Introduction

The B727 and Trident had S-duct installations where afterbody drag had been limited to skin friction levels by sufficiently low afterbody slopes. This had not been possible for the DC-10. The high-bypass-ratio turbofan with its lower thrust-to-airflow ratio produced a proportionately larger cross section than existed for the B727 and Trident. Secondly, the DC-10 evolved with an overall length restraint that limited the afterbody fairing length. As a consequence, both the S-duct configurations studied by Douglas and the Lockheed 1011 had afterbody geometries that exceed the criterion for flow separation and buffet.

Although an aft engine installation presented an inherently more complex problem than a pylon-mounted engine, Douglas established that the aft engine configuration should provide performance and operational characteristics equal to the wing-mounted engines and should have the highest degree of installation commonality possible. In addition, high dispatch reliability along with rapid and convenient accessibility for line maintenance were also prime design requirements. All these requirements had to be met while maintaining high structural integrity and low weight, and without compromising aircraft controllability. The accompanying chart summarizes the design requirements for the aft engine installation.

Fig. 01. Aft-Engine Configurations

MD studied numerous aft-engine installation configurations (Fig. 01). Weight, drag, inlet loss, flow distortion, inlet/engine compatibility, accessibility, maintainability, vulnerability to engine-disk failure, engine growth, commonality with wing-mounted engines, and reverser effectiveness were considered. MD conducted extensive wind tunnel tests at actual Mach numbers and high Reynolds numbers during the program's study phase. These studies showed that only a straight inlet was capable of meeting the DC-10 design requirements.

Fig. 02. Straight Inlet Performance

The straight inlet configuration (Fig. 02) had inherently good flow distortion characteristics and was sized to accommodate all anticipated General Electric CF6 and Pratt & Whitney JT9D engine airflow growth.

Good crosswind capability was achieved by locating the inlet forward of the vertical tail and using a conservative lip thickness and diffuser angle. Blow-in door weight and complexity were avoided by this design. The problem was simplified because straight ducts did not develop additional adverse internal pressure gradients. The inlet was located above the fuselage so that fuselage boundary layer air did not enter the inlet. Fuselage length growth by up to 50 feet was been provided in the selected vertical location.

Fig. 3. Straight Inlet Wind Tunnel Tests

Straight-inlet flow distortion and inlet loss were measured by wind-tunnel tests in the NASA Ames 12-foot low-speed pressure tunnel and in the NASA Ames 11-foot transonic wind tunnel (Fig. 3). These tunnels were selected because of their high Reynolds numbers; the models tested were 23% and 13% of full scale, respectively. Inlet were tested for all important airplane attitudes, flight Mach numbers, engine airflows, and crosswind conditions.

Fig. 4. Inlet Total Pressure Loss

Figure 4 shows inlet total-pressure loss as a percent of free-stream total pressure (inlet mass-flow ratio) for a compressor-face Mach number schedule typical of airplane operation. The compressor-face Mach number was 0.52 at takeoff and 0.59 at cruise. The inlet loss included in the airplane performance was the increment above that demonstrated with the engine on a test stand using a bellmouth. The measured performance was generally better than that used for the airplane performance, and was considerably better at cruise conditions. At high free-stream Mach numbers, where the inlet mass-flow ratio was moderate, the measured loss was the same as a calculated value based on the assumption of flat-plate skin friction. This loss, the minimum possible, was a demonstration of the inlet excellence. The increased loss at lower free-stream Mach numbers, which was due to the higher local inlet lip velocities, and not due to a flow separation at the lip, was similar to that of the wing-engine inlet.

Fig. 5. Inlet Flow Distortion at Cruise	Fig. 6. Inlet Flow Distortion at Takeoff

Although inlet total pressure loss was important, it was inlet flow distortion that must receive the greatest attention. At high-speed cruise (Fig. 5), the inlet distortion was due only to a small peripheral boundary layer, and at no time does the total-pressure distortion even approach engine manufacturer's limits. The distortion was 4% and was not affected by changes in airplane attitude or yaw within the airplane operating limits. Figure 6 shows the crosswind effect on total-pressure distortion for takeoff power. The inlet will have a tolerance to crosswind comparable to that of the wing engines and will be capable of operating in crosswinds up to 50 knots without exceeding the engine manufacturer's distortion limits. This performance has been achieved with inlet lips thickened on the sides. This permits the inlet to be placed nearer to the vertical tail, thereby keeping the duct as short as possible. General Electric evaluated distortion patterns and found them to be well below the engine compressor stall distortion limit.

Fig. 7. Drag Reduction Schemes

Improper afterbody fairings can result in supercritical velocities and flow separations, which create drag and buffet problems. MD carefully considered these potential difficulties during DC-10 aft-engine installation design. The horizontal and vertical surfaces were located to prevent high-velocity region superposition. The fuselage fineness ratio and upsweep were chosen to provide separation-free flow characteristics. No large diverging regions were permitted in the channel between the fuselage and engine. The pylon trailing-edge angle was conservatively selected to ensure satisfactory flow conditions in this region (Fig. 7).

