How much power are you losing to friction?
#1
01-26-2012, 02:25 PM
Join Date: Sep 2010
Location: Chicago, Illinois
Posts: 4,424
How much power are you losing to friction?
So a couple years back I had been doing some work on one of my race motors, when the conversation turned to parasitic losses.
The majority of the energy released in combustion in a piston engine is lost to heat, friction and light, but often it is hard to gather really just how much and put even a ball park number on it.
Well during a particular oil thread months ago on this board I posted up the shorthand I did for an individual engine I built using its specs and the specs available for a particular 10w30 dino oil I used to break it in as they were the most comprehensive specs available at the time.
Here are the excerpts from the FF thread I brought them up in as part of an exchange between myself and another poster:
Anyways, I found a paper done by R.I. Taylor an Engineer for Shell that uses an early 1990's Mercedes Benz M111.940 2.0L 4cyl with lower compression, lower BMEP, and looser tolerances than our Fits and compares it to the 3.0L F1 V10s from the late 90's.
http://www.eng.auburn.edu/~jacksr7/SAE2002013355.pdf
To understand some of the differences in what allows the losses from friction in the 18000rpm F1 motor compared to the numbers from the 2.0L OE motor here is a brief write-up from 1999 by Csaba Csere:
Yanking the Crank at 18,000 rpm
Some of the more pertinent bits I'll quote here:
So with all that context, it actually confirms the realism of my numbers. Which is why, in fact oil choice can make a significant impact on not only performance, but fuel economy and longevity.
The conclusion wound up finding that the F1 motor, with a special oil blend, and all the advancements and tricks of the trade was still losing as much as 60kw from just the valve train, bearings and piston bores. Not accounting for other pumping losses for oil, coolant and intake/exhaust cycles, as well as accesories driven from the motor and drivetrain losses.
Every little bit counts. Just some food for thought.
The majority of the energy released in combustion in a piston engine is lost to heat, friction and light, but often it is hard to gather really just how much and put even a ball park number on it.
Well during a particular oil thread months ago on this board I posted up the shorthand I did for an individual engine I built using its specs and the specs available for a particular 10w30 dino oil I used to break it in as they were the most comprehensive specs available at the time.
Here are the excerpts from the FF thread I brought them up in as part of an exchange between myself and another poster:
Some things cannot and should not be dumbed down for consumption by the lowest common denominator. Oil is one of them.
Fluid dynamics is not something you can just put in layman's terms and illustrate the mechanics of it in an accurate fashion.
You want an example of how a simple change in oil can create a significant difference in fuel economy and power (which are both related by the way, one often follows the other)
The following was originally intended for one of my boosted gas 2.0L's
There is something called Bearing Operating Condition(BOC) which can be one of 3 states, but you really only want it operating in the 3rd state, fully developed hydrodynamic lubrication. Which correlates to a BOC of 35+. The formula to find BOC is as follows:
BOC = Viscosity x RPM x Diameter x K / Unit Load
Viscosity is in units of absolute viscosity. "K" is a value to convert RPM and Diameter into surface speed. Unit Load is the applied force divided by the projected area of the bearing (the insert width times the journal diameter).
First, viscosity. I can't find absolute/dynamic viscosity numbers for various grades of motor oil, so I took the Kinematic Viscosity of 10w30 (dino) motor oil at 100*C and multiplied it by it's specific gravity. This should give me the Dynamic Viscosity.
Second is the value "K". To find surface speed you multiply circumference by RPM. So Diameter x Pi x RPM = Surface speed, therefore K = Pi Right?
Lastly, unit load. I used 12,000 lbs of force, but this is relative to motoGP/Formula 1 type hyper-square engines, not boosted 4cylinders. I know our engine's see a BMEP nearly twice that of a 2.4L V8 Formula 1 car, so it is a conservative place to start.
I also had no bearings out of the motor to measure their width, so I just put a caliper on the main cap and eyeballed it. I came up with an area of 1.425, so a load value of 8421.
Put it all together, 8.4 x 9500 x 2.245 x 3.1415 / 8421 = 66
The optimal range between 35 and 50 so we are well out of that range.
