I did the test the other night. It didn't turn out so great, but I think it still gives you an idea of how much energy there is to capture. The outlook isn't so bright.
For the testing, I used the Paseo. I took exhaust pipe temperatures from just behind the cat. The exhaust manifold on the car comes out the front of the engine, then wraps around and goes toward the back. The cat is probably a good 5 feet from the head. The main problem I encountered was the new exhaust. It is so shiny, that my IR thermometer had problems getting consistent readings. I got as accurate of readings as I could. The coolant temperature was measured with the scangauge.
It took 3.5 minutes for the exhaust pipe temperature to exceed the coolant temperature. At this point the temperature does rise up above coolant temps pretty quickly. However, this is also just sitting and idling. If we were to be actually driving, both coolant and exhaust pipe would heat up much faster. Unfortunately, I have no way of testing this.
On the other end of the stick. I did take one temperature reading of the exhaust manifold right after the test with the engine still running. It was right around 580F, so roughly twice as hot as the pipe behind the cat. This I do think would be a viable area to extract heat from. However, the result would be the cat warming up slower and increased emissions.
Here is the graph I made. Temperature is in degrees fahrenheit. Time is the bottom scale and is in seconds.
