The following is extracted from “History of the Aircraft Structural Integrity Program” dated June 1980, published by the Air Force Flight Dynamics Laboratory, WPAFB, Ohio.

During the "cold war" of the 1950's, the national security of the United States hinged on the John Foster Dulles doctrine of "massive nuclear retaliation." The principal American de­ terrent force was the United States Air Force Strategic Air Command with its turbojet fleet of B-47 and B-52 bombers. Amer­ icans in the late 1950's were almost complacent about the mantle of protection which this bomber fleet afforded. This compla­ cency was shaken in the fall of 1957 with the announcement of "SPUTNIK," reflecting a startling Russian technical achievement and requiring a revision of how soon the Russians might be able to counter the Strategic Air Command. Despite the fact that the Communists adopted a more aggressive international attitude, the American retaliatory force still seemed an effective bal­ ancing force in deterring Communist ambitions. The complacency was completely shattered in the spring of 1958 when a crisis temporarily immobilized the entire B-47 fleet, threatening to drop the B-47 from the active inventory several years before the Air Force had counted on replacing it. Under a heavy cloak of secrecy, the Air Force and the aviation industry rallied to save the B-47 from the scrap pile. A tremendous engineering effort, identified as project MILK BOTTLE, rehabil­ itated the ailing bomber and gave it new life.

The importance of the B-47 to those concerned with the defense and security of the United States in the 1950's may be difficult to understand today, but at the time the B-47 was the numerical backbone of the Strategic Air Command. The first swept-wing jet bomber to be built in quantity, 2,041 were pro­ duced by three different manufacturers (Boeing, Douglas, and Lockheed) by the time the last B-47 rolled off the assembly line in February 1957. The Air Force had counted on using the B-47 as its medium-range strategic bomber for at least another seven years after 1958. This fact alone would have justified the apprehension which swept the defense establishment when this key aircraft threatened to become unusable years before its anticipated obsolescence date. The B-47 crisis, serious in itself, raised other questions which troubled American planners. Just how dangerous and far-reaching was the lack of theoretical and actual knowledge concerning structural fatigue? What did the B-47's demonstrated weakness imply for other current and planned high-performance manned aircraft, such as the eight- jet B-52, the supersonic B-58, the KC-135 jet tanker, and the futuristic B-70 and F-108?

To understand the relevance of the B-47 problem to poten­ tial problems with other aircraft in the USAF fleet, it is helpful to consider the history of the B-47, and the similari­ties and the uniqueness of the B-47 as compared to other mili­ tary aircraft. Built as an experimental multi-engine jet bomber in competition with the XB-46 and XB-48, the XB-47 was flying before bonafide military characteristics were established by Headquarters USAF. Its radically new design, incorporating six jet engines slung on pylons beneath an extremely thin and flexible laminar-flow swept-back wing, helped the aircraft achieve a performance considerably better than its designers had hoped. A top speed in excess of 600 mph allowed it to out fly almost every fighter in existence at the time. Because of this outstanding performance, the B-47 was selected by SAC as its new high altitude medium bomber in 1950.

The prototype B-47 had made its first flight on December 17, 1947, with a gross weight of 125,000 pounds, powered by six 4,000 pound thrust jet engines. The gross weight grew to 185,000 pounds for the B-47B and reached 206,700 pounds for the B-47E as a result of structural strengthening, equipment changes, and the addition of extra fuel capacity to increase the range. To maintain the desired performance, the engine thrust was in­ creased to 5,200 pounds on the B-47A, 5,800 pounds on the B-47B, and finally 6,000 pounds on the B-47E. The structural design of the B-47 series aircraft was accepted by the Air Force based on a static test of the B-47B in 1950 and a flight load survey demonstration of a B-47B from September 1952 to March 1954. No finite service life was specified for the B-47 series aircraft, so the number of flying hours to be expected during the life of one of the airplanes did not enter into the acceptance proce­ dure. It was procured, however, with the intent that it would not be replaced until 1965. In actual fact, most of the opera­ tional B-47 fleet was phased out by 1966, although a few air­ craft remained in service as late as 1969. The B-47 thus had a relatively long career, although its future looked dim in 1958.

