[bwilson4web]

Hi,

Lately I've been working on a new project:

The goal of this project is to see if we can convert the waste heat of Prius exhaust to usable energy. The maximum power out of the NHW11 traction battery is 20kW so there is no reason to design a higher power output. The vehicle overhead, electrical load is just under 500W so if we don't get at least this from the waste heat, it is useless. The rest have to do with technical details of the traction battery interface. For example, we can monitor the current drain and use that as the current output limit of the generator.

The first and most important question is whether or not we are trying to violate any laws of Thermodynamics:

We've made the assumption that an equal amount of engine energy flows out the exhaust. This in turn is 'cooled' by boiling water injected into the exhaust. The increased pressure then leads to higher, gas kinetic energy needed to spin the turbine.

One concern is the effect of increased, exhaust back pressure on the engine. The question is whether or not the increased back pressure during the exhaust stroke will lose as much or more power than is produced:

My expectation is the faster spinning turbine will generate more power than the engine back pressure loses because of the longer displacement. However, this experiment will provide metrics to determine if this is the case. Note that the turbine will not have a stock muffler but rather a short glass-pack.

So the overall architecture is to take an off-the-shelf turbocharger and convert it to a turbo generator:

In addition to the water injection, we also need pressure fed, oil lubrication and some sort of high rpm, generator.

Several years ago, I bought a turbo-charger to find out what we are dealing with:

Here the impeller, compressor has been removed so we could get an idea of what the high-rpm units look like internally.

The traditional approach has been a rotating magnetic rotor with a multi-field stator. However, my current thinking, a homopolar generator that spins a metal disk at turbine speeds in a strong, magnetic field. Past practices used brushes to pickup the electrical power from the rotor and hub:

Everything has been drawn to scale including the metal impregnated, carbon bearing/brush that is used to pull the axle current.

So this is a sketch of the first homopolar generator:

So using this as a baseline, we can start modeling the expected electrical output.

The voltage output is a function of the magnetic field strength, two radius, and rpm:

These are modest voltages and worse, the brushes provided additional drag on the rotating disks. However, the original ring magnet is out of stock so I'm going with one twice as thick with an expected increase in field strength. Possibly doubling the voltage.

Even if doubling the voltage from a stronger magnet, we're still looking at low voltages. Given there are two sides to the magnet, having two, disks connected in series would double the voltage and halve the current needed:

Using roller bearings on the rim with a load bearing, outer race also reduces the effects of imbalances. In the earlier design, the rotors were suspended on a single shaft and any imbalance would rapidly warp and tear the unit apart. However, the solid, outer bearing races limit the displacement.

I was surprised to find that the generated voltage is not dependent upon the relative motion of the conductor and magnet. Since the magnet is coated with nickel, rotating it will generate a voltage proportional to the strength of the magnetic field and rotational speed:

Note that rather simple, thin, aluminum disk models can be stacked on the magnet to investigate and optimize the internal and outside diameters.

Regardless of the voltage and current, there will need to be boost regulators to bring the power up to traction battery level. It is the boost regulators that determine how much power is drawn from the generator:


Based upon technology, we would prefer to see at least 1.6V or higher. The challenge is to test and find out what sort of voltage range we can expect.

Questions? Comments?

Bob Wilson

[GeorgiaHybrid]

Bob,

In a normal turbo setup, the increased manifold pressure that the turbo forces thru the system more than offsets the extra exhaust restriction that the turbo induces. How will this affect the ICE as far as that restriction goes without any benefit back to the induction system? It would seem that you will be generating a nice amount of power but I am concerned how much it will affect the engine efficiency/fuel consumption compared to an unrestricted exhaust.

[bwilson4web]

GeorgiaHybrid said:

In a normal turbo setup, the increased manifold pressure that the turbo forces thru the system more than offsets the extra exhaust restriction that the turbo induces. How will this affect the ICE as far as that restriction goes without any benefit back to the induction system? It would seem that you will be generating a nice amount of power but I am concerned how much it will affect the engine efficiency/fuel consumption compared to an unrestricted exhaust.

Yes, I have been concerned about this and plan to see what I can find to model backpressure effects on ICE power. My speculation is that the kinetic energy released by steam generation more than compensates for the back pressure losses. Worst comes to worst, the tailpipe-turbine becomes a power transfer device from the ICE-to-turbine.

Today, two reference books arrived and I will have to recalculate the expected voltages. Also, I've ordered the aluminum stock and magnet(s). Next I'll order some brush holders and lathe tools needed to make the "brass board". I really don't know how this experiment will work out but that is why we experiment. <GRINS>

One interesting report from the reference is I may be able to make the magnet part of the rotor. This is toallly new to me but apparently two independent groups confirmed this surprising finding . . . the stuff you learn.

Bob Wilson

[jdenenberg]

Bob,

Just remember to keep your balance. Vibration can be a terrible thing.

JeffD

[bwilson4web]

jdenenberg said:

Just remember to keep your balance. Vibration can be a terrible thing.

