Relooking at warm-up

I'm starting to look again at Prius warm-up:

• Parked until "closed loop", minimum fuel burn
• Driving 25 mph speed limit, "N" coasting

Parked until "closed loop" means:

1. set parking brake
2. hit POWER
3. when engine starts, shift into "N" - prevents extra engine power used to charge traction battery
4. when engine noise 'shifts' and fuel consumption drops, shift into "D" - engine autostops

I'm finding this protocol has unusually low fuel consumption because the engine does not work to also put charge on the traction battery while in 'open loop' mode. In "P" I'm seeing typically 0.60 gal/hr. But shifting into "N" often drops it to 0.40 gal/hr. When the engine shifts to 'closed loop' (i.e., the O{2} sensors are working) fuel consumption drops to 0.28-0.30 gal/hr. More interesting, shifting into "D", the car tends to stay in EV mode well above 9 mph.

• 80.9 miles @59.0 MPG -> 1.3712 gal
• 80.9 miles @58.6 MPG -> 1.3805 gal
• 0.0093 gal burned (est)

Cold weather will be here soon enough and the warm-up cost can be a significant hit. But less of a hit if we can optimize the 1.8L engine warm-up.

Bob Wilson

 

Well editing the last note did not work well, here are the data in table format:

	Column 1
0	[th]time[th]miles[th]MPG[th]gal
1	[tr][td2]0:00.0[td2]80.9[td2]59.0[td2]1.3712
2	[tr][td2]1:42.6[td2]80.9[td2]58.6[td2]1.3805

0.0093 gal.

Earlier that day with the car parked in "N":

	Column 1
0	[th]ICE C[th]gal/hr
1	[tr][td2]38[td2]0.43
2	[tr][td2]43[td2]0.38
3	[tr][td2]47[td2]0.38
4	[tr][td2]51[td2]0.35
5	[tr][td2]53[td2]0.25

	Column 1
0	[th]time[th]miles[th]MPG[th]gal
1	[tr][td2]5:24[td2]41.3[td2]57.5[td2]0.718
2	[tr][td2]5:26[td2]41.3[td2]56.6[td2]0.730

0.012 gal.

More data coming.

Bob Wilson

 

Actually, I'm keeping the engine when it is most inefficient, running in "open loop", from trying to charge the traction battery. The traction battery will regain charge later when it is in "closed loop", the most efficient mode.

We don't know, yet. This is why I'm trying to get some metrics.

After the engine shifted into "closed loop," I was surprised to find the car would 'creep' out of the parking lot, to the street on just traction battery power even though the speed exceeded 9 mph.

I'm still in 70s F weather so there may be a temperature variable I've not seen, yet. But for now, it seems to be pretty harmless and I like the fact that the engine is at peak efficiency.

Bob Wilson

 

After the engine starts, shift into "N" until the engine noise changes indicating "closed loop" and then shifting into "D" or "P" to stop engine:

	Column 1
0	[th]time[th]miles[th]MPG[th]gal
1	[tr][td2]0:00.0[td2]15.7[td2]83.6[td2]0.1880
2	[tr][td2]1:27.2[td2]15.7[td2]79.8[td2]0.1967

	Column 1
0	[th]time[th]miles[th]MPG[th]gal
1	[tr][td2]0:00.0[td2]26.6[td2]66.7[td2]0.3988
2	[tr][td2]1:29.2[td2]26.6[td2]65.2[td2]0.4080

Using "N" during catalytic converter warm-up:
0.0087 gal., 0.0092 gal., 0.0093 gal., 0.0120 gal.

Bob Wilson

 

I'm interested in the vehicle total energy flow, not specific modes:

In these tests, the car is not moving as the parking brake is set. Shifting into "N" simply means the engine is running in "open loop" mode to warm-up the catalytic converter and the two O{2} sensors.

I will be repeating these tests with the car in "P" to see if the additional load from charging the traction battery leads to a significant change in warm-up fuel consumption. I don't know but testing provides a clue.

Bob Wilson

 

Using "N" during catalytic converter warm-up:
0.0087 gal., 0.0092 gal., 0.0093 gal., 0.0120 gal.

This test kept the car in "P" during the warm-up cycle

	Column 1
0	[th]time[th]miles[th]MPG[th]gal
1	[tr][td2]0:00.0[td2]44.9[td2]62.0[td2]0.7242
2	[tr][td2]0:57.1[td2]44.9[td2]61.3[td2]0.7325

0.0083 gal

I need to run a couple more but it looks like the shorter warm-up interval leads to a lower, total fuel consumption. A couple of more runs and we'll know.

Bob Wilson

 

Using "N" during catalytic converter warm-up:
0.0087 gal., 0.0092 gal., 0.0093 gal., 0.0120 gal.

This test kept the car in "P" during the warm-up cycle:
0.0083 gal

	Column 1
0	[th]time[th]miles[th]MPG[th]gal
1	[tr][td2]0:00.0[td2]24.8[td2]63.9[td2]0.3881
2	[tr][td2]0:53.2[td2]24.8[td2]62.7[td2]0.3955

0.0074 gal

The car had only sat for three hours.

Bob Wilson

 

Using "N" during catalytic converter warm-up:
0.0087 gal., 0.0092 gal., 0.0093 gal., 0.0120 gal.

