(1) Fossil-fuel-based CHP cannot be a long-term solution on climate or energy because they still burn fossil fuels, and therefore still emit a lot of CO2. Reducing that by 20% or even 50% is not enough; we need to take steps that over the next 30-40 years will bring fossil CO2 emissions close to 0.
(2) Efficiency claims for CHP systems are frequently greatly overstated. Heat is lower-quality energy than electricity, and only at high temperatures does it become close to comparable. Efficiency claims for CHP systems that use high-temperature heat are not so far off, but CHP systems that make use of low-temperature waste heat have much lower thermodynamic efficiencies than usually claimed.
The inflated efficiency claims often lead to assertions that CHP is the "largest" or one of the largest potential solutions. But the number of applications that require high-temperature heat where CHP efficiency really is quite high are limited. And the modest efficiency gains with low-temperature waste heat use, which could be much more widely applied, don't lead to very much improvement in overall energy use. [Update: this statement is not quite correct - improving from an effective 6% to 12%, say, is a big improvement even if there's still room for a lot more.] The combining of heat and power production in CHP systems can reduce our fossil CO2 emissions by a few percent, but much more than that is needed in coming decades.
This is my third and final installment on this topic; the introductory discussions on energy and entropy, and heating "efficiency" provide some important background particularly on these efficiency issues and related claims about energy "reuse".
My complaints may seem pedantic, but the misperceptions the common claims induce have real consequences for policy and planning. Once you get into the details of applications the relatively limited applicability of the technology becomes clearer, but abuse of the terminology remains widespread. For example, in the useful 2001 "EDUCOGEN" document from the European Trade Association for the Promotion of Cogeneration (COGEN Europe) - available here or from the internet archive, there is a brief good discussion of the problem with just adding electric and heating efficiencies together:
The quality of heat is lower than the quality of electricity and it is decreasing with the temperature at which it is available. For example the quality of heat in the form of hot water is lower than the quality of heat in the form of steam. Consequently, one may say that it is not very proper to add electricity and heat ... It is true that sometimes a comparison between systems based on the energy efficiency may be misleading. Even though energy efficiencies are most commonly used up to now, a thermodynamically more accurate evaluation and a more fair comparison between systems can be based on exergetic efficiencies...and the associated mathematics is then confusingly (and not quite correctly) explained. But this caveat is then ignored in the remainder of the document where simple sums of electric and heating efficiencies are used with abandon. The document even prominently displays a sample diagram exhibiting this problem of adding electricity and heat in the first section:
In reality, if the heat used is at low temperatures, the true thermodynamic efficiency of a properly configured "separate production" system (replacing the boiler with a heat pump, for instance) might actually be higher than for the cogeneration system - and the total fuel use for the same output electric power and heat could be lower, a case we'll explore further below.
What is commonly referred to as combined heat and power (CHP) is really two distinct situations, depending on whether the heat use is at high-temperatures (fuel burned primarily for heat) or at low temperatures (fuel burned primarily for electric power). I will also distinguish a third case which is sometimes considered a form of CHP but strictly is not - combined cycle plants that produce only electric power, but using more than one heat engine cycle to recover more of the fuel energy in useful form. The following paragraphs define the notation I will use for these three separate cases.
(I-CHP) Industrial process heat - heat is the primary product, and electricity secondary (referred to as "bottoming" systems in the EDUCOGEN document). This can be applied to an industrial process that releases relatively low quantities of high-temperature waste heat, for example a steel mill, or a food processing plant. The CHP component typically generates electricity for the plant itself or to feed back into the grid. When the heat is not needed, no electric power would be generated.
(R-CHP) Residential or commercial low-temperature heating; electric power is primary ("topping" systems in EDUCOGEN). This is usually an electric generator (for example a large central steam turbine, or smaller gas turbine), which releases high quantities of relatively low-temperature waste heat. The CHP component here then uses that waste heat for residential and commercial heating (or possibly industrial processes that need only those lower temperatures). The heat is available only when the generator is running.
(E-CHP) Primary and secondary products are both electric power. The most common example is a combined-cycle gas turbine generator. Since heat engines are hard to optimize across large temperature differences, having more than one heat-engine cycle allows capture of significantly more of the energy from burning fuel, with a first stage (gas turbine) running at high temperatures and a second stage (steam turbine) at lower temperatures, in a typical configuration. There is no heat output, or true "CHP" component here at all. Waste heat from the final heat engine stage can be used for low-temperature heating purposes, but that is just another kind of R-CHP system (with a more efficient electric generator component).