The structural analysis of the B-47, including the stan­dard static test and the abbreviated flight load survey, proved that the item under test would support at least 150 per cent of its design limit load. However, it provided no assurance that the test item would survive smaller cyclic loads in the order in which actual flight imposed them. Thus, repeated cycles of less than maximum loads, including warping, twisting and bending motions, might do more damage than the direct application of much larger loads. Absence of precise information concerning these in flight loads, theoretical or actual, explained how an unexpected fatigue problem could suddenly threaten the life of the B-47 aircraft. Fatigue analysis was not a sufficiently exact science to permit accurate prediction and warning, but it was clear to laboratory scientists that several factors in the use of the B-47 would shorten its expected life. Unfortunately, they could not identify 500, 1500, or 2000 flight hours as the danger point; nor, except in a general way, could they pin­ point the aircraft members which were receiving the most severe stress from additional loads. They could only argue that B-47's performing low altitude missions and those with high flight times be carefully inspected for external signs of stress. Problems with the B-47 could be anticipated from several key factors. The growth in gross weight aggravated the already severe problem of the reduced structural weight to gross weight ratios of modern aircraft design. The increase in engine thrust, due to larger engines and a water-injection modifica­ tion which provided a 17 per cent increase in takeoff power, intensified the strains on the fuselage and wings. Rockets used for takeoff assistance permitted shorter takeoffs. How­ ever, they also gave the aircraft structure a solid "boot" from an unexpected direction . Less tangible questions were being raised about the effects on the aircraft structure of acoustic noise and heat emanations from the jet exhaust, since these fatigue effects were just beginning to be recognized and investi­ gated. More important factors were the changes in service usage of the aircraft.

Designed to be a high-altitude bomber, the B-47 was used largely in that manner for the first ten years of its exis­ tence. However, during the last half of 1957, with Air Proving Ground approval, SAC began employing the bomber extensively at low altitudes. These low-level missions included a struc­ ture-wrenching low-altitude bombing system maneuver (LABS) for low-level delivery of nuclear weapons, and a strenuous "pop-up" bombing run. This operation also caused additional stresses due to the atmospheric turbulence encountered below 1,000 feet. In addition, the increased range due to strategic commitments required more frequent refueling missions, each of which cre­ ated unusual strains in the airframe due to the maneuver loads required to stay within the allowable bomber-to-tanker relative positioning limits during refueling. Repeated takeoffs and landings in training exercises also added extra, unplanned stresses due to the dynamic aspects of landing, taxi, runway roughness effects, and braking conditions. The additional structural loads imposed by all of these effects were difficult to measure. In any case, so long as the B-47 performed its varied missions satisfactorily, it was hard to justify the funds expenditure that a serious investigation of structural loads would require.

Conclusive evidence that a structural crisis had been reached came on 13 March 1958 when two B-47s broke up in midair in separate incidents. Near Homestead Air Force Base, Florida, a B-47B disintegrated at 15,000 feet, three minutes after takeoff. Its center wing section failed approximately at buttock line 45. The aircraft had a total flight time of 2,077 hours and thirty minutes at the time of the accident. The same day a TB-47B broke up at 23,000 feet over Tulsa, Oklahoma, after the bottom skin plate of the left wing failed at buttock line 35, causing the left wing to break off at the same point. This plane had flown a total of 2,418 hours and 45 minutes. While Air Force and contractor agencies were in­ vestigating these two accidents, three more occurred, indicating that the crashes of 13 March were not isolated events.

These successive accidents further served notice that the flaws might show up in almost any B-47, not just those with over 2,000 flight hours. On 21 March, as a result of over­ stress from a pull-up, a B-47E disintegrated in midair near Avon Park, Florida. This aircraft had a total flight time of only 1,129 hours and 30 minutes. Next, a B-47E seemed to explode at 13,100 feet just prior to a refueling rendezvous near Langford, New York, on 10 April. This aircraft had a flight time of 1,265 hours and 30 minutes. The final tragedy in this series occurred on 15 April when another B-47E, with a total flight time of 1,419 hours and 20 minutes; took off into a storm from MacDill Air Force Base, Florida, and dis­ integrated shortly afterwards. The pilot was believed to have encountered gusts of 80 to 100 miles per hour. One of these accidents was ascribed to the pilot exceeding the aircraft structural limits in a pull-up, but the remaining four were clearly due to structural fatigue failure.

The parties most immediately concerned in the crisis were the Boeing Airplane Company, which manufactured the B-47, the Strategic Air Command (SAC), which was the principal user of the B-47, and the Air Material Command (AMC), which was responsible for this and other in-service aircraft. The mag­ nitude of the threat posed by these fatigue failures also quickly involved the Air Research and Development Command ARDC) and its Wright Air Development Center (WADC). WADC was intimately involved from the first, especially the struc­ tural experts of the Aircraft Laboratory and the metallurgical scientists of the Materials Laboratory. The National Advisory Committee for Aeronautics (NACA) also participated actively in the efforts to meet this emergency.

The immediate problem was to keep the B-47's flying, since national security considerations and an approaching summit meeting in Geneva forbade a lengthy grounding, much less a complete discard of this substantial portion of the bomber fleet. Efforts to correct the structural problems by splicing, reinforcing or replacing the affected members were the first response to the crisis. Such stopgap structural corrections would at least permit SAC to continue using the aircraft, though with restrictions on speed, weight and inflight maneuvers. A logical continuation of the emergency corrective program would verify the interim "fixes" and extend them to the point of guaranteeing an "adequate" service life for the B-47.

