Browsing by Author "Van Eck, H."
Now showing 1 - 1 of 1
Results Per Page
Sort Options
- ItemDevelopment of a one-dimensional code for the initial design of a micro gas turbine compressor stage(Stellenbosch : Stellenbosch University, 2020-12) Van Eck, H.; Van der Spuy, S. J.; Von Backstrom, T. W.; Stellenbosch University. Faculty of Engineering. Dept. of Mechanical and Mechatronic Engineering.ENGLISH ABSTRACT: The use of micro gas turbines for once-off and often unique applications means that a generic, rapid turnaround design and performance evaluation process is required. This thesis aims at developing an application based One Dimensional (1D) design and flow analysis program to be used as an initial design tool for radial and mixed flow compressors. The application was developed in MATLAB®. The code is primarily based on the mean line flow and loss models developed by Aungier (2000). For verification, 18 test compressors were developed using the 1D Application (App). These test compressors cover a wide range of design velocities, mass flow rates, as well as meridional exit angles (mixed flow compressors). Predicted performance results were validated using Numeca/FINE™ Turbo software. Initial comparisons between 1D and CFD results did not match very well. The 1D software generally over-predicted compressor performance, i.e. total-to-total efficiency and pressure ratio. The 1D software further provided fairly poor compressor choke prediction. Consequently, empirical models to correct these deviations were derived and implemented into the 1D software. Following these corrections, the results predicted by the modified 1D application compared well with the CFD results obtained. Initial 1D mean line results presented a mean choke prediction difference of 14.98% for the 18 test compressors, with a maximum difference of 39.88%. In the updated 1D App code the mean difference was reduced to 1.59% with a maximum difference of 4.97%. Initial 1D mean line results presented a mean total-to-total pressure ratio prediction difference of 9.31%, with a maximum difference of 17.27%. In the updated 1D App code the mean difference was reduced to 1.24% with a maximum difference of 5.01%. Total-to-total efficiency predicted by the initial 1D mean line code presented a mean difference of 11.23%, with a maximum difference of 15.09%. In the updated 1D App code the mean difference was reduced to 0.74% with a maximum difference of 1.57%. As a final means of validation, two more test compressors were designed using the 1D App. Performance results compared well with the CFD results.