cogeneration@home

(UPDATED 2005-03-29:  See below.)
(note:  this topic was handled briefly in the appendix of "Where to go from here?")
There is one universal truth about heat engines:  none can convert 100% of its input energy into work (in the physics sense).  Some fraction of the input energy is always going to be converted to low-temperature heat which must be exhausted to the "heat sink", which is typically the environment (see Entropy Blues).  This exhaust heat is often called "waste heat".
One feature of waste heat is that it isn't always necessary to waste it.  If there is a need for heat at or below the exhaust temperature of the engine, the engine's heat exhaust may be usable as a heat source instead of burning fuel directly.  This system of generating power is called cogeneration.
Cogeneration effectively uses fuel energy two or more times:  the first time as high-grade (high-temperature, low entropy) heat as input to the engine, and the waste heat is used a second time as lower-grade (lower temperature, higher entropy) heat for other purposes.  These other purposes include industrial process heat, distillation of water or alcohol, input to absorption chillers, or space heat.
If both heat and electricity are required at the same time, cogeneration can achieve much higher overall efficiences than the separate generation of heat and electricity.  For instance, a cogenerator which takes 100 kWh of fuel to produce 30 kWh of electricity and 70 kWh of heat can eliminate 70 kWh of fuel that would otherwise be burned in a furnace to provide heat.

Very high effective efficiency

If the efficiency of cogeneration is rated in terms of additional fuel burned versus additional energy produced, it can be extremely high.  A modern home furnace may burn natural gas and capture 95% of the energy as space heat.  A gas-fired cogenerator with an electric efficiency of 30% and an overall efficiency of 90% will lose 10% of the total energy of the gas, but for each unit of electricity it uses only 1.33 additional units of gas; its effective efficiency is a whopping 75%.  If the heat losses can be reduced to the same 5% as the furnace, this figure increases to 86%!  This contrasts very favorably with centralized powerplants which may burn gas with perhaps 50% efficiency, or coal at efficiencies as low as 33%.  If fuel is being burned to produce low-temperature heat, cogeneration is a good way of getting electricity with very little extra fuel and fuel cost.
Domestic cogeneration (also called combined heat and power) is one form of distributed electric generation.  Like standard furnaces and water heaters, domestic cogenerators burn fuel and produce low-grade heat; unlike furnaces and water heaters, they turn part of their fuel into electricity.
Some attractive technologies have impediments to their adoption.  Gas turbines are inefficient in very small sizes and fuel cells remain too expensive for common use.  A building which needs ~390,000 BTU/hr of heat and can use or sell 60 kW of electricity can use a Capstone ICHP system, but this is about ten times too big for most homes.  For this and other reasons small cogenerators use internal combustion engines, either Otto (spark ignition) or Diesel cycle.

Broad applicability, no availability

Just about every home which needs winter heat and is not all-electric is a possible site for cogeneration.  Heating fuel is delivered to most homes either by pipeline or by truck.  Home-heating oil is almost identical to number 2 diesel fuel, and engines can be run quite well on either propane or natural gas.  Homes appear to be natural sites for cogeneration.
Yet for all the availability of fuel, small cogenerators are very scarce.  Perhaps this is due to a lack of true markets in electricity, where homeowners with an excess could sell as easily as they buy.  Regardless of the reason, devices which look like they ought to be common household appliances are not for sale.
In the absence of a market in small cogenerators, it is difficult to determine what they would cost in volume production.  Superficially similar systems have per-kW prices all over the map.  Examples:

• A gasoline-powered Generac portable generator rated at 7550 W continuous retails for $1150 at Home Depot, or $152/kW (2005 dollars).
• A German FCEV study compared the cost of fuel cells vs. conventional powertrains.  The study estimated a cost of $2450 ($39/kW) for gasoline up to $6170 ($89/kW) for a diesel parallel hybrid, with CNG and conventional diesel in between.  This study used 1997 dollars; 2005 figures would be about 20% higher.
• A California study found that a gas-fired, 500 kW reciprocating cogenerator would cost $1075/kW installed.  The page is dated 2002; the figure would be about 8% higher in 2005 dollars.

