[Target 50–60M, Actual 76M Incl 73-87 on Materialss]

[target 50–60m, actual 76m incl 73-87 on materialss] Integrative design for radical energy and materials efﬁciency at lower cost IIASA internal seminar Laxenburg, 02 September 2019

M OUN KY T C A I O N

I N E STIT U T Amory B. Lovins Cofounder and Chairman Emeritus, RMI (www.rmi.org) Independent contractor, Lovins Associates LLC [email protected]

Thank you for this opportunity to “re-mind” you about how to design whole systems for radical efﬁciency in using energy, water, metals, and other resources. The practice I’ll summarize applies orthodox engineering principles but asks different design questions in a different order and therefore gets very different answers. I recently presented this material as a six-day intensive course, and my team is developing various other tools to help turn this approach from rare to common. * Clean watts are the easy part

I won’t be talking today (rather tomorrow) about energy supply. Modern renewables now provide two-thirds of the world’s net additions of electric capacity, thanks to their powerful business case. Our bigger challenge is capturing modern negawatts. * Reduced energy intensity has had 30× the impact of renewable growth (United States, 1965–2018p, not weather-normalized, EIA data) 250

Primary energy use if at 1975 eﬃciency and structure 200

1975–2018p savings Energy saved by reduced intensity from intensity reduction: 150 2,589 qBTU

100 Primary energy use, 1965–1975 Actual primary energy use

Growth in renewable energy use 1975–2018p growth in total renewable output: U.S. primary energy use (quadrillion BTU/y) U.S. primary energy 0 87 qBTU 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015

Few people realize that the world’s biggest energy “source,” bigger than oil, is the energy saved just since 1990, two-thirds by more-eﬃcient end-use technologies. If the United States’ * total primary energy * demand had kept growing in * lockstep with GDP since 1975, the US would have have used this much energy. Instead, the US * cut that use by more than * half, saving cumulative energy equivalent to 25 years of current use. Meanwhile, * renewable output doubled —yet with 30x less cumulative * impact than the savings. Renewables get virtually all the headlines, because they’re visible, while energy is invisible, and the energy you don’t use is almost unimaginable. Yet saved energy is avoiding about twice as much carbon each year as renewable growth—and the potential savings keep growing. * increasing whole systems,notaspilesofparts—canoftenmakeverybigenergysavings cost I originallythought,atathirdtherealcost. Today thatlooksconservative,because“integrativedesign”—optimizingbuildings,vehicles,andfactoriesas could drop72%in50years.*Sofarit’s dropped58%in43years. Yet justtheinnovationsalreadyaddedby2010*cansave Around 1975,USgovernmentandindustryallsaidthe*energyneededtomake adollarofGDP couldneverdrop.* A yearlaterIhereticallysuggestedit returns. *

Index of U.S. Primary Energy Per Dollar of Real GDP Heresy Happens US primaryenergy intensity, 1975–2017 0.25 0.75 1.25 0.5 0 1

1975 Actual

1990

Foreign Lovins, 1976 Fall , airs ﬀ A

2005 Government and Industry Forecasts, ~1975 Forecasts, Industry and Government 2020

less thansmallornosavings,turningdiminishingreturnsinto

Reinventing Fire 2011 , Reinventing 2035 2050 another threefold,twicewhat How low can we go in the energy limbo?

steelasophical.com

“…85% of energy demand could be practically avoided using current knowledge and available technologies”

Cullen J Allwood J (2010) Theoretical eﬃciency limits in energy conversion devices, Energy 35(5):2059–2069, doi:10.1016/j.energy.2010.01.024 Cullen J Allwood J Borgstein E (2011) Reducing Energy Demand: What Are the Practical Limits? Envtl Sci Tech 45(4):1171–1718, doi:10.1021/es102641n

Professor Allwood’s group at Cambridge University says global energy use in 2005 was only about one-ninth as efficient as physics allows—that is, Second Law efficiency averaged ~11% (vs. ~14% in the US)—so including passive options too, they said * “85% of energy demand could be practically avoided using current knowledge and available technologies.” I think that’s a bit conservative, but whatever the right number is, integrative design gets us closer, cheaper, as I’ll illustrate with diverse examples that start in other sectors to set up some concepts I’ll then elaborate for industrial systems. [Of course, apparent thermodynamic limits can often be evaded by redefining the desired changes of state. Rather than just improving lighting equipment, we can open the curtain to admit daylight. Rather than developing a better car, we can design our cities so people are already largely where they want to be and needn’t go somewhere else. Rather than developing a more-efficient cement plant, we can switch to biomimetic or natural materials, or change our business models and reward systems to wring more structural performance from less concrete, or even refine urban design and societal values so we build less and need less.] * Geological reserves are a small part of resources

Schematic comparison of reserves One of many variants of the canonical McKelvey diagram used by the US and resources (by NERC for British Geological Survey and worldwide Geological Survey) Orebodies are limited. Energy efﬁciency isn’t (practically).

Economic geologists know that a mineral’s “reserves”—identified deposits profitably extractable with current technology and price—are only a small part of the resource base. * Most energy analysts also narrowly define reserves of energy efficiency, like the bright-green zone in these mineral resource definitions. But the actual energy efficiency reserves are severalfold larger than are now typically recognized and captured. The missing majority is hiding in plain view, exploitable by integrative design.

But this geological analogy breaks down on cost. Orebodies are finite assemblages of atoms, while energy efficiency resources are infinitely expandable assemblages of ideas, depleting only stupidity—a very abundant resource. * A major scientific paper on integrative design https://doi.org/10.1088/1748-9326/aad965

Chinese and Japanese translations are at https://rmi.org/insight/how-big-is-the-energy-efﬁciency-resource/

That’s documented in a year-old peer-reviewed paper called “How big is the energy eﬃciency resource?” Its evidence across all sectors shows that unlike oil or copper, most new energy-eﬃciency reserves cost less than current savings, because they come not from adding more or fancier widgets but from using fewer and simpler widgets—more artfully chosen, combined, timed, and sequenced. Before explaining how do we do this magic, let’s start with a little mental calisthenic. * Edwin H. Land (1909–91) “People who seem to have had a new idea have often just stopped having an old idea.” 不忘初心 Bù wàng chū xīn Shoshin wasuru bekarazu 初心忘るべからず Don’t forget original mind

–Avataṃsaka Sūtra, མདོཕལཔཆ, 華嚴經, 대방광불화엄경

One of my early mentors, the inventor Edwin Land, said, “Don’t undertake a project unless it is manifestly important and nearly impossible.” He also said, * “People who seem to have had a new idea have often just stopped having an old idea.” * Asian tradition similarly urges us to seek original mind, beginner’s mind, child mind—opening ourselves to new ideas by shedding all assumptions and preconceptions. * The Nine Dots Problem

So in that spirit, here’s an example from CalTech’s late great aerodynamicist Paul MacCready: For decades, textbooks on creative thinking have posed this problem as “Find the solution that connects these nine dots with just four lines without lifting your pen from the paper.” So you’re supposed to try this...or this...solution that won’t work... * The Nine Dots Problem

...until you think “outside the box” (which is where the expression comes from). But one day, a student startled her professor by saying she’d solved the problem with just three lines. Gee, four was hard enough! How do you do it with just three lines? Dots are inﬁnitely small. Hmmm...wait a minute, these are rather plump dots, and you don’t actually have to go through their center, so if your paper is wide enough... * The Nine Dots Problem

...you can do this! The students then started to feel liberated, and started solving the problem with just one line! Here are a few of their many solutions...* origami solution

If you’re Japanese, you might think of the origami solution....* geographer’s solution

Or if you’re a geographer, you might use a very long line....* mechanical engineer’s solution

Or if you’re a mechanical engineer, a tool-using critter, you might take a tool called a scissors to cut out the dots and impale them….* statistician’s solution

Or if you’re a statistician, you might crumple up the paper, and if you stab it with the pencil over and over again enough times, eventually you’ll go through all nine dots at the same instant. The solution I liked best came from a nine-year-old girl who said, “You didn’t say it had to be a thin line...* “wide line” solution

...so I used a really thick line!” Thus as Paul MacCready said, this “tyranny of the word the”—ﬁnd the solution with four lines—puts us back in the box and keeps us from being more creative in ﬁnding more elegantly frugal solutions. So with beginner’s mind, never having built a house before, and therefore not knowing what was impossible, 37 years ago I did the conceptual and energy design of the owner-builder house…* Lovins House, Old Snowmass, Colorado (1983)

…where Judy and I live 2200m up in the mountains near Aspen, and * temperatures used to drop to –44˚C, with up to 39 days of continuous midwinter cloud. Yet we have no heating system. Omitting the heating system subtracted slightly more construction cost than was added by the superinsulation, superwindows, and ventilation heat recovery that eliminated the heating system. This house helped inspire the German, then European, passive house movement. The central * atrium, seen here in a February snowstorm, has produced * 77 passive-solar banana crops. An analogous approach also works ﬁne in Bangkok; nearly everyone on earth lives in climates between Bangkok’s and mine. * Integrative design gives many beneﬁts from each expenditure: this white arch [point] has 12 functions^ but only one cost. *

^ Supports greenhouse glazing, supports roof purlins, distributes varying cantilevered loads, mounts atrium lights, acoustics, esthetics, thermal mass, controls atrium’s solar gain seasonally (for N–S thermal balance), collects hot water, collects hot air, distributes daylight, vents excess heat. The atrium collects energy in seven ways: heat, light, hot water, hot air, recovered condensate’s latent heat of evaporation and negentropy of mixing (it’s distilled water), and photosynthesis. The recurving (reentrant) walls have ten main benefits: more strength and stiffness per unit material, one set of forms (flip over for opposite curvature), more visual solidity (look out the window and see inside & outside of the 40-cm wall simultaneously), control over site and angle of entry for heat and light, acoustics, esthetics, more thermal mass, better exterior aerodynamics (hence less heat-robbing turbulence and no whistling noises on windy nights). US office buildings: 2–5× best-efficiency gains in 5 years (site energy intensities in kWh/m2-y; US office median ~293)

...21 (–93%) ~277➝173 (–38%) ...➝108 (–63%) …and in Germany, 2010 retrofit 2010–11 new 2013 new 284➝85 (–70%) ...51 (–83%) (office, gallery, apt.) 2013 retrofit 2015 new Yet all the technologies in the 2015 example existed well before 2005!

