On balance, Mach 2.2 not only looked the better bet but it was all that the British industry dared attempt. (Later the Americans spent vast sums on Mach 3 and gave up). Archibald Russell at Bristol was given the job of heading the design team for a Mach 2.2 SST to carry 130 passengers non-stop from London to New York. His report in March 1960 showed that the best slender-delta design, which had by then clearly emerged as the most efficient, would weigh 385,000 lb and need six Olympus engines of 25,000 lb thrust each. This was more than had been expected, but instead of boldly going ahead the committee chickened out and instructed Russ to redesign to only 270,000 lb using four engines. He was reluctant to do this; he maintained the original Bristol 198 design was correct. How right he was can be shown by the fact that today Concorde takes off at 408,000 lb!
At this point everyone was too busy merging with rivals to do much designing. Sydney Camm called the process Mixomatosis, but it was clear that the internecine warfare must cease if we were really to compete with the USA. Bristol Aircraft were merged with Vickers-Armstrongs (Aircraft) and English Electric Aviation, which then took over Hunting Aircraft to form BAC (British Aircraft Corporation). Bristol Aero-Engines combined with Armstrong Siddeley, de Havilland Engines and Blackburn Engines to form BSEL (Bristol Siddeley Engines Ltd). It was a time of great turmoil, especially at the financial and managerial level, but BAC and BSEL quickly became fine teams with a great spirit. The SST remained under Russ, and the Olympus under me.
Over in France the nationalized giant Sud-Est under Georges Hereil, which had produced the successful Caravelle short-haul jet, had lately merged with Sud-Ouest to form Sud-Aviation. Chief designer Pierre Sartre and his right-hand man Lucien Servanty thus became opposite numbers to Russ and Dr Bill Strang at Bristol. This was suddenly important when in 1961 the British and French governments demanded that Russ and Sartre work out a common SST design. The French wanted only a short range, but Russ knew the North Atlantic was essential for an SST to sell in the world market. The final compromise of November 1962 saw the French adhering adamantly to a short/medium-range SST weighing 220,500 lb and BAC pushing a long-range (but not transatlantic) version at 262,000 lb. Fortunately pressure from world airlines shifted the entrenched French position and enabled the design to go ahead, initially at a weight of 326,000 lb, and powered by four of a substantially new type of Olympus, the Mk 593. Under the terms of the inter-government agreement we were required to collaborate with the French national engine company, SNECMA, but this was no problem because both SNECMA and Bristol had in fact survived and flourished as engine companies on the strength of the Jupiter licence taken out in 1920 by Gnome-Rhône, SNECMA’s main ancestor. SNECMA was still busily making Hercules engines under Bristol licence in 1955.
There has been much criticism of the management structure of the Concorde programme, which will not be repeated, and on the airframe side there were frequent strained relationships; but from the very start we had the happiest partnership with SNECMA. Their two top engineers, Michel Garnier and Jean Devriese, quickly became BSEL’s friends, and I cannot recall a single technical difference that was not settled by a single short meeting. The official work (and money) split was Bristol 60 per cent and SNECMA 40 per cent, my team being responsible for the basic engine and the French for the new and complex jetpipe, thrust-reverser, noise suppressor and the convergent/divergent final nozzle which in cruising flight provides a major proportion of the thrust.
At take-off the thrust comes from the engine itself, as in any other turbojet aircraft. At supersonic cruise conditions the situation is very different. First, the air inlet must convert the relative speed of the oncoming air into high pressure in the most efficient manner. The air approaches the Concorde inlet at about 2,000 ft/sec, roughly the speed of a rifle bullet, and must be slowed to a relative speed of 500 ft/sec by the time it reaches the engine intake face. It is first very suddenly slowed to Mach 1, about 1,000 ft/sec, by shockwaves in the inlet, and then diffused subsonically by an expanding inlet duct leading to the engine. This all happens in about one-hundredth of a second, during which time the air pressure rises fivefold, an extremely important factor in achieving the desired thrust and efficiency. Yet at takeoff the inlet has to be opened wide to admit the full airflow at zero forward speed, while at cruising speed it must be closed down to provide the greatly reduced airflow needed. This calls for very large movable flaps, ramps and doors, the control of which is scheduled by an automatic electronic system.
