Nitpick: jets drink avtur, not avgas.

Practically speaking, they don't like to carry needless fuel, but at the same time they need at least enough to get from A to B, hold for a while, divert to C, hold for a while, and then land with the legal minimum reserve of 45 minutes usable fuel - plus enough to account for the winds aloft deviating from the forecast.

Google tells me that common burn rates on a 737 in cruise are 30-40kg/min (though they will of course vary hugely given which model, the payload, the weather and any ATC flight level restrictions). Minimum arrival fuel is typically about 3 tonnes. Running some numbers I found under a rock - assuming the extras for holding and diversion aren't needed - a typical flight might take off with 6500kg + 35kg/min enroute time, and land on maybe 5000kg. (This assumes that the price of avtur is largely the same at both ends, and that you aren't pulling sneaky tricks like fuelling up for two or more flights worth at once so you can speed up your turnaround, which is common practice on short-haul ops.)

YVR-YUL on a great circle route is about 2000 nautical miles. A 737-700 ideally cruises at M0.78; approximating the numbers horribly I make that maybe 4.15h cruise time (+0.25h climb and 0.33h descent already factored in). Plugging 249 minutes into the above fuel calculation gives us 8715kg of cruise fuel, a take-off fuel weight of 15.2 tonnes with an expected in-flight consumption of 10.2 tonnes.

Now, the density of avtur is about 0.8 at 15°C, so our hypothetical 10.2t of fuel consumed would measure at 12750 litres.

Thanks for the clarification of figures. I was giving a most-pessimistic figure for the plane, and assuming that the tanks were completely emptied. What I'd not checked was the actual range, which is somewhat further than I'd realised.

So, that means the plane is about half as thirsty as I thought.

Interesting and largely irrelevant factoid from my bygone days in the Airline booking industry:

Japanese domestic carriers, which tend to carry large numbers of passengers on short island-hops, use custom 747s with modified undercarriage to allow them to land with a full load of fuel.

The fuel consumption comes from the two locos (each quite heavy itself, of course) each having a 6000 litre tank, and 6 fuelling stops.

There is a point that could be raised about how empty those tanks are when refilling is required, but if it could do it on 4 stops, say, then why refill 6 times?

I do think it's an example of the worst possible performance for a train. Why it appears quite that thirsty, I can't say, though I suspect the locos are optimised for pulling power rather than speed: some of the freight trains coming the other way were a mile long, with only a pair of locos pulling, so it may be that the whole train was vastly overgeared.

(For speed, it's limited to 70 mph top speed.)

http://www.oetr.on.ca/oetr/uploads%5CNIC%5C201%5C2009A%5C201_2009A_supprt4.pdf has some fuel consumption figures for VIA Rail's F40PHs. In particular, Table 5 has a consumption figure of 208.8 litres/hour, which would suggest 40,000 litres for 2 locos for 100 hours.

OTOH, this appears to be based (Table 12) on a duty cycle of (approx) 50% idle, 20% Notch 1 (min non-idle), 20% Notch 8 (max) and 10% other powers, and probably reflects time stabled between trips as well as time on the road with a train.

Indeed, FIgure 1 suggests about 500litres/hour for a loco producing 3000HP (the F40PH's rated power).

So I'd guess that 50-60,000 litres would be about right.

To support passenger service, the F40PH is equipped with a secondary electrical generator known as the head end generator. The HEP unit generates three-phase AC power at 480 v AC 750 A or about 500 kW to provide power to the cars for lighting, heating, and air conditioning. Because head end power must be delivered to the rest of the trainset at a constant frequency, the prime mover in these locomotives must turn at a constant high speed (900 rpm) while delivering head end power (even if the locomotive is standing still). Power to the traction motors is controlled by varying the field excitation of the main (traction) generator.

So the engine basically runs full speed all the time, unless the whole train is turned off. That's not going to help efficientcy at all.

Full speed, but not full load.

Yes. Still, there's going to be all that friction internal to the engine.

Hmm, I wonder if the use of twin engines means that the one of the two doesn't have to do this.

I suspect they're running on one engine at a time under "no load". The one larger ship (a cruise ferry on the Stockholm-Helsinki route) that I've looked inside more closely (A Silja Line ship, can't remember the exact ship, though) had four massively huge 12 cylinder diesels, running between 1 and 4 constantly, at the peak-efficiency RPM, driving one or two generators through gearboxes.

As I understand it, they run on a single engine while docked, one engine on each generator while at sea, but having the option of pulling more power if they need it. They also have accumulator banks for produced, but not used, electricity.

We were told that the drivers work 8 hour shifts, and are changed for a new set after that shift. At such times, the train stops to let them off. (The drivers then wait for the next train to come through (2 days later) before then doing their next shift.)

