For a moment, let's look for the most benign explanation of VW's actions. Suppose, for example, that DPF regeneration is an occasional operating mode that has high levels of NOx, particulates etc. One relatively benign explanation would be that VW does not want the EPA test results to be influenced by a random, but very significant, DPF regeneration. They could rationalize that the EPA knows about this, and that if the EPA cared, they would design on-road, full cycle certification testing protocols.
Furthermore, common practice in EGR use has been to avoid its use at full power, in order not to compromise performance. This has applied to gasoline powered engines for years.
The engineering teams involved in engine performance may have been assured that company policies required meeting EPA standards and that the EPA had a choice of how to do the testing. They may have been told that aftertreatment measures would handle any excessive NOx generated by low EGR rates under certain conditions.
I have carefully plotted the B1S2 oxygen sensor signal from my 2010 Jetta TDI (measuring group 38, field 4) and from my daughter's Golf TDI on typical 7 mile commuting trip. It shows a value of 0.126 about 5% of the time while the deNOx process is taking place. The signal exceeds this value (i.e. more than 0.126 and less than 0.5) about another 5% of the time. A look at the detailed waveforms shows the NOx storage and reduction cycles with only occasional interruption over a period of seconds, not minutes. Each long string of 0.126 values is preceded by a short rich burst and followed by a drop in O2 signal. This means that during typical cycling the NOx storage catalyst is not completely reduced (i.e. not all the NOx is removed with each cycle).
Another thing I noticed is that the percentile graph for O2 signal for values 0.126 and above follows a very stable and characteristic form, with percentiles from 0 to 99. It starts flat at 0.126 to a percentile of about 56, then rises in a curve convex upward to about 0.49 at a percentile of 99.
I interpret the positive readings as the time when the late post-combustion fueling leads to a rich (reducing) exhaust. If the light-off in the NOx storage catalyst were instantaneous, I think the amount of time and the magnitude of the rich readings would be less. I suspect that VWs models use the light-off time as an indicator of how much NOx is on the catalyst.
Apart from emissions associated with DPF regeneration, VW should not have any trouble programming the parameters for the NOx storage catalyst to achieve the required goals. Maybe the problem is with the DPF regenerations. As is, regenerations are done with EGR off. They could keep EGR on during regeneration and then heat up exhaust with early post-combustion fuel injection. The engine would produce less NOx because peak combustion temperature would be lower.
