The full picture
What happened
Before departure, the system for measuring the amount of fuel worked with a malfunction, so the stock was determined manually by volume and density. The calculation required to convert liters to kilograms, but a pound-per-liter coefficient was applied. The number looked plausible and did not cause the process to stop.
At the height, one engine failed, then the second. The generators stopped working, some of the displays went out, and the emergency turbine retained the minimum necessary pressure in the hydraulic system. The crew chose the former Gimli airbase, which the commander knew by service.
The plane was above the target trajectory. The commander used a slide to increase the resistance without overclocking. At the landing, the front stand was formed, but all 69 people survived. On the former lane at that moment there were car events.
Context and mechanisms
Detailed account
Background, technical details and decisions are examined separately to avoid reducing a complex event to a single explanation.
Metric system within a non-metric world
The Boeing 767 C-GAUN was one of the first aircraft of this type in the Air Canada fleet. He brought a two-membered cabin, electronic displays and a calculation of the mass of fuel in kilograms. The rest of the fleet and much of the airline's ground practices were still using pounds. The transition was not in one day: the aircraft, documentation, personnel and the usual coefficients were in different systems.
The route began in Montreal on July 23, 1983, with a stop in Ottawa and a follow-up to Edmonton. There were 61 passengers and eight crew on board. Before departure, a malfunction of the FQIS fuel quantity indication system was known. A specific combination of failures and actions performed created uncertainty as to whether the aircraft could be produced and how to accurately determine the stock.
Regularly, FQIS measures the fuel in the tanks, calculates the mass and shows it to the crew. With an inaccessible indication, the quantity was determined by measuring lines through special points of the tanks: they received a volume in liters, then multiplied it by density. The procedure looks simple, but requires the correct unit of density and independent verification of the result.
For jet fuel, a value of 1.77 pounds per liter was used. It was taken as kilograms per liter, although the correct metric density was about 0.8 kilograms per liter. The number 13,597, derived from 7,682 liters, was recorded as kilograms. The mistake did not look absurd and therefore went through several participants in the preparation of the flight.
Why several checks did not stop the error
The flight required about 22,300 kilograms of fuel. Since the calculated residue was mistakenly considered equal to 13,597 kilograms, an addition of approximately 8,700 kilograms was determined. Then this mass was again transferred to liters for the tanker. In fact, the original balance was expressed in pounds, so the aircraft received significantly less than half of the required additional mass.
Calculations were repeated by ground specialists and the crew, but all used one wrong coefficient. This is not an independent test: if the second person repeats the same method with the same unsigned unit, he confirms the arithmetic, but not the original assumption. The protective barrier should differ in the method - for example, compare the total volume with the expected consumption or require an explicit signature of kg/L.
The uncertainty increased the documentation. The staff did not have a simple, unambiguous and well-established instructions for the new metric aircraft with a broken regular system. Responsibility was shared between service, refueling and crew, but no one formed a complete picture. The plane took off with a number that looked consistent on paper.
That is why the investigation did not reduce the event to the inattention of one tanker or pilot. The error was made possible due to the corporate transition between units, the admission of flight with a critical fault, unclear procedures, and the repetition of a general erroneous assumption. The system created a situation in which professional people could get the same wrong answer.
Silent failure at 41,000 feet
In cruise flight at an altitude of 41,000 feet, low fuel pressure warnings appeared. At first, they could be mistaken for pump failure: at height, fuel is able to flow to the engine alone, and a separate pump failure does not mean an immediate stop. But soon the left engine stalled, and the crew began a decline with the intention of leaving for Winnipeg.
Then the right engine stopped. The main sources of electricity and hydraulic power disappeared, many screens went out. In the stream came ram air turbine — a small emergency turbine. It drove a hydraulic pump and maintained minimal controllability, but its efficiency depended on speed. The crew could not slow down too much, even when approaching landing.
The co-pilot estimated the range by vertical speed and altitude. The plane didn't make it to Winnipeg. Commander Robert Pearson recalled the former Canadian Air Force base Gimli, where he had previously served. The co-pilot knew her area. There was no pre-arranged navigation data for the operating civil airport, because the site had already been converted.
Gimli's choice was not a romantic insight, but a practical use of available knowledge of the area. The alternatives were further away, the engines could not be restarted without fuel, and each minute of uncertain course reduced the achievable zone. The cockpit recorded the target, the flight attendants prepared the passengers, and the plane continued planning to the old lane.
Wide-body liner slide
When we got to the airport, it became clear that Boeing was too high. It was impossible to remove the thrust - it did not exist - and the early release of the chassis and mechanization gave a difficult predictable energy consumption. The commander used a straight slide: he turned the aircraft relative to the incoming stream, creating great resistance and a rapid decline without the same increase in forward speed.