So there are 10 stops for driver changes. All other crew are based out of Winnipeg, with staff working non-stop out to Vancouver and back, or to Toronto and back, and living on the train. Any particular employee works one side or the other of the country, on a 6 days on, 8 days off shift pattern.

I do wonder whether the fact the carriages are 1955 vintage stainless steel makes them much heavier than modern ones would be. The outside surfaces did feel pretty damn solid, with no sign of any denting or distorion after half a century's use.

Where's that figure of 72,000 litres come from?

We called in at the railway museum in Winnepeg during our stopover, and were shown a film about the trains. The train is pulled by 2 diesel engines, the tanks on each engine hold 6000 litres and the tanks are filled 6 times to make the journey. (6,000+6,000)x6=72,000

I think the repeated assumption that the tanks are empty when fuelled here is basically flawed and sounds more like the kind of claim Michael O'Leary would make... *cough*

Firstly, the tanks won't be empty at each refuelling stop. In fact, I'd guess they'll be half full or more, for the obvious reason that if for some reason there are supply problems and fuel is unavailable at one point, you want to be able to have enough to get to the next fuelling point. You also want a hefty surplus, as in places like Canada there are real risks of being stuck in the middle of nowhere courtesy of unexpectedly heavy snowfall or mechanical breakdown, and you'll want to keep the train supply available for anything up to several days because in the sort of climate Canada has people could freeze to death otherwise.

I'd suspect that while it's highly unlikely they could cross the continent on a single tank of fuel, it's simply good practice to top off the tanks when the opportunity arises. Also, the weight of a full tank of fuel is irrelevant compared to the weight of the train as a while (hell, in some cases the heavier the loco the better) - unlike aircraft, where a significant amount of fuel is burned simply carrying the contents of the fuel tanks around. The actual fuel consumption for the trip I have no idea of, but I'd be willing to bet it's a hell of a lot less than 72000l.

You're certainly right that, compared to a 70 tonne take-off weight 737-700 carrying 20 tonnes of fuel, a 600 (?) tonne train carrying 10 tonnes of fuel is insignificant.

The claim by the railway museum is that the tanks are 'refilled'. Not 'topped off', but 'refilled'. Given the traffic on the route, I don't think you're going to find anywhere that is that far from a refuelling place: the train has stops listed at roughly 100 km intervals, even if the majority aren't used. So your hypothesis that it's got to worry about running out of fuel seems a bit of a stretch to me.

(Unlike the case of the Gimli Glider, a train running out of fuel has a rather unexciting failure mode.)

So, while I'd agree that the tanks won't be empty when the refueling points are arrived at, I think timill's guess of an actual consumption of 40 kl is probably about right.

It is, for a train, hideously inefficient. The carriages are 1950s vintage, built like tanks. (Which may be why it requires two locos to pull them, when the mile-long freights we were passing only ever got up to three.) The locos are somewhat newer, but still not exactly new. The whole setup dates to an era when fuel was dirt cheap. And the number of passengers for that number of coaches is miniscule. All in all, if there's a less fuel-efficient train running anywhere in the world, I'd love to know where.

(I stress 'passengers' there - because there's the train crew being carried too, and that's another 21 people.)

Trainsets are very heavy -- they have to be for the smooth steel wheels to provide enough friction on the smooth steel rails to drive a very heavy vehicle at a reasonable speed. It takes a lot of energy to get a trainset up to "cruise" speed and keep it there.

A long-distance passenger trainset with, say, 10 carriages plus one or two locomotives can weigh over 300 tonnes; in contrast a 737 aircraft typically weighs less than 70 tonnes at takeoff and it gets noticeably lighter in flight.

Given that most of the route is single track with passing places, it also has to get up speed again quite frequently.

(In Canada, passengers give way to freight.)

Our trainset was 2 engines and 19 carriages.

With 19 carriages and 2 locos it's probably about 500-600 tonnes all-up weight, so that's about 2.5 tonnes of vehicle weight per passenger (assuming 250 passengers in total) compared to about 0.5 tonnes vehicle weight per passenger on a 737. The fuel load for the trip is maybe 500 tonnes (it's unlikely that the tanks were empty before each refuelling stop) or about 2 tonnes of fuel burnt per passenger-trip. At UK street prices that would be about 2000 quids worth...

A quick Google shows that a 737 cruising at 510-520mph and 35-40,000 feet requires about 5000lbs of thrust to maintain airspeed. Getting up there from the ground requires a lot more thrust of course, but at the cruise altitude air pressure is about 25% of that at the ground, hence the air resistance is a lot lower. At Concorde's cruise altitude of 60-65,000 feet the air pressure is a lot lower, about 5-6% of sea level pressure which explains how it could fly Transatlantic supersonic on one load of fuel.

The fuel load for the trip is maybe 500 tonnes?

I think that may be an order of magnitude out: 50 tonnes seems a more reasonable estimate for the weight of <72 kl of fuel.