Yeah Tas, I probably should have read the paper in more detail before throwing down all that skepticism. After re-reading it, it's a pretty awesome study.
I can concur that the benefits of the moderate charge dilution (EGR) appear to be from lower gas temps = lower heat loss, and allowing for greater injection advance gives better work transfer from the cylinder pressure. The heat loss relation they mention in conjunction here is actually up in the middle of the paper.
I suppose I would have preferred if they kept fueling constant and allowed the IMEP to vary though. IMEP is dependent not just on peak cylinder pressures, or locations of peak pressure, but really the entire shape of the pressure curve through the entire compression and expansion stroke. Varying fueling is going to change the combustion signature, and even if total IMEP is matched, the shape of the curve might be different, thereby affecting heat loss and work conversion throughout the cycle. If you kept fueling constant, then you allow the cylinder pressure curve to vary as it naturally would from the different SOI and EGR parameters, and have a fairer look at effect upon fuel consumption, in my opinion. Yes, they are measuring instantaneous fuel consumption and including that into the fuel conversion efficiency calculations, but if the combustion had varied significantly to meet the same IMEP, than it may not have been optimal or fully representative of how the engine would run on its own. Tricky stuff, for sure. Probably trade-offs either way.
Earlier in the paper they show two figures, effect of charge dilution upon cylinder pressure with FIXED injection timing, and effect of injection timing on cyl Pressure with fixed dilution (10% O2).
IMEP goes up when theres more area under the curve after TDC. Because IMEP is proportional to brake power output (with constant friction and pumping losses), you can see from figure 4 that if one cannot adjust their injection timing, increasing EGR definitely will hurt economy. On the other hand, in the second plot with high EGR, if you can re-advance your timing, this effect is nullified. I think this sums up the whole debate on whether EGR helps economy or not: ONLY if you can advance your injection timing to take advantage of it.