Similar problems arise at the nozzle end. In cruising flight the jet has to reach the speed of sound at the minimum-area throat of the nozzle, after accelerating down a converging section; downstream of the throat the rules are reversed, and to accelerate the now supersonic jet the nozzle has to extend to a large final diameter. There must also be a reverser to slow the aircraft after landing and the most effective possible noise suppressor. SNECMA held complete responsibility for the back end, but the complicated inlet system needed the collaborative efforts of the two engine and two airframe partners.
In jet engines such as the turbojet or turbofan the components at the front, where the air pressure is rising, are subjected to a useful forwards force, while downstream of the combustion chamber, where the pressure is falling, everything tends to be blown out of the jetpipe by an adverse rearwards force. It is the difference between these forces that propels the aircraft. But at Concorde cruising speed one gets an unusual distribution of the various forces: from the variable inlet ramp, -12% (rearwards); from the air intake system, +75%; from the engine itself, +8%; and from the convergent/divergent nozzle, +29%. The total, of course, is 100%, but what emerges is that the inlet is ten times as important in thrusting the aircraft along as is the engine, while the final nozzle is almost four times as important. We therefore had to pay very great attention to the inlet and nozzle, though as the engine was the sole producer of thrust at take-off we were not let off the hook!
Shell carried out tests on the Concorde fuel system which showed that at cruising speed the fuel would gradually be heated to 160°C, at which point it would be on the point of throwing down waxes and varnishes which would quickly be fatal to the sliding parts in the fuel control system. Likewise the oil would heat up to 300°C, at which it became acidic and attacked the bearings. But if the speed were to be reduced to Mach 2.0, 1,320 mph, the air temperature would be reduced to only (13.2 squared, minus 50) 124°C, or 36°C cooler, giving an acceptable margin of safety for the fuel and oil and also for rubber and plastic components. It also promised much lower attrition of the airframe fatigue life. It had little effect on aircraft range, and added only a few minutes to transatlantic block time.
Russell readily approved our proposal to reduce cruising speed from Mach 2.2 to Mach 2.0, but the French were very difficult. However when Pierre Young and I argued the case before the Concorde Committee of Directors in Paris the change was agreed. Today, in fact, the usual figure you will see displayed on the digital speed indicator in the passenger cabin is 2.05.
In the meantime, Concorde grew heavier, as most aircraft do. Increased weight means more fuel, which means more weight, which means both more fuel and also bigger engines, which again means more fuel. More fuel also means a bigger wing, which means more weight, and these adverse spiralling effects are difficult to fight. It was especially vicious in the case of Concorde because there could be no reduction in either payload or in nonstop Paris to New York capability (about 200 miles further than from London). It is for this reason that we finally got a working aircraft at a weight of 408,000 lb, which in turn forced us to work harder on the engine to obtain more thrust.
We started by giving the Olympus 593 a zero-stage on the front of the LP compressor to pump more air with a higher pressure-ratio, as well as a redesigned turbine with aircooled rotor blades to allow a higher operating temperature, but there is a limit to this path and we were running out of steam at around 30,000 lb thrust. The Concorde, however, was obviously heading for a minimum requirem
ent of 36,000 lb per engine, and so as a last resort we persuaded SNECMA to incorporate partial reheat (afterburning) in the jetpipe, which could easily give a 20% thrust boost at take-off. Reheat was common, but only on military aircraft. The drawbacks were extra complication in the variable propelling nozzle, possibly higher fire risk and certainly extra noise; on the other hand, more thrust would mean more noise however it was produced. The aircraft design team were reluctant to accept reheat, but it was too late to go back to Russell’s six engines. SNECMA produced a superb reheat and nozzle system, and today, though the captain usually informs the passengers when he is switching on reheat to start his transonic acceleration, through the once-feared sonic barrier, it is all a non-event and Mach 2 is as smooth and quiet as subsonic flight.
In 1960 Bristol Aero-Engines was merged with Armstrong Siddeley Motors Limited to form BSEL, as noted earlier. Sir Arnold Hall, once Director of the RAE, came from the Hawker Siddeley board to become managing director. In 1966 BSEL was acquired by Rolls-Royce — probably an inevitable move of which, in principle, I thoroughly approved. At a personal level I was less happy and decided to retire on my 60th birthday, in September 1967.
