Power vs Torque Questionnaire -RESULTS

[Charlatan]

As a mod it seems the thread has been well behaved and Machin is more than capable of explaining simple physics.
...

This seems to be the very problem, as soon as you bring in "torque" you stray away from simple physics, where the only relations you need is Power equals Force times Speed and Force equals Mass times Accelleration. Everything else is irrelevant.

[Tim.Wright]

This seems to be the very problem, as soon as you bring in "torque" you stray away from simple physics, where the only relations you need is Power equals Force times Speed and Force equals Mass times Accelleration. Everything else is irrelevant.

The last I looked any physical system can be represented using energy methods or force/moment methods. Both are equally valid, both are used for different jobs. Saying one is "straying from simple physics" is ignorance in the extreme. By the way, torque is a force. So if you think F=ma is valid then so it T = Ia.

In every vehicle model I have ever built, the engine/powertrain is defined using a torque curve. The traction control, engine control and differential control algorithms that I have seen and worked with all use engine torque to calculate wheel forces in their internal modelisation. When I want to calculate the yaw torque due to the differential locking I use torques/forces.

For dumbing down the longitudinal performance of a vehicle into a single number so any muppet can understand it - I use power. I also use power and drag to calculate the top speeds.

[Miguel]

The red curve in the chart above is my "power" curve, and can be converted from (Nm x RPM) to watts by dividing by 9.55. Which gives me a peak power output over 30 seconds of about 620 watts.

Extremely OT: Machin, from a pretty untrained eye, and without knowing your weight, your numbers look really impressive to me! That's almost 50% the peak power of a WT sprinter!

The last I looked any physical system can be represented using energy methods or force/moment methods. Both are equally valid, both are used for different jobs. Saying one is "straying from simple physics" is ignorance in the extreme. By the way, torque is a force. So if you think F=ma is valid then so it T = Ia.

Warning: nitpicking ahead. Technically, if you really, really dig, energy methods are more fundamental. Quantum mechanics are explained and developed from a Hamiltonian point of view, quantum field theory (and, I believe, sting theory) use the Lagrangian formalism to develop Feynman diagrams. Getting classical forces in a quantum system requires dealing with expectation values. I'm essentially unqualified to talk about how people deal with general relativity.

In any case, and even if power can give you the maximum acceleration at a given point (P=dE/dt), dealing with torque and forces seems the clear way to go when computing car dynamics.

[Charlatan]

...
For dumbing down the longitudinal performance of a vehicle into a single number so any muppet can understand it - I use power. I also use power and drag to calculate the top speeds.
...

That's not dumbing down anything, it's simple physics.

A) A given Power will yield a resulting Force at a cetain speed, as Power equals Force times Speed.
B) Deduct air-resistance Force at the same Speed.
C) The amount of Force left, if any, will translate into Accelleration as Force is Mass times Accelleration.

That's all there's to it really, no need to involve torque whatsoever.

[Tim.Wright]

That's all there's to it really, no need to involve torque whatsoever.

Apart from all the real.world examples that I listed?

[Richard]

That's all there's to it really, no need to involve torque whatsoever.

As you rightly say, the various units are mutually dependent so it is possible to skip one of them because it is inferred by the other units. We'd all agree that's the quickest way to work out the acceleration.

However sometimes it is convenient to use a different unit when working on different problems. I'll not take this off topic by debating example units used in my world. I can imagine analyzing ABS is best undertaken with torque because the engine power is irrelevant. Then if you want to extend that to traction control it might be easier to express the engine input as torque so the numbers can be easily compared with the brake and traction forces.

[Stradivarius]

Apart from all the real.world examples that I listed?

As I have tried to point out earlier, there is a difference between evaluating the performance of an existing car and developing a new car from scratch. The former can be done knowing just a few parameters, while the latter requires a lot more knowledge and information which is not necessary to know in order to evaluate the performance.

It's important to appreciate the difference between designing somenthing and knowing its performance or capacity. Some small bridges are not capable of supporting the weight of the heaviest vehicles driving on the roads. If you know that the bridge can support 10 tonnes, and that your car weighs only 2 tonnes, then you have all the relevant information you need to conclude that you can safely cross the bridge. But in order to design a bridge that can support 10 tonnes, there is a lot of other parameters you need to know, like all kinds of dimensions and material data which are simply irrelevant for a driver who wants to know if can cross the bridge or not with his vehicle.

