Engine shutdown over the Atlantic – why training pays off

TWA began flying the two pilot, twin-engine Boeing 767 in late 1982. The 767 was the first Boeing wide-body to be designed with a two-pilot crew digital glass cockpit. New automated electronics replaced the role of the flight engineer by enabling the pilot and co-pilot to monitor aircraft systems directly. Boeing’s philosophy maintained that the autopilot flies more efficiently than the pilots. Pilots were expected to use the autopilot as much as possible. This troubled me.

Why?

TWA 767 airplane
The Boeing 767 was a ground-breaking airplane, with only two pilots and a glass cockpit.

Every six months, captains are required to take a proficiency check in the simulator or airplane. Maneuvers such as stall recovery, steep turns, emergency descents, visual approaches, non-precision approaches, missed approaches and crosswind landings are all flown manually and are required on the check.

On a beautiful summer day in Germany in 1985, I was excited to command TWA Flight 745 from Frankfurt to St. Louis. The flying time was scheduled for 9 hours and 27 minutes. Flights over eight hours required a relief pilot and he was sitting in the cockpit jumpseat. There was a full load of passengers and fuel.

Several hours into the flight, everything was operating smoothly. In the cockpit, the relief pilot was standing between the pilot seats discussing our international operations. We were about an hour southeast of Iceland over the North Atlantic Ocean when a red emergency light illuminated. It was the left engine oil pressure light indicating the oil pressure was low and the left engine oil quantity gauge indicated zero.

Shutting down an engine while cruising at 39,000 feet in the middle of the ocean was not on my wish list. The EMERGENCY 767 CHECKLIST procedure for an ENGINE FAILURE has 10 steps to be followed. The engine was secured according to standard operating procedures. Although the 767 is able to safely fly on one engine, the airplane is unable to maintain level flight at higher altitudes and must descend. In order to descend, we must divert off course so as not to interfere with aircraft that may be below us. Our diversion procedure at the time called for a turn off course in the direction of the nearest suitable airport.

Our procedure in an emergency is to aviate, navigate, and communicate. I disconnected the autopilot and trimmed the rudder for flight on one engine. We descended the airplane and navigated toward Iceland. The non-flying pilot communicated with Oceanic Control that we were declaring an emergency and descending and diverting to Keflavik, Iceland. We also had to communicate with TWA dispatch to advise them of our emergency. The passengers and flight attendants could tell something was amiss and I advised them of our plans to land in Iceland.

Flight to St. Louis
Not the route for an engine failure.

An engine failure requires a strong muscular leg effort on the side of the operating engine to keep the rudder in balance. Fortunately, the pilot can relieve the pressure on his leg by applying rudder trim to balance the airplane. TWA’s procedure at the time was to fly on one engine with rudder trim until five miles from the runway, where pilots were required to remove the rudder trim and use leg muscles to keep the airplane aligned with the runway. This means the approach must be flown with an out-of-trim rudder.

TWA’s training included an engine inoperative ILS approach procedure. Having demonstrated proficiency in this maneuver only months ago, I was confident of flying a successful single engine ILS approach. However, Murphy’s law was lurking. We were informed that the ILS at Keflavik was not available and was shut down for maintenance. We would have to do a non-precision VOR approach to an altitude well below the prescribed safe landing minimums. The autopilot was not approved for a single engine non-precision approach. I would manually fly the approach.

A non-precision approach is a misnomer in that it requires more precision flying because of the lack of vertical guidance that the ILS glidepath provides. In the 1980s, a non-precision VOR approach was flown as a series of step-down fixes. After reaching one fix, reduce thrust and descend to a set altitude and then level off by adding engine thrust until reaching the next fix. At the Final Approach Fix start descent to a Minimum Descent Altitude (MDA) and level off until you see the runway or execute a missed approach. Managing the rudder trim on a multiple step down approach can be very challenging.

This approach would be far different from a normal approach. The 767 was designed without a fuel dumping system, which is used to reduce the landing weight. An overweight landing with a higher landing speed was our only choice. Also, the approach had to be flown at 20 knots faster because only partial wing flaps were to be used – full landing flaps create too much drag and severely limit the aircraft’s ability to climb, if necessary. This approach speed would be close to 200 MPH.

Keflavik airport
Keflavik is about the only good option over the North Atlantic.

TWA Flight 745 was cleared for the VOR Runway 20 approach. We crossed the Final Approach Fix at 1500 feet AGL, reduced power to descend to the published minimums of 440 feet above the ground and leveled off. Level flight required more engine power and more rudder trim. This was an emergency and using captain’s authority I was ignoring the prescribed procedure of removing all rudder trim.

