Flow in transonic compressors

Abstract

Introduction B the term transonic compressor we mean an axial flow compressor in which the inlet flow Mach number, relative to the rotating blades, varies from below unity at radii near the inner casing or hub to values substantially above unity at radii near the blade tips. Usually the flow is diffused to subsonic velocities in the flow passages formed by the rotating blades, generating part of the desired pressure rise. A further pressure rise is generated in a stationary blade row downstream of the rotor in the process of removing the swirl imparted by the rotor, but the flow Mach numbers relative to the stator blades are generally below unity. Thus, the flows of principle concern here are those in the rotor passages. Axial flow compressors have been the subject of research and development efforts for at least 40 years, and the transonic compressor for nearly 30. The earliest systematic studies in the United States were carried out by the Lewis Research Center of the NACA about 1950, and the technology was quickly developed for application in the first generation of turbofan engines. Because it offers higher pressure ratios per stage and larger airflows per unit of engine frontal area than the subsonic compressor, the transonic compressor has become a major component of all modern aircraft engines. It has received a proportionately large share of the huge research and development effort expended on these engines. The opinion has often been proffered that it represents a highly developed technology which will not yield significant gains from additional research and development. I argue to the contrary, that there is much scope for improvement in the design of these important components, and that the capability for making these improvements is rapidly coming to hand. This position rests on several facts. First, there are large inefficiencies in transonic compressors due to the three-dimensional character of the inviscid and viscous flows and their coupling. Second, the design methods currently in use do not deal rationally with these loss mechanisms. Third, until recently there has not been a capability for measuring the actual flow in the rotor passages to provide feedback to the design process. The designer has had to rely on overall performance measurements and timeaveraged measurements between the blade rows for verification of the design process. We now have several diagnostic methods capable of providing detailed information about the flow in the blade passages. These include laser velocimetry, gas fluorescent density measurements, and timeresolved pressure measurements, as well as the time-honored hot-wire anemometer. We also have a capability for numerical computation of the three-dimensional transonic flow, inviscid at present but with viscous effects in the near future. Turbomachinery should be one of the major beneficiaries of the advances in this area as was discussed by Chapman in last year's Dryden Lecture. Taken together, the problem of the transonic compressor and these new engineering tools represent an opportunity quite comparable to that presented by recent improvements in high-speed airfoil sections, together with three-dimensional computational techniques in external aerodynamics. My objectives in this paper are to describe the role of the technology and its status, discusss some recent progress toward understanding the flowfield, and indicate the directions which I feel future work should take. I do so with the full knowledge that some of my colleagues in industry and government have much more experience in the design of compressors than I, and that I may not fully represent their viewpoint. My hope is that the view expressed here will prove complementary to that of practicing engineers, and have some beneficial effect on future fan and compressor designs. This review will deal mainly with the question of efficiency. There are many other important problems in the design of transonic compressors, including flutter and forced vibration, stall and surge, foreign object damage, and noise. Some will be mentioned as they impinge on the issue of design for high efficiency, but only the latter will be discussed in detail. Apart from the necessity for focusing on one of these many issues in order to achieve the depth appropriate to a Research Lecture, this choice is motivated by the perception that efficiency is the