These systems are not directly comparable.  Here is a table of costs and properties: 

System	Size (kW)	Engine cyl.	Engine speed,  RPM	Fuel	Emission controls	2005 $	2005 $/kW	Notes
Generac	7.55	1	3600	Gasoline	None	1150	158	No cooling jacket or exhaust heat recovery. Inadequate quieting. Engine speed too high for long life.
Vehicle	62.8	4?	2000-3000	Gasoline	Catalyst	2940	47	Presumably includes radiator and exhaust.
Vehicle	68.5	4?	2000-3000	Diesel	(not stated)	5100	77	Presumably includes radiator and exhaust.
Vehicle	59.6	4?	2000-3000	CNG	Catalyst	3430	58	Presumably includes radiator and exhaust.
Large cogen	550	(not stated)	(not stated)	NG	Catalyst w. ammonia NOx reduction	639,000	1150	Presumably includes full heat recovery.

The comparisons are interesting.  The 550 kW cogenerator's price is off the chart; clearly it is a low-volume product with few economies of scale, and is not comparable to the other examples.  The one-lung engine in the Generac and vehicle powertrains are high-volume products.  Curiously, the cost of the vehicle powertrains all run roughly 3-5 times the retail price of the Generac, despite their outputs being nearly ten times as great; their cost appears to scale with cylinders rather than power.  The complexity of a 4-cylinder engine is roughly 4 times as much as a 1-cylinder engine.  Complexity and expensive components would appear to account for the extra cost of CNG over gasoline (costly tanks, high-pressure plumbing), and diesel over CNG (high-precision injectors and injector pumps).  It seems fairly safe to posit that cost is roughly proportional to cylinder (parts) count.

What do we need?

What are the properties required of a domestic cogenerator?

• Long life:  Should run 10 years with only normal maintenance.
• Serviceable:  Should be easily refurbished or replaced as maintenance or obsolescence requires.
• Quiet:  Should be suitable for installation in a basement utility room or crawlspace.
• Low pollution:  Should restrict pollution to marginal increments over conventional heating plants.
• Standard form factor:  Should fit in the footprint of a standard heating plant and be otherwise suitable for retrofit applications.

There are some difficulties unique to cogenerators which must be solved:

• Pollution regulations:  Federal law imposes more stringent pollution limits on generators connected to the grid than those which are not.  CNG-fired piston generators require catalytic NOx reduction with ammonia injection to meet these requirements.  Anhydrous ammonia is clearly not an acceptable material for use in domestic heating plants, for reasons of both safety and law enforcement.  NOx controls are primarily required to limit production of photochemical smog on hot days.  It may be reasonable to give waivers to heating plants which only operate in chill or cold weather.
• Grid regulation:  The utility must be able to limit backfeeding so that grid protection measures remain effective.  If the available technology for demand-side management is any indication, this should not present large difficulties.

The product

What's the 2010-model, million-selling home cogenerator likely to look like?

1. It'll come in one or more boxes, totalling about the size of a conventional furnace.
2. It will have a one- or two-cylinder, long-stroke engine operating at a speed well under 2000 RPM, perhaps as low as 1000 RPM.  This engine will operate on something like the Atkinson cycle, for improved efficiency and lower exhaust noise.
3. Maintenance will ordinarily be limited to annual changes of the engine air filter and an oil reservoir/filter cartridge.
4. The engine module will have a heavy, sound-absorbing housing.  This may be a light shell filled with sand or water on-site.  There may also be active noise cancellation for the intake and exhaust.
5. Gas- and propane-fired generators will use catalytic converters for emissions control.  Diesel generators (retrofitted for oil furnaces or where propane delivery is not available) will use catalysts to reduce gaseous emissions and filters to trap particulates.
6. The treated exhaust will feed into a condensing heat exchanger.  The cooled exhaust will be cool enough to exit through PVC pipes, like today's condensing furnaces and high-efficiency water heaters.
7. Engine thermal efficiency will be on the order of 35%.  Alternator and electronic losses will reduce this somewhat.  Good design will recover this energy as space heat, e.g. by using the house air to cool the electronic heat sinks.
8. Units will probably have some stand-alone capability, so that they can provide heat and essential electricity during winter grid outages.  This would require a small starting battery, some additional sophistication in the starting and power conditioning electronics, and a resistor bank (dump load) capable of handling excess power.

