Thermodynamics and Economic Flow

Volume II — Foundations: The Paradigm and the Science

THE PLATE

A woman walking away along a path through a glasshouse of palms, sunlight breaking through the canopy ahead of her. — Plate II.06The Glasshouse Behind the Turbine Hall.The turbine took the work out of the steam. The glasshouse is what was left, and what was left is warm, and warmth is not nothing.

THE LETTER

There is a sentence that people who have read a little physics like to say at dinner, and it is: the economy is subject to the second law of thermodynamics. It is true. It is also, said that way, useless — because it does not tell you which decision to make on Thursday, and because the people who most enjoy saying it rarely say what follows.

This chapter says what follows.

Here is the shape of it. Energy is conserved; that is the first law, and it is why an energy account never balances against anything interesting. A joule entering a building equals a joule leaving it, and you can run a perfect energy audit of a boiler and learn almost nothing about whether the boiler was a good idea. What is not conserved — what is consumed, destroyed, used up in the plain sense that a household means by used up — is the capacity of that energy to do work. That quantity has a name, exergy, a definition, a unit and an accounting. It is the economically meaningful quantity, and it is the one almost nobody counts.

Nicholas Georgescu-Roegen built an entire economics on this observation in 1971 and got the central thing right, decisively and before anyone else. He then proposed a fourth law of thermodynamics that is not true, and spent the last decade of his career defending it. Both halves of that sentence matter. The first half is why this chapter exists. The second half is why it is written the way it is — precisely, with the arithmetic done, and with the places the argument fails marked on the page rather than left for a hostile reader to find.

The reason to care commercially is narrower than the philosophy and much more useful. When you replace energy with useful work in a growth accounting, the largest unexplained number in twentieth-century economics — the Solow residual — stops being unexplained. That is a claim about a specific set of regressions, it has specific weaknesses, and we will do it properly.

Chapter I.01 did the energy-return arithmetic, the EROI table and the societal floor; that work stands and is not repeated here. That chapter asked how much energy comes back for the energy spent. This one asks the other half: of what came back, how much was ever capable of doing anything at all?

— The Editors

DISCOVERY

What is already working

The exergy account is not a proposal. It has been kept, in places, for fifty years, and where it has been kept it has changed what was built.

The American Physical Society, summer of 1974. A group of physicists spent a season on the technical aspects of energy use and published Efficient Use of Energy the following year. Their contribution was a definition: second-law efficiency, the ratio of the minimum work a task actually requires to the work actually spent on it. The move was to stop asking how much of the fuel's heat reached the room and start asking how much of the fuel's work capacity reached the room. The first question flatters a furnace. The second one does not, and it is the question that produced the heat pump industry.

Gordon Reistad's national accounting. In 1975 Reistad put the United States through that measure for the first time and found an overall second-law efficiency of roughly twenty-one percent against a first-law figure about twice as large. Two national audits of the same country in the same year, differing by a factor of two, and only one of them describing anything an engineer could improve.

Robert Ayres and Benjamin Warr, and a century of American data. Beginning in the 1990s, Ayres and Warr constructed something nobody had: a continuous series of the United States' aggregate efficiency at converting exergy into useful work — shaft power, motive power, process heat delivered at its working temperature, light. The series runs from 1900 to 1998 and it rises from about two and a half percent to about thirteen. That is the number this chapter's arithmetic is built on, and the reason it matters is that nothing else in the economy was compounding at that rate for that long in the background, uncounted.

Iceland, and the idea of a cascade. Roughly nine in ten Icelandic homes are heated by geothermal district heat, and the field at Svartsengi does something more instructive than that: the same geothermal fluid makes electricity, then heats a municipal water system, then warms greenhouses and aquaculture ponds, and what is finally left — too cool to do anything an engineer would respect — is the Blue Lagoon, which is a business. Four uses, in descending order of required grade, from one flow. Nobody had to be persuaded that this was virtuous. It was simply obvious once somebody thought of the resource as having a quality that degrades in steps rather than a quantity that is spent at once.

