Chapter 5
High-speed Cowlings, Air Inlets and Outlets, and Internal-Flow Systems
[139] These high-speed programs were superimposed on the low-speed foundations built up in the 1926-1936 decade. An understanding of the earlier work is important in reading this section and, therefore, the pertinent background will be reviewed briefly.
The exposed radial engine had been in use only a short time before designers began to be concerned about its drag and cooling problems. The most rudimentary knowledge of aerodynamics suggested some kind of rounded fairing to cover the engine. The first "cowling" of this kind was designed in 1922 by Col. V. E. Clark of "Clark Y" airfoil fame, for the Dayton-Wright XPS-1 airplane which was powered by the first Lawrance radial engine. Cooling problems were encountered and the lack of understanding of how the cowl worked and how much drag it saved discouraged others from using it (ref. 43). A successful foreign application of a cowling which completely covered a 50-hp air-cooled engine was made by one Piero Magni in 1926 (ref. 164). This well-shaped cowling employed a "blower-spinner," which provided satisfactory cooling. The design must have had very low drag but no data were given. The "blower-spinner" was rather obviously too great a complication to be considered for the large radial engines of the twenties.
By 1927 the drag problem of radial engines was widely perceived' as very serious. Several requests were made by attendees at the 1927 Engineering Conference to make a full-scale engine/cowling investigation the first work to be undertaken in the new PRT, which was then nearing completion. NACA fully concurred, having already in hand a request from the military for cowling work (ref. 165). Fred E. Weick, who had been hand-picked by G. W. Lewis to design and manage the PRT, laid [140] out a tentative test program and diplomatically submitted it to industry for comment and suggestions. As finally agreed upon, the program reflected the great concern of the time that cowlings might seriously inhibit engine cooling. Five of the seven shapes to be tested on a representative cabin fuselage with a Wright J-5 engine did not completely cover the cylinders; only one complete exterior cowling was designed, and it was tested with two inner-body shapes (ref. 166).
Perhaps the most important single test was the first run with the fully exposed J-5 engine. The simple scales of the PRT registered 85 pounds increase in drag due to the engine at 100 mph. For typical J-5 powered airplanes of that day this meant that up to as much as 30 percent of the engine power and fuel was being expended simply to provide cooling. There was now the strongest possible motivation to find an effective cowling.
Only the complete exterior cowling produced a really dramatic reduction in drag. Weick's test procedure was to make cut-and-try changes until the engine temperatures approached those of the fully exposed engine. Only two important changes were made in the original complete cowl-the exit area was increased severalfold by cutting 3 inches from the skirt and the inlet diameter was increased from 24 to 28 inches. Although the cooling was finally judged adequate, the barrel temperatures were still some 600 F hotter than for the exposed engine. After these changes to favor cooling had been made the drag was measured carefully and found to be 60 percent lower than that of the uncowled installation (ref. 166).
In retrospect there was nothing very remarkable about the cowling itself-an arbitrary, rounded external fairing tailored by straightforward cut-and-try changes to favor engine cooling. The underlying achievement was NACA's creation of the PRT, which made it possible for the first time to work with a full-scale engine/ propeller/ cowling in a wind tunnel and measure precisely the drag and cooling data. NACA was not the first to lay out the cowling and did not "discover" it in the usual sense of that word. They did discover its enormous drag-saving potential. The basic internal flow processes and the cooling mechanisms of the radial engine still remained obscure at the conclusion of the J-5 installation testing in the PRT, in spite of the fact that tolerable cooling had been achieved [141] by crude methods involving what was later recognized as a large excess of internal airflow and internal drag. Some of the first industrial applications of the cowling were less successful in solving the cooling problem, and employed very large exit openings to try to encourage cooling flow. In several cases the external drag advantage was apparently nullified by the huge internal drag of the large cooling flow. Lack of understanding of these internal drag effects caused much puzzlement. Rex B. Beisel of the Vought Company described the situation in his classic 1935 paper (ref. 167). "The first five years of industrial experience with the NACA cowling has brought forth a maze of contradictory data.....leading to confusion and suspicion."
Added to the real problems was at least one largely illusory one, the belief that the cowling would seriously impair the pilot's vision. The surviving pilot of an Army midair collision had claimed that the cowling had blocked his view. A consequence was the Anny's adoption of the Townend Ring for its P-12 and P-26 airplanes (ref. 41), and this in turn led to NACA's decision to flight-test several truncated cowlings on a Curtiss XF7C-1 airplane. It was clear in NACA's tests that the visibility issue had been overplayed and would not exist at all in many cases (ref. 168). Aside from this useful result these flight tests illustrated the primitive state of knowledge in 1929: the cowls were tested and compared with different exit areas and thus different internal drags. Furthermore, shutters were used in front of the engine to control part of the cooling flow, subsequently known to be one of the least efficient flow-control techniques.