All three of these situations are certainly improving the efficiency of use of primary energy, but they are doing it in quite different ways. In the case of I-CHP and R-CHP, as long as the installation of the secondary facility doesn't impact the efficiency of industrial heat use or electric generation, respectively, the waste heat use is "free" other than the cost of capital and maintenance, and there is real displacement of fossil fuel use. However, I-CHP is fundamentally limited by the number of industrial facilities that need those high temperatures, and that total is reduced whenever they improve their efficiency of use of that heat (so there is less waste heat to turn into electricity). R-CHP sounds like it could be applied more broadly, but the actual improvement in total energy use is much more modest, or may even be less than other technical alternatives as we will discuss below. The combined potential impact of both of these true CHP approaches is at the level of a few percent of our energy use and CO2 emissions - a piece of the picture, to be sure, but not huge.
In the E-CHP, combined cycle gas turbine case, the efficiency improvement over standard steam turbine systems is readily quantifiable and substantial. However it is not an add-on technology; these new generators are a replacement for traditional fossil-fuel power plants. Doing E-CHP may be a good short-term strategy for reducing CO2 emissions in the electric power sector, but it needs to be recognized that this means capital investment in new production, not just improving the efficiency of existing plants. That capital investment might be better spent going all the way to renewable electric generation capacity, rather than simply more efficient use of fossil fuels. Fuel cells might also be a better solution at least if their technology and cost improves a bit further, though those have similar fuel source issues.
IPCC and McKinsey views
CHP has been discussed in several reviews of potential energy technologies to solve our climate/energy problems. Working group 3 of the Intergovernmental Panel on Climate Change's 4th assessment report (IPCC AR4 WG3) mentions CHP systems at least in chapters 4, 6, and 7. Unfortunately section 4.3.5 (p. 284) and the associated figure 4.21 and table 4.4 abuses "overall efficiency" numbers in the same fashion that the EDUCOGEN document does, but at least table 4.4 also shows the electric-only efficiency for typical technologies, so those who understand the energy-quality problem can get a better idea of the real potential.
The discussion in chapter 6 (section 6.4.6, and table 6.1) is more muted - in particular note the importance of local climate for the usefulness of the (R-CHP) district heating technology being discussed - cogeneration of that form works best where the weather is cold for much of the year so the generated heat can actually be used. In the colder countries (CHP is particularly common in northern Europe - Russia, Finland, Denmark) building heat is needed during most of the year, so power production and heating needs may be reasonably matched. However, in warmer countries there will be large portions of the year where little or no heat supply is needed, and that waste heat is more fully wasted.
Section 7.3.4 of IPCC AR4 WG3 covers I-CHP-style cogeneration and other "heat and power recovery" technologies for industrial facilities. This section makes it clear this is just one of several ways to invest in industrial processes to improve efficiency of energy use. There is real potential for reduction of CO2 emissions here, but at the level of a few percent, not really addressing a huge fraction of our total problem.
A realistic assessment of the magnitude of potential from CHP solutions is given by the following chart of "CO2 reduction technologies" from McKinsey (this version comes from a 2008 Oak Ridge CHP report, about which more below):
In this chart the vertical axis is cost (those below the horizontal line actually *save money*) and the horizontal represents the magnitude of the solution. This chart represents solutions available with present technology to reduce US emissions by 50% in the next few decades; the two highlighted bars represent the contribution of CHP. The potential is real, but there are other solutions that would have a bigger effect and should perhaps take priority.
The Oak Ridge Report
A lot of the recent over-promotion of cogeneration seems to stem from a December 2008 report ostensibly from Oak Ridge National Laboratory: "Combined Heat and Power: Effective Energy Solutions for a Sustainable Future", by Anna Shipley, Anne Hampson, Bruce Hedman, Patti Garland, and Paul Bautista.
Of the authors only Garland is actually at Oak Ridge. The other authors come from commercial entities: Hampson and Hedman affiliated with SENTECH, Inc, Bethesda MD, while Shipley and Bautista are with "Energy and Environmental Analysis", Arlington VA.