Denmark. Copenhagen's district heating network reaches nearly all of the city's heat demand and is supplied substantially by plants that make electricity first and sell the tail. This is not an environmental programme; it is a century of municipal finance discovering that a stream of warm water leaving a power station is an asset with a buyer. The Danish Energy Agency has published the national accounting for decades, and the striking feature of it is how ordinary it reads.

The heat pump, at scale. The International Energy Agency's assessment is that heat pumps now meet on the order of a tenth of global building heating demand. A heat pump is not an efficiency improvement to a boiler. It is a different answer to the question, and it is the single clearest case of a second-law insight becoming an industry: it moves heat instead of making it, which is why it can deliver more heat than the energy it consumes without troubling the first law at all.

Cullen and Allwood's map. In 2010 Jonathan Cullen and Julian Allwood traced the global energy flow from fuel through conversion devices to the services people actually buy — warm rooms, moved freight, formed metal — and produced the map that makes the destruction visible as geography rather than as a scalar. You can see, on one page, where the work capacity of the world's fuel is annihilated. Most of it is annihilated in combustion, before anything useful has happened at all.

Seven cases, one pattern: in each, somebody stopped counting joules and started counting the capacity of those joules to do the specific job in front of them — and the decision changed. That is the whole of this chapter's method, and it costs nothing but a column.

THE ARITHMETIC

What works, what does not, and where the line sits

First, the quantity.

Exergy is the maximum work obtainable from a system as it comes into equilibrium with a defined reference environment — the dead state. For a flow of heat Q at temperature T, against a dead state at T₀:

        B  =  Q · (1 − T₀/T)

  B   exergy: the part of Q that could, in principle, become work
  T₀  the dead state — a choice, and every figure below moves with it

The bracket is the Carnot factor, and it is the whole of the argument compressed. Heat at eighteen hundred degrees carries a factor of 0.8634 against a ten-degree environment: nearly all of it is work waiting to happen. The same joule as room warmth at twenty-one degrees, against a freezing outside, carries a factor of 0.0714. Same joule. Fourteen times less capacity to do anything.

Two consequences follow, and they are why exergy is the economically meaningful quantity rather than a physicist's refinement.

Exergy is destroyed and energy is not. The Gouy–Stodola relation says exactly how much: the work lost in any real process equals the dead-state temperature times the entropy generated, W_lost = T₀ · S_gen. An energy audit of a real process always balances. An exergy audit never does, and the gap is the finding.

Exergy has a quality dimension that energy lacks. A megawatt-hour is not a megawatt-hour. Electricity is pure exergy. Hot water at sixty degrees is fifteen percent exergy and eighty-five percent environment.

Second, the worked case that makes it concrete.

A condensing gas boiler heating a room. Ninety percent efficient, which is a good boiler.

  methane, lower heating value              802.3   kJ/mol
  methane, chemical exergy                  831.65  kJ/mol   (ratio 1.037)
  heat delivered at 90 % first-law          722.07  kJ/mol
  Carnot factor, 21 °C room, 0 °C outside     0.0714
  exergy delivered                           51.55  kJ/mol
  ------------------------------------------------------------
  second-law efficiency                       6.20 %
  exergy destroyed                          780.10  kJ/mol   = 93.80 %
  entropy generated, W_lost / T₀              2.856  kJ/(mol·K)

A ninety-percent-efficient appliance that destroys 93.80 percent of what it was given. Both numbers are correct; only one of them is about whether you should have bought it.

The same job with a heat pump at a seasonal coefficient of performance of three and a half. Electricity is pure exergy, so one unit in delivers three and a half units of heat at the room's grade:

  exergy out per unit of electricity   3.5 × 0.0714  =  0.2499
  second-law efficiency                             =  24.99 %
  ratio to the boiler                               =   4.03 ×

The first law cannot tell these two devices apart in any way that matters. The second law says one is four times the other. Every argument about heat policy in the last twenty years is downstream of that line.

Third, the aggregate — and the claim that earns the chapter.

