Peter Gray/FORT WORTH and WEST PALM BEACH
It was a dark and stormy night. I was in the jump seat of a Sikorsky S-61N helicopter. We were in the cruise and all was well. Suddenly, all hell broke loose. One engine ran up quickly and went slightly over the turbine temperature limits (T5), along with the compressor speed and torque. The other engine dropped off to a low power setting and we lost some rotor RPM, that vital function that keeps the helicopter safe in the air and its occupants alive and well.
The commander of the aircraft who was flying was a junior captain; his first officer a very senior captain. The first officer, without consulting with his captain, immediately took charge of the malfunction, diagnosed a run-down of the low engine and advanced its manual throttle to restore power. He had not noticed and brought to his captain's attention, and neither did the captain check that the free turbine power rpm (Nf) on the high power engine was at zero. As he restored the power of the low engine with the manual throttle and tried to match torques, he was unable to do so and we then had a gross rotor overspeed.
There was also now gross confusion in the cockpit and, in the meantime, the high engine was still operating above its maximum permissible T5 and was probably experiencing unusual stress which might well cause it to fail a burn-out fairly soon.
At this point, the simulator instructor, sitting behind me in his cabin, brought the flight to a halt and we had a debrief. The event that had initially caused the engines to run up and down was failure of the signal (a mechanical drive shaft) between the Nf and fuel control unit (FCU). The FCU uses this signal to interpret rotor rpm. Sensing a gross loss of Nf, it immediately applied full fuel flow and power to the offending engine. Since the FCU does not sense T5, the engine overheated.
Unlike multi-engine fixed-wing aircraft, however, which do not sense what the other engines are producing, in all multi-engine helicopters they do. So the good engine, sensing its mates full power, backed off. The big clue in this incident which would have produced an immediate and correct analysis was the lack of the Nf indication of the high engine. That engine should have been put into manual throttle, at which time the low, good engine would have restored its power and all would have been well.
There are almost always several causes of such incidents and accidents. In this case, the primary cause was the failure of the crew to correct the low rotor rpm and excessive T5, then to each analyse and communicate what they thought the problem was and discuss subsequent actions. The captain should have overridden his co-pilot and demanded time to do just that. The secondary cause was the failure of the company's management to provide training in cockpit resource management (CRM).
IMPOSSIBLE TO PRACTICE
The above malfunction cannot be simulated in the air on the actual aircraft. It is only discussed and one hopes the crews understand. Furthermore, on a training and testing flight, the training captain has to act as co-pilot since he has to be at the controls to pull back engines and introduce other malfunctions during pilot recurrent competency checks. He is also the aircraft commander, so any lack of CRM training would not be apparent.
Failure of a power unit in a multi-engined aircraft at critical times during take-off and the approach to landing at maximum take-off or landing weight in both helicopters and fixed-wing aircraft is almost impossible to practise safely on the actual aircraft. If you do, you inevitably have to use contingency power on the remaining engine if it is a twin or put other parts of the aircraft under severe stress.
This eats into its life, which does not amuse commercial directors. Furthermore, if the pilot under test does not get it right, you could have a heavy landing or, after a continued take-off, the aircraft going downhill instead of up. Imagine an acceleration/stop exercise to just before critical decision point, when the take-off should be aborted and the aircraft stopped within the length of the remaining runway, in a Boeing 747 weighing 500,000kg (1.1 million lb) and travelling at well over 130kt (240km/h) and think of the wear and tear on the whole suspension system. The aircraft has been approved to do it by the certifying authorities, but no company is likely to practise the manoeuvre in the aircraft, and neither would an operator practise an engine failure in the S-61N at night, at maximum weight, during an offshore platform take-off, just after rotation when the aircraft is committed to continue the take-off.
