustained testing within the available timeframe. With approval from both the industrial supervisor at Siemens Energy AB and the academic supervisor at KTH Royal Institute of Technology, the scope was revised to a comparative steady-state numerical investigation under controlled and repeatable boundary conditions. The work comprised two studies. Study 1 established mesh independence for both geometries using compressible adiabatic CFD across three refinement levels. The selected baseline meshes contained 829,621 fluid cells for the legacy reducer and 913,301 for the redesigned reducer. At the most demanding pressure condition, the legacy reducer consistently exhibited a localised supersonic vena-contracta pocket (Ma = 1.010–1.016), whereas the redesigned reducer operated near-choked (Ma = 0.989–0.990), confirming that the flow-regime difference was not a numerical artefact. Study 2 comprised six steady-state conjugate heat transfer simulations for both geometries at outlet gauge pressures of 80,000, 40,000, and 0 Pa, with an inlet condition of 1 bar(g) and 673 K. Under uninsulated conditions, the legacy reducer delivered slightly higher outlet static temperatures, by 7.80 K, 4.75 K, and 3.36 K across the three cases, because its external surface area was 3.35 times smaller and therefore lost less heat to the surroundings. However, the redesigned reducer consistently showed 9.5–14.3% lower external heat flux per unit area, indicating reduced local thermal loss intensity associated with the smooth contraction. At the supercritical pressure-ratio condition, the redesigned reducer reached Ma = 0.990 while the legacy reducer reached Ma = 1.006, confirming a clear flow-regime asymmetry between the two designs. Additional insulated reference simulations at the 0 Pa case reduced the outlet-temperature difference from 3.36 K to 1.24 K and reversed it in favour of the redesigned reducer, showing that the uninsulated thermal deficit of the redesigned geometry arose predominantly from its larger external surface area rather than inferior internal flow behaviour. The results show that the redesigned reducer provides the superior internal flow field and lower heat-loss intensity per unit external area, but that its longer body introduces a thermal penalty under uninsulated operation. External insulation is therefore identified as the key practical measure required for the redesigned geometry to realise its full thermo-fluid advantage in the Pulsatorn rig.