active injection control of rotating stall, some results were also reported for the
use of injectors in a passive steady manner. These results showed comparable
successes in using the two very different control schemes. In Ref. [44], with the
same injection rate of 3,6% of the compressor main flow, the steady injection
gives 4,3% extension of the stability boundary towards the lower mass flow and
the controlled injection for zeroth, first and second modes of pre-stall disturbances
gives 7,8% improvement. Three years later since the investigation of Ref. [44],
on the same test facility, Suder et al. [45] published their passive control results
giving the reduction of the stalling mass flow by 6% using steady flow injection of
2% of the compressor flow rate. Unlike the traditional empirical design approach
of passive control by trade-off for the improvement of compressor performance,
detailed study was performed in Ref. [45], both experimentally and using steady
flow CFD simulation, for the blade passage flow parameters, e.g. the redistribution
of the incidence and blade loading towards decreasing at the tip under various
injection configurations.
Judging the development of air injection control as described above, several
lessons can be learned. Firstly, since the benefits of active and passive injection
are comparable, it could imply that the unsteady characteristics of the compression
system have changed to its improvement even though the control measure is a
steady one. This proposition dictates a mechanism of unsteady response, which
would exist in the compressor when the steady injection acts on the unsteady flow
process of rotating stall. Secondly, in contrast to all the steady flow consideration
when the amount of injected air needed is on the order of a few percentage of the
compressor main flow in order to realize the effect on the mean flow parameters
such as the incidence or blade loading, the amount of steady injected air considered
for being effective only on the unsteady flow characteristics can be dramatically
reduced. Thirdly, with such tiny amount of injected air, the steady compressor
characteristic with no injection should not be influenced and thus could keep its
shape unchanged, a problem that the traditional passive control scheme suffered
for years. Based on the above argument, a new approach, steady micro air injection
from the casing, was recently proposed in Ref. [46] and experimentally verified
that the injected flow of only 0,05% of the compressor main flow is able to trigger
the unsteady response and lower the mass flow rate at stall for up to 5,8%. The
mechanism of unsteady response was demonstrated and proved through the tests at
various injection configurations, plus wavelet analysis of measured pre-stall flow
disturbances and numerical computation for the behavior of tip clearance vortex
at near stall condition. The computation is now further extended to the operating
point of full stall and Fig. 3, taken from Ref. [47], gives a visualized flow picture
at 80% span from hub for the whole blade row annulus, showing that the micro
injection can keep the operation stable while, at the same compressor flow rate,
the rotating stall already takes place for the compressor without injection.
The micro injection scheme is an example which shows how a new approach
has been evolved from its past development and how it benefits from the past
experiences. It must be stressed that in the case of micro air injection, the design
philosophy remains unsteady (active) in spite that the control action could even
be steady (passive). Moreover, this active control philosophy should be directly
related to the unsteady blade flow physics on which the nature of stall is routed.
In this context, the active control in terms of long length scale waves is
actually unable to effectively interact with the flow of short length blade scales,
and therefore unable to realize the original goal of increasing the pressure rise
ISSN 0236-3941. Вестник МГТУ им. Н.Э. Баумана. Сер. “Машиностроение”. 2006. № 2 121