Our “integrative design” to retrofit the * Empire State Building saved 38% of its energy with a 3-year payback. Three years later, our * cost-effective Denver retrofit saved 70%, making this half-century-old federal complex more efficient than the * then best new US office—which in turn is * less than half as efficient as our own passive, net-positive, no-mechanicals office, using one-sixth the normal energy in the coldest North American climate zone. [In milder Seattle, the Bullitt Center uses a fourth less energy [~38 kWh/m2y] than ours, and] * Now this Bavarian building is using three-fifths less energy than ours! Yet these technologies all existed over a decade ago; what * mainly improved, doubling best efficiencies in five years, is not so much technology as design—the way we choose and combine technologies. * Integrative Design in Retrofitting the Empire State Building

The Empire State Building retroﬁt * remanufactured all 6,514 windows onsite into superwindows that pass light but block heat, plus... * Empire State Building retroﬁt sequence

$8.7M

Minus $17.4 $2.4M

$5.6M

$2.7M $4.4M Annual Savings $4M

Windows Radiative DDC VAV Lighting Avoided Chiller Barrier Controls AHUs & Plugs Plant Retroﬁt

...better lights, ofﬁce equipment, etc., cutting the maximum cooling load by one-third. Then renovating smaller chillers instead of adding bigger chillers * saved $17 million of capital cost, paying for most of the other improvements * and cutting the payback to three years (or less than one year if we’d counted non-energy beneﬁts to the owner or tenants). A major energy-service company had also offered a three-year payback—but with dis-integrated design yielding only one-sixth the savings! *

5x-more-efﬁcient new Indian commercial buildings

Infosys’s 1.5 million m2 of 22k-m2 oﬃce blocks (2009–14) in six cities: Energy Performance Index fell 80%, to 66 kWh/m2-y with capital cost 10% to 20% lower than usual, and comfort better

Courtesy of Peter Rumsey PE FASHRAE (Senior Advisor, RMI) and Rohan Parikh (then at Infosys, Bangalore, now at McBERL)

Similarly, in six muggy Indian cities, 1.5 million m2 of oﬃces integratively designed by Rohan Parikh’s team use up to 80% less energy than the Indian norm, with 10–20% lower construction cost, [60% less cooling capacity,] yet superior comfort and satisfaction. Glarefree daylighting is delivered throughout by contract: if workers complain of glare and demand blinds, the architect doesn’t get paid. * Benchmarking a big new oﬃce (~10,000+ m2, semitropical climate, no PVs, USA; ~2012 Japan; 2015 1,451-m2 RMI Innovation Center; ~2012 India Normal Better Best delivered MJ/m2-y 1,100/1,737 450–680/566 100–230/126/182/158–194 del. el. kWh/m2-y (EPI) 270/203/~200–400 160/195 20–40/35/51/<75 (25 cooling)

lighting W/m2 as-used 16–24/12 10 1–3/2/1/<1.6

plug W/m2 as-used 50–90/12 10–20 2

glazing W/m2K center-of-glass 2.9 1.4 0.3–0.5/0.43/1.1

glazing Tvis/SC 1.0 1.2 >2.0 perimeter heating extensive medium none/none roof α, ε 0.8, 0.2 0.4, 0.4 0.08, 0.97/0.1,0.9 2 m /kWth cooling 7–9 13–16 26–32+/∞/20–26 (750–1000sf/TR) cooling syst. COP 1.85 2.3/2.0–2.7 6.8–25+/–/>6.4 (<0.55 kW/TR) relative cap. cost 1.0 1.03 0.95–0.97/1.11/0.85–0.90 relative space eff. 1.0 1.01 1.05–1.06/1.01

Japan standard: median of 40 buildings, Energy Conservation Center of Japan; better: average of six SHASEJ Junen Award-winning buildings; best: the most efficient of those six buildings (Nissei Yokkaichi Building, 293 MJ), now Takanaka Higashi Kantō 2015 retrofit, ~126 MJ); data courtesy of Urabe-san, CRIEPI, via Asano-sensei, Todai; 2 W/m2 lighting is Shimizu Building 2012. India: empirical Infosys new- office performance data from Rohan Parikh; standard estimate from Indian designers—100 of the 200–400 (nom ~250) is cooling. India:

This chart summarizes technical efficiency benchmarks for big new office buildings in three efficiency categories—normal (i.e. rather poor, typical of >90% of the stock), better, and best. White data are US, blue Japan, green our own office in Western Colorado, and red India.

Comparing the left with the far-right column, you can see that best US practice cuts total and electrical needs by 5–10x, as-used lighting power density by 5–24x, and plug loads by even more. Glazings improve by an order of magnitude and become almost perfectly selective in sorting light from heat. Add superinsulation (not shown) and a heat-rejecting roof, and you need 3–5x less cooling. Then you can make the cooling system ~4–13x more efficient, or even eliminate it altogether. Mechanical savings help reduce capital cost by several percent and use the space ~1–6% more efficiently. Whatever exists is possible. * “Energiesprong” unsubsidized mass retrofit of public housing

Before: 6 Dutch units, each with annual energy bills ~€1.5–2k

After: net-zero-energy, expected soon to be financed just from energy savings; made affordable by industrializing the manufacturing: retrofit originally cost €150k/unit, now €75k (15% subsidized), self-financing target ~€65k, long-term goal €40k

In Holland, industrializing mass retrofits to net-zero-energy, and streamlining their finance and soft costs, is now getting cheap enough to finance entirely from saved energy, while extending building life and improving amenity, health, and value. The whole retrofit has even been demonstrated in Britain to be installable in a single day while you’re at work. *

[“Outsulate” superinsulation including new modular roof with integrated solar PVs (eliminating gas connection and bill), air-tighten, replace all windows, provide heating/(cooling)/hot water with an efficient air- to-water heat pump in a 3-m3 module, renovate kitchen and bathroom, retrofit lights; total retrofit cost ~€40–60k/unit (already fell from €130k to €80k in 2 y).]

] ]

] ] kWh

kWh “Tunneling through the cost barrier” in peer-reviewed studies of ambitious European building retroﬁts

US 2018 average retail prices (2010 $/kWh): residential $0.11, commercial $0.09 Sources: BP, except IEA for enewable heat. Electricity is shown at its heat value, 3.6 MJ/kWh, not at its primary input to an equivalent thermal power plant. Primary-to-final losses are notEuropean reflected. retrofitted building IPCCsavings AR5 WG3reported pp 702–704 2006–13 (2014) reports that (IPCChigh-ambition AR5 WG3 European p 703), new 3%/y (left) and retrofit (right)real discount buildings rateshow over no significant 30 y. increase in theNote cost that of saved the betterenergy cases up to ≥ 90% savings. showSome virtually examples no do rise show in highercost up costs, but they needn’t:to >90% whatever savings exists. Some is possible. cost more, but they needn’t.

The Intergovernmental Panel on Climate Change reported five years ago [in 2014] that for diverse building types and climates, the best European newbuilds on the left and retrofits on the right—both with bigger energy savings toward the right—are saving up to at least 90%, without raising the cost per unit of saved energy. The better projects are all highly cost-effective. The big vertical cost scatter shows the business opportunity to conform inferior projects to best practices. * Why systems?

But thinking that energy-efficient design is about choosing and installing energy-efficient equipment is like supposing that if you simply toss good ingredients in a pot and heat it, you’ll get a tasty dish. Actually, efficient systems, like good cooking, result from whole-system design. Even the finest ingredients won’t make make a great meal unless we * use a good recipe and a skilled chef to combine the right ingredients in the right * sequence, manner, and proportions. The Right Steps in the Right Order

So in design, as in cooking, it’s vital to do the right things in the right order. * The right steps in the right order: lighting

1. Improve visual quality of task 2. Improve geometry of space, cavity reﬂectance 3. Improve lighting quality (cut veiling reﬂections and discomfort glare) 4. Optimize lighting quantity 5. Harvest/distribute natural light 6. Optimize luminaires 7. Controls, maintenance, training

For example, most lighting retrofitters start by installing * higher-efficacy sources and better controls. But the Illuminating Engineering Society’s Fundamentals book rightly tell us * first to improve the visual quality of the task, * then the geometry and cavity reflectance of the space, * then the lighting quality (cutting veiling reflections and discomfort glare), * then optimize illuminance, * then harvest and distribute natural light, and * only then optimize the luminaires and the * controls, maintenance, and training. This sequence can often save an order of magnitude more energy with better visual performance and esthetics. The right steps in the right order: space cooling 0. Cool the people, not the building 1. Expand comfort envelope 2. Minimize unwanted heat gains 3. Passive cooling 4. Active nonrefrigerative cooling 5. Superefficient refrigerative cooling 6. Coolth storage and controls

Result: ~90–100% less energy, more comfort, lower capex, higher uptime

How does RMI’s Innovation Center in Basalt, Colorado eliminate its cooling equipment and keep you cool in 11 ways and warm in 10 ways, none of which involves blowing hot or cold air at you? I once asked my hostess in Japan why she didn’t heat her house, and she replied, “Why should I—is the house cold?” People have nervous systems and comfort sensations, but buildings don’t, so we should keep people comfortable, not buildings. Therefore our building delivers task comfort, like task lighting. Everyone has a * Hyperchair® whose touchscreen or smartphone control delivers comfort to individual requirements, practically eliminating discomfort complaints [the 42% fraction dissatisfied in recent Berkeley surveys of 50,000 officeworkers]. The Hyperchair’s 3.6 W of silent fans and two 7-W electric car-seat heaters are powered by a laptop battery. We * expanded the range of indoor drybulb air temperature to maintain ASHRAE comfort from 64 to ≥86˚F / <18 to ~30˚C, by exploiting all six classical comfort variables—from excellent ceiling fans to superwindows that slash the radiant loads. We * rigorously reduced unwanted internal and external heat gains, and recovered 93% of ventilation heat or coolth. Like an Aspen skier, dressed in a down jacket and sunglasses, this building is nearly twice as airtight as a PassivHaus, superinsulated, and * passively cooled by active exterior shading and natural ventilation, including automated night-flush to cool phasechange walls. We didn’t need such other passive options as ground-coupling, groundwater-coupling, geothermal heat pipes, or a seasonal-storage icepond. We also didn’t need * active nonrefrigerative cooling—evaporative, desiccant, absorption, adsorption, and hybrids—let alone * refrigerative cooling, whose efficiency we can triple (I’ll tell you how later) or * fancy storage and controls. Even in humid climates, this comprehensive approach can * save around 90–100% of cooling energy with better comfort, lower capital cost, and higher uptime. * Superefficient big refrigerative HVAC too (100,000+ ft2 water-cooled centrifugal, Singapore, turbulent induction air delivery—but underfloor displacement ventilation [UFDV] could save even more energy)

Element Std kW/t Best kW/t How Best vaneaxial, 0.2–0.7 kPa (less Supply fan 0.6 0.061 w/UFDV), VAV 120–150 kPa head, efficient ChWP 0.16 0.018 pump/motor, no pri/sec 0.6˚–1C˚ approaches, optimal Chiller 0.75 0.481 impeller speed 90 kPa head, efficient pump/ CWP 0.14 0.018 motor Big fill area, big slow fan at CT 0.1 0.01 variable speed 1.75 (COP 0.588 (COP TOTAL Better uptime & comfort, 2.01) 5.98, –66%) less capex or 0.52 (incl. 0.41 chiller) = COP 6.8 = –70% w/dual ChW temperature (4.5 & 12˚C)

