02-11-2014, 04:14 PM | #1 |
First Lieutenant
41
Rep 389
Posts |
F30 Driver Adaptive ECU?
I found this thread over at e90Post
http://www.e90post.com/forums/showthread.php?t=899052 I find a lot of things similar between the e90 and the f30 so I was curious whether or not the f30 shared the same adaptive ECU/DME technology as the e90. The major reason for my curiosity is that on 2 separate occasions when I took my car on long road trips, the DME seems to adapt to the higher speed driving (70-80mph) on open freeways for longer intervals as opposed to the traffic jammed, short, day to day commutes to work in NYC. I noticed when I get back into the city from these trips, the car would seem to pick up harder, drive smoother and have better overall throttle response than it had prior to the trip; even when sitting in heavy traffic. I would assume the car gets conditioned to the stop and go traffic in NYC and once it's allowed to open up a little in less traffic intense highways up north, it learns the new driving style (I rarely fell below 75 mph on my trip in almost 150 miles of straight highway). Anyone have any input on this or is this all a placebo effect? If this is infact the car relearning the new driving pattern, is there anyway to reset the DME without having to make these long trips? I didn't RTFM so sorry if this info is in there somewhere, figured I just jump on the forums and have someone else answer this for me |
02-11-2014, 04:56 PM | #2 | |
Captain
251
Rep 640
Posts |
Quote:
__________________
2018 Audi RS7 4.0T
2006 BMW e46 330Ci Convertible ZHP 2010 Audi S5 4.2L [Sold] 2014 BMW 335i xDrive M-Sport [Sold] 2010 Audi A6 3.0T [Sold] |
|
Appreciate
0
|
02-11-2014, 05:53 PM | #3 |
Lieutenant Colonel
50
Rep 1,730
Posts |
I believe most modern engine controls use model predictive control (MPC) techniques, which fundamentally is a on-line optimization approach. This contrasts to older PID based techniques, which might rely on gain scheduling through "maps", based on octane, temp etc. These maps are generated by calibration engineers who drive the cars *a lot* in varying conditions.
In MPC, the cost function (model based with parametric weights) used to define the open-loop optimization can change based on feedback or other sensor measurements, so it's definitely more "adaptive". The need for this comes from advancements in actuators which allow much greater controllability, but are often multidimensional and cannot be decoupled. There is also a lot more computation available in modern ecu's, and MPC is very computationally expensive. Not to mention the desire to improve emissions, fuel economy and of course performance lol. The people who create these tools are brilliant and entire academic communities are interested in the problems. |
Appreciate
0
|
02-11-2014, 10:01 PM | #5 | ||
First Lieutenant
41
Rep 389
Posts |
Quote:
__________________
2013 EB2 335i M-Sport 6MT // BM3 Stage 2 // VRSF DP // AA CP
|
||
Appreciate
0
|
02-11-2014, 11:51 PM | #6 |
Lieutenant
19
Rep 503
Posts |
Most of the times, systems like this in cars do not "adapt" in the idealistic sense. They most likely are programmed with a set (let's say 3) different "scenarios" of throttle mappings. For example, one for an aggressive driver and one for a hyper miler and one in between.
Then, the car adapts by putting you in one of those 3 categories. Free reign adaptation would drastically increase the cases BMW would have to test (e.g. combinations of shift speeds, engine torque output, etc). |
Appreciate
0
|
Post Reply |
Bookmarks |
|
|