This cogenerator will have perhaps two cylinders.  Its cooling system will be roughly comparable to a vehicle's radiator.  The alternator will have fewer parts, less bulk and lower operating loads than an automotive transmission.  It will not need a separate starter.  It will have electronic controls on the order of a vehicle.  The exhaust system will be more sophisticated, including not merely muffling, catalytic aftertreatment and possible filtration but heat exchange and handling of condensate.  The fuel supply is either piped in or assumed to be in place, so there are no additional expenses there.  Its cooling fan will be the house air circulation blower.
It appears that the entire cogenerator system will have roughly the same complexity as half a 4-cylinder car engine, and ought to be able to sell for roughly half the price:  $1500 for natural gas or propane, $2500 for oil-fired diesel.  This purchases a system capable of producing 5 to 10 kW continuous, plus 30,000 to 60,000 BTU/hr of heat (roughly 6000 BTU/hr/kW).

What does it get the owner?

Apparently, one hell of a lot.  The average gas-heated home in the USA uses 50 million BTU of gas per heating season; at current retail prices of $.60/therm, it costs about $300 a year.  If an owner replaces an older, 80%-efficient furnace with a cogenerator with 60% thermal efficiency and 35% electric efficiency, gas consumption rises by 1/3 to 66.7 million BTU and annual fuel cost rises by about $100.  However, the cogenerator also produces 23.4 million BTU (6840 kilowatt-hours) of electricity; at a retail price of 8 cents/kWh, this electricity is worth about $550.
The typical household uses about 1 kilowatt average; over a 180-day heating season, one would expect the house to use only about 4300 kWh of electricity.  The cogenerator would produce considerably more electricity than the household could consume, and it seems unlikely that the utility's other customers (who might have their own cogenerators) could be expected to consume it.  What else could the cogenerator's output help to displace?
Why not gasoline?  If the household has an electric or plug-in hybrid vehicle, the cogenerator could feed it during the heating season.  If the electricity was fed to a vehicle using 250 Wh/mile, the excess 2520 kWh could drive the car for 10,080 miles.  If the car would otherwise have gotten 35 miles per gallon of gasoline, that is a savings of 288 gallons worth nearly $590 at today's pump price.
The final question:  Where do you get the increased 1/3 in natural gas consumption?  Everything comes at a price, and in areas where electricity is generated by burning coal the likelihood is that some of it would come from petroleum, as fuel oil instead of gasoline.  If about 1/4 of the cogenerators switched to fuel oil, the average household would burn 16.7 million BTU of oil and 50 million BTU of gas.  At 19,110 BTU/lb this would be 847 pounds of fuel, or 114 gallons.  This displacement would not occur in areas where electricity is generated with natural gas.

Final accounting

One unresolved issue is how much fuel is displaced at the electric powerplant.  If the house draws power from a coal-fired powerplant with a heat-rate of 10200 BTU/kWh, the saved 4320 kWh would displace 44 million BTU of coal, or about 2 tons.   The savings in CO2 emissions would be 7.3 tons.  A gas-fired powerplant with a heat rate of 6800 BTU/kWh would reduce its emissions by a mere 1.69 tons.  Last, I'm going to assume that heating oil costs, and is taxed, the same as motor fuel so that there is no tax subsidy of electric vehicles via oil-burning cogenerators.  Last, for simplicity's sake, I'm going to assume that the chemical formula of both gasoline and fuel oil is C_nH_2n.