Useful work U is primary exergy E times the aggregate conversion efficiency f. Growth rates add:

        U = E · f        →        g_U = g_E + g_f

Ayres, Ayres and Warr's series gives f, rising from 2.5 percent in 1900 to 13.0 percent in 1998 — a multiple of 5.20 over ninety-eight years, or 1.682 percent a year, compounding, invisibly, for a century. United States primary energy over the same span went from 9.6 to 94.8 quads, a multiple of 9.88, or 2.337 percent a year. So:

  useful work growth   =  1.682 + 2.337  =  4.019 %/yr
  useful work multiple over 98 years     =  51.4 ×
  real GDP at 3.2 %/yr, multiple         =  23.0 ×

Useful work grew at 4.019 percent a year while output grew at 3.2. Now ask what share of output growth a factor can carry — elasticity times the factor's growth, divided by output growth:

  energy, at its cost share of 6.0 %      0.06 × 2.337 / 3.2  =   4.4 %
  useful work, elasticity 0.3             0.30 × 4.019 / 3.2  =  37.7 %
  useful work, elasticity 0.5             0.50 × 4.019 / 3.2  =  62.8 %
  useful work, elasticity 0.7             0.70 × 4.019 / 3.2  =  87.9 %

Robert Solow's 1957 accounting attributed about 87.5 percent of the growth of American output per man-hour to technical change — a residual, which is to say a name for the part that the measured factors did not explain. The elasticity of useful work that would absorb that residual exactly is 0.697. Ayres and Warr's estimates land near there.

That is the result: the largest unexplained quantity in growth economics is, on this accounting, not unexplained. It is the compounding of thermodynamic conversion efficiency — a century of engineers getting better at not destroying the work capacity of fuel — showing up in the output statistics as a mysterious free lunch because the input that was improving was never on the books.

Fourth: where this fails, stated before anybody else states it.

The wedge. An elasticity of 0.5 against an energy cost share of 6.0 percent is a wedge of 8.3 times. Under competitive factor pricing, output elasticity equals cost share; that is not a stylistic assumption, it is the first-order condition. So either factor markets misprice energy by nearly an order of magnitude, or the elasticity is an artefact of the functional form that produced it. The thermoeconomic literature has argued the first — that energy is priced at its extraction cost and not its productivity — and the argument is not absurd. It is also not settled, and a programme whose central number requires one of the two pillars of production theory to be wrong should say so on the first page rather than the fortieth.

Prediction. This is the honest negative that matters most, and it is about the programme rather than any one number. Thermoeconomics is a superb accounting frame and a weak forecasting instrument. It has no out-of-sample record that beats a naive extrapolation. Its aggregate efficiency series are engineering reconstructions assembled from device-level assumptions, published without error bars, and not independently replicated at national scale by a second team using different assumptions. Its fit degrades after the 1970s as services and information output grow, which is precisely the period a forecaster would want it for. Use it to understand what happened. Do not use it to tell you what GDP does next; it does not know, and neither does the person quoting it at you.

The scalar has no preferences. Exergy cannot price anything by itself. Two goods of identical exergy content differ in value by orders of magnitude, and nothing in thermodynamics says which. A theory of value built on exergy alone fails immediately on any object whose worth is informational, aesthetic or positional — which is most of a modern economy's output by value.

The threshold, as a number. An exergy account earns its keep only where the exergy bill is large enough to move a decision. Take a twenty-percent exergy saving as a good outcome:

  energy at 30.0 % of cost  →  total cost moves  6.00 %
  energy at 10.0 % of cost  →  total cost moves  2.00 %
  energy at  2.0 % of cost  →  total cost moves  0.40 %
  energy at  0.5 % of cost  →  total cost moves  0.10 %

Below roughly two percent of cost, the exergy account is bookkeeping rather than management — the saving disappears inside the noise of every other line. That is the boundary of this programme stated as a threshold rather than as a caveat, and it excludes a great deal of the service economy. Know which side of it you are standing on before you build the report.

Fifth: Georgescu-Roegen's fourth law, and why the physicists were right to reject it.

Georgescu-Roegen's 1971 argument was that the economic process is entropic — it takes in low-entropy matter and energy and returns high-entropy waste, irreversibly. That is correct, and it was a real contribution, because the economics of the day modelled production as a circular flow with no thermodynamic direction at all.