Several related actions comprised the special engineering effort to restore the B-47 to its former usefulness. Major General Thomas P. Gerrity, Commander of the Oklahoma City Air Material Area (OCAMA), identified his immediate concerns as "inspection criteria, flight restrictions, additional instru­ mentation, and a further test program" An official inquiry into the two accidents of 13 March was already under­way. By 23 March, restrictions on flight maneuvers were being urged by WADC. On 4 April ARDC agreed that "continued, un­ restricted operation of the B-47 fleet was hazardous." By 11 April, specific limitations such as 360 knot indicated air speed and 1.5 g maneuvers were in effect. Formal restrictions were laid down on 25 April, applying to all B-47's except those previously inspected for cracks at all critical points. Low- level flying, except for takeoff and landing, was banned. Air­craft gross weight could not exceed 136,000 pounds (without external tanks) or 185,000 pounds (with full external tanks). Maneuvers were to extend no further than 1.5 g's in a 30 degree bank. Maximum indicated air speed was to be 310 knots, with continuing restrictions on stalls, buffet, flights through tur­ bulence, and touch-and-go landings. Finally, there were addi­ tional rules concerning the conditions under which refueling could be accomplished.

Operating under these flight restrictions, SAC was able to continue to fly training and operational missions with a minimum of hazard until each aircraft could be inspected and retrofitted. The aircraft inspections were initially based on the investigations of the two aircraft which crashed on 13 March, 1958. This investigation led to Technical Order 1B-47-1016 published on 16 April, which incorporated inspec­ tion of buttock lines 35 and 45 (BL-35 and BL-45) with in­ spection of wing station 354 (WS-354). This was quickly re­ scinded on 22 April after investigation of the 10 April crash showed that accident to be caused by a failure at fuselage sta­ tion 515 (FS-515). As a result, T.O. 1016 was replaced on 25 April by a requirement that all aircraft be inspected at all four of these critical points. Other directives quickly followed (Technical Orders 1B-47-1020 and 1B-47-1022) which contained temporary measures which would at least keep most of the B-47's flying. On aircraft without cracks, cer­ tain critical holes at WS-354 and BL-45 were reamed out and received oversized bolts. The aft wing-to-body fittings at FS-515 were also reamed out and received oversized pins. This pin, weighing approximately 25 pounds, was sized and shaped like a milk bottle, eventually resulting in the name "Project Milk Bottle" for the inspection and retrofit project.

The ultimate "fix" for the B-47 wing was incorporated in Technical Order 1B-47-1019, which appeared on 29 May 1958, along with the kits required to reinforce the wing root. The work called for in these three technical orders (1019, 1020, and 1022) comprised the phase of the B-47 rescue work identi­ fied as Project Milk Bottle. This endeavor eventually encom­ passed structural modification of 1,622 B-47 aircraft. The first half of May was a build-up period: the project crested in August and by 1 January 1959, all B-47's had been inspected and reworked at least once .

The $62 million cost of Project Milk Bottle did not necessarily assure SAC that it now had no fatigue problems with the B-47. By the beginning of April 1958, the parties concerned were already in general agreement that only cyclic testing could provide valid proof that the B-47 fixes would guarantee an "adequate" service life for the aircraft. To formulate valid test cycles, however, it was first necessary to define the environment in which the aircraft would per­form. This meant identifying the number of takeoffs, land­ ings, accelerated climbs or letdowns, abrupt turns, low alti­tude missions and flights through turbulent weather conditions. Of necessity, the spectrum for cyclic testing had to be based on available data. At the same time, it was recognized that gaps existed in this information and programs were being under taken to collect data to fill this need. In the meantime, cyclic tests of the B-47 would indicate whether the engineering repairs had restored a useful service life to the aircraft. In addition, the tests would be a preliminary step in the larger structural integrity program which the B-47 crisis had spawned. The engineers would be interested in when cracks first appeared during the cyclic test, where the cracks were located, and how fast they spread under continued applications of a spectrum of loading.

By mid-May, 1958, a canvass of available facilities and interested parties resulted in a decision to cyclic-test the B-47 to destruction at three independent establishments--the Boeing plant at Wichita, Kansas; the Douglas plant at Tulsa, Oklahoma; and the NACA laboratory at Langley, Virginia. Boeing began its testing early in July, Douglas began late in July, and NACA started almost a month later. However, neither the Boeing nor the Douglas aircraft had received the "1019 fix" so testing on both was stopped temporarily in order to install this modification. The Langley airplane had undergone this modification before its cycle test began.

