First, there's a thermodynamic efficiency limit (from Carnot) that means that waste heat, if it's at relatively low temperatures, simply cannot physically be turned into very much electricity. If your hot exhaust is around 100 degrees C, water's boiling point, you'll never get more than about 25% of the heat energy back as electricity, and that would be lucky.

Second, waste heat is waste. Before using waste, every industry should have an interest in reducing waste. If high temperatures are needed for an industrial process, then let's keep the heat in, don't let it out as waste! Insulation or smart use of fluids to retain and recycle the heat will do far better than an ad hoc electricity generation scheme in reclaiming the waste.

There are a few industrial processes that generate their own heat that needs to be removed - but then you're just turning chemical energy into other forms anyway which is what a typical generator does. One give-away in the article that this doesn't make sense: talk about using this for gas-turbine electric generators. Those are already turning chemical energy into electricity, and compared to coal and nuclear power they're already very efficient (because they bypass the thermal conversion step converting chemical directly to mechanical energy). In fact, combined-cycle gas technology is proposed as a significant way to improve the heat output of coal-fired generators, taking advantage of the efficiency of gas turbines. The added electric energy in capturing the heat output with additional steam turbines is unlikely to be very much, and will be capital intensive when we have a lot of other things we need to invest in.

Second, while it seems a bit awkward for me to comment on the post in general, I do want to respond to apsmith for a few clarifying points.  Bear in mind that we have done these projects and they are generating real electricity, and paying real dividends to all involved based on the capital investment - this isn't just a crackpot theory.  However, the concerns you raise are pretty common, and warrant a response.

1. You're absolutely right that you can't do much with low temperature. But what's remarkable is that there's so much high temperature waste heat out there.  We have looked at all the data on EPA stacks (the database mentioned in the report) and found 40,000 MW of potential opportunities based ONLY on those stacks that have exhaust gas temperatures > 500F and total potential power outputs of > 5 MW.  The latter cut is a commercial screen we use to try and avoid including those projects that are going to be too expensive on a $/kW basis to warrant their commercial pursuit.  However, this is admittedly an artifact of our business model.  If we take off the 5 MW screen, we get to 50,000 MW.  Note that our model factors in first & second laws of thermodynamics, first by making sure that we only recover steam at volumes and pressures that can be justified by the stack temperature and volumetric flow rate, and second by calculating steam-to-power conversion efficiencies based on 70% isentropic steam turbine cycles, all assuming an inlet pressure dictated by stack temperature.  (For the non techno wonks, this all translates into an assumption that you make a lot less power per Btu of available heat from 500 degree gas as you do from 1000 degree gas.)  There is a big remaining uncertainty with respect to the accuracy of the EPA database itself, but the spot checks we've done thus far suggest that it's as likely to be high as low, so the average is probably OK.

2. It's a nice idea that factories could recover this, but not practical for two reasons - one technical, one commercial.  On the technical side, much of the exhaust heat coming from facilities is a function of the temperature of their process.  A steel blast furnace fuses coke with iron at very high temperatures - drop that temperature down and you don't get the chemical conversion to make steel.  Similar with the coffee roaster cited, who has an interest in maximizing the throughput of their beans, and lowering the temperature of that oven is going to affect both flavor and time to cook (as anyone who's ever tried to cook a turkey for twice as long at half the temperature can attest).  Since those processes run at high temperatures, the exhaust is also at high temperatures.  And while some of those processes can be recuperated (e.g., take the hot exhaust around and preheat the inlet air to the process), this is (a) not always technically possible and (b) not necessarily economically optimal if one can first make power, since the former is worth more per Btu.  (This is not to suggest that recuperation isn't good, but rather that high-grade energy is more valuable when used to make a high-grade product like electricity, after which point you can still use the residual energy to recuperate.)  And this brings us to the commercial point.  And industrial facility that makes widgets is really good at making widgets, and if it's well managed has employees who are really focused on making better widgets.  However, they are not focused on making energy.  They're not dumb.  Quite the contrary - they're actually really good at what they do.  But they are focused on their core business.  The fact that the waste heat is there doesn't mean that they've failed, but it also doesn't suggest that the fact that it's there means they shouldn't use it.  This is a classic example of the need for outsourcing, no different in concept than the decisions that a million other businesses make to outsource everything from their photocoping services to their benefits plan administration.  The difficulty in the electric side (as many of my prior posts here have pointed out) is that the regulatory environment is strongly tilted to make it difficult to outsource electric plants - a point McKibben made much more eloquently than I can.  And so long as the opportunity exists - and so long as we remain responsible about GHG reduction - this is a massive pile of low hanging fruit.  (As Amory Lovins has sometimes said, "when you're surrounded by low-hanging fruit, shake the damn tree.")

