Airline Flight Planning in the Computer Age

In the early 1980s Qantas began to look for a replacement for its cumbersome and labour intensive ‘deck-log’ system. This system involved storing route information on large supplies of paper ‘deck-logs’ at all of Qantas en route ports. The main disadvantages of this system were the cost and logistics of replacing the entire stock of on hand deck-logs if something as small as a VOR frequency was changed. Instances also occurred of crews being presented with out of date logs and not discovering this fact until into the cruise. Potentially very embarrassing.

In 1982 Qantas signed a contract to use the Continental Airlines Phoenix system which at that stage was still under development by a Continental subsidiary company called Systemone.

The half-million dollar contract was signed after an evaluation of six other systems including Lockheed Jetplan and the KLM system. Qantas also investigated the possibility of developing its own in-house system, but baulked at the possible ten-year time frame before all the required features could be incorporated. The Phoenix system promised to provide a complete, ready to use flight plan including all relevant en route information from a worldwide database, alternate information including PNR, depressurisation and engine out details in an easy-to-use format.

It took another 18 months and at least a further one million dollars to customise the program for Qantas particular needs. The introduction onto the line was timed for the advent of the 767 into the Qantas fleet. This was partially due to the fact that Phoenix uses the same Jeppesen database that the 767 utilises in its Sperry Flight Management System.

The coal face of the Qantas CFP system is on the second floor of the new Flight Operations building at the Mascot Jetbase. It is manned 24 hours a day, 365 days a year by a rostered staff of 34 flight dispatchers. These dispatchers look after all flights originating in the Asia Pacific region, from Bombay to Tahiti, from Tokyo down to Christchurch. There are another 7 dispatchers in London who look after all flights departing Europe. A further 13 dispatchers are employed by Qantas’ Los Angeles handling agents, AOS, who oversee the flights departing the West Coast and Honolulu.

Talking to Tony De La Haye, whose duties as the Training and Standards Manager of the Flight Operations Services Department include ensuring that all dispatchers are capable of producing accurate, error-free flight plans, we find that becoming a dispatcher can be a long, demanding task.

“Dispatchers must know, not only the Qantas system but also the details of several other systems used by the twenty-three customer airlines that we service here at the Jet Base,” says Tony.

Prospective dispatchers must have at least the PPL theory subjects and are initially employed:is dispatcher assistants. To make the step up to fully fledged dispatcher, the SCPL Met and Commercial Navigation subjects must be passed and the first of two in-house courses must be completed. The course lasts four weeks and once concluded enables the dispatcher to work on simpler SYD MEL type plans plus all plans for foreign carriers. When sufficient experience has been gained, the dispatcher completes a further six week course and is then equipped to handle any plan on any route in the Qantas network including the complex European routes and the ultra-long SYD LAX flights, plus any occasional charter that comes up.

“The quickest that any person has reached this standard is about two years while the average time taken is closer to four years,” relates Tony.

To gain an idea of how the system operates on a practical basis, let us watch the series of events that culminate in the production of the flight plan for the QF11 7 47-400 non-stop service from SYD to LAX.

Approximately three hours before scheduled departure, the dispatcher rostered to handle the flight {they can deal with anything up to ten flights a day) begins to collate the information he will require. This is in the form of sector notes which are all contained in the Jeppesen database. These include details on the route, available alternates, possible problem areas such as volcanoes, and special instructions such as fuelling guidelines peculiar to the route being flown. An example of this might be instructions to fill the tanks to the maximum available instead of the amount actually calculated by the program. This is done in order to enhance the integrity of the schedule and is common on long haul flights.

About ten minutes later the dispatcher reviews what is known as the planner messages. These are the TAFs and NOTAMS for the destination and any alternates. From this the dispatcher determines the minimum fuel requirements for arrival at destination. About forty minutes later the system automatically begins to compile what are known as key words for the plan. These are words peculiar to each of the different programs, Phoenix or one of the customer programs, that are used for the compilations of different portions of the plan. It is at this point that the computer begins an analysis of all the different possible tracks in order to find the one with the minimum fuel burn. There are ten tracks for this route stored on the computer plus a possible further ten if the computer is instructed to utilise random routing.

