Exaggerated Atkinson Cycle

As we all know, the key determinant in any internal combustion engine's efficiency is its expansion ratio. This is because this ratio denotes how much the mixture of fuel and air is allowed to expand after combustion, and thus how much energy is actually converted to work vs lost to heat.

A common misconception is that the thermal efficiency of an engine is determined by its compression ratio, which makes sense since the mechanical, static compression ratio is equal to the expansion ratio. However, the key thing to note is that while expansion ratios increasing beyond, say 15:1, will continue to yield increases in thermal efficiency, increasing the compression ratio past that point leads to much higher frictional or "pumping" losses.

The diesel engine is known as achieving around 40% thermal efficiency vs. 25% for the otto cycle. Besides perhaps a few percentage points that are accounted for by the fact that diesel engines are not throttled and lose no energy to vacuum, this increase comes purely from the increase in expansion ratio. And what's more, this increase in expansion ratio is not even overly profound (avg diesel compression/expansion ratio is ~16.5:1-18:1 vs up to 12.5:1 for direct injection gasoline). Part of the reason diesel engines run out of steam so early in the RPM range is because the pumping losses from compressing the mixture increase drastically as the revs climb.

The Atkinson Cycle engine works by physically closing the intake valves late to allow a certain volume of air to flow back out. This lowers the "dynamic" compression ratio, and increases thermal efficiency because the expansion stroke is larger than the compression, but at the cost of reduced power density. If the engine is a 3L, and we end up with a maximum of 8/10 original volume left after the Atkinson cycle, you only effectively have a 2.4L.

I propose a highly exaggerated Atkinson Cycle; one that would have an abysmal effective volume and not so great power density, but extremely high thermal efficiency (~70% or greater). What's more, there's really nothing proprietary here; this could be done easily on existing internal combustion engines. It just sounds like a really bad idea until explained.

In my highly exaggerated Atkinson Cycle Engine, the static compression ratio (thus expansion) is 75:1. However, the Atkinson Cycle is implemented in such a way that the engine can only ever hold 2/10 of its original volume; thus a dynamic compression ratio of 15:1 maximum at full load. To throttle the engine, there would be cam phasing such that up to say 95% of air sucked in is expelled back out during compression. All of this means that even at maximum load, the engine is working extremely little at compressing the charge compared to what it harnesses as work.

On the surface, the engine would appear to have characteristics of extremely poor power density. 2/10 original volume means a 2L engine will now only hold 400ccs. However, because of the stratospheric expansion ratio, I predict thermal efficiency in the ~70% range. So while the power density on the surface is only 2/10 original, because the efficiency increases 3-fold compared to an Otto Cycle, the power would also. This would mean around 60% power density vs. an average engine of the same size (which is already character, and most importantly 3x the thermal efficiency.