If nonetheless you do have a big water-cooled centrifugal chiller system, here’s how LEE Eng Lock in Singapore trebles its eﬃciency with lower capital cost and better comfort. His end-to-end system COP at the Singapore design hour is 6.8 with or 6.0 without dual chilled-water temperature. That’s a 66–70% measured saving from supposedly good normal practice. Notice the order-of-magnitude savings in the supply fan, pumping loops, and cooling tower by redesigning components, including pipes and ducts, to minimize friction and optimize performance. * Low-face-velocity, high-coolant-velocity coils

Correct a 1921 mistake about how coils work

Flow is laminar and condensation is dropwise, so turn the coil around sideways, run at <1 m/s (<200 fpm): 29% better dehumidiﬁcation, ∆P –95%; smaller chiller, fan, and parasitic loads

There is one trick in the system design: rethinking cooling coils. Mr. Lee’s design corrects Willis Carrier’s 1921 misinterpretation of his lab data. He thought airflow through a cooling coil was turbulent and water condensed in a film, but the late Prof. Sam Luxton in Adelaide found experimentally that actually the airflow is laminar and condensation is dropwise. Turning the usual deep, dense coil around sideways to make a shallow, sparse coil, then blowing the air through it at <1 m/s, <200 fpm, with one-fourth the face velocity, preserves the droplets and their extra surface area. This increases dehumidification 29% per unit of sensible cooling, cuts the airside pressure drop 95%, and shrinks the evaporator load and the entire cooling system. [I understand Carrier has such coils in the lab with production intent.] * Sequence of integrative building design

• Deﬁne the end-use (why cool a building if it can’t feel hot?)

• Optimize the building as a system: costly windows reduce total construction cost

➡ Efﬁciency shrinks or eliminates HVAC; saved capital cost buys the efﬁciency

• Start saving downstream, at the point of use, shrinking capital cost upstream

• Do the right steps, in the right order, at the right time

And get rewarded for excelling in these achievements, via Integrated Project Delivery

and Performance-Based Design Fees!

So what do all these building examples teach? * Start with the end-use effect you intend; * optimize buildings as whole systems, not as a bunch of components (so the key to cutting construction cost is to use expensive glazings); * in particular, pay for the efficiency, largely or wholly or more, through the HVAC shrinkage it causes; * cut the most capital cost by saving from downstream to upstream, as I’ll describe later; * optimize the sequence and timing of design steps; and * reward design professionals for what they save, not what they spend. * Oak Brook Tower retrofit design (1992) 19,000 m2, 20-year-old curtainwall office near Chicago (hot & humid summer, very cold winter); dark-glass window units’ edge-seals were failing, as happens each ~20 y

• Replace not with similar but with superwindows • Let in nearly 6x more light, 0.9x as much unwanted heat, reduced heat loss and noise by 3–4x, cost $8.4 more per m2 of glass • Add deep daylighting, plus very efficient lights (3 W/m2) and office equipment (2 W/m2) • Replace old cooling system with one 4x smaller, 3.8x more efficient, $0.2 million cheaper • Capital savings more than repay all extra costs • 75% energy savings, cheaper than usual renovation: nominal simple payback ~ –5 months • Deep-retrofit portfolio tools: www.retrofitdepot.org

Oak Brook Regency Tower West, 1415 W. 22nd St., Oak Brook, Illinois, http://www.rejournals.com/wp-content/uploads/2013/07/OBRTExterior1.jpg

Timing matters too. When retrofitting a big glass office tower, superwindows plus efficient lights and equipment can shrink mechanical loads and systems by fourfold, more than paying up front for the efficiencies that shrank them. A fourfold efficiency gain in this old Chicago building could thus pay back in about minus five months—cheaper than the routine 20-year renovation that saves nothing—if you coordinate that deep retrofit with the routine renewal of the curtainwall façade. [Tools for analyzing a commercial real-estate portfolio for such opportunities are at www.retrofitdepot.org.] Deep retrofits of all our big buildings will take decades, so let’s right-time them to make the savings much bigger and cheaper. Start with tractive load, not powertrain

Energy content in fuel (~2010 Avcar) 0% 20% 40% 60% 80% 100% moving the engine loss driver idle loss tractive load driveline loss accelerating the accessory loss vehicle

aerodynamic rolling drag resistance • 6% accelerates the car, ~0.3–0.5% moves the driver

• Most fuel use is caused by mass

• Each unit of energy saved at the wheels saves ~5 (formerly ~6–7) units of fuel in the tank

Similar design logic applies to automobiles. * The propulsion system or “powertrain” loses four-fifths of the fuel energy before it reaches the wheels. But our savings should start at the wheels. Here’s why. Just a fifth of a modern car’s fuel energy reaches the wheels and moves the car. Of that * “tractive load,” nearly half heats the * air that the car pushes aside. Most of the rest heats * the tires and road. * So only the last * ~6% of the fuel energy accelerates the car and then heats the brakes when you stop. But 19/20ths of the mass you’re accelerating is the heavy steel car, so just 1/20th of that 6%, or about * 0.3%, of the fuel energy ultimately moves the driver—not very gratifying after 1-1/3 centuries of devoted engineering effort. Moreover, both acceleration and rolling resistance depend on mass, * which therefore causes most of the tractive load [in US autos, or ~90% in India where driving is slower]. / Automakers work hard to cut losses in the powertrain because that’s where most of the losses are—much as Willie Sutton said he robbed banks because that’s where the money is. But reducing powertrain losses is harder than reducing tractive load, because it’s had so much past effort. It’s also less rewarding, because saving one unit of energy in the powertrain saves only one unit of fuel in the tank—but saving one unit of energy at the wheels avoids 4–5 more units lost in getting that energy to the wheels, leveraging 5–6 total units of energy saved at the tank. Thus we should first reduce tractive load, then improve the powertrain—which shrinks for the same acceleration, saving more weight…and also saving capital cost to help pay for the lightweighting. * Migrating advanced composites from military and aerospace to automobiles

95% carbon composite, 1/3 lighter, 2/3 cheaper

How light can we make an auto without compromising crash safety? Using ultrastrong carbon-fiber composites, at least 70% lighter. A 787 Dreamliner is half carbon-fiber composites by weight, but automaking needs roughly a thousandfold higher volume and lower cost than aerospace. That’s a big gap. But in the early ’90s I met Dave Taggart at the Lockheed-Martin Skunkworks. There he’d led the * clean-sheet design of a 95%-carbon advanced-tactical- fighter airframe that was * 1/3 lighter but 2/3 cheaper than the 72%-metal base design. That was too radical, so Dave quit, and one bounce later, 19 y ago [2000], I hired him to lead the complete virtual design, with two Tier One auto engineering firms, of something I’d invented nine years earlier... Radically simplified manufacturing Revolution (2000) “We’ll take two.” — Automobile magazine Midsize SUV, all-wheel-drive Full virtual design, World Technology Award, 2003 5 adults in comfort ! full-scale pusher 2 m3 of cargo 0–100 km/h in 8.3→7.2 s Very sporty handling 857→now ~700–740 kg Superior crash safety 3.56 L gasoline/100 km (67 mpg, realistic on-road, with a ~1-L (!) hybrid engine 2.06 “L”/100 km with H2 fuel cell (114 mpge, realistic on-road) Intl. J. Veh. Design 35(1/2):50–85, 2004, https://www.rmi.org/insight/hypercars- hydrogen-and-the-automotive-transition/

14 body parts, ~95–99% less tooling cost no body shop, little or no paint shop ~80% less automaking capital 2/3 smaller powertrain

…a carbon-fiber midsize hybrid SUV we designed 19 years ago with two European Tier Ones. Its * airframe-inspired body—suspended from rings, not built up from a tub—had just 14 parts, each made with one low-pressure dieset, saving ~95–99% of the tooling cost. Each part can be lifted in one or two hands with no hoist. The biggest part, on the side, I can briefly lift with one finger. * The parts snap precisely together for bonding, self-fixturing and detoleranced in two dimensions, needing no robotic body shop. Laying color in the mold can nearly eliminate the paint shop. There go the two hardest, costliest steps in automaking, * saving ~80% of manufacturing capital. * That plus the two-thirds-smaller powertrain pays for the carbon fiber, making the ultralighting approximately free. * Reinventing the wheels Hypercar Revolution midsize concept SUV (2000) Toyota 1/X carbon-fiber concept PHEV sedan (2007) 3.6 L/100 km on-road (gasoline) or 2.1equiv (H2) Prius size, 1/2 fuel use (1.8 L/100 km), 1/3 weight carbon-fiber structure, ≤2-y retail payback

BMW i3 4-seat electric, carbon-fiber passenger cell Bright IDEA 1-T 5-m3 aluminum fleet van (2009) 2013– mass-production, >150k sold for ~$41–45k ~2.4 Lequiv/100 km PHEV, 3–12×-eff., needs no subsidy 1.7 L/100 km, MY2019 247-km range (≥370 w/REx)

Seven years after we * designed that SUV, * Toyota designed this 70%-lighter carbon-fiber plug-in hybrid sedan. * 2013 brought to market this profitable * quadrupled-efficiency carbon-fiber electric car I’ll describe in a moment, plus another automaker’s 0.9 L/100-km 2-seater. But * even one-ton-lighter aluminum fleet vans, like this hybrid model we developed and road-tested a decade ago [2009], could save a fifth of US light-vehicle fuel at lower lifecycle cost with no subsidy. And * carbon-fiber autos can save far more oil than Saudi Arabia lifts, yet with simpler designs, can be made at normal cost. * sglcarbon.com A competitive carbon-fiber electric car, 2013–

https://www.autocar.co.uk/car-news/industry/bmw-set-make-more-extensive-use-carbon-ﬁbre

2013 BMW i3, http://www.superstreetonline.com/features/news/epcp-1303-bmw-i3-concept-coupe/ BMW MY2013’s ~120–150-kg carbon-fiber-composite passenger cell; mc 1,250 kg BMW’s sporty, 1250-kg 4x-efficiency i3 was profitable from the first unit, because it: • pays for the carbon fiber by needing fewer batteries (which recharge faster) • saves ~2.5–3.5 kg total for each kg of direct mass saved (Detroit says <1.3–1.5) • needs two-thirds less capital, ~70% less water, ~50% less energy, space, time • requires no conventional body shop or paint shop • provides clean, quiet, superior working conditions • delivers 1.9 Lequiv/100 km (124 mpge) on US 5-cycle test, 1.7 Ger., ~1.6 old US cycle • provides exceptional visibility, agility, traction, and crash safety