Sorry for the length, but I hope the clarification is helpful.

Finally someone is making lemonade out of lemons. If we had enough people with thermodinamic knowledge out there, this would be a basic part of the design of any heat producing or consuming plant.

I see taking advantage of decompressing gas similar to regenerative braking in cars.

What about water pumping? Have you detected any potencial in water that just pours into cisterns? or is that to small?

That hot water can then provide building heat as well.  Cogeneration makes the extra expense of a closed cycle/lower temp system competitive per kwh cost.

Once this system is in place, solar collectors could be installed on roofs and tubing to collect solar heat over parking lots that would add in a whole new stream of energy.

Heat energy recovery could have multiple sources of heat.  Furnace and boiler exhaust, extra solar heat from PV, solid oxide fuel cell/turbine exhaust, solar concentrators, for instance.

http://amazngdrx.blogharbor.com/blog John Schneider, Northern Wisconsin

An economizer is simply a water jacket with a big hole in the middle for exhausts to flow - think like a residential natural water heater, a tube around a tube.  The stack cannot restrict flow, especially in the case of large diesel motorships.  Depending on residence time and stack temperature, an economizer can make warm water, hot water, or sometimes even steam.  

Ships require water heating so as to warm the fuel lines and centrifugal filter, since much of the ship fuel is so thick to pump.  They also use it for "house water" such as for kitchen and lavatory warm water.  On a cruise ship the demand for hot water are extreme but when at sea the boilers are all turned off and waste heat from the main engines is used for all heated water.

Cool, huh?  I mean hot!  

This is not new in industrial application, either.  We all know about stories such as using excess steam for brewing beer, or Con Ed using steam for residential heating in lower New York.  Hope I didn't wander too far astray from the thread, though ...

Onward through the fog

Fun fact: Cruise ship power plants typically average about 50% overall efficiency.  The US power industry averages 33% overall efficiency - and is responsible for almost 40% of US CO2 emissions in the process.  

In other words, doing things like you mention are well proven on ships (and in those handful of projects that have been deployed in spite of our goofy electric regulatory model) could massively reduce our CO2 emissions and lower the cost of energy.  Which means that it's time to change the regulatory model.  (After all, I'm pretty sure that even the most fervent global warming deniers would prefer to pay less for their energy.)

Sean, the comparisons to power plant efficiency are not practical. The temperature level of power plant exhaust steam is so low that it is only directly useful for applications like district heating and since we insist on locating power plants away from population centers, there is little to gain.

1. I'd reiterate that I wouldn't characterize industry as "laggards" any more than I would characterize an industry that doesn't hire top notch, in-house HR benefits administrators as laggards.  They simply choose to focus on their core business, and the stuff outside their core business doesn't get attention.  What's different in this case is the fact that the regulatory paradigm hasn't allowed third-party providers to exploit the resulting niche.  For example, there are still a dozen states in this country where no one but the regulated utility monopoly is allowed to sell a kilowatt-hour.  That ban (mostly in the states of the southeast) makes it contractually impossible for a company to come in & recover waste heat to make power at the industrial unless they do it as a pure capital lease, but that's commercially useless.  In other words, the problem is the regulation, not the industrials.  The good news is that those rules are changing, but not nearly fast enough.

