Fuel Economy Baseline Testing of Insight 1

[jime57]

I've been doing some testing on the Interstate on my Insight 1 to try to establish a FE baseline for aero modification assessment. From what I'm seeing here in somewhat hilly Richmond, and in this nice summer weather, the baseline FE looks like 99-100 MPG at 55 MPH trip average using best hypermiling techniques.

A few observations:
1. 55 MPH is not the optimum speed for FE, but it is about as slow as one dares test if the question is aero assement. I think...........
2. If there are any hills of any degree, it is virtually impossible to maintain a constant 55 MPH and lean burn, therefore it seems that the best course would be to target a roundtrip average of 54 MPH, allowing a slight deterioration for the turn-around. This method would allow the speed to drift up and down by about 4-5 MPH while maintaining the 54 MPH trip average. I have been monitoring the trip average by use of a Scangauge.
3. Lean burn must be maintained at all or nearly all time if the data is to be relevant. A large mix of lean burn and standard mixture is not going to be stable or repeatable.
4. Any traffic holdup at all will ruin a test run. Just start over if there is any stop or pertubation.
5. The RT test course must be of some significant length, 35-45 miles, to allow slight driving errors or pertubations to be submersed in the larger data set.
6. An "uncomplicated" turn-around location must be selected since the turn-around has a significant effect on the data. I found a cloverleaf without lights which allows me to have a perdictable turn.

I'd love to have comments on these points. I'm moving along on my mod package, which I intend to test as a whole, grill block, underbelly, boat tail.
Like Metro, I am trying to execute the boat tail in cardboard as a first effort.

OBTW, has anyone tried to duplicate the Cd baseline of .25 by coastdown method?

Jim E.

[RobertSmalls]

Jim,

Lean burn is the trickiest part of this testing, and I have an unconventional solution. Rather than trying to stay in lean burn, insist on staying out of it. You may only get 55mpg (), but if you drive at 75mph, FE will be less twitchy. Also, aerodynamic drag is easier to measure at higher speeds.

Another option which I think you'll like better: Instead of measuring fuel economy at constant speed, measure speed at constant fuel economy. Nail your mpg meter to 80mpg (or whatever number), and hold your throttle foot steady even during NOx purge cycles. Every ten mile markers, record segment time and mpg.



I've recorded a few coastdown runs by videotaping my speedometer. Unfortunately, I haven't yet done a bidirectional, high/low speed run, and I haven't gotten any good numbers out of my data. I'll be out in the sticks this weekend, so I will try to make time for testing.

I might actually have access to an aerodynamically stock Insight for a little while, starting in a week or so. In which case, I will attempt to verify the 0.25 figure via coastdown.

[jime57]

[QUOTE=RobertSmalls;178597]Jim,

Lean burn is the trickiest part of this testing, and I have an unconventional solution. Rather than trying to stay in lean burn, insist on staying out of it. You may only get 55mpg (), but if you drive at 75mph, FE will be less twitchy. Also, aerodynamic drag is easier to measure at higher speeds.[/UNQUOTE]

That is an interesting approach that had not occured to me. I'll have to do a bit of testing to see if I like it better. You are certainly correct that 55 MPH is a bit low for aero testing. There are two problems that I see. In my hilly territory, I will probably get some lean burn on the downhills and that may distort the data also. My first reaction is that I should see if I can maintain leanburn at 60 MPH. Also, I'm always in competition over on the "other" site and I don't want to "trash" a tank.

[QUOTE]Another option which I think you'll like better: Instead of measuring fuel economy at constant speed, measure speed at constant fuel economy. Nail your mpg meter to 80mpg (or whatever number), and hold your throttle foot steady even during NOx purge cycles. Every ten mile markers, record segment time and mpg.[/UNQUOTE]

Also an interesting idea. I'll have to give that some testing.

I've recorded a few coastdown runs by videotaping my speedometer. Unfortunately, I haven't yet done a bidirectional, high/low speed run, and I haven't gotten any good numbers out of my data. I'll be out in the sticks this weekend, so I will try to make time for testing.

I might actually have access to an aerodynamically stock Insight for a little while, starting in a week or so. In which case, I will attempt to verify the 0.25 figure via coastdown.

I've heard from others that Cd by the coastdown method is very difficult, so I'll be anxiously awaiting your feedback

[3-Wheeler]

Either my car is less efficient or yours is much better....

I recently took the less traveled route on a 400 mile trip.

My average speed was 55 mph, and this was mostly flat roads, and outside temperature was hot that day at 95°F.