We know that because * BMW did it starting six years ago with this * carbon-fiber electric car that I drive. This i3 reportedly * made money from the first unit off the assembly line. Sandy Munro, the normally understated US master of automotive costing, called it the “most significant vehicle since the [Ford] Model T” and “the most advanced vehicle on the planet.” * Validating our 1990s claims, its carbon fiber is paid for by the batteries that its lightness saves (and fewer batteries mean faster recharging). Its integrative design * decompounds mass far more than usually assumed. Its * manufacturing is radically frugal, * confirms the elimination of conventional body and paint shops, and * is much better for workers. And overlooked synergies between ultralight materials and electric traction * quadruple efficiency without compromise and * with many driver advantages. * World’s fastest carbon tape layup is in the supply chain 2016 ver 4: two precise prepreg courses in <1 second up to 4 materials, automated coil change, 90˚ or 45˚ cutting materials throughput up to 490 kg/h (~1,000,000 components/y) structural performance 10–30% better than weave-based laminates

http://speautomotive.com/SPEA_CD/SPEA2016/pdf/et/et5.pdf http://www.dieﬀenbacher.de/en/company/public-relations/news/composites/new-possibilities-for-lightweight-construction-in-the-automotive-industry.html http://www.dieﬀenbacher.de/front_content.pho?idart+709&cjamge;amg=3

An RMI spinoff developed an even faster manufacturing process and sold it to a Tier One pressmaker. This 2016 version can make a complex, variable- thickness, optionally anisotopic, overmoldable 2x2m carbon-fiber part in one minute. It could become severalfold faster still. There are at least 16 rival processes, including 3D-printing of whole carbon-fiber cars. Let the competition roll. * 3.6×-more-efficient SUV (6.3× with 2000 fuel cell) can cruise at 89 km/h with the same power to the wheels that a normal SUV uses on a hot day to run the air-conditioner

35-kW fuel cell (small 35-kW 137-liter 345-bar H2 storage enough to afford early: load-leveling (small enough to package): ~32x less cumulative batteries 3.4 kg for 330-mi range production needed to reach needed price)

2017 Mirai (300-mi range, 5 kg 700-bar H2): 2× pressure because 2.2–2.6× heavier, 39% less efﬁcient

[Powertrain costs for both battery- and fuel-cell electric are in rapid flux, but since our original SUV design used a fuel cell (long before cheap batteries), let me emphasize how] Such radical vehicle fitness enables all kinds of advanced powertrain. Our SUV’s 2/3 lower tractive load made its H2 tanks 2/3 smaller for the same range, so 1990s off-the-shelf 345-bar [5,000-psi] carbon-fiber tanks were small enough to fit. [We didn’t need 700-bar tanks like Toyota’s two-ton Mirai. I’m in awe of Mirai’s doubled-power-density, 95%-cheaper-in-9- years [vs 2008 Highlander FCV-adv] fuel cell, but its stack and tanks could have been 2–3x smaller if put in Toyota’s concept 1/X rather than a Prius V platform.] The fuel cell also became 3x smaller [tradeable with the buffer battery according to their relative prices], so you could pay 3x more per kW. At a standard 80% experience curve, you’d then need ~32x less cumulative production volume to reach competitive cost, speeding the hydrogen transition by a decade or two [, using the integrative infrastructure solutions we described to the National Hydrogen Association in 1999]. * Decompounding mass and complexity also decompounds cost

Only ~40–50 kg C, 20–45 kWe, no paint?, radically simplified, little assembly,...

Exotic materials, low-volume special propulsion components, innovative design

To make a car half to two-thirds lighter, you must go repeatedly around the “design spiral” [as it’s called in naval architecture] or “design cycle” [as it’s called in aerospace]. / First you make the auto light and slippery to cut its tractive load by at least half, permitting smaller and more-advanced powertrain and smaller, lighter chassis components. That leaves more packaging space for comfort and more crush space for safety. Next, you keep going around the spiral, making components smaller as their structural loads shrink, because the less weight you have, the less weight you need. Many big parts then disappear: a good series hybrid can eliminate the transmission, clutch, ﬂywheel, driveshaft, U-joints, axles, diﬀerentials, starter, and alternator, each saving even more mass.

At first, the special materials, powertrain, and design may raise manufacturing cost. But after more mass-decompounding, you need so little carbon fiber and powertrain, and manufacturing the advanced-composite structures can get so much simpler, that those two savings pay for the carbon fiber, making the ultralighting roughly free, as BMW proved. * Hypercar® !

Designing our uncompromised, 4–6x more efficient carbon-fiber electric SUV in 2000 required us to organize the designers differently…. * The secret sauce: organizing designers differently

“If we are to achieve results never before accomplished, we must employ methods never before attempted.” —Sir Francis Bacon

The basic design used not a thousand-plus engineers but seven, all around the same table, and collectively responsible for dauntingly ambitious whole- vehicle requirements. Each engineer was responsible for one major vehicle system or function, but for those we deliberately wrote no requirements, because we didn’t want him to make his problem into her problem—we wanted to make the whole team design a highly integrated vehicle together. Two engineers weren’t comfortable without their very own requirements, so we replaced them in the ﬁrst week or two. Then it went great. Toyota asked us how we did it, we told them, and out came their 70%-lighter 1/X that you saw earlier. * Design to win the future, not perpetuate the past

Present design space New design space

Define the end point

Development targets

Risk management

Market introduction

Economic insight

Customer relationships

Technology introduction

Integration payoff areas

First production Foundation variant Platform

Design “in the future”

[To achieve such results,] You must also design in the future, not in the past. When the Soviet Union shot down * Francis Gary Powers’s U-2 spy plane in 1960, Kelly Johnson didn’t say, “I’m going to design a slightly better U-2”; he said, “I want to own the skies for decades, so * we’ll design a Blackbird [SR-71]; I don’t know how, but we’ll ﬁgure out.” And so they did—in about 13 months. Johnson understood that such an airplane was impossible within the conventional design context. He knew that * design is like a rubber band: if you try to stretch it too far from the conventional design space, you encounter more and more resistance, and eventually it breaks. But if you jump to the * new design space you really want, you can stretch the rubber band back as far as needed for technologies not yet ripe, and then as they mature, the rubber band relaxes to where you want to be. * Integrative vehicle design more than doubles potential fuel savings

$5,000

NRC High 2015 NRC High 2011 Bright IDEA PHEV Light Trucks 2014 BMW i3 (same test cycle 2015 Cars Fleet Van, Recosted 2019 (2018 5-cycle test) as rest of graph) 2004 Prius (2004 actual 2004 RMI $4,000 NRC High to 2017 EPA est.) State of the Art NRC Low 2015 2002 Cars Average Light Truck Light Trucks DeCicco & Ross 1995 Avg car 2002 $3,000 NRC High 2002 2017 ULSAB-AVC Hybrid Light Trucks (rough RMI estimate of initial and mature cost) 2017 NRC Low 2002 2000 Revolution w/AWD 2004 RMI Light Trucks Hybrid Powertrain State of the Art $2,000 Average Car NRC Low 2002 Cars DeCicco, An, & Ross 2001 Moderate & Average Cars 2004 RMI steel ICE OEM/RMI 2007 C-class design Average Car = market vehicle NRC Low 2015 Cars = virtual design 2004 RMI steel ICE $1,000 Light Truck = derivative of virtual design 1992 VX Subcompact = road-tested prototype 2000 Revolution 2002 ULSAB-AVC w/AWD ICE

Increase in Manufacturer’s Suggested Retail Price (2000 US$) Suggested Retail Increase in Manufacturer’s 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160

Miles per US gallon or equivalent (EPA adjusted, 55 city/45 highway cycle, 33.7 kWh el = 121 MJ = 1 USgal gasoline)

Source: https://doi.org/10.1088/1748-9326/aad965 (2018), with OEM/RMI 2007 C-class design and BMW i3 updated

These auto examples reveal a big policy lesson. With extra sticker price on the vertical axis and rated fuel efficiency on the horizontal axis, the official technology- by-technology analytic method underlying US and global efficiency policies yielded the * aqua National Research Council 2001 high and low supply curves of potential US light-truck and car efficiency ~15 years ahead, then their * dark-blue 2015 updates, catching up with * previously rejected independent analyses. But those official forecasts were soon embarrassed by actual market platforms like these from * Honda, * Toyota (a hybrid), and BMW (an EV); by a major automaker’s * light-metal gasoline-engine virtual design in collaboration with RMI [, which we could start mentioning a decade later]; and by a Porsche Engineering virtual design using * high-strength steel with a gasoline engine or RMI’s estimate for a hybrid variant. In 2004, we adapted the * base vehicles in our Winning the Oil Endgame analysis, based on our * 2000 Revolution SUV design, yielding these typical * light-truck and car values. And * here’s the aluminum commercial fleet van. So NRC’s component-based analysis misses the entire right-hand two-thirds or more of the design space. That is, highly integrative whole-vehicle design can at least triple, and at lower cost, the fuel savings that policymakers now expect. Analyzing auto efficiency by the part, not by the car, makes efficiency look severalfold smaller and costlier than whole-vehicle integrative design can achieve. So current efficiency standards are far more conservative than anyone thought, and electrification can be far cheaper and faster than today’s heavy platforms yet exploit. *

[Hypercar variants: A gasoline-engine version could save 58% of normal fuel use for 15¢/gal (2000 $), a gasoline-hybrid variant could save 72% for 56¢/gal, and a fuel-cell version with the costly stacks of 18 years ago could save 83% for $2.11/gal. NRC’s 1991 report is labeled “1992” because a 1992 revision made slightly less conservative assumptions. The 2002 ULSAB-AVC developed by 33 steel ﬁrms and Porsche Engineering was 2,200 lb., Taurus-class, 52 mpg, 5✩ safety, $9,538 production cost; its body-in-white was –52 kg and –$7. ] Ultramodern aeronautical technology embodied in a gliding bird

Courtesy of Dr. Paul MacCready (1925–2007) Founder and Chairman, AeroVironment, Inc.