You can imagine how long such a procedure would take to do manually. The computer, which is the Qantas mainframe, and not a computer dedicated to this task, completes these long, complex plans in a maximum of about two and a half minutes. Shorter plans can take a lot less depending on how busy the computer is. To do it, the computer uses the worldwide upper air database from the US National Met Centre at Suitland, Maryland, USA. This record of windspeeds, direction and temperature is one of two compiled throughout the world and recognised by ICAO. The other is collected at Brecknell in the UK. Both are updated every twelve hours. In the Australian region, grid point data from the Bureau of Meteorology is used in preference to Suitland. The Suitland Database is supplied to Qantas by high speed teletype and processed directly into the computer.

The program in most cases works backward. That is to say that the computer takes the zero fuel weight, adds the minimum fuel required at destination including the variable reserve and then uses this weight to work back through all the portions of the flight plan to arrive at a brakes release weight and a required fuel load. The fuel burnoff is modified by a correction factor for each individual aircraft in the fleet compared to the manufacturer-supplied data. This information is updated by the Qantas Performance Engineering Department about once a week.

At this stage the plan is more or less complete, however the dispatcher usually reruns it through again at least once or twice to check that the computer has delivered the best possible resolution of the plan. Minimum fuel burn is the main criteria but sometimes the plan must be checked for details such as constant Mach number cruise in US airspace instead of the usual decreasing Mach number economy cruise schedule.

A dispatcher summary is now produced which contains all the salient points of the plan on one sheet of paper. This is rechecked and the plan is held in the computer memory until fifteen minutes before the briefing time of one and a quarter hours before scheduled pushback. The system then automatically releases the plan to be printed for presentation to the operating crew. This late release of the plan is to allow for any last minute changes to the weather forecasts.

It is important to note that while the plan takes into account all aspects of the Qantas fuel policy, the operating Captain retains final authority over the amount of fuel that is ultimately ordered. It is well within the Captain’s perogative to take more fuel than the system calculates and it is quite usual to do so.

Allowance for the effect of carrying extra fuel is made in the fuel analysis section of the plan and it is usually considered unnecessary to print an updated plan incorporating the new figure unless the ZFW varies by 5,000kgs in the 747 or 2,500kgs in the 767.

The operating crew will then check that the printer has delivered all the portions of the plan and that all portions carry the same retrieval code so as to ensure that they have the correct plan and not part of someone else’s. The fuel order form is then incorporated into the document and is signed by the Captain and checked by another crew member. A hard copy of the plan is then kept on file and two copies are used by the flight crew.

What happens if the computer is on the blink? A backup plan can be ordered from Systemone in Los Angeles using either the Suitland or Brecknell databases. It won’t have all the features of the Qantas plan or be in the same format, and hence would only be used to avoid a delay to a service. Other methods of transmission to out-stations exist and include the AFTN network, telex or even a Fax if all else fails. If worse comes to worse and the plan simply cannot be sent through, out-stations hold copies of recent flight plans which could be used as a last resort.

An interesting feature of the Phoenix system is the ability to transmit updated fuel-time analyses to airborne AGARS equipped (currently 767 and 747-400) aircraft. This is very useful in cases such as unforeseen diversions or instances where en route winds vary significantly from the forecast and render the original planinaccurate. The current system is on trial and uses a representative ZFW to complete its calculations. Crews will be able to send actual ZFW to the computer in a future AGARS update. These details and the required route are sent to the computer via the AGARS link and an updated plan is automatically calculated and transmitted back to the aircraft where it is reproduced on the aircraft printer.

This can be a very quick process. A recent example of a request for a diversion plan to Melbourne from a 767 approaching fog bound Sydney was received, relayed to the Qantas mainframe (which admittedly was not particularly busy at the time), processed, relayed back to the transmitter and then to the aircraft . . . in three seconds! Just think about that the next time you are at the two and a half hour mark of the ATPL Flight Planning exam and still trying to figure out if you should work backwards or forward.

The main limitation of this feature is the need for the aircraft to be within 200nm of an AGARS transmitter. Eventually, all AGARS data will be transmitted by satellite. It is obvious that CFP systems can quickly recoup the high initial outlays required to install them. The Qantas Phoenix system and others like it provide fast, accurate and flexible flight planning for airlines that need every possible advantage in the cut-throat environment of modern airline flying.