Use of "N" for coasting

qbee42,

Interesting thought. I never though of that. It does make sense to me. But that is true if ALL valves stay closed when engine is freewheeling.

What if Valves operate the usual way except that fuel injection does not happen. Then I feel your argument would not be valid. If engine keeps taking new air in compressing it and recovering a part of it during power stroke, I feel that will still lead to extra losses. Valve timing and time it takes for valves to operate would lead to increased losses.

My question was really more like is valve operation controlled differently/ If crank shaft and cam shaft are mechanically connected then I do not think valve operation can be any different than rotation of engine.

 

You have asked a reasonable question with no simple answer. In a conventional engine with fuel shut off there are still pumping losses. The intake stroke reduces pressure in the cylinder at the cost of some work and the atmosphere pushes air into the engine. The compression stroke compresses the air and requires work to do that. Then the expansion stroke returns most of that energy but not all of it because the process isn't purely adiabatic, more energy loss. Then the exhaust stroke pushes the air out of the cylinder which takes more work.

My semi-educated guess is that if the all the valves stayed closed all the time the engine would turn the easiest. If they all stayed open all the time it would turn the next easiest. Anything in-between would take more power to turn the engine. A good thermodynamic analysis or some motoring dyno testing would get you a more definitive answer.

An opposite example of what you are looking for is the Jake brake on diesel engines. It changes the exhaust valve timing so that the air is compressed in the cylinder, then the exhaust valve is lifted at the start of the expansion stroke so the energy in the compressed air is wasted by blowing it past a cracked open exhaust valve at a high pressure drop. That results in a lot of extra engine braking

 

Following up to Tom's comment, Honda Insight-II uses valve closed method.
Integrated Motor Assist (IMA) - Wikicars
"Low-speed cruising - The valves of all four of the engine’s cylinders can be closed and combustion halted, the electric motor alone can power the vehicle."

Ken@Japan

 

Personally I would never use "N" for coasting, because (I believe) it goes against the function of regenerating the hybrid battery.

Other than that.. I would use "B" for coasting down a steep hill.

 

When forced to regen, you are certainly correct. However, regen involves major conversion losses, so maximum mpg is achieved with no regen. Unfortunately, we have to regen at times, so we must avoid N during those times. Luckily, we can get pretty close to an N glide just by feathering the throttle enough to hide the bar on the HSI screen.

Gliding beats regen coasting.
Regen coasting beats friction brakes.
Friction braking beats collision.
Collision stinks.

 

I agree.

 

I gathered some data this afternoon coasting downhill in "N":

	Column 1	Column 2	Column 3	Column 4
0	Time	MPH	RPM	MAF (gm/s)
1	02:18:02.381	69	1130	2.28
2	02:18:03.865	70	1152	2.28
3	02:18:05.334	70	1152	2.29
4	02:18:06.802	70	1152	2.48
5	02:18:08.271	71	1152	2.20
6	02:18:09.771	72	1152	2.50
7	02:18:11.240	73	1184	2.64
8	02:18:12.709	73	1179	2.29
9	02:18:14.177	74	1191	2.17
10	02:18:15.646	75	1184	2.29
11	02:18:17.115	75	1184	2.59
12	02:18:18.599	76	1184	2.68
13	02:18:20.068	77	1184	2.62
14	02:18:21.552	78	1184	2.14
15	average	73	1169	2.39

Captured using Auto Enginuity 8.3.1, generic powertrain from ZVW30.

So doing the math:

	Column 1	Column 2	Column 3
0	2.39	gm/sec air
1	14.7	stoich ratio
2	0.1625	gm/sec fuel
3	3	600	sec/hr
4	585	gm/hr fuel
5	2	659	gm/gal gasoline
6	0.22	gal/hr
7	73	miles/hr
8	331.7	miles/gal

Calculated MPG from MAF.

Bob Wilson

 

Both "D" and "feathering" are modes that Auto Enginuity does not provide enough data. The current version doesn't have injector timing so we can't tell if they are turned off. As for "D", well the Prius is converting some kinetic energy into battery charge but the real problem is I don't have injector timing. I can't tell if the MAF is showing a flow of air that is not getting any fuel.