The next frontier is to try to become as good designers as nature is—as the late great aerodynamist Paul MacCready taught us by diagraming some design features of a California condor. Besides the lovely aerodynamic features, I particularly liked the “fully integrated system” and “advanced manufacturing techniques.” * RMI’s latest >$40b worth of integrative design in diverse industrial projects—retroﬁts and newbuilds (solid = built, shaded = incomplete data, circle = not yet built)

5 normal = 1 4 ESCOs best? ESCO 0.8 3 0.6 2 0.4

simple payback (y) simple payback 1

relative capex capex relative 0.2 0 - 20 40 60 80 100 0 energy savings (%) 0 20 40 60 80 100 energy savings (%) ESCO Oil Mining 0il (CDU pumping) Data center Chip fab 68 chip fabs' HVAC LNG Naval hotel load Supermarket Chemical Yacht (el.) Supermarket Food (opex) partner firm

Retroﬁts Newbuilds

These examples from buildings and autos can help us cultivate beginner’s mind and radical efficiency in industry, which uses half the world’s energy and electricity. My team’s latest >$40 billion worth of industrial integrative designs typically found ~30–60% retrofit energy savings, paying back in a few years, and * in newbuilds, ~40–90+% savings with generally lower capital cost. That’s far better than the upper-left brown retrofit zone where most Energy Service Companies deliver dis-integrated design. Our better results come from rethinking industrial processes and redesigning basic elements like pump, fan, and motor systems. *

[One of the best ESCOs, Crowley Carbon, delivered the smaller brown dot for its 62 industrial retroﬁt projects in 22 countries, saving an average of 37% of primary energy with a 2.8-y payback without yet using most of our integrative-design techniques. We’re not aware of other practices that routinely deliver the large savings at lower capex shown on the right for newbuilds.] * Designing to save ~80–90% of pipe and duct friction— equivalent to about half the world’s coal-ﬁred electricity thin, long, crooked fat, short, straight

Typical paybacks ≤1 y retroﬁt, ≤0 new-build But not yet in any ofﬁcial study, industry forecast, or climate model

For example, in both buildings and industry, * better pipe and duct design can save ~80–90% of friction. If everyone did it, this could save roughly half the world’s coal-fired electricity, * typically paying back in less than a year in industrial retrofits and instantly in newbuilds. Just as my house paid up front for superinsulation by eliminating the heating system, so fatter pipes and ducts more than repay any higher first cost by shrinking pumps, fans, motors, and power supplies. * Designing to save ~80–90% of pipe and duct friction— by making them fat, short, and straight

Big pipes, small pumps Nonorthogonal layout, 3D diagonals, few & sweet bends

The methods are simple: use * big pipes and small pumps, not small pipes and big pumps, and * lay out the pipes ﬁrst, then the equipment. But such rearrangement of designers’ mental furniture remains largely unnoticed and unpracticed, because it’s not a technology; it’s a design method, and few people yet think of design as a scaling vector. * New design mentality, an example

No new technologies, just two design changes:

1. Big pipes, small pumps (not the opposite)

2. Lay out the pipes ﬁrst, then the equipment (not the reverse)

Let me amplify these two changes in the design process. * First, we specify big pipes and small pumps, not small pipes and big pumps. Friction in a pipe falls as roughly the fifth power of its diameter; so how fat should the pipe be to optimize friction? The textbooks say to make the pipe just as fat as will repay its extra cost over the years from the saved pumping energy. But that’s wrong, because it leaves out the capital cost of the pumping equipment. A pumping system’s pump, motor, inverters, and electricals must all be big enough to overcome the friction, so their size, and roughly their capital cost, will fall as about the fifth power of pipe diameter. Yet the capital cost of the fatter pipe rises as only about the second power of diameter. So when we optimize the pipe as a component, we pessimize the system. Instead, optimizing the entire system at once yields fat pipes and tiny pumping equipment, so the total capital cost goes down. The * second shift in design is even simpler and thus harder: we lay out the pipes first, then the equipment—not the reverse. Traditionally, we put the tanks, boilers, etc. in convenient or arbitrary places, then ask the pipefitter to come connect point A to point B. But by then A and B are far apart, other stuff got in between, they’re at the wrong height, they face the wrong way, and by the time the pipe snakes across the space (all dressed at neat right angles as they teach us in trade school), it has ~3–6x the friction it would have had with a straight shot. The pipefitters like this because you pay them by the hour, they mark up the extra pipes and fittings, and they don’t pay for your bigger pumping system or electric bill. But for you as owner, it’s better to make pipes fat, short, and straight, not thin, long, and crooked. * No new technologies, just two design changes

Fat, short, straight pipes — not thin, long, crooked pipes!

Beneﬁts counted “Bonus” beneﬁt also captured •7–12× less pumping •70 kW lower heat loss from pipes

Additional beneﬁts not counted •Less space, weight, and noise •Clean layout for easy maintenance access •Needs little maintenance, yet better uptime and longer life •More ﬂexibility for later debottlenecking

Count these too and save…~98%?

In a Dutch industrial pumping loop case by Ing. Jan Schilham, this approach * cut pumping energy by 7–12x with lower capital cost, and as a free bonus, it also saved 70 kW of heat loss with a two-month payback, because short, straight pipes were easier to insulate than long, crooked pipes. But belatedly I realized that * we’d left out eight additional benefits we could have captured, each justifying even bigger savings: the system is more compact, weighs less, and is quieter; those all have a value. It has a wonderfully clean layout for easy maintenance access, reducing maintenance cost, but it’ll need less maintenance anyway because there’s less stuff to go wrong, and uptime will improve. The pipes will last longer without so many elbows’ being eroded by liquid turning the corner. And there’s more flexibility for later debottlenecking if desired. I later estimated that had we properly counted and valued these extra benefits, we’d have saved not 86–92% of the pumping energy but * probably nearer 98%, and the capital cost would have been even lower. Indeed, adding some piping in my own house recently saved ~97% of the friction and pumping power. * 100% 100%

99% 1% 99%

The design changes we need are so simple you can grasp them just from a sketch or a photograph. In practically every building and factory, a critical pump with an in-place spare pump is normally laid out like this, so the flow always goes through two right-angle bends. That means friction. Why not lay it out * like this, so the main flow has no bends (and also fewer valves)? The same logic applies with identical pumps in tandem, a very common arrangement. * Retrofitted Low-Friction Piping Layout

Images courtesy of Peter Rumsey, PE, FASHRAE, Senior Fellow, RMI

In * California’s Oakland Museum, our colleague Peter Rumsey retrofitted * an efficient piping layout into the condenser-water pumping loop, cutting pumping energy by three-fourths with a 2–3-month payback—* and eliminating 15 pumps that will never again waste energy and maintenance costs. When he * also repiped the museum’s chilled-water loop and added variable-frequency drives, he doubled the flow and saved 85% of the energy. He simply asked the pipefitters to lay out the supply pipes as if they were drains. * Which of these layouts uses less capital and energy?

…or how about this?

return from tower to chiller • Less space, weight, friction, energy • Fewer parts, smaller pumps and motors, less

return from installation labor tower • Less O&M, higher uptime

Here’s how most big buildings pipe cooling-tower water back to the condenser. But if we lay it out instead as * Peter does, everything gets * better: less space, weight, noise, friction, and energy; fewer parts; smaller pumps and motors; less installation labor; less maintenance, higher uptime. The only obstacle is force of habit. We should bend minds, not pipes. * 100 Energy units

-70% -9% -12% -55% -20% Power Plant Power Grid Motor/Drivetrain Pump/Throttle Pipe

10% Delivered ﬂow

What do such savings mean for the pumps and fans that use half the torque of the motors that use over half the world’s electricity?

From the fuel burned in the power plant to the end use, many successive losses compound, so only a tenth of the energy in the fuel comes out the pipe as ﬂow.

But now turn those compounding losses around backwards... * 10050 Energy units

-70% -9% -12% -55% -20% Power Plant Power Grid Motor/Drivetrain Pump/Throttle Pipe

5 % Delivered ﬂow

...into compounding savings (from right to left), and every unit of flow or friction you save in the pipe saves ten units of fuel, cost, emissions, and “global weirding” at the power plant. And as you go back upstream, the components get smaller and cheaper, so the total capital cost goes down. * U.S. drivesystems’ 1986 retrofit potential, assuming the same flow delivered with the same friction—no downstream savings

1600 40

1400 TW-h 30

20 1200 10 1000 0 800

TW-h baseline baseline

600 bad rewinds bad rewinds

gross oversizingphase unbalance 400 1400 high-efficiency1200 models improved power supplies eff., motor, & brush losses 200 1000 800 0 TW-h 600 400 loss loss local 200 specific savings Input to after net

after 6% 0 improved drivetrains mechanical drivepower distribution distribution adjusted for after motor- at the meter HVAC saving HVAC

baseline all motor types asynchronous motors bad rewinds gross phaseoversizing unbalance synchronous motors DC motors higher power factor high-efficiency models poor supply waveform turning off idling motors deviation in supply voltage better siting & maintenance ASD & power factor controller Source: The State of the Art: Drivepower, RMI/COMPETITEK, 1989 compressor-only fast controller improved timing of winding heater

So while you’re making the pumps, fans, and motors ~5–10x smaller, you can also do 34 other things to the motor systems that collectively save ~3–5x more energy than just the usual two improvements (high-efficiency motors and ASDs), and cost ~5x less per saved kWh, because if you do the right 7 improvements first, you get 28 more as free byproducts. Our 1989 systems analysis of this potential examined induction motors in red, synchronous in yellow, and DC in green. Our system boundary was from the electric meter to the input shaft of the driven machine, so we counted the efficiency upstream in the wires and controls, in the motor itself, and in the mechanical torque transmission downstream. Altogether we could cumulatively save nearly half the total 1986 US drivepower energy, as shown in blue, with paybacks typically < 1 y at 5¢/kWh. This matches EPRI’s estimated savings but is twice IEA’s, and much cheaper. A third of a century later, that synthesis merits updating. * If a 10x pumping savings sounds to you incredible, just consider that at this instant, your heart is pumping blood ~10x as efficiently as typical industrial pumping systems move water and other liquids. If your ~100,000 km of fractal blood vessels had the design and friction of standard industrial piping, you’d need a heart bigger than your body—very inconvenient. But in fact, your 1/3 kg, 1.5-W heart suffices because your blood follows nature’s standard design —laminar vortex flow.* Biomimetic hydrodynamics Australian sea-captain/ naturalist Jay Harman at paxscientific. com noticed Fibonacci spirals all over nature. Why?

He began to

Images courtesy of Jay Harman and Pax Scientiﬁc (San Rafael CA) imitate them.