[Richard]

I think there was also a cultural thing. Macin proved that two black boxes with appropriate gearing had a very similar performance at the wheel. However one of them was diesel which triggered a knee jerk response based on the emotion of driving a diesel engine.

So to move this on a bit with a rhetorical question, is that knee jerk response valid? Is there any difference in the responsiveness or drivability of a diesel with the same torque curve at the wheel as a petrol engine? After all Macin's analysis says they should be identical.

[machin]

I think there was also a cultural thing. Macin proved that two black boxes with appropriate gearing had a very similar performance at the wheel.

To be clear; Autocar magazine proved that "two similar cars with similar power curves (but hugely different crankshaft torque), had very similar performance" (Mercedes also claim the same thing). This is irrefutable.

All me and others have been trying to do is explain why this is so. Even if anyone disbelieves the explanation, the fact still remains that the two cars have very similar performance in terms of both acceleration at all speeds, and top speed.

[mrluke]

Responsiveness is really a ratio of available power : resistance forces.

You can achieve a much higher rate of change of acceleration with a car that has lots of power available close to the typical cruising rpm.

So in a very highly strung NA car you may drop it down a gear or two to get into the required portion of the power band whereas in a big capacity V8 (or modern diesel) whatever gear you are in you can just put your foot down and go. Also see electric cars which have a pretty much flat power "curve" whereby throttle position correlates better with available power than the rpm range does.

[Charlatan]

That's all there's to it really, no need to involve torque whatsoever.

As you rightly say, the various units are mutually dependent so it is possible to skip one of them because it is inferred by the other units. We'd all agree that's the quickest way to work out the acceleration.
...

When you have said it in so many words, isn't this the time for you as a moderator to cap the discussion here and now?

[Richard]

I see your post chopped out my caveat.

So "Yes" for calculating acceleration, but there are circumstances when torque is a useful expression of what is going on.

[Charlatan]

I see your post chopped out my caveat.

So "Yes" for calculating acceleration, but there are circumstances when torque is a useful expression of what is going on.

Kindly elaborate on that last part?

[strad]

you guys sure like to try to pick the fly crap out of the pepper.

[Stradivarius]

I figured it could be illustrative to give an example to demonstrate what information is necessary to know in order to evaluate engine performance, and thereby also demonstrate what information is not necessary to know.

The following data is given about a car:

The mass of the car is 1600 kg. The car has 6 gears (forward) and the ratio between successive gear exchange ratios are [0.55 0.6 0.7 0.8 0.85]. So the gear exchange ratio of 2nd gear divided by the gear exchange ratio of 1st gear is 0.55. The gear exchange ratio of 3rd gear divided by the gear exchange ratio of 2nd gear is 0.6, etc. So when shifting from 1st to 2nd gear, the engine speed immediately after the gear shift will be 55% of what it was immediately before the gear shift. When shifting from 2nd gear to 3rd gear, the engine speed will drop to 60% of what it was immediately before the gear shift and so on. The engine power curve is as follows:


Finally, the top speed of the car is 280 km/h and this is achieved at the peak power which occurs at 92% of maximum engine speed.

With this information and nothing else, it is possible to calculate how the car will perform in terms of engine-limited performance. As a first approach I am ignoring aerodynamic drag and I am also ignoring friction in the drive train as well as rolling resistance and I am not considering the rotational energy of the wheels and other rotating parts. With these simplifications it is possible to develop the speed vs time curve which should be a good measure of the engine performance when the mass is given.

The first step towards getting to the desired result is to develop the engine power as a function of speed:


This can then be used to develop the speed vs time curve:


Here I have set an upper limit for the acceleration of 8 m/s^2 in order to account for the limited traction of the car, hence the linear behaviour at low speed where it is not the engine that limits the acceleration. I am thinking about also including a simple model for the air resistance, assuming the drag is proportional to the square of the velocity so that the terminal velocity is reached at 280 km/h. I could also include the moment of inertias of the wheels in order to arrive at more accurate results, but I think this simplified apporach is just as valuable in this discussion.

My point with doing all of this is that here we clearly have all the information we need to calculate how the car performs. But we do not have any information about the torque and it is also impossible to determine the torque from the information available. This could be a car with a peak torque of 600 Nm or it could be a car with a peak torque of 300 Nm or something completely different. There is absolutely no way to tell which it is. This proves that torque is indeed an irrelevant parameter. You may use the torque to find the information you need, but you don't need the torque as this exercise demonstrates.