Adaptive algorithms

[robert joseph dingli]

Hi Matt,

My own research (PhD) is into very similar systems.  In fact a recent paper I
preseneted at FISITA 94 in Beijing in October was all about self tuning AFR 
control of a NG engine.  (It was a modified Ford EA26MPI engine actually.)
Work by some of my colleagues is also focussing in port NG injection for
taxi use in the local market.  (Ford EECV (eeech!) EFI system).

I was wondering whether the system was put into a vehicle or even a real
world stationary application (as our work was) and, secondly, what sort
of feedback did the adaptive system use.  My own work is similar in concept
to that done at Ohio State and other places using crank speed fluctuation 
to calculate engine torque for performance measurement.

I'd be very intered to get references to your work as it sounds very
interesting.

> I have had experience with a crude speed density system (speed pressure 
> actually, since we don't measure temperature).  We have applied it to two 
> port fuel injected natural gas engines.  A VW 1.7 and a Hercules G1600 
> industrial 163 cu in engine.  I say crude because we were free to be 
> crude.  Our main focus was an adaptive, thermal efficiency maximizing 
> algorithm, based on small perturbations in spark timing and manifold 
> pressure (at constant fuel per cycle).  (Two recent SAE papers and one ASME 
> paper cover it pretty well).  We run it with a simple fuel control and then 
> constantly, synchronously, adaptively fine tune to set pulse width and 
> spark timing for best efficiency.  This system can compensate for changes 
> in fuel composition or system wear or other changes which are normally 
> difficult to measure or correct for.  It is exciting to see how it really 
> works!  Anyway, our crude speed pressure sets pulse width as a linear 
> function of MAP and that is all.  It runs fine for our purposes.
> 
> By the way, I'm graduating soon with a PhD in mechanical engineering.  Does 
> anybody have any openings within the next year or so?  I could use a real 
> salary.
> 

I'll be in the same position in about a year.  Good luck!

> Later,
> Matt

Robert
-- 
----------------------------------------------------------------------
             Robert Dingli           r.dingli at ee.mu.oz.au

Power and Control Systems                 Thermodynamics Research Lab
Electrical Engineering                    Mechanical Engineering
   (+613) 344 7966                           (+613) 344 6728
  University of Melbourne, Parkville, 3052, Victoria, AUSTRALIA
----------------------------------------------------------------------