Łukasz Borkowski
Combustion Engineer · Propulsion Engineering
Warsaw, Poland
Combustion Engineer specializing in CFD-based combustor design and emissions prediction for aircraft gas turbines. Has hands-on experience with reacting flow simulations using flamelet and EDC combustion models, and with correlation of CFD results against test rig emissions measurements. Focused on lean premixed combustor development to meet ICAO CAEP/8 NOx limits.
Expertise
- combustor CFD with reacting flow
- NOx and emissions prediction
- lean premixed combustion
- thermoacoustic stability
- experimental combustor rig testing
Technologies
Work History
2024-11
Combustor rig test support — instrumentation planning, CFD pre-test predictions for 8 test conditions, and post-test correlation. Led the CFD-test delta analysis and root cause investigation for two outlier test points.
Challenge: Two test points showed 30% higher CO than CFD predictions. Investigation revealed that the fuel atomizer was partially blocked on the test rig — a hardware issue not captured in CFD. The lesson was that CFD-test mismatch is not always a model error.
Learned: Always investigate hardware condition before attributing CFD-test discrepancy to model error. Post-test teardown inspection should be a mandatory step in any rig test correlation study.
2024-03
Thermoacoustic stability screening for the new lean premixed combustor using ANSYS Fluent acoustics module. Identified two potentially unstable acoustic modes and assessed fuel injection timing sensitivity.
Challenge: Thermoacoustic instability is inherently transient — RANS steady-state analysis cannot capture it. Had to switch to URANS with a time step fine enough to resolve the 150-400 Hz unstable frequencies. Compute cost was 40x higher than steady analysis.
Learned: Thermoacoustic screening cannot be done with steady RANS. Budget for URANS or LES from the start if lean premixed combustors are involved — steady analysis will miss the instability modes that can cause catastrophic combustor damage.
2023-08
NOx emissions correlation study — compared ANSYS Fluent CFD predictions against emissions rig test data across 10 operating conditions (varied FAR, P3, T3). Identified systematic bias in NOx prediction at high T3.
Challenge: CFD underpredicted NOx at T3 > 820K by 18-25%. Root cause: the flamelet library was generated at standard pressure without accounting for NOx formation rate pressure dependence above 30 bar. Regenerated flamelet tables at representative P3 levels.
Learned: Flamelet tables for high-pressure combustors must be generated at representative operating pressures. Using atmospheric-pressure flamelets for a 30+ bar combustor systematically misrepresents NOx chemistry.
2023-02
CFD analysis of a TAPS (Twin Annular Pre-Swirl) combustor using ANSYS Fluent with the flamelet/progress variable (FPV) combustion model. Assessed NOx formation distribution and pattern factor at combustor exit.
Challenge: FPV model convergence was extremely sensitive to initial conditions — starting from cold flow without pre-conditioning diverged in every attempt. Developed a staged ignition sequence: start with simplified EDC model, then switch to FPV once the flame is established.
Learned: Reacting flow CFD for combustors requires a careful cold-to-hot solution strategy. FPV models in particular need a well-established flow field to initialize from — the right startup sequence is not documented in solver manuals and must be developed empirically.