I had a passenger along (my Mom) and she had the window down about half way.

At around 150 miles into the trip, the Insight logged 87 mpg.

This is basically lean burn almost the whole way, except for the small towns we went through, where speeds are varying.

Jim.

[3-Wheeler]

Originally Posted by jimepting

....I've heard from others that Cd by the coastdown method is very difficult, so I'll be anxiously awaiting your feedback...

You are absolutely correct about coast-downs being somewhat of a challenge.

I drive a reasonably flat section of road on the way to work. According to Google Earth, the road is flat within several feet, up until a slight 5 foot rise at the end.

There are now well over 50 coast down tests accumulated from the winter to late spring, using a hand held GPS on this section of road, always going the same direction.

I like math and use it allot at work, and when analyzing the data, it's absolutely amazing to me that the coast down testing reflects every single dip in the road with uncanny realism.

Unfortunately all these dips have to be accounted for in the data stream. Even a foot rise in the road shows up in the dV (delta-Velocity) testing as a slight change in velocity, which another way of saying, an acceleration.

If the road bumps are not accounted for, then the decleration data due to air and Crr are skewed, and one can easily be mislead or come to the wrong conclusion regarding Cd of the car.

After spending many hours analyzing the data, the numbers are still revealing their secrets!!

I have a ways to go before making any concrete conclusions however.

________________________________________________

Regarding baseline runs, I would follow Metro's lead and simply drive at a constant speed, and simply monitor the fuel used over a stretch of road, just like Metro did when he extended his tail.

Driving to a constant MPG does work, but since the air velocity changes, comparing apples to oranges is harder when trying to figure out how much better the aero tweak worked out.

Jim.

[groar]

Originally Posted by jimepting

2. If there are any hills of any degree, it is virtually impossible to maintain a constant 55 MPH and lean burn, therefore it seems that the best course would be to target a roundtrip average of 54 MPH, allowing a slight deterioration for the turn-around. This method would allow the speed to drift up and down by about 4-5 MPH while maintaining the 54 MPH trip average. I have been monitoring the trip average by use of a Scangauge.

In Scenic, which is showing real time fuel consumption, I can see that a very light change on the right foot can change the FE by 15% (10% is easy) and change the speed by only 1-2 mph. This is a technique I do use, noticeably on the highway or whenever I'm at a too high speed to pulse & coast.

If I had a cruise control and a real time fuel consumption in Megane then I would use them to test my mods. Currently I'm using Coast Down Testing but I can see that another car or a light change in the slope can influence the deceleration. I have the feeling that the effect is higher on the FE.

Denis.

PS Anyway, what ever the method used, I wish you good results. I'm often frustrated to not be able to tell by how much each mod is improving my car as I have more and more the feeling that my butt-o-meter tells me what I want to hear. Being able to reply to other people with numbers (other than consumption in daily commutes) is also important.

[jime57]

Originally Posted by 3-Wheeler

After spending many hours analyzing the data, the numbers are still revealing their secrets!!

I have a ways to go before making any concrete conclusions however.

________________________________________________

Regarding baseline runs, I would follow Metro's lead and simply drive at a constant speed, and simply monitor the fuel used over a stretch of road, just like Metro did when he extended his tail.

Driving to a constant MPG does work, but since the air velocity changes, comparing apples to oranges is harder when trying to figure out how much better the aero tweak worked out.

Jim.

I'm wondering is you see any convergence of the data toward the magic .25 Cd that the basic car is supposed to deliver

I think that Metro has an easier problem of it. He doesn't need to deal with lean burn. As you know, lean burn is vastly more efficient than regular burn. It seems to me that one must decide on either very fast testing with no lean burn, or slow enough testing to maintain lean burn. A mix of the two is thoroughly random and unperdictable. That is the basic reason I argued for an average speed for the test trip. I have done a bit of testing of testing just so I would have some confidence in my results. I'll analogize a bit. If one makes length measurements with a tape measure which can stretch a bit, of what value is the data???? I do plan to run some tests at 60 MPH ave. just to see if I can maintain lean burn. I think I can.

[jime57]

Originally Posted by groar

If I had a cruise control and a real time fuel consumption in Megane then I would use them to test my mods. Currently I'm using Coast Down Testing but I can see that another car or a light change in the slope can influence the deceleration. I have the feeling that the effect is higher on the FE.

Denis.

I had pretty good testing results with an Echo using an after market cruise control. Cruise control isn't the best FE way to drive, but it sure makes aero tesing a lot easier. I use a proven, flat course near Memphis, TN and did two way RT tests. Got lots of interesting data.