2. Disagree that the comparison to power plants isn't practical.  This is another point where thermodynamic theory gets confused with environmental and economic reality.  The relevant metric at the end of the day is how much fuel we burn, and how much useful energy we get out.  If I can get to 60% efficiency in a cogen cycle, it is no different economically or environmentally than if I get there through a power-only cycle, since the net impact is an equivalent reduction in fuel use (e.g., whether the fuel is displaced from the inlet to a power plant or the inlet to a steam boiler is immaterial).   And while the location of many power plants does indeed make heat recovery impractical, this is an argument for poor plant siting, not for immutability.  I could build a salt factory in the middle of the desert that would be really expensive to run and ship it's product to market - and the only conclusion I'd take from that is that I was really stupid to have built my factory in the desert.  Ditto on the power plant, except that our regulatory model doesn't allow us to penalize plant builders for putting the plant in the wrong place.  However, there are quite a few plants that are adjacent to industrials which could be readily modified to raise their exhaust pressure and provide useful thermal energy to neighboring plants, along with much greater overall energy efficiency, since we would now recover the full heat of vaporization from those plants.  Our failure to do so is again a regulatory failure, since - under current rules - a regulated utility has no economic incentive to make that change, since they would have to pass along any resulting savings to their customers.

I guess I have to take back some of my earlier comments. I didn't intend to imply industrial processes should run at lower temperatures to be more efficient - obviously they have constraints on that. And I realize now the waste stream may be more than just heat - it would typically include byproducts from the industrial process that they probably don't want to feed back in to raise byproduct concentrations at the start (you'd probably make the coffee beans taste bad if you were recycling the hot air too much). So if a process really necessarily has a hot exhaust stream, generating electricity off it isn't a bad idea.

Still how much of the 40 GW total is power plant exhaust? Because I have a hard time believing there's really much opportunity there - these guys are already making electricity, and fuel costs aren't zero for them - if there was a simple way to be more efficient, why haven't they taken it? Is it just that fuel is still too cheap?

Anyway, 40 GW is still less than 9% of the 460-500 GW average electric power production in the United States, so as I said at the start, it's not going to be a huge contributor. A good thing to do, but it doesn't solve all our problems!

Couple responses:

1. Capex is very project dependent, but as a rough average we usually assume $2000/kW.  The range is pretty broad about that average, but even at $3000/kW, these projects can pencil pretty well by virtue of the fact that (a) they have no marginal fuel expense and (b) they don't require any upstream transmission, since all (or at least most) of the power is used locally.  Both make these vastly more economic than any centrally generated power.

2. We've excluded most of the power plant exhaust.  The 500F test knocks out all the steam cycles, which is about 70% of US generation.  You do end up with a fair amount of power gen left in there for simple cycle gas turbines (e.g., convert them to combined cycles), but much of this is excluded from the total due to a low capacity factor - our gas turbine fleet is off-line for too much of the time to think about getting too much of a annual revenue stream from a steam tail.  That said, one of the interesting pieces that remains is the large number of small gas turbines that sit on natural gas pipelines running compressor stations.  Those run pretty steadily, using a slice of the pipeline gas to generate mechanical power to boost pipeline pressure, and you can add 5 - 7 MW at most of the gas turbine (as opposed to engine)-driven compressor stations throughout the country on that catch.  But again, this isn't a power gen opportunity so much as a gas turbine opportunity.  The specific answer to your question though is that the vast majority of the opportunity is on industrial stacks.

3. You're absolutely right it doesn't solve all our problems - it's just a big pile of low hanging fruit.  Do recognize though that these processes tend to be pretty steady, so as compared to the ~1000 GW of total installed generation, this offsets considerably more than 40/1000s on a MWh basis.  Add in another 135 GW of potential cogen that's out there and then recognize that the 40 GW is almost certainly low due to the ability to boost temperatures above that presently reported, and it's not unrealistic to think about displacing 30% or more of fossil fuel associated with power gen in this country just with economically-attractive energy recycling projects.

Is capacity factor an issue? I don't think industrial facilities run 24/7 and there must be considerable downtime on occasion for re-fitting or putting in new products.

Lifetime must also be an issue - power generators are typically planned for 30-50 year life-spans, but industrial facilities may have much shorter typical lifetimes. I don't really know the numbers there.

But to get financing for these you must be able to argue the numbers with those putting up the money - the article didn't seem to mention that as a difficulty, just getting electric regulatory approval. Is that really the biggest stumbling block?

