1.  Lord Jauncey has asked for an explanation of why I consider the discrepancy between the figure of rotor speed (NR) found on the aircraft NR gauge and that postulated by the Boeing simulation is so important. I regret that to do this properly requires me to draw up a simple explanation of how a helicopter rotor works and its relationship to engine power and the collective (or thrust) lever.

  2.  Paras 1-4 of the attached give some very simple explanations of how a helicopter rotor works. Paras 5-7 raise some of the points of discrepancy between the simulation and what was found in the accident.

BASIC AERODYNAMICS

  1.  Any heavier than air flying machine or aircraft flies by using the difference in air pressure above its wing and below its wing. To do this you shape the wing in a certain way so the air has to go faster around the top of the wing than around the bottom of the wing. It is called the "venturi effect" and works like the airflow in a carburettor. This reduces pressure over the wing.

  The difference in pressure is called lift and acts upwards and counteracts the aircraft's weight which acts downwards. To increase the lift you either need to get the wing going faster through the air or to tilt the leading edge of the wing upwards to increase the distance the air has to travel over the top surface of the wing.

  Eventually if you tip the leading edge of the wing up too much the airflow over the top of the wing breaks away and the wing ceases to produce lift and the aircraft falls under the influence of gravity. This is called stalling.

  2.  In order to fly lift must be greater than weight. In a fixed wing aircraft you must run along the runway until the airflow over the wing is fast enough for lift to overcome weight and then you will take off. In practice the pilot will also tilt the nose of the aircraft upwards thus also tilting the wing to get an extra bit of lift. In a helicopter you have a rotor to get lift. The rotor is in effect a hub with two, three or four long thin wings attached to it. The rotor spins so the wings move even with the helicopter body stationary. This gives the helicopter its unique ability to hover. The whole rotating assembly of the hub and blades is called the (rotor) disc.

  3.  For structural reasons (and unpleasant problems when the rotor tips get near the speed of sound) there is a maximum rotational speed for a helicopter rotor, and for reasons I will explain in para four there is a minimum speed. Thus we need to keep the rotor speed constant (usually 100 per cent plus or minus 2 per cent of design speed). So we are constrained to use the method of tilting the leading edges of the blades to increase or decrease lift. This is done by raising and lowering the collective (or thrust) lever which pivots (the technical term is "feathering") all the blades at the same time. This not only increases lift but also "drag". This is caused by the greater disturbance of the airflow as the blades tilt up.

  To overcome drag we need more engine power. In old fashioned helicopters the pilot had a motor cycle twist grip on the collective lever and he twisted this to increase throttle as he lifted the lever. Modern helicopters have automatic systems to do this. FADEC is one of these and it keeps the rotor speed (NR) to about 100 per cent plus or minus ½ per cent.

  4.  The lower speed limit of the rotors is also fixed. The rotor blades are held on by hinges at the roots and are free to flap up and down freely. Only when the rotors slow down to about half speed do mechanical arms (droop stops) come out to keep the blades up to stop them hitting the ground. The reason for this hinging of the blades is a phenomenon called dissymmetry of lift.

  If a helicopter goes forward at 100 kts and the tips of the rotors are going round at 500 kts rotational speed, the tip speed of rotor A will now be 500 kts plus wind speed—600 kts airspeed. The retreating blade B will only be at 400 kts airspeed. Therefore there will be a massive difference in lift (due to the differing speed of the rotor tips through the air) and the rotor and helicopter would roll uncontrollably to the retreating side. We overcome this by hinging the blades at the root (this is too complicated to go into in detail here). Thus the rotors are only held out flat by centrifugal force and as soon as the helicopter lifts from the ground weight acting on the hub will act against the centrifugal force and will pull the blades into a cone.

  If the rotor speed slows down centrifugal force will be reduced and the angle of the cone will steepen until ultimately the blades will meet at the top and "clap hands". I would suggest that the minimum rotor speed is somewhere between 80—85 per cent before the situation becomes irrecoverable.

THE SIMULATION

  5.  It can be seen from the above that maintenance of rotor speed is by far the most critical criterion for safe helicopter flying. The accident investigators found the NR indicator at 100.5 per cent category two (Evidence appears positive). The Boeing simulation postulated 91 per cent. The figure of 91 per cent is the minimum transient NR permitted when engine power is not applied in the Flight Reference Cards.

  The collective (thrust) lever was fully up when found in the wreckage; this means that there was maximum lift being demanded by the rotors and thus maximum drag. The job of the FADEC was to keep the rotor speed at 100 per cent at all times. The only way for the rotor speed to be dropped to 91 per cent as postulated in the simulation was for the engines to have been giving at least maximum power—in fact emergency power. Even this was not enough to keep the rotors at 100 per cent in the simulation.

  6.  In the AAIB report 7.3.4 the power plant summary shows that the engines were at matched power at an intermediate level. This ties in neither with the position of the collective lever as found or more particularly with the simulation. (If there had been a partial runaway or other malfunction and one engine was running up and the other compensating by running down this could possibly cover the case.)

  7.  A great weakness of the simulation is that it assumed "steady flight conditions" (AAIB para 8 page 60). If the aircraft was having control or engine problems this would not have been the case. Although I have drawn particular attention to rotor rpm I do note that the simulation ground speed is 11 kts faster than that found (147 kts). As I suggested I have grave doubts about the Chinook at 37,700 lbs weight actually being able to carry out the climb as suggested by the BoI and stay within the aircraft's power limitations. I assume that the BoI must have had a Chinook go and try to reproduce the flight profile as suggested by the Boeing simulation. I cannot find a record of that being done. Simulator trips in the flight simulator at Farnborough at that time would not have been representative—it was very inaccurate as far as power requirements were concerned. Perhaps one of the other witnesses on 7 November could confirm that actual flight trials to prove the simulation were carried out and with what results.

Robert Burke

31 October 2001