He then went further. In the late 1970s he proposed a fourth law: in a closed system, available matter degrades irrevocably into unavailable matter, so complete recycling is impossible even with unlimited energy. Matter matters too, he said, and he meant it as physics.

It is not physics. Bianciardi, Tiezzi and Ulgiati showed in 1993 that complete material recycling is thermodynamically possible given a sufficient energy flux through an open system — and Earth is emphatically open, receiving 173,000 TW against human primary power of 19.0 TW, a ratio of 9,105 times, computed in Chapter I.01 and not re-derived here. Robert Ayres, otherwise the fourth law's most constructive reader, concluded the same in 1999: false as a law, important as an economics. The physicists' objection was not pedantry. A law that is not a law gets a whole programme dismissed by people who would otherwise have listened, and it did.

And here is what the correction actually reveals, which is more interesting than either side of that argument.

The minimum reversible work to separate a species present at mole fraction x from a mixture is R·T₀·ln(1/x). Do it for copper.

  from 0.5 % ore     13.13 kJ/mol  =  206.7 kJ/kg  =  0.0574 kWh/kg
  from seawater      57.93 kJ/mol  =  911.6 kJ/kg  =  0.2532 kWh/kg
  dilution between them                             7.042e+07 ×
  rise in the thermodynamic minimum                      4.41 ×

Dilute a resource by seventy million times and the entropy cost of recovering it rises by four. That is what a logarithm does, and it means the entropy of mixing is very nearly free.

Now count the mass.

  seawater handled per kg of copper     4,000,000,000 kg
  ore handled per kg of copper                    200 kg
  throughput ratio                         20,000,000 ×
  pumping alone, at 0.10 kWh/tonne            400,000 kWh per kg
  against a thermodynamic minimum of           0.2532 kWh per kg
  ratio                                     1,579,724 ×

The barrier to recovering dispersed matter is not entropy. It is mass throughput — and entropy grows as ln(1/x) while throughput grows as 1/x. Georgescu-Roegen was right that dispersal is the binding limit of a material economy and wrong about which mathematics builds the wall. The wall is real; it is hyperbolic, not logarithmic; and because it is a handling cost rather than a law, it is a cost curve, which means it moves when technology moves and can be designed around in advance by not dispersing the material in the first place.

That last clause is the entire industrial programme, and the empirical record confirms it. Aluminium, from ore:

  thermodynamic minimum, alumina to metal      6.34  kWh/kg
  modern smelting cell                        14.00  kWh/kg   → 45.3 %
  whole chain, bauxite to metal               58.6   kWh/kg   → 10.8 %
  remelting clean scrap                        0.70  kWh/kg   →  5.0 % of the cell

Recycling aluminium saves 95.0 percent of the smelting energy — not because aluminium is virtuous but because the atoms in a can were never dispersed. The United Nations Environment Programme's survey of sixty metals found eighteen with end-of-life recycling rates above fifty percent and thirty-four below one percent. The split does not track price and does not track policy. It tracks whether the metal ends its life in a lump or in a dispersion — a design decision taken years earlier by someone who was not thinking about it.

Copper's own second-law efficiency, mine to metal, is 0.63 percent: 0.2067 MJ/kg of minimum against roughly 33.0 MJ/kg spent. That is not a scandal. It is headroom, and headroom only becomes visible once the denominator is agreed.

DREAM

What becomes ordinary

In the economy that has absorbed this, the phrase waste heat has quietly gone out of use, the way long distance did. There is no such category. There is heat at a grade, and every grade has a buyer, and the question asked of any thermal stream leaving any process is not whether it can be disposed of safely but who is next in the cascade.

Industrial sites are laid out by temperature. The high-grade process sits at the head of the flow and the things that need less sit downstream of it in descending order — a kiln, then a dryer, then a district network, then a glasshouse, then a pond. Planners draw the site as a staircase rather than a grid, and the staircase is visible from the road.