But the good news is that the returns on these projects are good enough to attract the financial community, provided you can put the deals together - and provided the regulatory environment doesn't put the interests of regulated utility shareholders above the public they serve.  

Re: capacity factor, there is obviously a variation in industrials between 1 shift/day, 5 day/week facilities to 24/7 ones.  Through a happy coincidence though, the ones that produce a lot of thermal energy tend to operate pretty steadily (in no small part because of that thermal energy.  If you're running coke ovens, you can't turn them off without them collapsing on themselves as the walls cool & shrink.  If you're making raw metals, you've got big vats of goopy molten stuff somewhere that is a pain in your A*& if it cools down... so they tend to be damn near 24/7 operations.)

Re: lifetime, you'd be surprised.  You need visibility out 15+ years to finance the projects, as otherwise the facility is just paying debt service rather than saving money.  But to a substantial degree, that's attainable.  Remember that no one really wants to have a steel plant/chemical factory/petroleum refinery/sugar beet refinery/etc. for a neighbor.  Which means that once they're built, they don't move.   To put it another way: one doesn't have to worry about whether or not Gary Indiana is still going to have steel mills 30 years from now.  That said, there is clearly a risk that the facility goes out of business before the contract term expires, but this is a financing risk, not a technical one.  It does, however, go to the crux of the regulatory problem.  If you want to build a really expensive, really dirty, really stupid central coal plant that is miles from the load and requires lots of T&D to connect, the public will bankroll that investment and guarantee your equity returns through the magic of utility ratemaking.  But if you instead want to build a fuel-free, locally-sited power plant that produces no carbon, lowers the cost of energy and eases congestion on the power grid, your equity is anything but guaranteed - and banks raise exactly the question you do as to how long it will be there.  Which creates a truly massive subsidy to the worst technologies.

Which, as you might guess, is why I keep coming back to the need to reform the regulatory paradigm.

AP, you are a skeptic in the true sense of the word.  I have consistently faced your cousins over the last eight years trying to raise money to prototype a newly invented engine that will convert waste heat to power more economically than any current technology.  Sean has done a great job responding to your questions, but let me add a little more information to the pot.

First, some comments on the scope of the opportunity.  Glass ovens generally operate at around 2500F.  Because they are fuel-fired, they must have an exhaust - you can't just keep the heat inside forever.  The exhaust temperature in the flue is between 1500-2000F.  You won't believe (I didn't until I saw it) that glass manufacturers generally let this exhaust go up the stack.  Why?  Two reasons: one, management of low margin businesses rarely allocate scarce capital to energy efficiency projects - they need the capital to keep their technologies current; two, the economics of collecting the heat and re-routing it back to the incoming gas burner air just aren't good enough to warrant the cost - simple economics.

Another point to make is that glass ovens are NEVER shut off - it costs too much to heat them back up again.  OK, every few years they will shut down for maintenance.  Most capital- intensive commodity-type industries - glass, metals, cement, ceramics - are pretty much the same.  This is where most of the waste heat is, so the capacity factor of waste heat power plants can be very high.

My company, ReGen Power Systems, has designed a modified Stirling-cycle engine to convert industrial waste heat to power.  The group may be interested to know that Robert Stirling, the inventor of the Stirling engine, actually patented an economizer as part of his engine.  The real invention behind his engine was the re-use of lower temperature heat rather than exhausting it as steam engines did.  Today's Stirling engine incorporates what is called a regenerator, but it is really just an economizer.  The reason that Stirling engines are often considered attractive is because they retain this heat and, therefore, have a higher Carnot efficieny than other engines.

We will have engines that operate at two temperatures - 250C (490F) for exhaust heat as in the glass case, and 100C to condense low pressure steam.  In the latter engine, the latent heat of condensation is the energy used by the engine, rather than any significant temperature differential.  We see very large opportunities for it in the other commodity businesses that produce so much of the waste heat Sean refers to -paper, oil refining and petrochemicals.

This all sounds so easy, you're thinking, why didn't we already do this?  In our case, our inventor figured out how to operate the engine at these lower temperatures.  No one else has, and until oil gets to 100 bucks, who cares?

Sean, great job in your analysis of the nation's smoke stacks and in promoting waste heat to power.

Dick