In San Rafael CA, Pax Scientiﬁc’s founder Jay Harman has developed this new kind of hydrodynamics, displacing centuries of supposedly mature practice. An ocean captain and a keen observer of shapes in nature, Harman noticed a ropy Fibonacci structure (like the one at the left) in vortices like water draining out of a bathtub, so he froze such a whirlpool in acrylic, realized it must be the minimum-energy shape for the air-water surface, brilliantly inferred that perhaps a rotor with that shape could move ﬂuids with minimum energy, and did so in these * fan and pump designs. So the rotor at the lower right emulates the water vortex at the upper left. * Simplified Water Purification

A growing family of such products can often improve pump and fan eﬃciency by ~20–30%, and has many other applications we’ve barely begun to imagine. * One successful early use is the big tanks that hold puriﬁed municipal water before it’s sent out for use. That water must keep circulating to keep its disinfectant active. * Biomimetic hydrodynamics

That circulation normally needs a giant rotor like the one on the left, driven by hundreds of kW. The ﬁst-sized version in the center photo, driven by a 25– 50-W motor, outperforms it by setting up ring vortices that require almost no energy to keep going, as illustrated on the upper right. Since these biomimetic rotors’ performance doesn’t depend on scale nor on Reynolds number, there’s even more than we thought to be saved in the ~25–30% of global electricity that runs pumps and fans! Modern pumps and fans are marvels of Victorian technology—but learning from nature’s 3.8 billion years of design experience can teach us even better how things are made, how they work, and how they ﬁt. * Start saving downstream for data centers

First debloat software and ensure Then cut utility …then cut support …then cut IT equipment’s that every computation cycle is losses by ~50% overhead by 90% internal losses by 75% needed

Similar logic applies to big data centers. * Two-thirds of the fuel fed into the power plant is lost in the plant and grid. * Half the metered electricity is lost in the cooling system and uninterruptible power supplies—together, half the facility’s capital cost—before reaching the servers. * Half the server energy doesn’t reach the chips because it’s lost in inefficient, usually very underloaded power supplies and in many fans to take heat that largely shouldn’t be there off the motherboard into the room so we can do dumb things with it. The * next problem is severe underutilization of computing resources, partly through insufficient virtualization. The resulting energy flow is about to vanish, so let’s * magnify it. Next comes bloatware running many unnecessary threads and processes and making simple tasks very complex because CPU cycles were cheaper than programmers’ attention, and someone else bought the energy. * Downstream of all that you may even have inefficient business processes. In all, a few hundred-thousandths of the original fuel energy actually delivers customer value. / Where should we start fixing this? Downstream: start at the end. * First write elegantly terse code, optimally compiled, with the goal that every compute cycle is a needed and wanted one. (I’d assumed this could save an order of magnitude in compute cycles, but recent tests suggest it’s two orders of magnitude. The shift to mobile devices now makes this valuable because efficient code stretches battery life.) * Then at least quadruple server efficiency—now even more—and the servers * will need far less cooling and power supply, both of which can be done in smarter ways. * We could even save half the utility losses by using fuel-cell trigeneration, cheaper than the uninterruptible power supply it displaces. / Multiply these savings from downstream to upstream and you get at least two orders of magnitude energy savings. In the actual installation for which we made this diagram, the client rejected most of our recommendations, so we were only able to triple efficiency at the same capital cost, but our partner EDS said that had they all been adopted, we’d have saved ~95% of the energy and half the capital cost. * RFAB from the air

92 acres 1.1 million square feet 284,000 square feet of cleanroom Texas Instruments’ RFab (2005) Capacity for 1,000 employees 40% less energy, $230 million cheaper Paul Westbrook, The Joy of Efficiency, July 2019 www.joyofefficiency.com 40% less energy to process a wafer pattern than Spreading such methods cut TI’s specific energy use TI’s previous best plant (6 miles away, 10 y older) 62% in 12 y, water 56%, greenhouse gases 57%

In another energy-intensive high-tech industry, Paul Westbrook and I co-led the 2004 charrette that made Texas Instruments’ semiconductor wafer fab 40% less energy-intensive with 30% or $230M lower construction cost—which is why it was built in Richardson TX rather than in China. Its system efficiencies eliminated one of the two utility floors. Paul’s new book The Joy of Efficiency updates the * proof of RFab’s performance and TI’s * 12-year spread of its methods to cut company-wide chipmaking energy intensity by 62% so far, water 56%, and greenhouse gases 57%. RMI’s next chip-fab design for another client showed how to save roughly 2/3 of its energy use and half its capital cost while replacing all 22,000 tons of chillers with any of three natural cooling methods that can yield ≥100 units of cooling per unit of electricity. * Optimizing solar power plants across the value chain

Factory assembly Shipping Handling

Installation Interconnection

Integrative design can also transform energy supply technologies. Thus RMI’s SHINE team wrung out half of photovoltaics’ Balance-of-System cost (which was then 60–90% of total cost), enabling DOE’s SunShot program. We then led two more rounds of whole-system redesign with industry. The third round recently yielded a * 2x8-module solar “packet” or solar “Lego block” design. This highly integrated industrial product has 98% fewer parts and 10% less land-use, all optimized for shipping, materials, civil works, electrical and mechanical installation, and windloads. This could unlock massive beneﬁts across the value chain, and match or beat wholesale power prices without needing transmission lines. Such a community-scale (~0.1 to a few MWe) groundmount PV system can feed ~2.5¢/kWh unsubsidized power straight into the distribution system. That’s below SunShot’s 3¢ target for 2030. Now you can buy similar equipment in Australia and it’s being ordered by co-ops in Texas. [We published this work to foment competition in exploiting it. When we used that tactic with Hypercars, our $2–3 million R&D investment leveraged >$10 billion of industry commitments in seven years. Now these latest solar innovations are starting to come into the market.] * Heat pumps are getting even more interesting

9–20 kWt, 200 krpm DHW heat pump >60% of Carnot efﬁciency COP=6–15 for △T=13–31K (www.bs2.ch: COP 12.3 for △T=10K)

PV-powered Tesla Gigafactory saved 6–12 months and 98.5% of energy in a key process by eliminating gas

Heat pumps are advancing nicely. Some produce 160–>200˚C [EDF can produce 130˚C steam at COP 5.2; Viking’s commercial units pump 90 to 160˚C; ECN and partners including Dow are developing 200˚C thermochemical and thermoacoustic heat pumps]. But it’s at least as interesting to find the many processes that need low temperatures and lifts, yet now use fossil-fueled boilers and furnaces. This Swiss miniature 200-krpm domestic-hot-water heat pump beats 60% of Carnot efficiency [raising COP from <3 to 6–8 for 56 or 41F˚, i e 31 or 27K, △T, i.e. 55 or 70˚F input and 111˚F output] by optimizing for the small lift often actually required. * JB Straubel’s first design decision about Tesla’s all-electric, all-PV Gigafactory was that it would have no gas pipe. His colleagues thought he was nuts, because the plant would need process heat, and everyone knows you do that with gas. But with no combustion, getting the air permit took a half-hour over-the-counter— not a minimum of six months…if it’s uncontested. What’s 6–12 months worth in the battery business? Then going all-electric drove serious innovation. Redistilling a vital solvent, normally done with 1 MW of gas boilers, took only a 15-kW electric heat pump, because the 1.5K lift allowed near-Carnot efficiency—a 98.5% energy saving. Surprisingly many process heat loads are like that. A European process-heat retrofit similarly saved 92%. * Negamuda: Safeway ice cream

New process: Old process: After each batch, immediately send next batch through, After every ﬂavored batch, clean out all piping to prepare for saving the “mixed batch” as a special new product. making the next ﬂavor (“clean in place”).

Net New Product Margin = Average product margin (across all ice-cream product lines) + Average raw material cost (previously lost in clean-out) + Avoided waste disposal cost (from not having to pay for disposal) + Avoided time & materials (from now not cleaning piping frequently) + Incremental production achieved at zero capex (from capacity-factor boost) + Incremental product margin (from price premium) – Lost margin on cannibalized products

Now, is that delicious, or what?

But again, the most powerful interventions come not from finding a better widget but from starting with beginner’s mind and asking a different design question. For example, a new environmental manager at Safeway was asked to figure out what to do with the nasty fatty waste left from cleaning out pipes in their ice-cream factory. Explain. * Negamuda: an Arctic industrial diamond mine

Old process: New process: Mine a kimberlite pipe with tens of $b of diamonds Eliminate the mine and mill, yet improve recovery and conventionally—drill, blast, dig, load, haul, crush, separate (air dramatically reduce cost. Probably well worth switching jets activated by diamonds’ X-ray ﬂuorescence ﬂash). Any big diamonds are pretty likely to be broken in blasting or crushing immediately, a third of the way through the surface-then- before you even know they’re there. underground mining plan.

Diﬀerent questions: • How does the mine assay diamonds in kimberlite? (Caustic fusion—dissolve the kimberlite in hot NaOH so only diamonds remain.) • Create a sandworm (Dune sciﬁ novels), with hot NaOH in its gut, that burrows around in the carrot-shaped kimberlite pipe and poops out diamonds. Start it up and neutralize behind it with cheap e.g. H2SO4. • Sandworm will need fancy and costly alloys, but even if made of pure unobtainium costing several orders of magnitude more than conventional metals, still much cheaper than the current mining method—even counting zero salvage value for the abandoned mining and milling equipment no longer needed.

Not yet done because the visionary CEO who loved the idea got hired away.

In a second case, I was able to ﬁnd substantial conventional improvements at a nearly new high-Arctic industrial diamond mine. But then I got curious about grade-tonnage distribution, asked to see the curve, and asked how they assayed the diamonds in their kimberlite deposit. This * revealed a promising way to…Explain. * * Negamuda: a large platinum-group mine

Conventional improvements: New question:

End-to-end improvements in each step: drill, blast, load, hoist, Where can improvements yield the greatest leverage for haul, mill, concentrate, smelt, reﬁne. health and safety, resource recovery, capex and opex, and Can save ~43% of total energy with a few years’ payback. strategic advantage? Answer: ergonomics & info at the face.

Highest-leverage interventions are about the people more than the equipment or process: Miner ergonomics: 30x-lighter/lux-h cableless headlamps (lithium/LED), phase-change vests (cool the miners, not the air), minimized latent loads, white-painted walls with eﬃcient indirect lighting, continuous hydration + electrolytes, quieter/cleaner/less scary working conditions: light lights, cool heads, cool people Separately provide coolth and dryness, ventilation and ﬁltration and coolth; no underground diesels; exhaust underground heat not by blowing coal-electric-cooled air several km down and sideways, but by rising heat pipes Key information to the miner at the face: backpack (or smaller) assay equipment, incentivize hoisting ore not tons

And in a third case, my team found that * 108 kinds of retroﬁts (43% of which required no capital) could save >2 TWh/y, >5 ML/y diesel, and >2.2 MT/y CO2/y, but that * the highest-leverage interventions were not about the mining equipment or process but about the miners. * Explain with *s. * Or get way out of the box (there is no box)

◊ All-electric mine, but save most of the electricity by very eﬃcient use ◊ Get 5 MW by lowering W African iron ore to port with an Austrian reverse-ski-lift, recovering 92–93% of the gravitational energy—capex is similar to railway’s, builds faster, easier maintenance ◊ Get more power by dewatering and lowering perched water table down to dry port through a hydroelectric turbine ◊ No power plant, highly reliable, no grid electricity, no fuel, no emissions ◊ Resupply by ski-lift and dirigible—no railway, no road, secure, probably cheaper, easy cleanup

A different mine, for * iron up a West African mountain, later offered a different opportunity. * Living near a ski town, we’d noticed the Austrian ski lifts and wondered if they could be equipped with buckets, not seats, and run backwards to lower the heavy iron ore down to the port. Sure enough, the same vendors offer that option, recovering 92–93% of the gravitational energy as electricity. * We could then make still more electricity by running the perched water table, a nuisance up at the mine, down to the dry port through a hydroelectric turbine, then selling the water. Together, these two generators (even without PVs) could run an efficient all-electric mine with no fossil fuels, emissions, or grid connection, and * it turned out to be attractive to resupply by a combination of the ski-lift and dirigible. More beginner’s mind. / * Typical areas for big industrial savings

◊ Rightsizing everything (if we designed 747s this way…) ◊ Thermal savings and integration ◊ Innovative and distributed power systems ◊ Designing friction out of fluid-handling systems ◊ Superefficient drivesystems ◊ Water/energy integration ◊ Superefficient and heat-driven refrigeration ◊ Advanced controls

In summary, we’re all used to looking for big industrial savings in these familiar eight places. Explain w/*s. That’s a good menu, but…* Designing for breakthrough industrial energy efficiency: an eightfold way

1. Vision, intent, model, strategy, This approach lets you: and culture first: why do it? 1. Capture multiple benefits 2. Task elimination before task 2. Make them compound 3. Demand before supply 3. Free up the most capacity 4. Downstream before upstream 4. Avoid the most capex 5. Application before equipment 5. Eliminate the most waste & harm 6. People before hardware 6. Make the most profit 7. Passive before active 7. Do the most good 8. Quality before quantity 8. Have the most fun

…I’d encourage you also to apply eight simple methods and priorities, each with important beneﬁts explain with *s. * If a problem can’t be solved, enlarge it.