When the tests began, the wing seemed to be the B-47's critical structural element, but one month of accelerated test activity uncovered a new fatigue danger point: the fuselage longeron at fuselage station 508. On 8 August, after the application of 1,275 equivalent flight hours, both upper fuselage longerons on the Boeing test aircraft failed during a 90 percent limit load test. The fuselage skin had shown warning cracks but the longeron collapse was still a surprise, especially since the aircraft had accumulated a total time (actual plus simulated flight) of only 3,442 hours. The fuselage was replaced, retaining the original test wing, and the cyclic test was started over again on 8 September. On 16 September, with a total time for the replacement fuselage of 2,156 hours, a crack again appeared at fuselage station 508. On the same day a service B-47 with 2,900 hours of actual flight time was also discovered to have longeron . cracks. Meanwhile, on 13 September, the same fate almost befell the test aircraft at Douglas. Crack detection wires, installed after the first Boeing test mishap, prevented a complete failure, but the Douglas disaster repeated Boeing's, even to the location of the crack and the timing (the Douglas aircraft had a total time of 3,022 hours).

Consultation led to a decision, to replace both upper fuse­ lage longerons on both the Boeing and Douglas aircraft. In addition, the replacement longerons incorporated a reinforce­ ment that Boeing had engineered after the August 8 failure. This was accomplished, and Boeing and Douglas resumed testing on 13 October 1958.

In early November, the Boeing test aircraft failed at buttock line 45 after a total of 5,872 actual and simulated hours on the wing. After repair, tests resumed on 4 December. In mid-December, a 27 inch crack appeared in the aircraft skin at right buttock line 32. The aircraft's total time was now 6,397 hours, of which 4,626 were simulated flight hours on the longeron modification. Boeing replaced the damaged skin panel and tested until 21 December, stopping with a total time on the test aircraft of 6,922 actual and simulated flight hours. The decision to discontinue the test short of actual destruc­ tion stemmed from two considerations; the longerons had suc­ cessfully withstood 5,151 equivalent test hours, and the wing had been extensively spliced and patched so that it no longer resembled the actual B-47 configuration. Therefore, no really useful information would derive from future tests. A teardown inspection followed so that test results could later be compared to other B-47 structural integrity findings.

The aircraft under test at Langley developed fuselage skin cracks at 4,243 total hours. After repairs, this B-47 held up until cracks appeared in the steel splice plates of the 1019 modification at 5,468 hours and again at 5,818 hours. A crack was also discovered in the splice plate at wing sta­ tion 179 at that time. This was counted as a major failure, and the Langley test was ended. The total time on the longeron repair was 4,550 hours, but Langley personnel reported that their "bird" was tired and that it was starting to develop a rash of small cracks.

Thus, the Douglas B-47 was left to carry on the burden of destruction testing. Though fuselage skin cracks necessitated repairs, this airplane went on to pile up a total of 6,425 hours before three cracks were discovered in the web of a rib at wing station 258. These cracks were stop-drilled, a tem­ porary measure to arrest the cracks and keep them from spread­ ing. At 7,145 hours, cracks similar to those which had ended the test life of the Langley B-47 appeared on the Douglas air­ craft. Stop-drilling arrested these cracks, but the process had to be repeated twice more at 7,845 hours and at 8,195 hours. In each case, the cracks would have produced wing failure if they had gone undetected. By using these stopgap measures, the Douglas B-47 survived approximately 10,000 hours fatigue testing before the rightwing lower skin panel failed and brought an end to the cyclic testing in February 1959. Significantly, the point of collapse was the wing skin closure panel at wing station 175, less than four inches from the point at which the Langley aircraft had finally come apart.

Insofar as possible, all three test aircraft received identical cyclic tests. Strategic Air Command mission profiles served as the basis for a composite series of simulated mis­ sions designed to duplicate the varying loads imposed on the B-47 in operational use. The stresses represented those on an aircraft with a takeoff gross weight of 170,000 pounds, flying missions averaging just over six hours. Firm decisions from the cyclic tests were not made immediately because investiga­tors wanted to correlate these results with those from other portions of the expanding structural integrity program. The three fatigue test aircraft had, to a considerable degree, proved the reliability of the 1019 wing modification and the longeron repairs. By November 1958, it seemed that a guaran­ tee of 3,000 hours was certain, and that further evaluation of the test results might boost this figure to 5,000 hours. Fur­ ther extension of the B-47's service life was questionable. Colonel R. D. Keator, Chief of the WADC Aircraft Laboratory, was not especially encouraging. In December, he agreed that it might be possible to guarantee the aircraft's useful life beyond 5,000 hours, but he indicated that such a gain could only be achieved by identifying new critical areas and engi­ neering "fixes" for them. The result might not be worth the effort, Colonel Keator suggested, as it could conceivably degrade the aircraft's performance.