Energy contracts are written in two columns. The first is megawatt-hours, as now. The second is megawatt-hours of exergy, computed from the delivered temperature with the Carnot factor and the site's own dead state, and the price sits against the second column. Nobody argues about whether a stream at sixty degrees is worth the same as one at ninety, because the tariff answers it arithmetically and the argument has no surface to occur on.

Every organisation of any size keeps an exergy account beside its energy account. It is one column wider and it is produced by the same system, and it is read differently: the energy account says what was bought, the exergy account says what was destroyed, and a management team that has looked at a destruction figure for four quarters running starts making different capital decisions without anyone mandating it. The account does the persuading, because what it reports is unambiguous and nobody has to be virtuous to read it.

Product design carries a dispersal note the way it now carries a bill of materials: for every material specified, where the atoms are at end of life — in a lump, an alloy, a coating, a solution, a plume. Specifying a metal into a dispersion is not forbidden. It is priced, at the cost of the throughput that recovery would require, which is a number the designer can see at the moment they choose.

Buildings are designed around the grade they need rather than the fuel they can reach. A district network runs at the lowest temperature its coldest customer can accept, because every degree of unnecessary flow temperature is exergy thrown away at the plant and paid for at the meter, and the network operator can say exactly how much. Low-temperature design stops being an environmental preference and becomes a procurement specification with a price attached.

Universities teach the second law in the economics faculty and not only in the engineering one, and the two faculties use the same worked examples. A graduate can compute a Carnot factor and a weighted average cost of capital in the same afternoon and is not thought unusual for it.

And the growth statistics are kept with useful work on the face of them. The residual is smaller, honestly labelled, and treated as what it is — the part not yet accounted for — rather than as a measure of ingenuity. Productivity debates become narrower and more tractable, because a large part of what was being argued about turns out to be measurable and was simply never measured.

None of this requires new physics, a new tax, or a change in anybody's values. It requires one more column, a temperature field on a meter that already exists, and somebody willing to write the report.

DESIGN

The structure that gets there

Four mechanisms, in the order they can actually be built.

One — the exergy column. Take the energy data you already have. Add the delivered temperature, which most meters already log and most spreadsheets discard. Multiply by the Carnot factor against a declared site dead state. Publish two numbers monthly: exergy delivered, exergy destroyed. Declare the dead state in writing and never change it silently, because every figure moves with it and a quietly revised dead state is how an exergy report becomes worthless. This is a week of work and it is the whole foundation.

Two — the temperature-matching rule in procurement. One sentence, added to the capital approval template: no proposal may destroy exergy of a grade higher than the task requires where a lower-grade source is available on site. This is the rule that kills the electric resistance heater and the fresh-fired steam raising for a drying duty. It is enforceable by an engineer with a thermometer, which is the test of whether a policy is real.

Three — the cascade. Order the site's thermal demands by required temperature, descending, and match them against available streams in the same order. Where the match is imperfect, a heat pump can lift a stream by a grade for a known cost — that is what a heat pump is for, and the arithmetic above prices it. The cascade is a design artefact, one page, and once drawn it is obvious where the next capital pound goes for several years.

Four — the contract. Cascading only holds if the downstream party can rely on the stream, which means a contract with a term, a floor and a grade specification. This is the Operationalize This movement and it is where most cascades die: the physics works, the engineering works, and nobody would sign a twelve-year offtake for a stream whose owner might reconfigure the plant.

What each mechanism is allowed to cost. The column is a week of analyst time and no capital. The rule is a paragraph in a template and no capital. The cascade ladder is a day with a site engineer and a spreadsheet. Only the fourth mechanism spends money, and by the time it does, the first three have told you precisely how much is worth spending and on which stream. That ordering is the design: three free moves that produce the evidence, then one paid move that the evidence has already justified.

The governance, in one line. The exergy account belongs to finance, not to sustainability. It is a cost-of-goods question and it should be reported by the people who report cost of goods. An exergy figure produced by a sustainability function is a claim; the same figure produced by the controller is a number.

The sequence. Column first, because it is cheap and it tells you where to look. Rule second, because it stops the waste before you have spent anything. Cascade third, because it needs the first two to be credible. Contract last, because by then you know what you are selling and what it is worth.