Fuel	Old  consumption	New  consumption	Δ consumption	Cost/unit	Δ cost	CO2 emission  per unit	Old emission	New emission	Δ emission
Electricity, coal-fired	4320 kWh	0	-4320 kWh	$0.08/kWh	-$345.60	3.4 lb/kWh	7.34 tons	0	-7.34 tons
Electricity, gas-fired	4320 kWh	0	-4320 kWh	$0.08/kWh	-$345.60	0.78 lb/kWh	1.69 tons	0	-1.69 tons
Natural gas (coal-fired electric)	500 therms	500 therms	0	$0.60/therm	0	11.52 lb/therm	2.88 tons	2.88 tons	0
Natural gas (gas-fired electric)	500 therms	667 therms	+167 therms	$0.60/therm	+$100	11.52 lb/therm	2.88 tons	3.84 tons	+0.96 tons
Gasoline	288 gallons	0	-288 gallons	$2.00/gallon	-$576.00	19.4 lb/gallon	2.79 tons	0	-2.79 tons
Fuel oil (coal-fired electric only)			+114 gallons	$2.00/gallon	+$228.00	19.4 lb/gallon	0	1.10 tons	+1.10 tons
TOTAL, coal-fired					-$693.60		13.01 tons	3.98 tons	-9.03 tons
TOTAL, gas-fired					-$821.60		7.36 tons	3.84 tons	-3.52 tons

Conclusions:

• The cogenerator can reduce electricity expenditures on the order of $350 per year.
• In combination with an electric or plug-in hybrid automobile, motor fuel costs could be cut by as much as $576/year, offset by $228 in fuel-oil costs.
• It appears that winter carbon emissions of the average household could be reduced by more than 2/3.
• In areas where the marginal electric plant is a gas-fired combined cycle plant with a heat rate of 6800 BTU/kWh, cost is cut by over $800/year and the net carbon emission is cut by almost half.
• Raw rate of return on the cogenerator, assuming the vehicle is already owned, is 33% to 50% per year.

This analysis does not include other improvements which could easily be levered in.  For instance, any building with central air already has most of the hardware required to heat with a heat pump; using a heat pump, external electricity can displace fossil fuel.  Wind power is widely available in the winter across much of the nation and could displace fuel for lighting, heating and transportation at the same time.  The potential savings in gas consumption is immense, and cogenerators could easily compensate for variations in wind intensity on a minute-by-minute basis.

Other issues and savings

The widespread use of cogeneration would improve many other things as well.

• Load on the electric transmission and distribution system would be slashed during the winter, along with its losses.
• Vulnerability of the electric distribution system to disruption by e.g. ice storms would be greatly reduced.
• Manageability of the electric grid would be greatly improved, as utilities could request electricity to be converted to heat in dump loads or fed to the grid with a response time of seconds.
• Households would have a supply of electricity during winter grid outages as long as heating fuel held out.  In combination with a plug-in hybrid car, the house would effectively have a complete uninterruptible power supply.
• In areas where electricity is gas-fired, net gas consumption would be reduced and natural gas prices would fall.  This would reduce total costs in a way not captured by the above accounting.

Some people have been asking if we can afford to adopt measures like cogeneration.  Unless I'm terribly wrong, it looks like we can't afford not to.
(As always, corrections cheerfully accepted.  I'll credit them to the originator.)
UPDATE 2005-03-29:  Tom DC/VA thinks that my estimate for cogenerator cost is low, by a factor of two to four.  I can't see how a mass-produced piston cogenerator in the 5-10 kW range could cost more than a 60 kW automotive powerplant; the engine itself would have about half the parts count, and the cooling system is roughly the same complexity but can be made more rugged (cheaper materials, lower warranty costs) because it does not have vehicular weight constraints.  The only things that might do that are emissions regulations much more stringent than vehicles (already noted), or difficulty in muffling.
Randall Parker writes that wind's lack of schedulability is a problem.  I agree, but I don't think that this is a reason to write it off; its value should be measured by what it can save.  More on this later.  ¶ 3/17/2005 07:54:00 PM