There are only two scenarios regarding heatsoaked startups.
1/ Fuel has evaporated and isn’t there for the injectors - you can’t start. It isn’t going to run until you have cleared the blockage.
2/ The computer fuels the engine at startup incorrectly because the sensor is returning a high (heatsoaked) temperature when in reality, the “correct” temperature is lower. More fuel is needed than the computer is giving it.
It then runs rough at startup. This isn’t a blockage. This is now definitely a sensor problem. The air and/or coolant sensors are returning too high a temperature and the fueling is being leaned out as a result. There is nothing you can do like feathering the throttle, because that is just leaning it out more. What you need to do is fool the computer into thinking it is colder than the sensors are reporting it to be. You can do this by pulling the coolant sensor out of its socket. This mimics a stone cold freezing temperature and the computer pumps loads of fuel in. You now feather the throttle to return the air:fuel ratio to a combustible value and the circulating coolant soon returns the sensor to a sensible value that reflects coolant temperature not heatsoaked temperature. You can now plug the coolant sensor back in and the car will behave. This is exactly what is going on in a smaller way when you report that it “runs rough for the first twenty seconds”.
I’ve observed exactly this on my extensively datalogged car when it was stuck heatsoaking for six hours on a summer’s day after a hot shutdown. What you see on the log is that the coolant temperature drops before it starts going up. This means the actual temperature was lower than the sensor was reporting at startup and it only moved to the non-heatsoaked value once coolant had started circulating in earnest. Coincidentally, that’s when the rough running peters out.
Heatsoak post shutdown. Look in the archives for my post on that which was posted in the xjs section. See Solving heatsoak - a worked example and Vapor Lock...a different view?
Don’t overthink this, because you cannot drop the temperature below that of the thermostat opening temperature, because that’s how thermostats work. What you actually want is for the coolant sensor to be accurately reporting the temperature, rather than get hung up about electric fans or coolant circulation. Whether you run the engine for 45 seconds to swap the coolant in the rad with that in the block, or whether you go some fancy electric route doesn’t much matter.
If you want to be scientific, use an ohmmeter to measure the coolant sensor ohms at shutdown and plug in a resistor no lower than that value into the loom (or use higher ohms to reflect cooler coolant) and restart your heatsoaked engine. My guess is that the engine will start, subject to it being a “1/” problem, not a “2/”.
If you genuinely have a boiled fuel rail due to heatsoak, then no amount of wrapping or insulation will help, long term. You will need either persistent airflow, intermittent fuel pump flow or a periodic swap of the water in the rad with the water in the block.
kind regards
Marek