Date of Award

3-2-2021

Publication Type

Doctoral Thesis

Degree Name

Ph.D.

Department

Mechanical, Automotive, and Materials Engineering

First Advisor

Jeff J.D. Defoe

Keywords

body force, CFD, fan/compressor modeling, loss prediction

Rights

info:eu-repo/semantics/openAccess

Creative Commons License

Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.

Abstract

This dissertation proposes a new method of modelling turbomachinery blade boundary layer and shock losses using the body force method. Body force methods are used to model fan/compressor performance at a lower computational cost than unsteady Reynolds-Averaged Navier-Stokes (URANS) computations in non-uniform inflows. Most loss modelling approaches in the literature require calibration. Some recent work has shown the use of non-calibrated methods for entropy generation calculations. However, recent non-calibrated methods cannot estimate flow losses with boundary layer separation. In this dissertation, an artificial neural network has been developed and trained to analytically relate the blade geometry and flow regime to the boundary layer momentum thickness at the trailing edge. The trailing edge momentum thickness is used in a body force loss model that accounts for the relative total pressure drop. This model is capable of predicting the loss at off-design conditions. The accuracy of the model is over 90$\%$ in 2D cascades. The model is then applied to the NASA rotor 67 compressor blade row. The model captures the high entropy generation near the tip region for uniform and non-uniform inflows. For non-uniform inflow, it predicts the isentropic efficiency to within 1$\%$ compared to a URANS computation.

Share

COinS