What to do when there is no counterparty. This is the common case and it is not a dead end. A site with no neighbour can still lift its own low-grade streams with a heat pump and consume them internally; the arithmetic above prices that without any external party at all. Failing that, the tail can be stored — a water tank is the cheapest energy storage ever built — and used to flatten the site's own morning demand. The cascade is a principle about matching grade to task, and a site of any size has more than one task.

DESTINY

How it holds when nobody is pushing

Three things make an exergy discipline self-sustaining, and they are unusually strong ones.

The physics does not change. A financial ratio has to be re-argued each time the business model shifts. The Carnot factor does not. A cascade designed correctly in 1975 is still correct, which is why Danish district heating has outlived every political configuration that built it.

The counterparty keeps it alive. Once a greenhouse, a neighbouring plant or a municipal network depends on your tail stream, the cascade has an external advocate who will notice immediately if it degrades. That is a far more durable guardian than an internal metric.

The saving recurs. Exergy improvements are structural: a heat pump installed delivers its four-times advantage every winter without anyone remembering why it is there.

Now the failure modes, named plainly.

Exergy fetishism. Optimising the scalar instead of the service. It is possible to build a thermodynamically magnificent plant that makes a product nobody wants. Exergy is a constraint and an efficiency measure; it is not an objective function, and a firm that treats it as one will lose to a firm that does not.

The dead-state drift. Every exergy figure depends on a reference environment, and an unremarked change in that reference can move a whole series. If the account is ever used for compensation, expect the dead state to become negotiable. Fix it in writing and audit it annually — that is the single cheapest control in this chapter.

Rebound. An efficiency gain that lowers the cost of a service raises demand for it. The empirical literature puts direct rebound for household heating in the low tens of percent — real, bounded, and not the refutation that either side of that argument likes to claim. Count it. A saving projected without a rebound allowance is a saving that will be disputed the first time somebody checks the meter.

The lock-in of the cascade itself. A well-designed cascade is a set of interdependent plants, and interdependence resists change. The greenhouse that made your cascade economic is also the reason the upstream process cannot be replaced for eleven years. Write break clauses, and price them.

DELIGHT

What it feels like

There is a specific pleasure in the moment the exergy column first appears beside the energy column and somebody who has run the site for twenty years leans over and says that cannot be right. It is right. It has always been right. It was simply never written down, and the gap between what he knew in his hands and what the report had been telling him closes in a single afternoon.

Then there is the walk. Every cascade has one: a path between two buildings where the flow is visible — a plume on one side and something warm and alive on the other. People take visitors on it. Nobody has to explain why it is good, because the glasshouse is full of tomatoes in February and the explanation is the tomatoes.

And there is a quieter pleasure that belongs to this chapter alone. Once you can see the Carnot factor, you cannot stop seeing it. The kettle, the shower, the resistive heater in a hotel bathroom burning pure work capacity to raise a room by four degrees — the world becomes legible in a new dimension, and it is not a depressing legibility. It is the feeling of having been handed a sense. Almost everything is being done at six percent of what it could be, which is another way of saying that almost everything can still be made much better, and that the headroom is measurable, and that it is sitting there in the open.

OPERATIONALIZE THIS

At the level of finance

The instrument is a waste-heat offtake agreement on an exergy-indexed tariff, financed as a self-liquidating facility against the savings stream.

The situation, costed. A process stack rejecting 12.0 MW of heat at three hundred degrees, of which 9.0 MW can be recovered and delivered to a district network at ninety degrees, against a site dead state of ten degrees.

  exergy in the stack       12.0 × 0.5060  =  6.072 MW
  exergy delivered           9.0 × 0.2203  =  1.983 MW
  cascade exergy efficiency                  = 32.7 %
  first-law recovery rate                    = 75.0 %

Note the two numbers. 75.0 percent is what a vendor will quote; 32.7 percent is what you actually captured of the work capacity. The vendor is not lying. The vendor is answering the first-law question, and you are buying against the second-law one.