This report first overstates the present and future potential for CHP by including all those large central E-CHP plants whether or not they provide heat to any customers. But the following is the principle misleading quote in this report, from p. 6:
While the traditional method of separately producing usable heat and power has a typical combined efficiency of 45 percent, CHP systems can operate at efficiency levels as high as 80 percent. The great majority of US electric generation does not make use of the waste heat. As a result, the average efficiency of utility generation has remained at roughly 34 percent since the 1960s. The energy lost in the United States from wasted heat in the power generation sector is greater than the total energy use of Japan. CHP captures this valuable wasted energy.
As noted in Part 1 of this series, the "valuable wasted energy" in present electric generators is almost entirely wasted on the high-temperature side of the process and is not available for waste heat recovery. If they are proposing E-CHP systems to replace present electric generation, that's a big infrastructure commitment. If they are proposing adding on R-CHP-style heat "recovery" to present power plants, the actual utility of that low-temperature waste heat is quite minimal (see part 2), and the 80% overall efficiency number (and particularly the "total energy use of Japan") is very misleading.
From p. ii in the appendix we find that 53% of "CHP" capacity in the US now is combined-cycle gas turbines (E-CHP, except for a small fraction which provide heat also in R-CHP fashion). If they displace coal plants these really do reduce CO2 emissions in part from their higher efficiency, and in part from their use of natural gas rather than coal as the fossil fuel source. When assessing the potential for CHP systems the report counts both the fuel-switching and efficiency improvement from combined-cycle gas turbines, treating this as purely a benefit of cogeneration. It's not - in fact E-CHP really is not strictly CHP at all. Combined-cycle gas turbines are a good technology for electricity generation. But that goodness is quite a different matter from the benefits of other other forms of CHP, and conflating the different systems gives a poor picture of the actual scope of benefits.
Claims in Gore's "Our Choice"
Al Gore's otherwise useful new book "Our Choice" has one chapter (Ch. 12, "Less is More") that relies extensively on this Oak Ridge report and related sources, and seems particularly confused by the abuse of efficiency numbers and conflation of the different types of CHP. In particular the misleading claim about central station electric generation efficiency noted above and the improved efficiency you can get from combined-cycle plants (attributed once again to cogeneration) is repeated over and over:
Overall U.S. electrial generation converts only 33 percent of fuel to electricity, but combined heat and power (CHP) plants extract more than twice as much useful energy by using energy twice.
the absurdly wasteful way in which we now generate most of our electricity needlessly doubles fuel use, cost, and CO2 emissions. These old electricity-only plants are long beyond their planned lives, propped up by their right to pollute. It is time to replace them with distributed CHP plants.
And highlighted in the middle of p. 257:
Sixty-five percent of all the energy consumed to generate electricity in the United States each year is lost. -- Carnegie Mellon University
According to the National Science Foundation, the typical large power plant has an average efficiency of 30 percent, with additional losses of electricity in transmission. By shifting to CHP, coupled with distributed generation, we could achieve an average efficiency o 80 percent, with virtually no transmission losses.
Incredibly, this distorted use of regulatory power makes it more profitable for most electric utilities to completely waste two thirds of the energy in the coal, gas, and other fuels they burn. Virtually all of this wasted heat could be economically recaptured at the point of generation and used efficiently for additional electricity - thus displacing the need for more coal and more generating capacity.
These claims of "using energy twice" and "recapture" are simply wrong because the waste happens at the front, not the back of the process, and cannot be recovered from the existing waste stream (see part 1). Gore echoes and exaggerates the claims of the Oak Ridge report. Higher efficiency conversion of chemical fuel to electric power is possible, but it requires complete, wholesale replacement of the existing plant, it's not an upgrade of equipment, not simply swapping in more efficient turbines or some waste-hot-water system. With fuel cells you can in principle get close to 100% conversion, triple the typical efficiency of current generation, but a fuel cell power plant is completely different from any combustion-based source. And a combined-cycle gas turbine (E-CHP, much-touted in the Oak Ridge report, which seems to be where Gore gets his "twice as much useful energy" from), which is not strictly CHP in the first place, is similarly quite different technology from the typical coal-fired generator. That means the central question is not so much efficiency but the capital cost and expected lifetime for installation of these new technologies. If the reduction in CO2 emissions can be achieved more cheaply by building more wind and solar production or by increasing end-use efficiency, we should be just closing the old coal plants, not replacing them with new fossil fuel technologies.
Ironically, Gore makes a point along these lines in discussing California's experience (p. 249):
The experience in California and a few other states has proved that the cost of efficiency measures that avoid electricity use is typically far less than the cost of building new generating capacity.
