only be released at the present time by the numerical computation of the unsteady
internal flows.
Several authors have tried such numerical simulation with different schemes
but the general trend is from two-dimensional to three-dimensional and from
inviscid Euler to viscous Navier-Stokes computation schemes [37–39]. The
numerical computation was mostly performed with multiple blade passages even
entire blade rows. In addition to the characteristics of developed stall such as the
number and propagation speed of stall cells, internal flow structure during stall
and its inception can be described which could clarify the role of tip clearance
vortex in the stall via short length-scale inception process [38, 39]. To further
explain this phenomenon, from the recent results of numerical computation in
Ref. [40], Fig. 2, (
a
) shows a comparison of 2D (naturally without tip clearance)
and 3D (with clearance) computation in obtaining different type of stall precursors
and different time elapse of their existence, Fig. 2, (
b
) shows a comparison of 3D
computation with and without tip clearance in resulting to the different stability
limit of compressor operation.
It should be noted that the computation results for the role of tip clearance
vortex in stall inception were obtained for the case of tip critical low-speed
compressors. Many challenges remain for cases of high-speed compressor in
describing the internal flow structure of stall and its inception. It is also a
challenging task to find the link of stalling process with the blade design
parameters, especially for multi-stage compressors. Gong et al. [41] attempted
this task by a new 3D computation scheme coupling the system model and the
model of blade row flow, represented by a body force field in terms of the
prescribed blade’s pressure rise and turning characteristics. In the low-speed
multi-stage environment, the method succeeded in capturing the development of
both long and short wavelength compressor instabilities. For the short wavelength
disturbances, the method is able to depict their occurrence on the negatively
sloped part of the overall compressor characteristic proving the experimental
result in Ref. [22], and to show the suppressing effect on their growth by closing
the rotor-stator gaps which was later experimentally proved in Ref. [32]. The
similar approach was also given by Nakano et al. [42] for numerically studying
the instability inception in high-speed multi-stage compressors.
Therefore the numerical simulation, along with its further development, should
offer many possibilities for studying the phenomenon of compressor instability,
from the detailed fluid flow mechanism in blade passages to eventually the stability
prediction of the matched or mismatched compressor system. As will be discussed
in the next section, the numerical simulation, in complement with experiment, can
also be used in studying the control effect to the compressor instability.
Active vs. Passive Control Strategy.
As described above, along with the
emerging technology of active control for improving the compressor stability,
the traditional means of passive control also obtained its new development. It is
interesting to follow this development taken the air injection control in the tip
region of compressor annulus as an example. The design principle of the injection
could be either active or passive. In a low-speed four-stage axial compressor
experiment, Day applied an active control scheme and succeeded with 4–6%
improvement of stability boundary at the expense of injected air maximum 1%
of the compressor flow rate [14]. Years later, in the works of Behnken et al. [43]
on a low-speed single-stage compressor at Caltech and Weigl et al. [44] on a
transonic single-stage compressor in NASA, while the main purposes are still
ISSN 0236-3941. Вестник МГТУ им. Н.Э. Баумана. Сер. “Машиностроение”. 2006. № 2 119