Fig. 8. Straight Inlet Drag Tests

Aft-engine installation drag was measured in the NASA Ames 11-foot transonic wind tunnel and in the NASA Ames Research Center 7-foot transonic wind tunnel (Fig. 8). The Reynolds numbers were 13% and 6.6% of full scale, respectively. To eliminate any influence of the model support on airflow over the model, a twin boom support was used in place of a conventional fuselage sting or a blade-type support. The measured drag increment obtained by subtracting the drag of the wing-fuselage combination from that of the complete configuration (wing, fuselage, vertical and horizontal surfaces and engine installation) is shown in Figure 9. Also shown are the calculated skin friction drag and the drag used for airplane performance. The latter includes an allowance for interference. Measured drag was essentially that due to skin friction, and the interference was small. Such interference levels could only be achieved by no flow separation or shock-wave loss.

Fig. 9. Aft Installation Drag	Fig. 10. Airplane Handling
Fig. 11. 22.8% Wind Tunnel Model	Fig. 12. Landing Approach Roll Due to Rudder.

The double-hinged rudder system provided minimum control speeds 10% lower than those of present four-engine jet-transport aircraft, allowing full use of the DC-10's high lift capabilities (Fig. 10). A 106 knot ground minimum control speed without nosewheel steering (icy runway) provided minimum field lengths approximately 3,800 feet under standard day conditions at sea level (Fig. 11).

Roll tendency due to the rudder (Fig. 12) represented a small fraction (15% during approach) of the available lateral control, which was approximately equal to Douglas DC-8 and DC-9 jet transports. Consequently, the higher rudder has no significant effect on handling qualities.

A joint program to develop the aft-engine reverser was conducted by McDonnell Douglas, General Electric, and NASA in the NASA Ames 40 by 80-foot full-scale wind tunnel (Fig. 13). The 22.8% scale model used a General Electric J85 jet engine to simulate the aft engine. Several cascade reverser arrangements were tested, and a configuration was developed that provided satisfactory reversing, rudder effectiveness, and pitching-moment characteristics. This was achieved by directing the reversed fan flow laterally in the upper cascade sectors and downward in the lower sectors. This configuration ensured good directional control throughout the landing or rejected-takeoff ground roll.

The vertical location of the aft engine above the airplane center of gravity produces pitching moments due to engine thrust that counteract those of the wing engine, resulting in a very small change in trim with increased engine power during a go-around. (Fig. 14)

Fig. 13. 22.8% Wind Tunnel Model	Fig. 14. Go-Around Control Column Force

Power Plant Installation

The aft and wing engine installations were nearly identical . The pylon beam extending aft from the vertical stabilizer base provided a top engine mount interface identical to the wing location. This permitted many aircraft-to-engine services to be also identical. The cowl doors and thrust reverser were arranged identically, thus providing the same degree of engine servicing accessibility.

MD gave careful consideration to systems and components accessibility in the engine installations and nacelle. Hinged engine access doors and reverser halves provided access to all engine parts and accessories that required inspection, maintenance, or servicing. These hinged components were all attached to the aircraft structure and were not part of the demountable power plant (Fig. 21). The door arrangement was identical on both the wing and aft-engine nacelles. Because the thrust reverser and fan cowling components were not part of the demountable power plant, the cost of the Quick Engine Change kits was greatly reduced, and spares logistics provisioning was correspondingly improved. In addition, the size of the demountable power plant was minimized. The aft-engine demountable power plant was within the 8-foot maximum permissible shipping width, and the wing-engine demountable was brought to this dimension by simply removing the nose cowl (Fig. 22). The aft engine installation design was particularly influenced by the need for convenient scheduled and unscheduled maintenance. The DC-10 configuration provided built-in access to the aft engine at a line station without the need for ground support equipment. With this arrangement, the lower fuselage tailcone section below the engine was reached from a common step ladder, and an integral ladder allowed entry into the tailcone. Normal servicing (such as engine oil replenishment) was accomplished from within the tailcone.

In the event a component failure required access to the engine accessory section, the upper tailcone doors and fan cowl doors were opened. All dispatch-critical power-plant components could then be readily serviced or replaced from the work platform built into the fuselage tailcone. Doors in the platform were arranged to permit lowering accessories directly down through the tailcone. The heaviest accessory requiring replacement for dispatch was the engine starter, which weighed approximately 40 lb (Fig. 23).

Engine removal and replacement was an important consideration in the aft engine design. The hinged engine cowling sections and fan reverser halves mounted from the pylon eliminating these components from the demountable power-plant assembly. The aircraft tailcone was hinged to swing downward and provide a clear path for direct engine lowering. The demountable power-plant assembly could then be removed by a simple hoisting action. Figure 24 shows a bootstrap hoist intended for use at a field station where limited ground support equipment was available; a fixture designed specifically for engine replacement and heavy engine maintenance provided a convenient means for scheduled maintenance activities and engine replacement at a main maintenance base. Figure 25 shows such a stand, similar in design to equipment currently in use on other large aircraft. This stand is equally usable during wing or aft engine replacement.