Then he goes on to get things wrong again in the very next paragraph:
In the global economy as a whole, using fuel twice by simultaneously generating electricity and thermal energy offers the largest opportunites for energy efficiency savings.
Aside from the false implications of "using fuel twice", this claim is simply not supported by the numbers - see the McKinsey graph above for instance. There is a small potential for near-term CO2 reduction from CHP, but it is not a major long-term solution. The problem seems to be that Gore is actually believing the invalid overall efficiency numbers here.
Gore's main section on cogeneration/CHP is a mixture of truth and confusion, starting on p. 252:
One of the largest opportunities for efficiency gains - an opportunity that is present in electricity generation and in many industrial sectors - involves the capture of waste heat. Indeed, most industrial facilities using large amounts of heat can profitably capture their wasted thermal energy and reuse it in their own processes - or sell it for use in nearby buildings for space heating and cooling. They can also simultaneously use their waste heat to generate electricity on-site - thereby significantly reducing their purchases of electricity from utilities, and thus sharply reducing CO2 emissions.
Aside from the claim about it being "one of the largest opportunities", this is largely correct, and describes the real (but in totality modest) efficiency improvements that I-CHP and R-CHP systems can provide. But notice that this paragraph is not actually addressing the 33%-efficient power plants that were the subject earlier. Finding a way to use the relatively low-volume high-temperature waste heat from a steel mill or food processing plant is a good idea - though reducing the flow of that waste heat in the first place would be a good efficiency measure too.
And that is typical of the claims about CHP. The focus is on an essentially meaningless efficiency number, rather than on the actual benefit provided by the technology. Yes, CHP can provide some increase in benefit from the same level of consumption of primary energy. But there are other ways to provide increased benefit as well - improved insulation, lower resistance in wires, replacement of the basic underlying technology, etc. The question is cost vs benefit, and whether the investment required matches the long-term path we need to be taking.
We need to be on a path to essentially eliminate our use of fossil fuel over the next 30 to 50 years. Gore has elsewhere called on the US to eliminate fossil fuel use in electric generation by 2018. CHP systems still can make sense with non-fossil fuel sources: biomass or gas from landfills and digesters. But installing new fossil-fuel-dependent CHP systems simply adds to the capital investment that faces obsolescence from their elimination.
Thermodynamic efficiency of cogeneration
The efficiency of E-CHP systems is straightforward to calculate if their output is purely electric power: the nominal ratio of kWh produced to fuel Btu's is, when converted to common units, exactly the thermodynamic efficiency of the system. Efficiencies of 55% or more are not unusual for combined-cycle gas turbine plants of this sort, and that is a real improvement over tradiational coal, nuclear, or other steam turbine systems that are limited to about 40% thermodynamic efficiency, and in practical use are closer to 30%.
The tricky thing is to understand the actual thermodynamic efficiency of systems that output both heat and power (I-CHP and R-CHP). Let's return to the EDUCOGEN example mentioned at the start, here's the figure again:
For an I-CHP-type system the process heat application is typically at high temperatures, say 500 to 1000 degrees C. The corresponding thermodynamic limit on nominal efficiency ranges from about 125% to 150%. So in the "separate production" case with a boiler as presented, the 80% nominal efficiency is actually more like 60% thermodynamic efficiency (60% of maximum nominal), and total thermodynamic efficiency for the "separate production" system is closer to 48% than 58%.
In the cogeneration case, the 55 units of heat production needs to be divided by the maximum nominal efficiency, giving more like 40% thermodynamic contribution from the heat component, or 70% total thermodynamic efficiency, rather than the 85% cited.
In this I-CHP case the cogen system certainly does improve thermodynamic efficiency of fuel use, but the final real efficiency of about 70% is considerably less than the 85% originally cited.
For R-CHP systems, the heat application is at low temperatures, perhaps at most the 50 C of typical home hot water needs. For such low temperature heat the thermodynamic limit to nominal efficiency is about 650%. So the 80% nominally efficient boiler for the "separate production" system is actually operating at only 12% of the thermodynamic limit, and the 55% heat component in the cogen case is more like 8.5%. That means the "separate production" system here is actually running at 24% overall thermodynamic efficiency, while the cogeneration system is at 38.5%.