From here we can use the Stribeck Curve to find our friction coefficient, which I found to be .0057
So with an applied load of 12,000lbs, that gives us a friction load of 68.4 lbs. The diameter of the journal used is 2.245, so 1.1225 x 68.4 = 76.7 in/lbs or 6.39 lb/ft.
At 9500 this equates to 11.56 hp per main bearing.
That is 8600 watts of energy wasted per bearing, and that is only for the 5 main bearings!
That is not accounting for all the rockers, cam journals, cylinder bore contact patches, rod journals or wrist pins. Or the oil pump its self.
So maybe you should bone up on the basics, BC.
What units are used for measuring viscosity?
How does temperature affect viscosity?
What are the real differences between weights of comparable group oils at a given temperature?
These all are key to understanding what the hell I am trying to tell you.
None of this is made up or conventional knowledge. When I tell you something, think about it don't just throw it back in my face or I won't bother explaining it further.
There is method to my madness and I wouldn't be sharing if it didn't have merit for Joe Commuter perusing this board.
If you actually give a damn about learning the science involved for oil go here and read everything two or three times and absorb it:
http://www.stle.org/resources/lubele...n/default.aspx
Fluid dynamics is not something you can just put in layman's terms and illustrate the mechanics of it in an accurate fashion.
You want an example of how a simple change in oil can create a significant difference in fuel economy and power (which are both related by the way, one often follows the other)
The following was originally intended for one of my boosted gas 2.0L's
There is something called Bearing Operating Condition(BOC) which can be one of 3 states, but you really only want it operating in the 3rd state, fully developed hydrodynamic lubrication. Which correlates to a BOC of 35+. The formula to find BOC is as follows:
BOC = Viscosity x RPM x Diameter x K / Unit Load
Viscosity is in units of absolute viscosity. "K" is a value to convert RPM and Diameter into surface speed. Unit Load is the applied force divided by the projected area of the bearing (the insert width times the journal diameter).
First, viscosity. I can't find absolute/dynamic viscosity numbers for various grades of motor oil, so I took the Kinematic Viscosity of 10w30 (dino) motor oil at 100*C and multiplied it by it's specific gravity. This should give me the Dynamic Viscosity.
Second is the value "K". To find surface speed you multiply circumference by RPM. So Diameter x Pi x RPM = Surface speed, therefore K = Pi Right?
Lastly, unit load. I used 12,000 lbs of force, but this is relative to motoGP/Formula 1 type hyper-square engines, not boosted 4cylinders. I know our engine's see a BMEP nearly twice that of a 2.4L V8 Formula 1 car, so it is a conservative place to start.
I also had no bearings out of the motor to measure their width, so I just put a caliper on the main cap and eyeballed it. I came up with an area of 1.425, so a load value of 8421.
Put it all together, 8.4 x 9500 x 2.245 x 3.1415 / 8421 = 66
The optimal range between 35 and 50 so we are well out of that range.
From here we can use the Stribeck Curve to find our friction coefficient, which I found to be .0057
So with an applied load of 12,000lbs, that gives us a friction load of 68.4 lbs. The diameter of the journal used is 2.245, so 1.1225 x 68.4 = 76.7 in/lbs or 6.39 lb/ft.
At 9500 this equates to 11.56 hp per main bearing.
That is 8600 watts of energy wasted per bearing, and that is only for the 5 main bearings!
That is not accounting for all the rockers, cam journals, cylinder bore contact patches, rod journals or wrist pins. Or the oil pump its self.
So maybe you should bone up on the basics, BC.
What units are used for measuring viscosity?
How does temperature affect viscosity?
What are the real differences between weights of comparable group oils at a given temperature?
These all are key to understanding what the hell I am trying to tell you.
None of this is made up or conventional knowledge. When I tell you something, think about it don't just throw it back in my face or I won't bother explaining it further.
There is method to my madness and I wouldn't be sharing if it didn't have merit for Joe Commuter perusing this board.