That remains a concern as balancing a rapidly rotating part it is not enough to handle just the disk but also the shaft. Any out of round condition or imbalance has to be counter balanced. Worse, there are potential resonances that could tear it apart.

I'm open to suggestions.

Thanks,
Bob Wilson

[bwilson4web]

I've updated the design to show the 3" magnet in all sketches; and updated electrical model and; a simplified test configuration. I am still working on the back pressure model and thinking about how to handle balancing.

Bob Wilson

[bwilson4web]

Version 2.02 includes:

• exhaust backpressure model - rough, without quantitative numbers, the expectation is the turbine will have a greater displacement than the piston and generate more power than is lost by the engine exhaust stroke.
• roller bearings for brushes - the use of roller bearings and the outer bearing races will limit imbalance displacements that otherwise would stress the center shaft. This does not eliminate the need to balance the rotors but any unbalanced forces will be constrained by the bearings.
• voltage multiplier, boost circuit - given the likely 1:305 V step up, there are challenges in finding low voltage MOSFETs as well as how to scale and bring the electrical power up to usable voltages. A two-stage, 1:18 boost circuit should handle the requirement. Granted, it won't be terribly efficient, we are dealing with what otherwise would be lost energy.

Bob Wilson

[jdenenberg]

Bob,

Have you considered having the generator putting out AC and using a step up transformer? A multi-pole alternator would put out high frequency AC and the transformer wouldn't need much iron. The overall system might have better efficiency.

JeffD

[bwilson4web]

jdenenberg said:

Have you considered having the generator putting out AC and using a step up transformer? A multi-pole alternator would put out high frequency AC and the transformer wouldn't need much iron. The overall system might have better efficiency.

I had but felt the build would be excessive due to the absence of a ready set of stators. But this morning I was thinking about auto alternators.

One possibility I had rejected was to get an auto alternator and tapping the AC side of the diodes. I understand others have done this and seen fairly high voltages 50-100 VAC unregulated. Typical alternator specifications:

• 100A @14VDC -> 1.4 kW
• 200A @14VDC -> 2.8 kW

Efficiently coupling the high rpm turbine to the alternator would be a challenge but it may be solvable. At least then I would working with auto-grade electronics.

I was also attracted to the homopolar generator because of the potential for high power in a small package. But boosting up the low-voltage remains an interesting challenge. I even sketched out a stack of insulated rotors to put in series. Every time the voltage goes up, the current goes down reducing the slip-ring/roller bearing current loads.

BTW, I also realized the series coupling between the two rotors can also become a coil around the permanent magnet and boost the field. This was inspired by a U. of Texas sketch of a high-power, homopolar generator that used the generated current to also drive the field. Of course their application was to put out a couple of megajoules to spark some fusion experiments. It offers and interesting variation where a relatively small (aka., cheap) magnet serves only as an initiator and as the current increases, it drives a much stronger, magnetic field coil.<grins>

One of the good things about experimentation is it brings a dose of reality to "PowerPoint" engineers. I'm not ready to give up on the homopolar design. The question is whether or not what is essentially arc welder currents can be tapped without melting the parts and that remains an unanswered question.

Bob Wilson

[bwilson4web]

UPDATE: It looks like I can get a GT15 whose turbine parameters more closely match what I need for this project for $150+shipping. After I do my initial tests with the T04E-50, I may elect to go with a GT15 turbine.

I may have found the "hard" technical problem when I did a mass-flow analysis:

The problem is in cruise state, the combustion exhaust gas, expected amount of heat, and the water it could boil are not even in the mass-flow vs pressure ratio of the turbo, T04E-50, that I have. Even looking at smaller turbos, we're closer but still out the mass-flow rate range. When accelerating, the engine exhaust flow typically triples and larger BUT these are transient events and this project is designed to handle normal cruise power levels.

The turbo graphs are in SAE units so I did the calculations using North American units. I need to pickup up my metric data and recalculate to verify I haven't made some mistake. I'll probably calculate the same heat units per unit time and see if they match. But stepping back from the problem, it begins to make sense.

A typical turbo charger is designed for peak-power, to handle maximum acceleration and gas-flow events. In my case, I'm looking for heat recovery, getting some fraction of the exhaust heat back as useful work. This appears to require a substantially smaller turbine, perhaps a model aircraft size, with a pressure-ratio, waste gate configuration.

In layman's terms, the turbine I have is too big by over an order of magnitude for the Prius exhaust. Instead of having enough kinetic energy in the flow to transfer to the turbine blades, the gas will simply 'take the bend' and not give up enough kinetic energy to efficiently extract energy.

One good thing is look at the typical rpms, 50-150 k rpm. This would increase the homopolar generator voltage by a factor of 3-10 from what I'd calculated before. This also substantially reduces the current and moves the voltage boost to a much nicer range. However, it rules out automotive alternators, typically 7,000 rpm, unless some gearing mechanism is added.

I have to recheck my figures and possibly look for an alternative turbine. I may go forward with testing at least to confirm in my own mind that these numbers are right. Regardless, insights have been gained.

Bob Wilson