1/ to be an efficient heat exchanger you need to maximise the area of heat exchanging surfaces, so thin and flat beats tubes.
2/ it is an INEVITABLE by product of reducing the temperature of the exhaust gas that you promote the formation of soot, so the more efficient your heat exchanger, the quicker it will soot up.
3/ if you know you are going to pull heat from the exhaust gas, you want the engine to produce as little soot potential as possible, so a clean fuel and maximised compression ratio, turbos are used as much for emissions as power in road vehicles, the turbo makes a good exhaust gas cooler, but the other side of the turbo changes the combustion process enough to make the exhaust gases less prone to sooting, vehicles with blown turbos aren't just down on power, they soot a lot.
======================
6/1 on full chat is 6 bhp, at approx 750 watts to the bhp that's 4500 watts, going by the rule of thirds you're getting 4500 watts or work, 4500 watts of heat to lube oil and coolant, and 4500 watts of heat in exhaust gas, those are MAX numbers, at 2 bhp load the numbers for the other two are different...
the same rate of fuel consumption in a good burner will produce 13500 watts of heat.
you are trying to pull heat out of 4500 watts of heat in exhaust gas, getting even half of it out will cost a lot of time, effort and money.
HEAT has always been piss easy to get, you can make a burner that will burn just about anything, even dirty waste engine oil, which is free.
turning HEAT into WORK has always been the engineering challenge.
put the same time, money and effort into making the HEAT into WORK stage, eg your engine and whatever it is driving, more efficient will pay you bigger dividends.
===========================
the last exhaust gas heat recovery plant I saw had an actual operating efficiency of about 12% when the motor was at full load, and below 2% when it was on light loads, given the rule of thirds this is 4% of the heat energy in the fuel at full load, and less than 1% at light loads.
it's not that hard to boost overall fuel efficiency in an engine by 4%, because you're doing a specialist single application system, not a made for the market road vehicle that might find itself in a desert summer or arctic winter.
Flywheels have an energy conversion efficiency of better than 80%, often nearer 90%, so doubling your flywheel mass will smooth the loads on the motor and improve overall fuel efficiency for a given power output at fixed RPM,
everyone is playing with small high RPM flywheels, because they are trying to make compact stuff with high energy density, often for use in a vehicle, we, as stationary bods, don't have any of these limitations, and low speed flywheels only have rim speed to consider, windage and bearing losses are small.
-------------------
bottom line here is don't waste too much time and money chasing diminishing returns, and ignoring other areas where you really can make savings.