This is where simulators come into their own, provided their fidelity in terms of handling qualities and reactions to pilot and other inputs are close enough to the aircraft, and that the visual clues are adequate, faithfully following the aircraft movements and what the pilot should see. The company can train and test its pilots on all those malfunctions and emergencies they cannot or dare not practise in the air. Training and practice can also be given going into unusual airports such as the old Hong Kong field or doing an instrument approach in a helicopter into Aberdeen or an offshore platform when the weather is right on minimums. Simulators are also helpful in getting to know a modern glass cockpit, its strengths, weaknesses and failure modes and what to do about them. Poor crew understanding has caused accidents in the past.
In the fixed-wing world, spending $10-15 million on a simulator could be a small proportion of the purchase price of the actual aircraft, so it makes good economic sense to buy a simulator. In the rotary wing world, the price of a simulator will often exceed the cost of the helicopter. Unenlightened managers will use that to instruct the training department that instead of purchasing a simulator, they will buy an aircraft which, when it is not being used for training, can be hired out to earn extra income.
Before entering Flight Safety International's (FSI) Sikorsky S-76 simulator, having that very day flown and evaluated the actual aircraft, an S-76C+, I was shown some of the training equipment. FSI has designed one of the best emergency and malfunction checklists I have seen. The inside cover contains a picture of the advisory, malfunction and emergency warning panel, each light having a page reference number. You then go to that page in the checklist. The immediate items to be done from memory are in a shaded box, which leads into a yes/no diagram - a flowchart.
Another useful classroom aid is the animated walkround inspection system. This starts with a projection on to the screen of the whole aircraft. The instructor can then zoom in on all the items in the checklist, showing where they are and how they are checked. Panels can be opened and closed just like on the aircraft and you do not have to clamber up the side to check the main rotor head. The US Federal Aviation Administration has approved the use of this system as an alternative to teaching it on the actual aircraft.
IN THE S-76 SIMULATOR
The drawbridge came up so we could use the motion system, which has the usual six axes, the same as the helicopter - pitch, roll, yaw, heave, surge and sway, more axes than most other things that fly. As I installed myself into the captain's seat, I looked around. It was unmistakably an S-76. The dusk-type visual system and motion were activated. The instructor behind me had put me on the threshold of a runway, motors and rotors turning and I came to a very unsteady hover indeed, trying to stay in line with the strip - not like the real thing at all.
My hover height varied considerably, almost alarmingly, my heading and position over the ground similarly. I relaxed as much as I could, trying not to move the control, but to no avail. Fixed-wing simulators do not have such a problem - but fixed-wing aircraft do not hover. The fixed-wing pilot has enough cues from the visual system and representative feel from the controls and motion system to land and take off in even the worst weather conditions and experience few differences from the real thing. Indeed, the fidelity is so good that many fixed-wing machines are approved to give all the transition training and testing without actually having to fly the aircraft. The first flight in the aircraft is a revenue one, with a training or line check captain alongside. One of the reasons for this fidelity, in my view, is that the simulator manufacturer gets a comprehensive data package from the fixed-wing aircraft manufacturer detailing the exact flight characteristics which it can then accurately copy. I do not know why this rarely happens, if at all, in the helicopter world. Nor, importantly, does the fixed-wing pilot need so many close-up detailed visual cues.
The rotary-wing simulator manufacturer often has to make do with limited data so has to design its own or rely on pilots experienced on type (all of whom will give different answers) and cope with the helicopter's characteristics of hovering, sideways and backwards flight, turns on the spot, hover taxiing, operations into confined areas, on to rooftops and the like. Although the modern computers which run the visual and motion systems are much advanced and react within, typically, 80-150millisec to pilot inputs, there are still discrepancies, delays and differences in feel and reactions, and insufficiently detailed visual cues with insufficient texture. Seeing blades of grass would help enormously. The manufacturers have done their best by applying, for example, scuff marks on the runway.
One company lightens the weight of the chassis in an effort to achieve a better reaction from the motion system and reduce reaction time to 20millisec. It is important to remember, however, that the primary purpose of helicopter simulators is not to teach pilots to hover.