The lack of a solid framework of understanding is evident also in other programs of the early thirties (ref. 169). The basic cooling problem was visualized as how to divert or deflect a part of the cooling-air flow toward hot parts of the engine. This was consistent, of course, with the excessive cooling flows that characterized the early installations. In this concept a great deal of effort was expended on "deflectors" of the type tried by Weick. The curved "shell" baffle emerged from this work; however, it was conceptually more a refined deflector than the ultimate tight-fitting baffle which actually contacted the fins and formed an outer wall for the finned heat transfer channel. As late as 1935 the loose-fitting baffles were still being tested by NACA (ref. 170) in spite [142] of the fact that by that time the work of Vought and Pratt and Whitney had evolved the "pressure-baffle" concept. A NACA investigation of a tight-baffled cylinder was finally carried out in 1936 (ref. 171). Useful NACA work relating to fin design and fin spacing was also carried out in the thirties.
The cylinder cooling work was one part of a three-pronged program often described in agency literature in the 1931-1934 period; the second part was aimed at finding the best cowling shape, and the third was verification of the concepts developed in parts 1 and 2 by tests of an actual cowled engine in the PRT. An R-1340 Wasp engine was borrowed from the manufacturer in 1932 for part 3. This program ran into difficulties in each of the three areas and was never completed as planned. Contributing factors were the loose, inefficient, shell-type baffles employed in the engine which were obsolete before testing was completed, the fact that only the climb condition of engine operation could be simulated in the 100-mph PRT, and the use of a very short nacelle on which the flow was prone to separate from the afterbody so that much of the cowling drag data were useless. In the report of the engine tests finally issued in 1937 (ref. 172) the negligible differences in engine cooling in climb with the four different cowlings is shown, but nothing is concluded in regard to the relative drags of these cowlings, and no reference is made to the initial ambitious objectives of the program.
During the 1931-1934 period the Vought group under R. B. Beisel conducted a program of wind tunnel and flight investigations of cowling and cooling problems which provided definitive enlightenment on the key issues. They established the high cost in drag of the large excess of cooling air that flowed through unbaffled and loosely-baffled engines, and aided by Pratt and Whitney, they evolved the idea of "pressure baffling" in which all flow through the engine is blocked except through tight-fitting baffles in contact with the after-quarters of the cylinders. They invented cowl flaps to vary the size of the exit opening and so to regulate efficiently the airflows to the minimum value required for cooling in each condition of flight. The internal flow system of the NACA cowling was at last understood and criteria for efficient engineering design and operation were established (ref. 167).
Beisel's group had periodic contact with the Langley finned-cylinder [143] and PRT work, but they credit British studies, particularly those of Pye (ref. 173), as a principal source of their inspiration. In 1934, Langley was called on to test, in the Full-Scale Tunnel, the first double-row engine installation with Vought's pressure-baffle system. The first NACA references to this system are found in the 1934 Annual Report. Vought was able to demonstrate that a properly baffled cowled engine had lower operating temperatures than the fully-exposed engine-an accomplishment believed to be physically impossible in the early years of cowling development.
In 1934, G. W. Stickle and M. J. Brevoort who had been working on the R-1340 cowled-engine/nacelle tests in the PRT were transferred to T. Theodorsen's group. In reviewing a prospective report covering their PRT work, Theodorsen had pointed to the blunt nacelle afterbody as the probable source of the drag anomalies that had been observed, a perceptive speculation that was later verified. Theodorsen's interest in the many problems of cowling and cooling was now aroused and he supervised plans for a comprehensive new investigation. NACA's unfortunate experiences with the R-1340 tests together with Vought's recent successes provided a framework for better planning. Following Vought's lead, a wind tunnel model employing a dummy engine was used. However, one cylinder was heated electrically so that cooling tests could be made, and a propeller with a 150-hp electric drive was included. Complete pressure distributions and smoke flow studies were very important special features of the program.
This solid full-scale investigation answered virtually all the remaining questions (ref. 174). The Vought findings were corroborated and extended importantly. It was found that large-scale turbulence, induced more or less automatically in the front of the cowling, was the mechanism that cooled the front of the cylinders. The internal flow was analyzed by evaluating the efficiency of the cowling considered as a pump. Unfortunately, the unbaffled engine showed the highest pump efficiency; however, it also had by far the highest pumping power requirement due to its large air flow. The tightly-baffled engine had the lowest pump efficiency but also the lowest power absorption because of its low flow. The cooling drag penalties were consistent with the Vought results. Regrettably, no reference is made to Vought's work in the NACA reports.
[144] By the end of 1936 it was obvious from the Vought and Langley investigations that the radial-engine cowling had virtues that were unsuspected in the beginning when it was thought of only as a device for external drag reduction. With proper design the cowling would enhance cooling rather than inhibit it as originally believed. Through physical containment of the internal flow, the cowling made the enclosed baffled engine in essence equivalent to a ducted radiator as far as cooling was concerned. By application of basic heat transfer principles the overall drag-power cost of cooling the radial engine had now been reduced to low levels, comparable to those of the liquid-cooled engine.
After 1937 and continuing through World War II, there were many NACA contributions to cowling and to fin and baffle design. The so-called "cooling correlations" were evolved relating engine operating conditions, fin temperatures, and cooling flow requirements. With the framework of basic understanding now firmly established, these later contributions were for the most part sharply focused and valuable. We will be concerned in this chapter only with the high Mach number projects relating to the NACA cowling. Of greater long-term significance were the high-speed investigations of generalized inlet, outlet, and internal flow systems, with which this chapter is chiefly concerned.