At 440 feet above the ground, minimums for the VOR approach, we were still in the clouds and could not see the ground. We descended to 300 feet and started to see some ground. The control tower was watching us on radar and the tower advised us to go around because they thought we were too low. We were right on course and I told the non-flying pilot to report that we had the field in sight. We did not see the runway but I was confident we would see it soon. We descended to 250 feet and the runway approach lights came into view. We started down and crossed the end of the runway. As I reduced the power for landing, the relief pilot removed the rudder trim and we made a normal landing using a time-tested technique.

This adventure led to some procedure changes. TWA pilots were advised to hand fly at least one approach on every flight to maintain manual flying proficiency. Also, the requirement to remove all rudder trim prior to the final approach fix was removed. One more change since the Iceland adventure some 30 years ago: the new Boeing 777s and Boeing 787s have automatic rudder compensation during engine out operations. It is called thrust asymmetry compensation. This greatly reduces the workload of the pilot flying the aircraft and most importantly, it keeps the aircraft in proper trim.

Pilots must retain proficiency in hand flying the aircraft. Thirty plus years after this diversion, there are still many cases of veteran pilots losing control of the aircraft when hand flying. Two cases come to mind: an Air France Airbus A-330 crashed in the south Atlantic when the relief pilot hand flew the airplane into an aerodynamic stall and an Asiana Airlines 777 crashed during a manually flown visual approach in good weather at San Francisco. Once again, aviate, navigate, communicate.

Finally, about seven weeks after the Iceland incident, I received a report detailing the cause of the oil leak. A metal bolt, not part of the 767, was found in the oil system drain pan. The bolt had torn a rubber seal that allowed the oil to leak from the engine. How it got there is a mystery. Could it be “Murphy’s Law?”

13 Comments

  • Great article. I entirely agree, there have been, as you correctly observe, a number of accidents where the pilot(s) were unprepared to fly the aircraft. I remember my flight instructor drilling it into me: “Fly The Damn Plane!” And training certainly came to my aid when I suffered and engine failure on take-off. In aviation, it’s definitely as case of ‘use it or lose it!’

  • I am concerned about twin engine flight over the Pacific, where aircraft could be 5 hours from a suitable landing airport

    • How safe is flying twin jets long distance over water? It is not safety, it is the money!

      WHY FOUR ENGINES ARE BETTER THAN TWO…………
      Richard Wyeroski, Dec 19, 2012 … 24 25 26
      Replies:515
      Views:140,549

  • How is it possible that a scheduled carrier diverts to an alternate which is below minimums, with the precision approach presumably notamed out of service, and proceeds to fly a non-precision approach, single engine, below minimums?

    The potential here for turning an equipment malfunction into a disaster is breathtaking.

    Or am I reading this wrong?

  • Yes, most aircraft accidents are a culmination of many factors, not just one. In this case, a mechanical failure due to a maintenance error, only one suitable airport to divert to, landing aids out of service for maintenance, and bad weather., if not for Captain Lawler’s skill, his crew and great training by TWA, this could well have been an accident. He was fortunate to have a third pilot in the cockpit, as was the United Captain on the DC-10 in Iowa. Taking the third man out of the cockpit by bean counters was a big mistake, as was flying over water on two engines. I shut down several engines on the B-747 over the water and we and the passengers hardly were effected!

  • It is an emergency. A heavy jet on one engine is no picnic. That airport was the closest suitable and the weather was unknown ahead. Below mins is dangerous and they had luck with them. An example of doing a non precision approach recently to an airport in Alabama resulted in an accident with a cargo flight and the crew killed and aircraft destroyed.

    Why the Cap didn’t use the auto pilot to help him with the approach is Intersting ?

    • Then it is a dispatch failure. If, indeed, Keflavik was the only available alternate for that portion of the ETOPS trip, then there was no suitable alternate.

  • The auto pilot is not capable of flying the aircraft on one engine, constant power changes and rudder trim changes are beyond the capability of the auto pilot, as is landing with a 30 knot crosswind. Management and the media preach how, “the autopilot does all the work, our pilots are overpaid bus drivers”. I think Captain Lawler proved them wrong. Having said that, on the Lockheed 1011 with all the engines running, the three automatic pilots backed up by three flight directors could land the airplane in zero- zero weather with a ten knot crosswind, The first time you saw the runway was after touchdown and the nose coming down. Kelly Johnson at Lockheed was “one helluva an engineer!)”

    • I stand corrected on the auto-pilot. It was an emergency and this cap fits in the sully category……he had no other airport to go to!

  • TWA actually got close to 5000 pilots in 1970, our inept New York flight operations management did not see the recession looming on the horizon, many were hired and then fired while in their training classes in January 1970, then the seniority list shrunk to about 3500.

Leave a Reply

Your email address will not be published. Required fields are marked *