MPG Data Analysis

So my day job is to keep manufacturing machines running as efficiently as possible, so I decided to apply this concept to my truck. I've collected MPG data on my tacoma for the last 5 years and so I decided to see if I could correlate it to any maintenance that I did.

I won't bore anyone with the actual stats, but what I found was surprisingly visible effects of certain operating conditions and maintenance events.

I first broke out the data by location, since one city was very hilly and the other relatively flat. That difference was obvious.

Then, the 3 circles indicate specific maintenance events.

1. (Farthest left circle) shows where I put the lift and bigger tires on.
2. (Middle circle) shows where I changed plugs/wires and replaced all u-joints.
3. (Farthest Right circle) shows the performance after my head gasket replacement. There was no signs of a head gasket issue during this time frame with the exception of losing a little bit of coolant every week. But the steady decline in MPG would indicate coolant leaking into a combustion chamber, causing inefficient combustion. I thought this was odd and sure enough 8 months after my head gasket job, there was some slight coolant in the oil, confirming what the numbers were suggesting.

I realize, there is variation in driving conditions and terrain but for the most part, the truck has been used as a daily driver that drives the same route everyday.

I think the next step is get a scan guage or something that can output more data into file logs automatically for me.

Does anyone else do this or am I alone on this one?

 

I am not doing this for the moment but I wish I could collect data from OBD II plug... Oh by the way nice control card haha. Is the yellow zone means that your MPG is out of control?

 

I'm a big ole data and statistics nerd. Thanks for taking the time to share this.

 

Maybe I'm too rich, but if you bought a Tacoma for MPG you're doing it wrong. I also think the Scan Gauge like oil analysis induces paranoia.

 

I didn't get the feeling that he bought a Tacoma for MPG, but rather he's just tracking the data of it. MPG is an easy stat to track and can definitely help you in the future when issues arise. Nothing wrong with keeping up with the data, can help keep things running efficiently as possible.

 

The colors are the sigma levels from the average. Green = 1sigma, yellow=2&3 sigma and red = out of control. I tried to follow shewart rules, but the data is not that good so I just did I quick visual of it.

I didn't buy it for the MPG either, I'm simply using it as a performance measurement. I would have chose HP or some other metric if i had the data available. I have a baseline MPG established, so I know if I start to deviate from that then I may have an issue. Even though the truck might not show any signs.

 

Reading the post, you found some oil in your coolant 8 months after doing the head gasket?

 

Yeah can you put a point on your graph on the exact time you did your head gasket job? Is the far right circle the MPG declining before you did the job, or after? I have a slowly failing head gasket so I have a little bit of a personal interest in this topic. Thank you!

 

I suppose burying your head in the sand and hoping it wont happen to you is a much better way to go.

 

Correct, I resurfaced the head and did a full valve job on it. There were some hairline fractures in the head, but I decided to go forward. It could be that these cracks have grown or that I didn't do a very good job of putting it back together... either way it's a slow leak and is finally surfacing with evidence of coolant in the oil.

The right circle is my declining MPG after the head job. Where the middle and right circles meet is when I did the head job.

 

Have you tried to get the r-squared value for a linear regression model? Lol

I'm guessing somewhere in the .70 range

 

Well what a bummer to find after all the work of digging into the top of the engine! If you are going to replace the heads would love to see the results afterwards.

 

SWEET!!!

1. Another Industrial Engineer
2. Who does 6Sigma analysis (on his truck, no less)
3. That lives in northern Oregon

We need to meet up for coffee, and I need to find you on LinkedIn!!!


By the way, take a look at the Bluetooth OBDII scanners. They are inexpensive and you can track a crap ton of data through the Torque app.

 

Let me get this right:

1. The lift kit makes sense.
2. The spark plug change improved MPG, which also makes sense (but that's a pretty big shift in the mean, and it tightened your range). Did you go with the same plugs, or try different ones? And how many miles did the old ones get?
3. You got WORSE MPG with a head gasket change??? Was it a bad install? EDIT: just read your post about the hairline fractures.

 

Since it's attribute data, the model is not as definitive as you might think. The variation for it is also unacceptable to really use any of the numbers lol. That's why I just stuck with a quick visual representation.

I doubt I'll replace the head just yet, but it could be coming in the future. I threw some stop leak in the coolant and it's slowed it down considerably. I figured once I start having issues with the stop leak clogging things up, then it'll be time for a new truck or engine rebuild anyways.

Ha what are the odds. I knew I couldn't be the only one. I'll give ya a PM!

I have a cheapo bluetooth OBDII that I use through Torque but cannot get it to read anything older than a 2010. Which one do you use?

 

Could've been bad install or some micro-fractures in the head that I ignored and got worse. I used different NGKs. The old ones were never changed by me, so there was at least 50k on them...

 

Same. I have this one:
https://smile.amazon.com/gp/product/B014MTAYK0/ref=oh_aui_detailpage_o02_s00?ie=UTF8&psc=1

It works, but the Instant MPG (calculated) is off, so I need to do a little more tweaking to it so it reads more accurately.

I'd like to start doing exactly what you're doing soon. I have an 05 that's pretty much all stock, so I want to use it as a baseline for measuring the impact of future mods (lift, tires, etc.). I drive the same route on I-5 everyday to and from work, so repeatability will be fairly easy. Only major variables would be weather, temperature, wind, and driving speeds (especially during rush hours).


By the way, what program are you using to calculate everything?

 

WOW! What a neat post, and job. May I ask how you got into the field?

 

Thanks. Hah, I don't have a real interesting story on how I got into it. I just got an engineering degree and then applied to a job and then I was in lol.

 

Thanks for info. Sounds like you have a good plan, I wasn't really planning on using my data for this, I was just tracking costs. I found it to be too much of a burden to keep track of everything without an automated way collection.

The program is JMP, it's what we use at work.