These tests kept the car in "P" during the warm-up cycle:
0.0074 gal., 0.0083 gal., 0.0108 gal.

	Column 1
0	[th]time[th]miles[th]MPG[th]gal
1	[tr][td2]0:00.0[td2]59.8[td2]62.5[td2]0.9568[tr][td2]1:10.0[td2]59.8[td2]61.8[td2]0.9676

0.0108 gal

The car sat for 12 hours. However, we're seeing significant variation in the fuel burn. This is likely due to an uncontrolled variable, initial SOC. A lower SOC, more load on the engine, more fuel burn. But there also appears to be a correlation with the duration of the engine run even in "P".

Bob Wilson

 

The protocol is:

• park car with trip meter showing less than 100 miles
• cold-soak
• setup cell phone in camera, MPEG mode
• power on car showing the trip meter: miles, MPG
• start cell phone movie record mode
• when engine stops, stop cell movie mode
  • miles / MPG = gallons of fuel (engine start)
  • miles / MPG = gallons of fuel (engine stop)
  • use movie editor to measure the elapsed time

Example, first frame when engine kicks off:


Last frame when engine stops:


I'm using iMovie, a Macintosh based application to edit the cell phone video clip. Once loaded, I clip out everything from before and after the engine runs. This gives me an elapsed time, 1:10.02, which I round to the nearest 10th second.

So this is what the data looks like:


Conclusion

Regardless of "P" or "N", about the same amount of fuel is burned to reach "closed loop" engine operation.

Next Phase

In the past, I've read recommendations to 'just drive' as soon as the engine starts. Now I'm thinking it makes more sense to wait until the engine stops or at least transitions to "Closed Loop" which the Scangauge II can detect.

The next question is the relative cost of reaching 25 mph, neighborhood cruise speed, in either "Open Loop" (pre-catalytic converter) or "Closed Loop." We need to find out what is the relative efficiency of accelerating to 25 mph (neighborhood speed). But we have a clue looking at the fuel consumption of the engine running in "N" while the coolant warms up:

	Column 1
0	[th]ICE C[th]gal/hr
1	[tr][td2]38[td2]0.43
2	[tr][td2]43[td2]0.38
3	[tr][td2]47[td2]0.38
4	[tr][td2]51[td2]0.35
5	[tr][td2]53[td2]0.25

0.43/0.25 = 172%

There are few options for a stock Prius. We can maximize coasting in "N" so the minimum fuel burn occurs when not accelerating. Once the engine coolant reaches 53C, drive normally. The other option is an engine thermostat temperature hack.

Originally developed by a Clive Burk posting in "Prius Technical Stuff," we can spoof a higher engine coolant temperature. On the NHW11, it allows the car to enter hybrid mode with engine-off operation. In the ZVW30, this might fool the engine controller into a more fuel efficient, fuel trim.

Bob Wilson

 

I'm OK with that.

How about adding some data points?

For now, 0.009 gallons is close enough for my purposes. So if I look at my typical summer commute, 10 miles @56 MPG, we're looking at ~0.1786 gallons. So the gas needed to get the catalytic converters running is about 5%. This is my normal 20 minute commute. This leaves one unknown, how much is burned to bring the engine coolant up to 55-65C when the "close loop" mode becomes efficient.

Still, if you can add some better data point on the cold-soak, warm-up to engine cutoff, it will help complete the picture.

Bob Wilson

 

"Open Loop" means the stored fuel-trim table is used for the fuel injector mixture control. It is often rich and poor MPG. Temperature and air density are used to make adjustments but there is no detection of the actual fuel-air mixture to tune the mixture.

"Closed Loop" means the O{2} sensors detect when the exhaust is rich or lean and provides an adjustment . . . to the injector timing to keep the mixture at the optimum ratio.

Bob Wilson

 

Ugh, not quite.

It looks like the additional load in "P" significantly shortens the catalytic converter warm-up. So instead taking ~100 seconds using "N", it takes just under 60 seconds to reach 'closed loop' operating temperature in "P." However, the total fuel burned is proportional to mass flow and fairly close:

• 0.0088 gal in "P" in 60 seconds (~0.60 gal/hr*)
• 0.0091 gal in "N" in 93 seconds (~0.40 gal/hr*)

The slightly higher fuel burn in "N" has at least these hypothesis for what happens in the additional 33 seconds:

1. The additional mechanical drag of the longer engine run burns more total fuel.
2. Higher thermal losses from the lower power and longer engine run requires more fuel.

But we are working at the limits of the small samples, just three each. Sad to say support for hypothesis of traction battery involvement would require metrics which were not part of these observations. However, there is a third, warm-up protocol that could add additional load and possibly reduce the warm-up interval.

One could set the parking brake and shift into "D." The energy flow arrows suggests the additional static load on the transmission should increase the fuel consumption rate and pre-heat the transmission, MG1 and MG2. A warmed transmission will have lower mechanical resistance improving moving fuel efficiency.

These tests have been based primarily upon using stock, Prius instrumentation. Anyone can replicate the protocol and add more data points. But to pursue more other hypothesis requires testing with more advanced instrumentation.

Bob Wilson

* - casual observations of ScanGauge II fuel consumption . . . not video metrics!