The economics.

  heat sold, 9.0 MW × 4,500 h            40,500  MWh-th/yr
  avoided-boiler value, 35.00 / 0.92      38.04  EUR/MWh-th
  revenue                              1,540,761 EUR/yr
  parasitic pumping, 1,575 MWh × 95.00   149,625 EUR/yr
  operations and maintenance             120,000 EUR/yr
  ---------------------------------------------------------
  net                                  1,271,136 EUR/yr
  capital cost                         5,600,000 EUR
  simple payback                            4.41 years
  return on facility                       22.70 %   vs WACC 8.00 %
  margin                                   14.70 points

The exergy-indexed tariff — the clause that makes this signable. The classic dispute in every heat offtake is grade: the buyer wants ninety degrees, the seller's stream drifts to sixty, and the contract says only megawatt-hours. At a flat 38.04 EUR/MWh-thermal, a buyer taking ninety-degree heat pays 172.69 EUR per megawatt-hour of exergy and a buyer taking sixty-degree heat pays 253.48 — the cooler stream is overcharged by 46.8 percent for identical money. So index the price to the grade:

        price(T)  =  base × f(T) / f(90 °C)

  at 60 °C:  38.04 × 0.6813  =  25.92 EUR/MWh-th

This is the novel term and it is the one that gets the deal done, because it converts a subjective quarrel about quality into a thermometer reading. Both parties can compute it. Neither can argue with it.

Balance-sheet treatment. The recovery plant is a long-lived asset; capitalise and depreciate it over the offtake term, not over the host process's remaining life, and say so to your auditors early — the useful-life conversation is one they have every year. Where the counterparty owns the network and you own the recovery, the capacity right is typically a right-of-use asset under IFRS 16 and the offtake a lease-like arrangement; get that determination in writing before the term sheet, because it moves EBITDA.

The counterparty. In order of ease: a municipal or investor-owned district heating utility, which has a demand book and a credit rating; a neighbouring industrial site with a drying or washing duty; a horticultural operator, which is the most delightful and the least bankable. Take the utility first, and use the greenhouse as the swing load that absorbs the tail.

Term, floor and break. Set the term just past the payback — five years' payback takes a seven-year term — with a take-or-pay floor at roughly sixty percent of nameplate, a grade specification enforced by the indexed tariff, and a break clause priced at the remaining undepreciated capital. Security is the savings stream and nothing else: if the heat does not flow, there is nothing to pay, and the contract should say that in one sentence.

The number that decides it. One line, on the front page:

     net annual saving from the cascade
   ---------------------------------------  >  WACC
     facility + verification + admin

Here, 22.70 percent against 8.00. The same inequality as Chapter I.01's shared-savings facility, because it is the same instrument wearing a thermometer.

The first ninety days.

Day	Action	Artifact
1–15	Log delivered temperature on every thermal meter	The exergy column
16–30	Rank all site thermal demands by required grade	The cascade ladder
31–45	Fix and publish the dead state; baseline the stack	The signed baseline
46–60	Identify the offtaker; draft the indexed tariff	Term sheet
61–75	Engineering estimate; confirm the parasitic load	Facility memo
76–90	Sign the offtake or the option on it	The executed agreement

APPRECIATIVE QUESTIONS

Twelve, for a room

Discovery — what is already working

Where on this site does one process already use something another process finished with — and who set that up, and how did they think of it?
Which of our machines has quietly outperformed its specification for years, and do we know what it is actually doing well?
When did somebody here last choose a piece of equipment on a reason other than purchase price, and turn out to be right?

Dream — what becomes possible

If every thermal stream leaving this building had a buyer, what would the site look like from the air?
If our management pack carried one number describing what we destroy rather than what we buy, what would we start doing differently within a quarter?
What would we design differently if, at the moment of specifying a material, we could see where its atoms end up?

Design — what we build

What is the highest-grade energy we currently spend on the lowest-grade task, and what would it take to stop this year?
Who would sign a ten-year agreement for our warm tail, and what would they need to believe about us?
What is the one temperature reading we are not logging that would change what we know?