If you actually give a damn about learning the science involved for oil go here and read everything two or three times and absorb it:
http://www.stle.org/resources/lubele...n/default.aspx
You're saying that at 9500 rpm in the particular engine you're using as an example there are frictional losses of almost 60 hp? This is a question, not throwing something back in your face.
When I was in HS an automotive engineer told me that at higher speeds it takes about a hp per mph just to overcome air resistance. Now that was before cars were more fluid so it's probably maybe 2/3 of that now. But still, that's where a lot of the engine's power goes at higher speeds- overcoming wind resistance.
When I was in HS an automotive engineer told me that at higher speeds it takes about a hp per mph just to overcome air resistance. Now that was before cars were more fluid so it's probably maybe 2/3 of that now. But still, that's where a lot of the engine's power goes at higher speeds- overcoming wind resistance.
Yes at WOT (full load) not accounting for all the other parasitic losses I am losing as much as 60HP at the crank on 10w30.
At cruise where load is much lower, at least on a modern car, I don't think it would even take 1 hp/mph. That will be dependent on gearing, rolling friction, drivetrain losses, and aero drag.
My cruising IDC's in my 2500lb, 0.29C/d Laser @ 75mph are usually less than 5% in 5th gear (very tall ratio), which would be about 290cc/min.
1450cc/min per injector times 4 injectors @ 5% IDCs = 290cc/min fuel delivered.
At stoich (which is what we see at cruise ~14.2-14.7:1AFRs depending on fuel) this is good for 5lbs/min worth of air flow, or about 50whp
So that may be about accurate. Seems a little high to me, but I can do some more math on that later.
At cruise where load is much lower, at least on a modern car, I don't think it would even take 1 hp/mph. That will be dependent on gearing, rolling friction, drivetrain losses, and aero drag.
My cruising IDC's in my 2500lb, 0.29C/d Laser @ 75mph are usually less than 5% in 5th gear (very tall ratio), which would be about 290cc/min.
1450cc/min per injector times 4 injectors @ 5% IDCs = 290cc/min fuel delivered.
At stoich (which is what we see at cruise ~14.2-14.7:1AFRs depending on fuel) this is good for 5lbs/min worth of air flow, or about 50whp
So that may be about accurate. Seems a little high to me, but I can do some more math on that later.
http://www.eng.auburn.edu/~jacksr7/SAE2002013355.pdf
To understand some of the differences in what allows the losses from friction in the 18000rpm F1 motor compared to the numbers from the 2.0L OE motor here is a brief write-up from 1999 by Csaba Csere:
Yanking the Crank at 18,000 rpm
Some of the more pertinent bits I'll quote here:
You'll notice that the rings are also rather thin. The compression rings are about 0.028 inch thick, and the oil-ring assembly is only 0.077 inch thick. Conventional rings are more like 0.063 and 0.125 inch thick. These thin dimensions not only reduce reciprocating weight but also help the rings maintain their seal through the pistons' frenetic gyrations.
The connecting rod and the piston pin are titanium, probably forged, making good use of that material's excellent strength-to weight ratio. The piston itself appears to be forged from aluminum with some darker coating to resist heat and reduce friction.
The connecting rod and the piston pin are titanium, probably forged, making good use of that material's excellent strength-to weight ratio. The piston itself appears to be forged from aluminum with some darker coating to resist heat and reduce friction.
That's why the engineers take every measure to reduce the weight of these components. Look at how little there is to this Ferrari piston. The piston skirts extend less than an inch below the compression ring. And for about two-thirds of the piston circumference (the nonthrust faces), there's no skirt at all!
At 18,000 rpm, which is the redline of a '99 Fl engine, each piston in the engine accelerates from a dead stop at one end of the cylinder to about 100 mph in less than one-thousandth of a second and then comes to a complete stop again less than one-thousandth of a second later. This start/stop cycle is repeated 600 times per second!
This is where the Ferrari Fl engine's short stroke and longish connecting rod come in. At 18,000 rpm, this combination of its 45.6mm stroke and 110mm connecting rod accelerates the piston at a peak rate of nearly 10,000 g. With the 14.9 ounce combined weight of the piston, rings, and small end of the connecting rod, such acceleration requires more than 9000 pounds of force, which is trying to rip the con rod in two or tear out the piston pin.
