BACKGROUND
Traditional engine thermodynamics report that 1/3 of the fuel combustion energy is delivered as mechanical, 1/3 the energy goes out the exhaust, and 1/3 in engine cooling. In the past, there have been several efforts to recover some of the exhaust energy such as a German effort to use a closed steam engine and several attempts at thermocouple generators. None has resulted in a practical automotive product but it is so tempting.
What we are describing is called a 'topping cycle' which has been used in power plants and large cargo ships for decades but not yet found in cars.
THERMAL ANALYSIS
The first step is to understand the energy source:
Part of the fuel chemical energy is released in the catalytic converter where the last of the hydrocarbons are oxidized with the nitrogen oxides. But the bulk of the energy comes from the hot exhaust gas that glows the exhaust manifold a dull red. This is the hot-side of the engine.
What we propose to do is inject water just behind the catalytic converter to increase the pressure and kinetic energy of the exhaust stream to turn a turbine-alternator. An open-loop, steam engine, the exhaust heat is converted to kinetic energy captured by the turbo-alternator where the exhaust temperature is over 100C to avoid condensation in the turbine.
The maximum Carnot efficiency is a function of the hot and cold temperatures and in this case could range from 33-57%. In real life, we won't achieve this efficiency but it represents the maximum. But this is a fraction of the assumed, 1/3 heat available in the exhaust.
We know the Prius power required as a function of speed so taking 1/3d as the starting point we can estimate the best case, power that might be achieved which ranges from a low of 2.2kW at 30 mph to up to 16.9 kW at 65 mph. Since our plan is to dump the extra power into the Prius traction battery and its power output is limited to 20 kW, 20 kW is the maximum practical power output of the turbo alternator.
Back-pressure loss vs turbo-alternator generation
There is a risk that the increased back-pressure losses would subtract as much power from the engine as is generated by the turbo-alternator. However, the turbine spins so fast that given equal pressure, the distance covered more than makes up for any loss of piston power on the exhaust stroke:
Or at least that is my understanding. Fortunately, we have further information from the turbine pressure-mass charts.
Turbine performance is mapped using the mass-flow versus the pressure ratio:
Sad to say, the total exhaust mass-flow including injected water is unlikely to reach mass-flow rates for normal turbine operation. Worse, my first turbine, a T04E-50, is sadly oversized for this experiment. Still, it may provide enough energy to test other aspects of the proposed design.
The T3-40 comes closer but we may have to look at much smaller turbines, perhaps those found in radio controlled, model aircraft. Fortunately, the Rankine cycle may allow less refractory turbine materials to be used.
TURBINE EXPERIMENTS
Several years ago, I bought an Ebay turbo charger to learn about them:
I could not find credible sources about turbo-charger engineering and mechanics. But I could buy one, take it apart and gain insights.
So one of the first steps is to understand the dimensions:
Yes, it would have been easier if mechanical drawings and technical specifications were available. But lacking them, buy one and 'blueprint' it.
So what I need is a high-speed, 16-20 kW generator, about 3" inches (76 mm) diameter and 1.5" (38 mm) in length. This is a challenge.
ALTERNATOR EXPERIMENTS
My first thought was to go with a high-speed, homopolar generator:
My thinking was the mechanically simple rotor would be easier to design and build than one built with a composite, magnetic rotor and external stators. So I bought a 3" diameter with 1" hole magnet to test to find out if this was practical:
According to the magnet vendor, the surface Gauss is 13,600 so using standard homopolar formula, the model looked like:
Sad to say, the expected voltages were too low and currents much too high for practical electrical circuits. But that is theory, how does the real part work?
So I decided to test the magnet at a slower speed to see what sort of voltages were generated in the nickel plating that is as close to the magnetic material as possible:
Any air-gap will reduce a magnetic field and I needed a metric.
The results were disappointing:
Based upon the generator voltages when spinning in the drill press, the surface, magnetic field is 0.477 T or 4,770 Gauss. Way to weak to generate useful voltages and power. But insights came from this effort.
A magnet has two poles so use both:
Two disks can be wired in series and double the voltage. But there is another problem.
The generated voltage is a function of the Lorenz force, the cross-product of the electron mechanical velocity and magnetic field. The fastest part of the disk is the rim so a better homopolar design is a rapidly rotating cylinder:
But a rapidly rotating cylinder needs high-speed, peripheral bearings and that led to a short-cut.
Ducted Fan, Brushless Motors
The RC community has ducted-fan, electric motors used in model jets. Some of these are rate up to 80,000 rpm and 2 kW. So I found a reasonably priced one and have ordered it for testing:
At first I was looking at three, high frequency, 10:1 transformers to step-up the output that can be boosted to traction battery voltages. But then I realized stepping the voltage up to traction battery voltages should well be within a switching power supply capability.
I may return to the high-speed, homopolar generator after rapid prototyping with this off-the-shelf motor. One fundamental question is whether we can drive the model motor enough to get its maximum, rated power out. There is no need to pursue an advanced, high-speed, homopolar motor if ordinary turbines can not even meet the 1.4 kW motor rating.
STATUS
I have the turbine but need to mount it and rig up an electric oil lubrication system. This is just plumbing and an electric oil pump . . . not a problem.
The turbine compressor shaft has to be trimmed down, a pre-stress nut configured, a shaft coupler, and motor mount fabricated. I have the aluminum stock and access to tools need to complete this build.
I've ordered the Schottky diodes needed to build the first stage, capacitor charger. My junk box has magnetics and I have an MSP-430 that should be able to drive a suitable IBGT/MOSFET for the boost converter.
By building a series of small steps, I should be able to bench test the electronics and build a suitable 'brass board.' Then mounted with the turbine and coupled to the normal exhaust, get metrics to see how practical this approach will be.
If testing shows enough promise, get a properly sized turbine, install the water injection system, and build the control electronics. Then I can proceed to system level testing and learn if there is enough recoverable exhaust heat to augment Prius performance. Right now, I don't know but I hope to find out.
Bob Wilson
Today
Other popular topics in this forum...