Prior to this Sir Arnold had become managing director of the Hawker Siddeley Group, and Warlow-Davies had succeeded him at BSEL. Poor Warlow-Davies reigned for only two years before suddenly dying, as did Lom, of heart failure. His successor was Hugh Conway, from Short Brothers. I disagreed with Verdon about his appointment, but I was wrong and my fears proved groundless. I found him a splendid man to work with, full of initiative, energy and good humour. In 1967 he insisted I stay on as a consultant.
As the Concorde programme progressed Pierre Young bore more and more of the responsibility for its engines. He had the advantage of having had a French mother and a boyhood in France, so he was totally bilingual. He was also a brilliant engineer and mathematician, able to deal with aircraft designers or airlines in a tough but completely fair manner. To him goes most of the credit for the absolute success of the Concorde’s propulsion system, which has no parallel in airline service. He was assisted by Leonard Snell, once a talented rocket engineer at de Havilland Engines, and Lionel Haworth who when at Derby had designed the Dart and Tyne turboprops.
Haworth had been a friend since 1938 and I was thrilled when he threw in his lot with us at Bristol. He did many enormous tasks in perfecting the design of the Olympus and Pegasus, bringing to bear equal proficiency in mechanical engineering, aerodynamics, vibration, material properties and accurate estimations of weight and cost. I shall never forget his eloquence in explaining to Pierre Sartre how to solve the unprecedented problem of installing the very long and rigid engine in the extremely long nacelle fixed to the highly flexible wing, which in any case varies in dimensions according to how hot or cold it gets in flight. He was a pupil of Rubbra’s, and there is no higher praise than to describe him as Bristol’s Rubbra.
By 1968 my life had changed. I was now a consultant to what had become the BED (Bristol Engine Division) of Rolls-Royce. Pierre Young was in sole charge of the Concorde engine, and Gordon Lewis was Chief Engineer of the brand-new RB199 augmented turbofan that was being designed in co-operation with MTU (Germany) and Fiat (Italy) as the engine for the MRCA, which today flies with three European air forces as the Tornado. Henceforward subsonic aircraft would invariably have highly efficient turbofan engines, quieter and more economical than turbojets, but the latter kind of engine was right for Concorde because of its very high speed. At 50,000 ft the Olympus 593 gives 10,000 lb of thrust, but as at this speed each pound of thrust is equivalent to 4 hp the total horsepower of Concorde is a staggering 160,000, or about as much as 160 Spitfires. Compared with Whittle’s engine of 1943, already a fairly mature machine, the Concorde engine weighs seven times as much and gives 25 times the thrust up to three times the speed, with a much lower specific fuel consumption.
Perhaps most interesting of all is the fact that the overall thermal efficiency of the engines of Concorde in cruising flight is about 43%, which is the highest figure recorded for any normal thermodynamic machine (obviously one cannot include nuclear power).
In its early days the Concorde promised to be even noisier than the extremely noisy 707 and DC-8, and something had to be done to try to reduce its nuisance at least down to that of these older aircraft. Acoustics, and especially noise nuisance, is an extremely complex subject, and I resolved to recruit the best experts I could get to form a Noise Panel to advise us. I approached Professor J. E. Ffowcs Williams, today Rank Professor of Engineering (Acoustics) at Cambridge but then a brilliant young Welshman at Imperial College. I appealed to him to take charge of our Noise Research Department, and he readily agreed despite losing financially on the deal because he had to give up a major consultancy to Boeing.
He quickly gathered Professor G. Lilley from Southampton, Professor Niels Johannesen from Manchester, David Crighton who became Professor of Mathematics at Leeds, and several other brilliant workers in the field. We at Rolls-Royce and SNECMA contributed our own experts. We hoped that, by more fully understanding the relationship between the noise and the source, we would be able to effect the greatest possible reduction in the nuisance caused by Concorde.