REALISTIC DUCK
I abandoned my attempt to hover and lurched into the air for a visual circuit, trying to keep my heading in line with the runway heading. Once safely airborne, I used the automatic flight control system to fly the aircraft, intervening only to trim the attitude, turn or descend. The simulator now behaved much like the aircraft. The dusk visual was realistic. I kept the runway in sight through my side window. Final approach was satisfactory to a slow run-on landing, though I had difficulty in judging my exact position in space once I was close to the runway. I managed to land, however, without breaking the "aircraft". Simulators will imitate a crash if you mishandle them, some tilting the nose fully up or down, applying full roll and putting the horizon vertically. This gets the adrenaline rushing even faster as you hang in your straps.
BELL 430 AT FORTH WORTH
I now looked forward to flying FSI's latest simulator for the Bell 430, commissioned in February and undoubtedly equipped with more up-to- date technology than all the other simulators I have flown. I had test flown the aircraft, so hoped to be able to make worthwhile comparisons.
Again the drawbridge came up and we were all set to go, engines and rotors running.
The whole cockpit was totally realistic. This simulator has a full daylight, visual system which can, of course, revert to dusk, night, mist, fog and cloud. It wraps round the front to 180¼ and I had to lean excessively forward to see the top and bottom. It is as good as any fixed-wing simulator I have flown, so I was expecting a better hover capability with so many more visual cues. In the distance I saw rising smoke - a nice touch to help give a sense of reality. The grass alongside the runway was, however, just an indistinct blur - no blades of grass here. I could see skid marks on the runway just ahead of me, but these, too, lacked texture.
I pulled up cautiously into a hover. I had no sensation of leaving the ground but just arrived in space in an almost uncontrollable hover. My first hover on the actual aircraft test flight report said: "The 430 came up into a near-perfect hover and I had very little work to do to hold our position accurately," so this was obviously not realistic simulation. My hover was getting so erratic that my vision occasionally blurred as the outside world moved around.
After some perseverance, it improved, but I had to work extremely hard. I made a tentative transition to forward flight, wanting to refer to the artificial horizon to give me a 10° nose down pitch and the torquemeter to apply 10%, but I had to concentrate so much on the outside cues to keep control that I was not confident enough to use the instruments much.
Once clear of the ground, I could relax and look around, the simulator flying much like the aircraft. I am told that I was not alone in experiencing such difficulties and that, after adequate practice, these problems are overcome. I performed a circuit. The visual system was excellent for this. My approach to the hover was straightforward and the hover better, but because of the lack of detailed texture and other references, I was still uncertain how high above the runway I was. I was confident enough to attempt some sideways, backwards and spot turns. Very hard work, but we made it. We left the circuit and went fast and low over the surrounding countryside. This part was quite realistic. I could clearly see masts and their cables. We headed for the smoke which was coming from a crashed helicopter in a confined area. There was another helicopter alongside the crash. This database area is used for training. We next visited a hospital which had two rooftop helipads. I attempted an approach to the hover and landing. I had to work hard to get an acceptable deceleration and angle of approach. I arrived over the edge of the pad, which was not very big, but there was massive over-controlling, particularly with the collective pitch lever. I persevered, but did not improve. My attitude changes started to go divergent so I pulled up and went round again. My next approach was better, but again I lost it at the bottom.
During our return to base, I asked for various malfunctions. The generator failures were quite straightforward, as was a number two hydraulic system failure. The electrical systems are quite complicated with various faults described in the flight manual, so it is comforting for pilots to be able to practise these as they actually occur, although this would be best done in a ground trainer if the failure does not affect the flight characteristics.
simulator advantages
Another big advantage that a simulator has over training in the actual aircraft is that after a failure or malfunction, not only are the primary effects given, but secondary and even tertiary effects are provided if there are any.