That's brutal enough, mechanically speaking, but if the V-10 engine had been designed with equal bore and stroke dimensions of 72.5mm and had a 130mm con rod, like the proportions in a Mustang GT V-8, then the peak piston acceleration at 18,000 rpm would be almost 17,000g accelerating the piston to a peak speed of about 160 mph. Even accounting for the smaller, lighter piston and rings, such a design would produce forces of about 12,500 pounds. That would tear apart either the piston or the con rod. And making these parts and stronger would just them heavier and increase forces ripping them apart.
This is where the Ferrari Fl engine's short stroke and longish connecting rod come in. At 18,000 rpm, this combination of its 45.6mm stroke and 110mm connecting rod accelerates the piston at a peak rate of nearly 10,000 g. With the 14.9 ounce combined weight of the piston, rings, and small end of the connecting rod, such acceleration requires more than 9000 pounds of force, which is trying to rip the con rod in two or tear out the piston pin.
That's brutal enough, mechanically speaking, but if the V-10 engine had been designed with equal bore and stroke dimensions of 72.5mm and had a 130mm con rod, like the proportions in a Mustang GT V-8, then the peak piston acceleration at 18,000 rpm would be almost 17,000g accelerating the piston to a peak speed of about 160 mph. Even accounting for the smaller, lighter piston and rings, such a design would produce forces of about 12,500 pounds. That would tear apart either the piston or the con rod. And making these parts and stronger would just them heavier and increase forces ripping them apart.
The conclusion wound up finding that the F1 motor, with a special oil blend, and all the advancements and tricks of the trade was still losing as much as 60kw from just the valve train, bearings and piston bores. Not accounting for other pumping losses for oil, coolant and intake/exhaust cycles, as well as accesories driven from the motor and drivetrain losses.
Every little bit counts. Just some food for thought.
#2
01-26-2012, 09:46 PM
I will bit. I have found info that I would like to share. I think a good oil with a good anti-friction additive will lower friction but its related to the bearing material and hows its made. The bearings in the Fit for example is not the same that is in a F1 motor.
http://dspace.mit.edu/bitstream/hand.../166145859.pdf The applies to all motors and is the most likely reason to friction.
Industrial Lubricants - Resources And Advice This explains the stribeck curve.
http://www.threebond.co.jp/en/techni...pdf/tech09.pdf
Trucks use a Group II and III oil and is under extreme loads most of the time. They hold up pretty good so they should be good in a car under the right set up. I did find out the black soot color is from the detergents that turn black when heated and not necessarily the soot.
There is new nano technology that help reduce friction to almost 0 but not sure how that works. http://www.lowerfriction.com/index.php <Thanks Tex
http://dspace.mit.edu/bitstream/hand.../166145859.pdf The applies to all motors and is the most likely reason to friction.
Industrial Lubricants - Resources And Advice This explains the stribeck curve.
http://www.threebond.co.jp/en/techni...pdf/tech09.pdf
Trucks use a Group II and III oil and is under extreme loads most of the time. They hold up pretty good so they should be good in a car under the right set up. I did find out the black soot color is from the detergents that turn black when heated and not necessarily the soot.
There is new nano technology that help reduce friction to almost 0 but not sure how that works. http://www.lowerfriction.com/index.php <Thanks Tex
Last edited by SilverBullet; 01-26-2012 at 10:39 PM.
#4
02-21-2012, 03:35 AM
#5
02-21-2012, 06:05 AM
Damn, the miniature V12 is crazy. 1220 hours to make it. Incredible detail.
SB, J's Racing has recently started selling nano additive. If works as advertised by M.K. it would be nice in both engine and tranny oils.
SB, J's Racing has recently started selling nano additive. If works as advertised by M.K. it would be nice in both engine and tranny oils.
#6
02-21-2012, 09:21 PM
I got side tracked but was reading that nano tech is not new. I have to find it but they have been working on this since the '40s.
Thread
Thread Starter
Forum
Replies
Last Post
