| Summary: | The life and performance of axial compressors are limited by the occurrence of instabilities
such as rotating stall and surge. Indeed, in the course of the design phase a
great effort is usually devoted to guarantee an adequate safety margin from the region
of instabilities’ onset. On the other hand, during its operating life, an axial compressor
can be subjected to several conditions that can lead to the inception of stall and its
dynamics. A few examples of possible stall causes, for the specific case of an axial
compressor embedded in an aircraft engine, are inlet flow distortion, engine wear or
shaft failure.
The shaft failure case can be seen as an exception, as a matter of fact, after this
event surge is a desirable outcome since it can potentially decelerate the over-speeding
turbine by reducing the mass flow passing through the engine. The possible occurrence
of surge and stall should be predicted and controlled in order to avoid severe damage to
the compressor and its surroundings. A lot of research has been carried out in the past
years to understand the inception and development of stall to achieve the capability
for predicting and controlling this severe phenomenon. Nonetheless, this problem is
still not well understood and unpredictable outcomes are still a great concern for many
axial compressor’s applications.
The lack of knowledge in what concerns inception and development of stall and
surge reflects in a lack of tools to investigate, predict and control these unstable phenomena.
The tools available to study stall and surge events are still not highly reliable
or they are very time consuming as 3D CFD simulations.
The doctoral research described herein, aimed at the investigation of the rotating
stall phenomenon and the derivation of the compressor characteristic during this unstable
condition. Following a detailed analysis of the tools and techniques available
in the public domain and the identification of their limitations, the development of a
FORTRAN through-flow tool was the methodology chosen. A distinctive feature of
the developed tool is the independency from steady state characteristics which is a
limitation for the majority of the available tools and its computational efficiency.
Particular attention was paid to capture various viscous flow features occuring
during rotating stall through the selection and implementation of appropriate semiempirical
models and correlations. Different models for pressure loss, stall inceptions
and stall cell growth/ speed were implemented and verified along with different triggering
techniques to achieve a very close to reality simulation of the overall phenomenon,
from stall inception to full development.
lel compressors’ technique that allows the correct modeling of asymmetric phenomena.
The methodology implemented has proved promising since several simulations
were run to test the tool adopting different compressor geometries. Verifications were
performed in terms of overall compressor performance, with simulations in all the
three possible operating regions (forward, stall and reverse flow), in order to verify
the tool’s capability in predicting the compressor characteristics. In terms of flow
field, the ability to capture the right circumferential trends of the flow properties was
checked through a comparison against 3D CFD simulations. The results obtained have
demonstrated the ability of the tool to capture the real behavior of the flow across a
compressor subjected to several different unstable conditions that can lead to the onset
of phenomena such as rotating stall, classic and deep surge. Indeed, the tool has
shown ability to tackle steady and transient phenomena characterized by asymmetric
and axis-symmetric flow fields. This document provides several examples of investigations
emphasizing the flexibility of the developed methodology. As a matter of fact,
within this dissertation, many examples can be found on the effect of the plenum size,
on the different transient phenomena experienced by the compressor when subjected
to multiple regions of inlet distortion instead of a localized region of low or high flow,
on the differences between temporary and stationary inlet disturbances and so on.
This document describes in detail the methodology, the implementation of the tool,
its verification and possible applications and the recommended future work. The work
was funded by Rolls-Royce plc and was carried out within the Rolls-Royce UTC in
Performance Engineering at Cranfield as three-year Ph.D. program that started in October 2010.
|
|---|