—Dwight David Eisenhower

I’d like to conclude by applying to industrial process heat General Eisenhower’s wise advice to make tough problems soluble by expanding their boundaries to encompass more options and more synergies, so they include what the solution requires.* The U.S. resource cycle (~1990): a massﬂow perspective

Durables Recycled (0.005%) natural nutrients technical nutrients grown (4%) Biotic Resources Durables (0.245%) Product (1.75%) Waste Natural capital 100% 25% Manufacturing Product in Use Ephemerals (1.5%)

Abiotic Manufacturing Waste (23.25%) Resources mined (21%)

Extractive Waste (75%)

• Total massﬂow/person-day, excluding ~4× more in water returned clean, is ~20× each person’s body weight. • Only ~2% of durable products get reused—just ~0.005% of total material inputs into manufacturing. • Thus ~99.995% of material inputs is wasted—surely the biggest business opportunity on the planet.

Source: Hawken & Lovins analysis, Natural Capitalism (Little Brown, 1999); data from Wernick & Ausubel, Ann. Rev. En. Envt. 20:463–492 (1995); data from ~1990,, graphed by Robbie McIntosh (RMI)

Wasted materials are extraordinarily pervasive. Around 1990, the US economy mobilized a * total massflow of ~20x each person’s weight, per person, per day—excluding ~4x more in water returned clean rather than dirty. Yet only * one-fourth of that vast massflow went into production; the * rest was lost in extraction. Of all inputs to manufacturing, * nearly all went to waste too; of the * products made, * 6/7ths were consumer ephemerals thrown away after one use or no uses; and of the durable products still in use six months later, * only 2% ultimately got composted or recycled. So of every million grams grown or mined, * only about 50 grams got fed back into nature or manufacturing, partly because so much of the waste was toxic. Thus the massflow in the US economy was roughly 99.995% waste. Does that sound like a fat business opportunity? * An alternative resource cycle applying the principles of natural

capitalism Durables, Ephemerals, and Existing Waste Recycled natural Regeneration nutrients technical nutrients

Biotic grown Durables Resources Product

Product in Use Waste Manufacturing (dematerialized, Natural capital long-lasting) Less Ephemerals

Abiotic Manufacturing Waste Resources mined Recycle Manufacturing Waste

Extractive Waste 1. Radical resource productivity in harvest & extraction upstream, manufacturing & use downstream. 2. Grow more, mine less, biomimetic production—with closed loops, longevity, no waste or toxicity. 3. Reward these shifts with “solutions economy” business models. 4. Reinvest resulting proﬁts in natural capital—the scarcest and most indispensable kind.

Source: Hawken & Lovins analysis, Natural Capitalism (Little Brown, 1999); data from Wernick & Ausubel, Ann. Rev. En. Envt. 20:463–492 (1995); data from ~1990, graphed by Robbie McIntosh (RMI)

A natural-capitalist economy delivering the same or better services would grow more and mine less of the reduced inputs, and use * radical resource efficiency to minimize losses in extraction and manufacturing. * Producing durable, dematerialized, and long-lasting products, manufacturing would * close loops and design out toxicity, so most of the dwindling waste flows can return to create more value. * A “solutions economy” business model speeds these shifts by rewarding both providers and customers for doing more and better with less for longer. Some of the increased profit * can be further leveraged by reinvesting it in regenerating nature to help the most scarce and precious forms of capital, nature and people, flourish once more. Now let’s dig further into how to use these and similar demand-side methods to decarbonize most industrial process heat.* How will we replace fossil fuels for process heat?

Traditional starting point: Redesign process heat – less – cooler – cleaner

Integrative-design starting point: – Design out the need – Substitute for the product – Save the product – Align incentives: reward doing more and better with less for longer

…leaving little process heat to decarbonize

Industrial energy is dominated by process heat to make basic materials like cement, steel, and aluminum, so products and processes are being redesigned to use less heat, cooler, and cleaner. But we should start at the end, with demand not for energy but for the materials themselves, and even before that, for the services delivered by using those materials: need less, use other, use less, use longer, use again, align incentives. Twenty years ago, our business book Natural Capitalism found that US materials ﬂow around 1990 was 99.99% wasted. So here’s how RMI is starting to think about this opportunity, using mainly cement and steel examples. * 8 neglected prior paths to decarbonizing process heat Design out the need for concrete, steel, etc

Chemical microreactor Mobility-as-a-service 3D printing, local manufacturing

https://www.researchgate.net/ﬁgure/Rendered-top-view-of-the-scale-up- microreactor-die-Metallization-is-shown-in-gray_ﬁg3_277494576 3D-printed metal chairs by John Briscella,”continuum3,” from Zach Andrews, designboom.com, 25 Mar 2018

Onsite (not remote) building services estimatefares.com New Urbanist design

https://www.greentechmedia.com/articles/read/how-much-does-a-rooftop-solar-system-with-batteries-cost#gs.dVVQ0FDG

First, we can change how we deliver the services that now need energy-intensive materials. The remote infrastructure that provides * six kinds of services to buildings, and the wires and pipes in between, can often be replaced by onsite techniques, at lower cost to society and probably to the developer, while saving steel, concrete, and plastics. Just turning impervious landscape into a sponge can manage stormwater without needing giant concrete pipes. Also, replacing separate specialized buildings with multipurpose buildings plus multi-use zoning can save both building space and travel. * 3D printing enables bone-like structures using no factory and little metal. * A chemical plant may use 80–90% of its capacity and energy to separate unwanted byproducts, but microreactors etched into stacked silicon wafers can often control the reaction so precisely that they make only pure product. Both these innovations also allow local manufacturing, displacing steel vehicles and concrete-and-steel roads and warehouses. * Shared and connected mobility likewise saves roads and parking. So does virtual mobility that moves just the electrons and leaves the heavy nuclei at home. And * designing cities around people, not cars, saves one-third of concrete and two-thirds of driving while greatly improving quality of life. * 8 neglected prior paths to decarbonizing process heat

Solutions-economy business models deliver service, not stuﬀ

Just like mobility-as-a-service, which now has 50 million drivers serving a billion ride-hailers, Natural Capitalism’s solutions-economy business model can lease structural performance rather than selling tons of cement, so the provider and customer both make more money by using fewer tons. The provider can also reward structural engineers for delivering the best performance with the least material. When I suggested this to the Chairman of a giant cement company a few years ago, he said, “Good idea. I have 200 people working on that right now.” I need to check how they’re doing.

[The solutions economy is taking hold in building services like lighting. It’s already won in much of the chemical industry: Dow and SafetyKleen lease dissolving services rather than selling solvents. Mining may be next, leasing conductance services rather than selling tons of copper; the head of the world’s largest gold-mining company already proposed such a model for gold.] * 8 neglected prior paths to decarbonizing process heat Productive use: frugal structural design

Tension structures—~80–90% less material RPS, IPTC, FabWiki

Fabric forms—≥50% less material Schlaich Bergermann—see the remarkable book Leicht Weit Mark West, The Fabric Formwork Book, Routledge, 2016; CAST (Centre for Architectural Structures and Technology), Biomimetic structures—~60–90+% less material University of Manitoba, Winnipeg. See Hawkins et al’s 172-reference 2016 review, doi:10.1002/suco.201600117

Brennan O’Rear, 2018, from Cooper http://animalia-life.club/ http://www.miralon.com/how-its-used? Hewitt, Smithsonian Design Museum other/bird-bone-cross- hsCtaTracking=0ebe525d-8588-4308-9bef- section.html 88516ce9b6ce%7C8601a7f7-555d-48c9-85 Cheung et al, Science, doi: 10.1126/science.1240889 9a-7c44e2156835

Rewarding elegantly frugal design can often wring an order of magnitude more work from materials. * Substituting tension for compression structures typically gives better strength, esthetics, and cost with one-eighth the tons. * Fabric forms’ optimal shapes can save upwards of half the concrete in beams, sheets, and other common shapes to deliver the same or better strength and stiffness—and then the weight savings snowball because you need less strength to hold up less weight. The overall saving in making all kinds of buildings from concrete and steel is at least 2x. * Nature is rich in incredibly mass- efficient designs that we can imitate, from trees to bird-bones. * MIT’s self-assembling hook-together miniature lattice structures molded from conventional thermoplastics are as strong and stiff as solid elastomers but as light as aerogel. Carbon-fiber lattices and * carbon-nanotube structures can then save another 1–2 orders of magnitude—and now we can make nanotubes from sugar, water, and sunlight via the bacterial enzymes hummingbirds use to weave their nests. * Latest MIT/NASA version—59✕ lighter than a “dumb” airplane wing

Structure as strong/tough as rubber but ~268✕ less dense (5.6 kg/m3), made of thousands of identical injection-molded anisotropic parts, all covered by a tough polymer membrane of identical material, can yield any desired overall shape An optimized-shape airplane that completely and continuously adapts passively to match ﬂight conditions can thus be made stiff, strong, but scalable in manufacturing and in microrobotic assembly, needing no separate ﬂight surfaces 4.27-m-wingspan model in NASA’s high-speed wind tunnel worked better than predicted; applicable to wind turbines

N B Cramer et al 2019 Smart Mater. Struct. 28 055006, 01 April 2019, https://doi.org/10.1088/1361-665X/ab0ea2, http://mit.edu/archive/spotlight/shape-changing-plane-wing/_, http://cba.mit.edu/docs/papers/19.03.MADCAT.pdf

Those lattice structures just produced an airplane wing 59x less dense than normal. Eliminating moveable flight surfaces, its entire shape passively adapts like a bird’s wing to optimize continuously to real-time flight conditions. This opens revolutionary prospects for lightweighting, aerodynamics, and cost reduction. * 8 neglected prior paths to decarbonizing process heat Productive use: material efficiency