Destiny — how it holds

What would have to be true for this cascade to still be running when everyone in this room has moved on?
Who outside this company would notice within a week if our stream degraded — and is that a risk or the best guarantee we have?
What is the first sign we would see that the exergy account had started being managed rather than measured, and who would spot it first?

WORKS CITED

American Physical Society (Ford, K. W., Rochlin, G. I., Socolow, R. H. et al., eds) (1975). Efficient Use of Energy. AIP Conference Proceedings No. 25, American Institute of Physics.

Ayres, R. U. (1999). "The Second Law, the Fourth Law, Recycling and Limits to Growth." Ecological Economics, 29(3), 473–483.

Ayres, R. U., Ayres, L. W. and Warr, B. (2003). "Exergy, Power and Work in the US Economy, 1900–1998." Energy, 28(3), 219–273.

Ayres, R. U. and Warr, B. (2005). "Accounting for Growth: The Role of Physical Work." Structural Change and Economic Dynamics, 16(2), 181–209.

Ayres, R. U. and Warr, B. (2009). The Economic Growth Engine: How Energy and Work Drive Material Prosperity. Edward Elgar.

Bianciardi, C., Tiezzi, E. and Ulgiati, S. (1993). "Complete Recycling of Matter in the Frameworks of Physics, Biology and Ecological Economics." Ecological Economics, 8(1), 1–5.

Cullen, J. M. and Allwood, J. M. (2010). "The Efficient Use of Energy: Tracing the Global Flow of Energy from Fuel to Service." Energy Policy, 38(1), 75–81.

Daly, H. E. (1992). "Is the Entropy Law Relevant to the Economics of Natural Resource Scarcity? Yes, of Course It Is!" Journal of Environmental Economics and Management, 23(1), 91–95.

Daly, H. E. (1997). "Georgescu-Roegen versus Solow/Stiglitz." Ecological Economics, 22(3), 261–266. With replies by R. M. Solow (267–268) and J. E. Stiglitz (269–270).

Georgescu-Roegen, N. (1971). The Entropy Law and the Economic Process. Harvard University Press.

Georgescu-Roegen, N. (1977). "The Steady State and Ecological Salvation: A Thermodynamic Analysis." BioScience, 27(4), 266–270.

International Energy Agency (2022). The Future of Heat Pumps. IEA, Paris.

Kåberger, T. and Månsson, B. (2001). "Entropy and Economic Processes — Physics Perspectives." Ecological Economics, 36(1), 165–179.

Norgate, T. and Haque, N. (2010). "Energy and Greenhouse Gas Impacts of Mining and Mineral Processing Operations." Journal of Cleaner Production, 18(3), 266–274.

Reistad, G. M. (1975). "Available Energy Conversion and Utilization in the United States." Journal of Engineering for Power, 97(3), 429–434.

Solow, R. M. (1957). "Technical Change and the Aggregate Production Function." Review of Economics and Statistics, 39(3), 312–320.

Sorrell, S., Dimitropoulos, J. and Sommerville, M. (2009). "Empirical Estimates of the Direct Rebound Effect: A Review." Energy Policy, 37(4), 1356–1371.

Szargut, J., Morris, D. R. and Steward, F. R. (1988). Exergy Analysis of Thermal, Chemical and Metallurgical Processes. Hemisphere.

United Nations Environment Programme, International Resource Panel (2011). Recycling Rates of Metals: A Status Report. UNEP.

Young, J. T. (1991). "Is the Entropy Law Relevant to the Economics of Natural Resource Scarcity?" Journal of Environmental Economics and Management, 21(2), 169–179.

Note on figures. Every figure in this chapter is computed in lib/verify/II_06.py and printed with its inputs, units and sources by python3 lib/verify.py II.06. The solar flux ratio and the EROI table belong to Chapter I.01 and are cited here rather than re-derived. The aggregate efficiency series is Ayres, Ayres and Warr (2003); the growth-attribution arithmetic is the authors' own, built on that series and on published United States primary energy and real GDP growth, and it is an attribution exercise rather than a regression — the distinction is stated in the text.