An example in invisible precipitation with a low ambient air temperature, the pilot has forgotten to switch on the anti-ice, so the engines start to ice up. The pilot becomes aware of small engine fluctuations and a slight but increasing loss of power. He realises his mistake and switches on the anti-ice systems of both engines simultaneously. The ice melts rapidly and disappears down the engine, doing some damage. The engine vibrations worsen. The training captain can have one engine over temperature if he wishes or introduce any other tertiary effect at his command.
I asked for a number one hydraulic failure - the system that powers the tailrotor pitch control. In the Bell 230, the 430's predecessor, neither I nor the Bell test pilot had been able to achieve a good enough hover to land. I had found the 430 somewhat better and had achieved a hover and landing, although it was hard work. In the cruise, the feel of the simulator was entirely realistic with a little more pedal force required.
The flat, low rate of descent approach with minimum control movements was satisfactory and the subsequent landing I judged to be easier than the aircraft. We next tried some serious full authority digital engine control malfunctions - a vicious run up of one engine with heavy surging on both as the good engine tried to compensate for the fluctuations - impossible to demonstrate on the aircraft. The malfunctioning engine was identified, converted to manual throttle and the subsequent approach and landing was identical to that of the 430. Autorotation was next - to the ground, not something you would want to attempt in a multi-million dollar aircraft. But double engine or transmission failures do occasionally occur, so it was worthwhile to try one. I found it difficult to judge my flare height, so the instructor called out radar heights and I reacted accordingly.
I heard the tyres squeal as I plonked her down on the runway - a nice touch.
In anticipation of a tailrotor driveshaft failure, I set up a left-hand circuit - the worst case, as the aircraft wants to turn right. Again this is an extremely useful exercise that, without a simulator, you can only talk and think about. On the downwind leg, there was a bang, accompanied by a rapid swing and turn to the right - classic symptoms. I applied some left bank to stop the turn and kept my airspeed at 100kt, as briefed. A lot of left bank was required to get the aircraft to turn left through 180° on to the runway heading when I set up a steep 80kt approach, adjusting the collective pitch lever to keep the aircraft in balanced flight, heading straight ahead in line with the runway. Unfortunately, the rate of descent was too high to land straight ahead while in this configuration, so at about 150ft (45m), I did a gentle flare to wash off some of the speed, raised the lever progressively to reduce the rate of descent and, to counter the subsequent yaw to the right, wound off enough throttle to keep us pointing straight ahead, levelling and running on the aircraft at the bottom.
It worked out well, although the simulator crashed as it always has done during this exercise, the instructor told me. This is probably caused by the high rate of descent as we landed. This crash was not an adrenaline rusher - the visual went blank and the motion stopped. In reality, I felt we would have carried out a perfectly safe, albeit somewhat heavy, landing.
Full flight simulators make excellent instrument training and testing vehicles. You can practise approaching to all those airfields you use right down to company minima or even below, requiring a go around at the decision point. To put some pressure on the pilot, as examiners are required to do, you can give him an engine failure or other unwelcome event at this point. If he crashes, no harm is done, apart from that to his ego.
There has been the occasional occurrence when a particularly masochistic training captain has overwhelmed his candidate with an unrealistic number of simultaneous malfunctions/ emergencies. Ground aids can be failed at inappropriate moments, severe icing conditions entered, autorotations in instrument meteorlogical conditions carried out, recovery from unusual attitudes practised and similar exercises. If the pilot makes a mistake, say during an instrument approach, the approach can be stopped in mid- air, a debrief given, the machine wound back in space to just before the point of failure and the manoeuvre practised or tested again. The aircraft weight, wind speed and velocity and ambient air temperature, as well as many other variables, can be changed at will.
FIDELITY GOAL
Neither of the simulators have yet achieved that fidelity of handling that the fixed-wing world has when the machine is being flown low and slow or hovering. Research and development has been under way for decades to overcome the deficiencies. Nevertheless, they are both splendid training and testing devices.
Source: Flight International
