Masters Theses

Date of Award

12-1996

Degree Type

Thesis

Degree Name

Master of Science

Major

Mechanical Engineering

Major Professor

F. Shahrokhi

Committee Members

Roger Crawford, Roy Schulz

Abstract

The problem of water ingestion into gas turbine engines installed on an aircraft is known to have serious consequences. Reducing the stall margin, changes in component efficiencies, blade and casing wall clearance problems, combustor and after-burner flame-out are but a few of these problems [1,2], To study this problem, an analytical model (AFCWWI FORTRAN CODE) to simulate the thermodynamics of water evaporation in an axial-flow compressor, using idealized assumptions and conditions, was developed by Curtis Stephen Mashbum. The simulated compressor represents a compressor that was mounted on a constant speed compressor test rig, fitted with a bell mouth inlet, and a variable area throttle valve that was held choked for all test conditions. The analytical model of the compressor flow was a two-phase, one-dimensional model in which the compressor two-phase flow was assumed to be in equilibrium and isentropic. This thesis is an extension of Mashbum's study and attempts to simulate non-equilibrium conditions in the compression process by reducing evaporation rates of the water droplets in the compressor. This was accomplished by applying a bias of 2, 4, 6, 8, 10, 20, and 30 percent to the original water vapor saturation pressure curve and generating pseudo false water vapor saturation pressure curves which were then applied to cases of 2, 4, 6, 8, and 10% water ingestion rates. These new curves forced the AFCWWI FORTRAN code to pseudo-equilibrium conditions at each stage within the compressor. But since the new saturation pressure curves are biased, the reduced rates of water droplet evaporation actually represent non-equilibrium cases. The same engine, compressor conditions, and simulated mass fraction of ingested water used in the study performed by Mashburn were used in this study. In analyzing the data generated by reducing the evaporation rate of the water droplets ingested into the compressor, the following results were obtained: 1) The bias applied to the saturation pressure decreased the amount of water vapor throughout the compressor where 6% or less liquid water by mass was ingested. When more than 6% liquid water was ingested, the water vapor present at each stage decreased in the first five stages of the compressor, but increased in the last two stages. This change resulted from forcing the liquid droplet evaporation to lag the original equilibrium case which increased the air flow temperature and liquid present throughout the compressor. Since the air flow was not cooled through evaporation as much as the baseline, the increased temperature and liquid water present in the later stages resulted in an increase in water vapor, even with a bias subtracted from the saturation pressure. If full evaporation was achieved (which is evident in the 2% water ingestion case) the difference from the baseline for both the liquid and vapor present at each stage within the compressor goes to zero. 2) As the percentage of water ingested by the compressor increased, the total and static temperatures and pressures all decreased. Where 4% or more liquid water was ingested into the compressor, and the saturation bias increased, so did the difference in the total and static pressures and temperatures when compared to the baseline case. In addition, the difference in static density, blade row diffusion factor, and input power used to drive the compressor also increased. The only exception was the 2% water ingestion case. This case was different because with a bias applied to PSAT, it was the only condition which still fully evaporated the ingested water within the compressor. Once full evaporation of the ingested water had taken place, the value of the parameters just mentioned, when subtracted from the baseline case, decreased rather than increased. 3) The axial velocity increased when a bias was applied to the 2% water ingestion case and subtracted from the baseline. The difference in axial velocity then decreased for all cases where 4% or more of liquid water was ingested. 4) A good rule of thumb in ensuring normal operation within a compressor requires that a value of 0.6 for the diffusion factor be considered the upper limit for this parameter. During this investigation, the diffusion factor was marginal for the original unbiased 6, 8, and 10% water ingestion cases. The effects of water ingestion and the lags induced on the water droplet evaporation for this study had an unfavorable influence on this parameter. These unfavorable effects coupled with the fact that as the bias subtracted from the saturation pressure increased, the diffusion factor also increased, makes it likely that the compressor would stall. The following conclusions may be drawn from these results: i) Water ingestion effects dominated the bias effects. ii) The total and static temperatures and pressures decreased, the axial velocity increased the first three stages and decreased the remaining stages, and the blade row diffusion factor decreased the first three stages and increased the remaining stages. iii) Subtracting a bias from the baseline saturation pressure and generating new saturation pressure curves had the effect of forcing the evaporation to lag the baseline case resulting in increasing the MDOTL and decreasing the MDOTV throughout the system when compared to the baseline, increasing the air flow total and static temperatures and pressures until ingested water fully evaporated, and increased the diffusion factor where 4% or more liquid water was ingested.

Download
Files over 3MB may be slow to open. For best results, right-click and select "save as..."

Share

COinS