At the time the most advanced theory of jet noise had been produced by Sir James Lighthill, who concluded that the noise was proportional to the eighth power of the jet velocity. But one must take frequency into account. If the pressure fluctuations radiating from the source are harmonic, that is if they are repetitive and of equal magnitude up and down about the mean pressure, then recognisable notes will be heard. Middle C is a constant 256 pressure fluctuations per second, and for each octave above or below the frequency is doubled or halved. Low frequencies attenuate slowly with distance, whereas high-pitched sounds carry only a short distance, which explains the deep note of a foghorn.
The human ear is so sensitive it can hear a watch ticking or an Olympus in full afterburner, a range of sound energy in the ratio 1 to 10,000,000. We therefore measure noise power in Decibels (dB), units which are based on a logarithmic scale which makes this vast numerical range manageable. The simple formula is:
Thus, if a noise increases 100-fold, the dBs increase by just 20, the log of 100 being 2. Likewise, if we have two equal noises, and since the log of 2 is 0.3, the two together increase the dBs by 3 compared with either noise alone. The four engines of Concorde put out just 6 dB more than any one of them by itself. Alternatively, if one could replace one Concorde engine by four jets each of one-quarter the thrust, the noise of each would be reduced by 6 dB and raised to a faster-attenuating higher frequency. This was the basis of the multi-pipe silencers used on the first 707s.
Lighthill’s theory was based on the assumption that the jet noise was produced by quadrupoles at the periphery of the jet. A quadrupole is an ideal conception of four sources arranged very close together in square formation, each source in antiphase with its neighbours. Thus when one source is producing a positive pressure its two neighbours are producing equal negative pressure, the fluctuations all having identical frequencies. Emitted noise is then a maximum along the diagonals passing through the four poles and at a minimum along the axes at 45° to these directions, thus explaining the long-familiar fact that the maximum noise behind a jet engine being tested on an open airfield is at 45° to its axis, and a minimum directly behind it.
Many measurements had confirmed that noise is proportional to the eighth power of jet velocity, at jet velocities around 1,000 ft/sec. At only half this speed the noise varies as the fourth power, and — fortunately for us — at very high velocities of 2,000 to 3,000 ft/sec the variation is proportional only to the third power.
Moreover, because of the ear’s logarithmic sensitivity, it makes little difference whether, in the presence of a loud noise, we add or remove a second noise almost as loud. If we have a 110 dB noise, we can add six extra noises of 100 dB each and still only increase the total to 113 dB, which the ear hardly notices. Thus the
main objective for us was to find the magnitude, frequency and position of the sources of the loudest noise. These occur at the periphery of the jet as it issues from the nozzle, and extend about ten nozzle diameters downstream. At this distance, entraining fresh air has reduced the jet velocity by half, removing its importance as a noise source. Members of our panel explained all this in great detail, and organised laboratory and full-scale engine tests so that we could explore the ‘far field’ noise, which is what would impact on the people near the airports.
The sheer volume of theoretical studies and test results from our academic members in 1965-70 was prodigious. Their contribution enabled SNECMA, where the burden of responsibility chiefly rested, to design a most elegant exhaust system which, for minimum weight and performance penalty, matched the variable nozzle and thrust reverser exactly to the tremendous variation in flight conditions whilst also reducing noise until it was never worse than that of the subsonic jets, such as the 707 and DC-8. Indeed, as Concorde climbs away much faster and more steeply than those aircraft, its noise nuisance is markedly less.
Noise at airports was not the only issue exploited by the once vociferous anti-Concorde lobby. Another was sonic boom, the noise created by the shockwaves as they pass the ear of an observer. In the case of Concorde the boom is heard about 20 miles behind the aircraft. The mechanism is analagous to the bow and stern waves of a ship, and the double boom-boom is similar to that from distant thunder, but it was said that the phenomenon would squash greenhouses, wreck old buildings and break all the windows along the flight path. In fact, no evidence was ever produced to show that the passage overhead of a Concorde ever cracked anything, but except in the Soviet Union — where the view was taken that nobody minds thunder — SSTs have never been permitted to cruise at Mach 2 over land. Indeed, everyone will remember the way every trick in the book was tried to prevent the Concorde from opening a service to New York at all, though for many years Concordes have carried people in comfort across the Atlantic in 3½ hours with never a mention in the media.
Not Much of an Engineer Page 20