This maneuver is common for small gliders, but extremely unusual for a heavy Boeing 767. Because of the slide, the air flow approached the fuselage from the side, the readings of individual devices and sensations became unusual. The crew had to stop maneuvering at the exact altitude, align the aircraft along the strip and maintain a speed sufficient for the operation of the emergency turbine.
There was a family car day on the former lane. Part of the cover was used as a track, people and cars were right in front. The front strut was not fixed and formed after touching; the nose of the aircraft descended on the coating, creating severe friction. This helped to stop to the main group of people. There was a small fire that was quickly extinguished.
All 69 survived. The landing was made famous by the crew's skill, but the investigation left a more rigorous lesson: outstanding piloting was the last barrier after a series of preventable organizational failures. Security should not rely on the commander to be an experienced glider and accidentally know the abandoned site within the planning.
Sequence
Event timeline
Rather than isolated episodes, this was an unfolding system: each stage inherited the constraints of the one before it.
- 01
Incorrect fuel indication
Instead of the usual automatic reading, crew and ground personnel used manual measurement and mass calculation.
- 02
Invalid coefficient
The volume was multiplied by the density in pounds per liter, although the on-board system was expecting kilograms. The tanks received significantly less fuel than required.
- 03
The first pressure signal
The fuel system warnings were initially interpreted as pump failure rather than total stock depletion.
- 04
Both engines stop.
After the second failure, the cabin was partially de-energized. RAT went into the stream and retained control of the main surfaces.
- 05
Course for Gimli
The plane did not reach the major airport. The former base was closer and was familiar to the commander, although it was already used as a racetrack.
- 06
Slip and touch
By lateral gliding, the crew quickly reduced altitude. The nose strut was not fixed, formed during the run and created additional braking.
More than one cause
Chain of contributing factors
An incident develops when several weaknesses come together along the same path.
Transition between systems
The metric plane entered the park, where many documents, tools and habits were still based on pounds.
FQIS Malfunction
The failure of the standard measurement forced the use of a rare manual procedure, vulnerable to arithmetic, units and transmission of information.
Probable result
The error did not produce an absurd number. It gave a value that went through several human checks without independent verification of units.
Improvised landing
Knowledge of the terrain, traction-free planning, and sliding techniques have turned fuel exhaustion into a surviving accident.
From the flight deck
Decisions and constraints
Decisions are assessed against the information, altitude, time and aircraft condition at that moment, rather than what became known later.
It wasn’t just a mistake in arithmetic.
The system allowed exploitation under non-working indication, did not provide an unambiguous method of calculation and mixed metric and imperial units. The human error went through the organizational holes.
How to fly a plane without engines
The air raider rotated the emergency turbine. The lower the speed, the lower its performance, so the crew had to simultaneously save speed and control the trajectory.
Why did they use slip
An ordinary turn could take the plane away from the lane. The slide allowed to increase the resistance and vertical speed, keeping the direction to the site.
Summary
Investigation findings
- The aircraft was not sufficiently fueled due to the application of the wrong volume-to-mass conversion factor.
- The failure of the measurement system and ambiguous procedures created the conditions for error.
- Both engines stopped due to complete fuel consumption, not due to their own technical fault.
- The crew performed a successful landing without thrust at the former airbase; there were no fatalities.
After the investigation
What aviation learned
A lesson becomes useful when it leads to a procedure, training, a design change or a new safety barrier.
Units as part of the interface
Units of measurement should be explicitly signed in all forms, instruments and calculations, and conversion should be independently verified.
Restriction on release
Failure of critical indication cannot be compensated by informal procedure without accurate documents, equipment and responsibility.
Corporate Transition
The introduction of a new measurement system requires changing the aircraft, training, ground processes and verification culture at the same time.
Its relevance today
Unit of measurement - a full part of security
Gimli is often used as a funny story about pounds and kilograms, but its strength is precisely in the absence of one ridiculous mistake. All participants performed familiar calculations, the numbers looked reasonable, the signatures of units were not noticeable enough, and the new aircraft existed inside the old infrastructure. This is how the most persistent system errors occur.
The protection is dimensional analysis: each number is written together with the unit, the coefficients are not transmitted orally, and the result is checked in an independent way. If an aircraft needs twenty-two tons, a rough estimate of the volume should show that several thousand liters can not give such a mass. Rough verification of the order of magnitude is sometimes stronger than accurate arithmetic.
The landing proved the value of basic piloting and terrain knowledge, but did not justify operation with ambiguous indication. The modern system seeks to stop the chain before departure. An emergency turbine, glider skill, and empty lane should remain a backup, not a way to compensate for a weak refueling procedure.
This case is also important for interface design. A field without an explicitly specified unit, a coefficient that can be confused, and the absence of an automatic range check form a single error-prone system. A good interface doesn’t just show the number: it reports the dimension, the source, the expected range and warns of a physically implausible result. This principle is equally applicable to fuel statements, electronic tablets, medical dosages and any process where incorrect conversion looks arithmetically accurate.