cross-disciplinary approach, the value of team work, the system view on the
phenomenon and the interplay between modeling and experiment. It is therefore
not the author’s intention to make this article in full coverage of related topics.
Rather, the selected topics are those that emerged as the research efforts went
on and the research subjects were updated. Many of the material covering these
topics are from the results obtained in my institute during the same period of time.
One topic is the research on stall in its early emerging period, which appeared
in parallel with the theme of active control. Another topic concerns the effort
to extend our understanding of the dynamic characteristics of rotating stall in
compressors to its internal fluid mechanics in the blade passages. The third topic
discusses the recent trend of control strategy based on the learning of active vs.
passive methods. Before discussing the three topics, a brief history of the recent
development is given so that the overall picture of the still developing process
would not be missed. Finally the paper is ended with summaries and conclusions.
History of Recent Development.
The concept of “smart engine” first
appeared in the open literature was given by Epstein in 1986 [1]. He described
the gas turbine engine whose performance can be controlled through closed loop
operation from sensors to actuators. Among different possible applications of this
concept, the active control of aerodynamic instability in compression system, due
to its value in practical application and perhaps the better formulation of its basic
characteristics, was the first to put forward into practice. The basic idea came from
a joint work by the experts of compressor aerodynamics and anti-noise acoustics
[6] which states that the feedback control on the small disturbances preceding
rotating stall and surge can change the dynamic behavior of the compression
system, render a given unstable operating point stable, and enhance the operating
range. One of the preparatory works before this idea came into the light was the
modeling of compression system by Greitzer [7] and the modeling of post-stall
transients in compressors by Moore and Greitzer [8, 9] which later became the
theoretical framework for many of the active control study on the compressor
flow instability. This theory when applied to the active control proposes a linear
process of development into stall, which was then supported by the experimental
work of McDougall et al [10] and Garnier et al [11]. The early attempt was very
encouraging as can be seen in the two results for surge control [12, 13] and two
results for the control of rotating stall [14, 15], conducted in the two gas turbine
labs in MIT, US and in Cambridge University, UK. The extension of the stable
operating range by about 25% could be attained for surge and 4. . . 20% for
stall control. In a hope of dramatically improving the stability of compressors,
the research soon became the hot point among researchers and engineers of
turbomachinery, control and other disciplines, with high project investment and
thus large manpower involved.
While the research strategies were various among different groups, the
reactions in industry were also mixed. The cautious reaction came primarily from
the complexity of the active control technique applied to the engine environment,
which reduces its reliability and robustness. Technically, its practical application
was questioned because the ultra-short time period of sensing, processing and
actuating that are needed in the feedback loop of active control. Therefore,
concurrently with the emerging research of active control, the study on the
behavior of stall in its inception period rather than fully developed stage also
received a great deal of attention. The research resulted not only in evidences of
the linear development of stall as assumed in the modeling of [6], but also in the
114 ISSN 0236-3941. Вестник МГТУ им. Н.Э. Баумана. Сер. “Машиностроение”. 2006. № 2