I'm comfortable with the "N" data because it turns off MG1 and MG2. This means the engine controller has to continue powering the ICE to keep it spinning. So MAF flow can be used to measure fuel consumption.

Upon reflection, the fuel burned per hour is probably the minimum needed to turn over the ICE. Call this the internal engine overhead fuel burn. If we use a 12 gallon tank, at 0.22 gal/hr, the ICE could run constantly for a little over 54 hours. Call this the maximum ICE run time. By cycling on-and-off, the ICE might run a little longer perhaps, but other environmental effects such as heat loss begin to predominate along with vehicle overhead.

Bob Wilson

 

This makes sense. Otherwise MG1 has to provide the power to spin the ICE, which would be less efficient than powering the ICE directly.

In case this seems counter intuitive, to spin the ICE via MG1 when moving, the field in MG1 would provide a drag force, which in turn would make the ICE spin. The power consumed would be the drag force times the speed of MG1.

Tom

 

Fuel consumption is not the same as risk. In this test, I knew the ICE was running so there was no risk.

I can only discuss how it relates to my local driving situation. I use "N" when in the neighborhood, average speed 25 mph, maximum ~30 mph, during warm-up. There is a distinct reduction in fuel burn rate in "N" compared to "D." Also, sometimes I'll use "N" to approach a distant light on a flat section of road. But otherwise, I leave "N" alone and let cruise control handle speed management.

High-speed use of "N" is OK if you have instrumentation but otherwise, I would avoid it. Of course, it is your car.

Bob Wilson

 

Japanese folks also found the instantaneous fuel consumption is less in "N" than "D" during warming up.
However, we also found it takes longer to warm up the ICE and catalyst, then the total fuel consumption was the same between "N" and "D".

Ken@Japan

 

Bob, I'm assuming 1169 rpm is considerably above normal idle speed. Do you have any idea why that is?

A high idle speed while the vehicle is moving is neutral or clutch disengaged is common in a regular car but that is to keep the power brakes and steering fully operational. Doesn't seem like a Prius needs that.

 

The engine has to turn that fast to keep up with MG2 and MG1. The gearing is always engaged and always at the same ratio. MG1 could be spun faster so that the ICE could turn slower, but the whole point of spinning the ICE is to keep MG1's speed at a reasonable level.

Tom

 

Actually I was trying to capture from about 20 to 70 mph on another hill but Auto Enginuity has a maddening ability to overwrite an important data file you just worked so hard to create.

I did notice there was a slight increase in the ICE rpm as the speed increased but this was a fairly brief test. Having some usable data is better than no data at all. <GRINS>

Bob Wilson

 

I agree somewhat with this:

I found using "N" when parked in the driveway can save a little fuel with the NHW11 but I don't see it helping with the ZVW30 . . . not without instrumentation.

When feeling 'frugal,' I'll leave the driveway using EV and turn towards a gentle downgrade about 40 m. from the house. Then I'll disengage EV to get the ICE running and coast in "N" to the bottom. I've not done properly instrumented runs but it looks like I start the morning drive with a significantly improved MPG, ~40 MPG vs ~30 MPG.

Bob Wilson

 

I would like to see the complete data analysis.
I believe we see better MPG number during partial warming up period, then the final MPG number becomes identical after complete warming up.

Ken@Japan

 

I recorded this data with Auto Enginuity, a somewhat buggy but usable, OBD scanner software and tool:

• Version 8.3.1 - the latest as of a week ago
• "generic powertrain" - all I can get for the ZVW30
• CAN 'fast' - I suspect the data drops may be the software not waiting for the ECU data response. You'll see some "0" values in the data
• Added Google location notes and shifter positions - next time, I will try to add the shifter position to more accurately track it. The notes should be fairly close
• I have Nuvi trip data and could correlate to this file but that would take more time (aka., delay.) If the Nuvi ephemeris is needed, let me know and I'll try to load it later.

I did a preliminary analysis last night and saw deceleration showed higher rpm during the slow-down in "N" compared to acceleration down a grade showing slower relative rpm. I did not see a strict, vehicle speed to ICE rpm ratio except on the hill descent.

Bob Wilson