Note that the 38.5% thermodynamic efficiency of the cogen system in this R-CHP case is not very much of an improvement from the 36% electric-only efficiency of the electric component of the "separate" system. The fundamental efficiency problem with the separate system is the boiler or furnace, at that terribly low 12% thermodynamic efficiency. Replacing the furnace with a heat pump system with COP of 3 (300% nominal efficiency), then the purely electric system (36% efficiency in producing electricity) can also produce whatever quantity of heat is needed at 108% nominal efficiency, or about 17% of the thermodynamic limit. That's roughly twice the heat-production efficiency of the cogen unit (but of course it's producing electricity too).
Producing 55 units of heat with the electric + heat pump approach would require about 18 units of electricity. Input of 133 units to such a "separate system" would be all that was needed to produce 30 units of electricity and 55 units of heat; such a purely electric "separate" system thus has 75% of the efficiency of the cogeneration system in that particular mixed-demand configuration. But it also has considerably more flexibility in distributing output between heat and electric production, with thermodynamic efficiencies ranging from 17% (pure heat) to 36% (pure electricity).
Using a large central combined-cycle or fuel-cell system at 60% efficiency boosts that thermodynamic efficiency range to between 28% (pure heat) and 60% (pure electric). Unless the heat demand is far higher than the electric demand, such a system would require less input fuel than the proposed cogeneration approach. To produce 30 units of electricity and 55 units of heat, only 80 units of input energy to the 60%-efficient electric generator are needed, so such a system would be a 25% improvement over the suggested cogen example.
However, it would be very hard to beat the 70% thermodynamic efficiency of the I-CHP type cogeneration system; the criticism here is almost entirely a problem for low-temperature waste-heat R-CHP systems.
The point being, the touted high (nominal) efficiency of R-CHP cogeneration systems really is pretty good compared to typical existing fossil-fuel electric and furnace heat. But compared to state-of-the-art electricity and heating systems, it does not necessarily fare so well.
There is a slightly related problem with I-CHP systems - for electric production via a heat engine the high temperature T_H is set by the existing high-temperature waste heat stream. The ratio T_L/T_H gives the minimum waste fraction that cannot be converted to electric power, so the higher T_H is the higher the potential efficiency. These systems (as secondary facilities) may also have difficulty obtaining a high flow-rate of cooling water or other coolant, so T_L is likely to be higher than for systems that are primarily electricity producers. In practice the electric component of a I-CHP system is almost inevitably constrained to an efficiency of 30% or less. But that is still essentially free electric power, given that we have to burn fuel anyway for the primary industrial process in question.
In summary, combined heat and power refers to three distinct application conditions which I have labeled I-CHP, R-CHP, and E-CHP. The case for E-CHP (large central gas turbine combined-cycle systems) is very good as far as improved efficiency and reduced CO2 emissions is concerned. However this involves significant capital investment in replacement of existing central power plants, an investment that may not pay off as we eliminate fossil fuel use completely over the next few decades. In any case, provision of heat is not necessarily involved in these systems, they are a pure electric generator technology that can stand on those merits with no need to talk about cogeneration.
The case for I-CHP is probably even better than for E-CHP - these would go in industrial facilities that need high temperature heat for their processing, and electric power can be generated relatively efficiently as almost a free bonus. Other capital improvements to reduce waste heat could also be important, and those opportunities define the competitive position for I-CHP installations.
For R-CHP the case is much more ambiguous. These do reduce energy use compared to existing typical fossil-fuel systems. But they can actually be less efficient, consuming more fossil fuel and emitting more carbon, than state-of-the-art electric generators and heating systems, which also have the advantage of considerably more flexibility in distributing outputs between heat and power. In cold climates where there is a relatively steady need for heat year-round, R-CHP installations still make some sense. Just as with any technology, they may be more cost effective in total than more efficient alternatives. But even for cold climates there may be even cheaper demand-side alternatives. Most building heat just seeps away over a few days through walls, windows, and floors/ceilings. Improving insulation can cut those heat flows significantly, reducing the need for input heat from a CHP system.
Altogether CHP can provide a short or mid-term improvement in fuel use and CO2 emissions for the US and world, but at the level of a few percent improvement. It's not something to be neglected, but it is also not "one of the largest opportunities" out there. Use of the thermodynamically-based efficiency numbers I've suggested here rather than the meaningless "heat + electric" values typically used would be a much better guide to system comparisons. These systems are really not substantially solving the "waste heat" problem so often mentioned in this context, particularly not in the case of the R-CHP installations that the expression almost directly implies.