Uppsats
Design and Performance Analysis of a Centrifugal Compressor Impeller for Micro Gas Turbine Applications
Master-uppsats
KTH/Skolan för teknikvetenskap (SCI)
Publicerad: 2026
Språk: Engelska
Nyckelord
klicka för att sökaSammanfattning
This thesis investigates the design and performance prediction of centrifugal compressorsfor micro gas turbine (MGT) systems, a technology domain where compact, efficientenergy conversion is central to decarbonisation efforts. For MGTs in the 25–500 kWrange, rapid aerodynamic iteration is essential, yet traditional design approaches relyingon full Computational Fluid Dynamics (CFD) or proprietary industry tools are too slowor restrictive. Motivated by the industrial needs of Mitis, this work develops a modularone-dimensional mean-line framework capable of fast and physically consistent evaluationof centrifugal compressor stages.A general mean-line solver was implemented using conservation laws and empirical losscorrelations. The tool was validated against two reference cases: the canonical EckardtO-Rotor and the modern NASA High Efficiency Centrifugal Compressor (HECC). Acrossboth cases, the solver reproduced the characteristic performance trends, with pressureratio and efficiency deviations consistent with the expected accuracy bands of one-dimensional methods. The validated tool was then used to generate a compact impellerdesign tailored to Mitis’s requirements, yielding an exit shroud diameter of 0.087 m anda blade height of 5.5 mm.Steady Reynolds-Averaged Navier–Stokes simulations employing the SST 𝑘–𝜔 turbulencemodel were performed to verify the design. The mean-line tool predicted an impellerisentropic efficiency of 85.1% at a total pressure ratio of approximately 4.0, while thecorresponding CFD result showed an efficiency near 90% and a pressure ratio of 4.44.Flow-field analysis demonstrated attached suction-side diffusion, a coherent low-loss core,and entropy generation confined to the outer span region where tip-leakage effects appear,consistent with observations from established high-efficiency designs.To assess mechanical loading, the Tiainen hybrid thrust model was applied and comparedwith CFD surface integrations. Front-side forces agreed within 5%, while back-disk forcesdiffered by roughly 13%, a discrepancy consistent with the known sensitivity of cavitypressure fields. Flow diagnostics revealed a partially ventilated single-cell circulation inthe backside cavity, producing predictable axial loading with only weak sensitivity toleakage-channel variations.The work delivers two main contributions. First, a validated, modular mean-line toolsuited to the rapid design of small-core centrifugal compressors. Second, the creation andCFD verification of a new, aerodynamically robust impeller design for Mitis. Together,these developments establish a foundation for future optimisation within the company’scompressor design pipeline.
Information
- Författare
- Shankar Seetharam, Shivank
- Lärosäte / institution
- KTH/Skolan för teknikvetenskap (SCI)
- Publiceringsdatum
- 2026
- Uppsatstyp
- Master-uppsats
- Språk
- Engelska
- Nyckelord
- ⌕Computational Fluid Dynamics (CFD)⌕Centrifugal compressor⌕Micro gas turbine (MGT)⌕Mean-line analysis⌕Performance prediction⌕Reynolds-averaged Navier-Stokes (RANS)⌕SST k-ω turbulence model⌕Aerodynamic design⌕Empirical loss models⌕Axial thrust⌕Impeller backside cavity⌕Tip leakage⌕Slip factor⌕Turbomachinery⌕Impeller design
Utforska vidare
Liknande uppsatser
Uppsatser med liknande ämnen och nyckelord.
Master-uppsats, KTH/Skolan för teknikvetenskap (SCI)
Cocaud, Livia
Publicerad: 2026
Master-uppsats, KTH/Kemiteknik
Theologou, Konstantinos
Publicerad: 2026
Master-uppsats, Lunds universitet/Institutionen för energivetenskaper
Wu, Siqi, Gao, Chenzi
Publicerad: 2026
Master-uppsats, Lunds universitet/Institutionen för energivetenskaper
Chandhavar, Shreyas Gowrishankar
Publicerad: 2026
Master-uppsats, KTH/Bro- och stålbyggnad
Mohamed Ibrahim, Abdirahman, Saleh, Safia
Publicerad: 2026
Master-uppsats, Lunds universitet/Institutionen för energivetenskaper
Karlsson, Max
Publicerad: 2025