A lightweight bridge being 3D-printed A 13-m bridge supported only by 1-mm-thick carbon-ﬁber ribbons https://www.shapeways.com/blog/archives/35854-3d-printed-bridges-now.html https://www.researchgate.net/publication/275653741_Carbon_ﬁbre_stress-ribbon_bridge

http://chinaplus.cri.cn/photo/china/18/20190112/234804_2.html The artistic 3D-printed https://3dprint.com/184852/bridge-3d-printed-railings-china/ 12.5m stainless-steel bridge over Amsterdam’s Oudezijds Achterburgwal canal

Chinese 3D-printed bridge girder and railing https://www.designboom.com/technology/mx3d-metal-bridge-3d-print-update-04-03-2018/

3D printing of materials-frugal bridges is becoming popular in China, Holland, Dubai, and elsewhere. * This swooping 3D-printed steel bridge is about to span a famous Amsterdam canal. * And here the structural engineering faculty at TU Berlin tests a 13-m free span supported only by three carbon-ﬁber ribbons 1 mm thick. * 8 neglected prior paths to decarbonizing process heat Productive use: other materials

BMW MY2013’s ~120–150-kg carbon-ﬁber-composite passenger cell; mc 1,250 kg

http://tudalit.de/wp- content/uploads/ 2016/02/TUDALIT13.pdf

speciﬁle.co.za

We can also substitute low- or even negative-carbon materials, like * wood and bamboo for steel and concrete [—cheap smoked-bamboo structures [from Simon Veléz or Gunter Pauli] can be safe from insect damage for 500 years—or Pliny Fisk’s rammed earth, or the Anasazis’ adobe, or Hassan Fathy’s in-situ-fired ceramic buildings]. The cheap husk of the rainforest- restoring Indonesian sugar palm has lignin fibers with tensile strength comparable to stainless steel but 1/7th the density. Modern materials are powerful too. A thin sheet of carbon-fiber composite can save two-thirds of the wood in a gluelam beam or 30–50+% of the concrete in a structural panel. * Carbon fiber can profitably displace steel autobodies. * Carbon-fiber anticorrosion wraps save the 30–40% of concrete needed not for static strength but just to cover the steel rebar, which * carbon-concrete composites eliminate. And * the otter-resistant shell of the abalone is tougher than our best missile-nose-cone ceramics, but instead of being made in an incandescent furnace, it self-assembles in 4˚C seawater. Conch shell is 10x tougher still—1,000x tougher than the biomaterials it’s made of. [/ Ernie Robertson in Winnipeg listed three ways to make limestone into a structural material. You can cut it into blocks as the Greeks and Romans did—durable, beautiful, but tedious. Or you can calcine it at ~1,250˚C into Portland cement; that’s inelegant. Or you can grind it up and feed it to chickens. Hours later it reemerges as eggshell, which is lighter and stronger than the best Portland cement. If we were as smart as chickens, we might master this ambient-temperature technology.] * 8 neglected prior paths to decarbonizing process heat Productive use: Productive use: quality and design substituents

Taiheiyo carbon-neutral bamboo concrete theconstructor.org

Freedom Tower (541 m) Shanghai Tower (632 m) Rahimian and Eilon, 2012 academy.autodesk.com

https://www.indiamart.com/proddetail/fumed- silica-8346603562.html

http://news.rice.edu/2018/06/18/cementless- ﬂy-ash-binder-makes-concrete-green-2/

Besides * bamboo rebar whose growth fixes offsetting carbon, low- or negative-carbon substituents to dilute Portland cement range from otherwise wasted * silica-rich rice hulls and * silica fume to * fly ash. Some can markedly improve concrete’s performance. * New York’s Freedom Tower saved 40% of its cement through stronger concrete and mass decompounding [14ksi = 96 MPa at 28 days, 18ksi at 56 days; normal house concrete is ~3 ksi]. Now ultra-high-performance concrete can be more than twice that strong [>200 MPa w/steel (not carbon fiber) rebar and no CF jacketing]—~10x as strong as house concrete. * Twisting the 128-story Shanghai Tower, [China’s tallest and] the world’s second-tallest building, shrank wind loads and their structural systems 24%, cutting materials cost by $58 million! [Here’s another example of elegant structural engineering: Consultants on the Golden Gate Bridge seismic retrofit first advised massive and costly additions of reinforcing materials. An Indian engineer reportedly pointed out that this would worsen the risk by slowing the bridge’s mechanical period into a range resonant with earthquakes. Small, light fixes could simply shift the structure’s resonances out of the seismically relevant range.] Higher-quality steel and cement in China

Impact of Building Lifetime on Steel Production 900 ) s e n

n 800 o t

n o i

l 700 Cumulative reduction l i insteel production ( m

n 600 from longer building o i

t lifetime: 2050 Mt c u

d 500 o r P

e 400 e t S

l a

t 300 o

T 30 year Lifetime 200 50 year Lifetime

100

0 6–8% reduction in 2012 2016 2020 2024 2028 2032 2036 2040 2044 2048

Sources: Climate Policy Initiative 2013; Allwood et al., 2011, LBNL 2050 LEAP Model

China makes over half the world’s cement. China’s once-ubiquitous shaft kilns produced poor cement of such irregular quality that its quantity had to be tripled to ensure desired compressional strength—tripling the fuel savings from switching to modern technology. Making cement and rebar to European quality can make buildings last longer and can thus cut cut Chinese cement demand 6–10%, Chinese steel demand 6–8%, and global rebar demand 13% [, with similar beneﬁts also in consumer goods].

Better building materials also have important indirect beneﬁts. Their lower quantity combines with lower manufacturing energy and decarbonized supply to save more coal. Less cement, steel, and coal needs to be transported, freeing rail capacity that can help shift freight [both bulk spillover and intermodal] from road to rail[— China still moves much coal in big trucks]. That cuts road damage, saving cement and money to rebuild roads. Less cement means still less transport need, etc, while freeing more rail capacity reduces rail buildout and repair, saving more steel and money. Such virtuous circles save more carbon at lower cost. * 8 neglected prior paths to decarbonizing process heat

Productive use: improve materials quality

https://www.chemistryworld.com/news/sea-urchin-spines-inspire-elastic-concrete/3008390.article

DOI: 10.1126/sciadv.1701216

Like abalone or conch shell but even more flexible, sea-urchin spines use a self-assembling mesocrystalline structure that can give calcium silicate hydrate 40–100x the flexural strength of normal concrete. What can we do with flexible concrete or analogous new cementitious chemistries? or with biomimetic spider-silk weavable into cocoon-like rooms? [Zygote Quarterly 24 3:8–23, 2018] * 8 neglected prior paths to decarbonizing process heat

Productive use: closing loops

https://www.slimbreker.nl/why-smartcrushers.html

Finally, construction and demolition waste is the world’s largest nonagricultural waste stream. In the EU it’s about a half-tonne per person per year, yet it’s especially suited to recycling. A Dutch technique for smart crushing with just one-tenth the normal force can selectively liberate the ~40% unhydrated cement for valuable reuse in new concrete. Globally this might save ~2.5 GTCO2/y. And wollastonite cements like Solidia that stick to themselves let concrete roads be patched like asphalt instead of having to be torn up and replaced. * 20 paths to decarbonizing process heat (e.g. for cement) A fuller portfolio Eliminate need: onsite building services vs remote infrastructure, 3D printing/local manufacturing, chemical microreactors, telecoms vs roads, shared & connected mobility vs parking, urban form vs automobility (⅓ less concrete) Service, not stuff: Solutions-economy business models (structural services not tons of cement, mobility services,…) Productive use: Elegantly frugal structural design with appropriate safety margins, rewarding civil engineers for quality Use other materials: e.g., ultralight carbon-fiber cars for heavy metal cars, timber for concrete, adobe/caliche,… Increase substituents: fly-ash, ground glass, rice-hull derivatives, nano or fume silica,… for clinker, bamboo for rebar Improve materials-quality uniformity (3⨉ in cement by eliminating Chinese shaft kilns) Materials efficiency: e.g., fabric concrete forms (≥2⨉), tension not compression structures (~8⨉), Girshenfeld, Miralon Close materials loops: longevity, dematerialization, reuse, remanufacturing, recycling, downcycling,… Less onsite waste via ontime delivery (Cemex), tighter specs, Smart crushing/unhydrated cement recovery/reuse Capture significant knock-on effects such as reduced energy to transport cement, build roads & factories,… Cleaner stuff: Substitute carbon-free or -positive chemistries (Solidia, Calera, Novacem,…) Processes requiring less or no heat or (biomimetically—abalone shell) Processes requiring lower temperatures: olivine+steam, ecocement, Bang bacterial cement, geopolymers,… Make better: More-efficient processes, equipment, and controls Heat recovery and cascading, cogeneration: e.g., McKay’s Hong Kong dioxin-free municipal-solid-waste cogen Make cleanly: Fuel-switching: biofuels, bioprocessing byproducts, waste solvents, old tires, crop wastes,… Solar process heat (logical evolution for solar concentrators; can include cogen; feasible even with cloud) Renewable electricity for heat pumps—now 160˚C, soon >200˚C Renewable electric process heat or plasma Renewable hydrogen process heat or reductant

These and many other * diverse * demand-side * options for using less cement, steel, etc multiply to very large savings even before circular economy, more-benign products and * efficient processes, * fuel-switching, and * renewables. Such efficient end-use multiplies back upstream to materials extraction and production, leaving far less high-temperature process heat to decarbonize. Thus adding the missing first eight lines to the conventional 12-line opportunity space makes the problem much easier. RMI is exploring such options. / So I don’t think heavy transport or process heat are uniquely hard. Of course, none of this stuff is easy, and some subsectors like cement do have unique institutional barriers. Yet the “hard” sectors’ small number of major players and large combustion devices, and their operators’ relentlessly competitive focus on the business case, might even offer potential decarbonization advantages over messier, fragmented-market, multimillion-target sectors like light- duty vehicles and buildings. Perhaps the main thing that makes heavy transport and process heat hard is the belief that they’re hard. This little gallery of examples suggests that may be a losing bet. We’re excited to test that hypothesis as the next enlargement of industrial energy efficiency’s integrative design space. * “Only puny secrets need protection. Big discoveries are protect by public incredulity.” —Marshall McLuhan

M OUN KY T C A I O N

I N E STIT U T

www.rmi.org | [email protected] | +1 970 927 3129

If anything I’ve said has surprised you, just remember Marshall McLuhan’s remark that “Only puny secrets need protection. * Big discoveries are protected by public incredulity.”

In tomorrow’s public lecture, I’ll integrate this demand-side approach with the parallel supply-side revolution. Meanwhile, thank you for your good work and your kind and prolonged attention.*