Costs of Decarbonization. Geoffrey Heal

Transcription

1 Costs of Decarbonization Geoffrey Heal

2 Introduction In its submission to COP 21, the US expressed a desire to reduce its greenhouse gas emissions by 80% by mid century. Not a formal goal, rather an aspiration consistent with keeping global warming to less than 2 C. I go some way towards exploring this alternative, and look into whether the US economy could possibly move largely away from carbon-based energy by

3 Basics The US has approximately one terawatt (1 tw = Watts) of electricity generating capacity, producing about four billion megawatt hours (4 bn mwh) of electric power each year This is the US s largest source of GHGs: 30% of GHGs come from electricity generation, 26% from transportation.* These numbers cover all GHG emissions: for CO2 alone, electricity production accounts for 37% and transportation for 31% * In April 2016 transport emissions exceeded generation emissions for the first time 3

4 4

5 5

6 Power source Percent of capacity Percent of output Coal Gas Nuclear Hydro Biomass Geothermal Solar Wind Petroleum

7 Scenario Assume we need to build new non-fossil capacity capable of generating 66% of current total output. Assume this new capacity is divided 50/50 between wind and solar PV, so from each we need 33% of current output of 4 bn mwh/year. There are 8760 hours in a year, so this means that we need 0.33x4x10 12 /8760W=0.1506x10 9 kw of capacity from each fuel type. 7

8 Capacity Costs To work out how much wind or PV capacity we need to build to produce an effective capacity of x10 9 kw, we need to know the capacity factors According to EIA these were 32.5% and 28.6% for 2015 on average Hence we need to construct x10 6 kw of wind and x10 6 kw of PV. These numbers may be too large: capacity factors for both wind and solar PV have risen over the last decade and may continue to do so NREL gives median CF for wind built in 2015 as 38%: using this we need only 396.3x10 6 kw of wind 8

9 Capacity Costs I assume that wind farms cost $1700/kW So x10 6 kw of capacity will cost $0.788 tn. The cost of solar utility-scale installations I take to be $1.91 per watt Hence the cost of x10 6 kw of solar PV capacity will be $1.005 tn. total of $1.793tn Using $1500 and $1.25 it is $1.353tn Using $1500 and $1.25 and 38% wind CF it is $1.252tn 9

10 Grid Extensive use of wind and solar, whose costs are lowest in specific areas of the country, might require the addition of another 25% of current 200,000 mile transmission capacity. This means 50,000 miles and at an average of $2m/mile this would cost $100 billion. 10

11 Storage Currently most grid-scale energy storage in the US takes the form of pumped hydro power stations Battery storage capacities are typically measured in megawatt hours (mwh) when used in the grid, or kilowatt hours in cars. (A Tesla model S battery has a capacity of kwh depending on the options chosen.) Battery storage has historically in the region of $400-$500 per kwh 11

12 Storage Consider a wind turbine with capacity of 2mW, a typical turbine. Assume a capacity factor of 32.5%: then on average it produces 24x0.65mWh daily, 15.6mWh/day. At a capacity cost of $1,700/kW it will cost $3.4m. At $300/kWh a battery large enough to store one average day s output will cost $4.68m A 10 mw solar installation would cost $19.1m, produce on average 24x2.86 mwh daily At $300/kWh a battery to store this would cost $20.6m 12

13 Storage At the promised price of redox flow batteries, $150/kWh, the costs to store a day s output from a 2mW wind turbine or a 10mW solar farm are respectively $2.3m and $

14 Storage The US consumes 4 bn mwh/year, of which in our scenario two thirds would be from renewables. This means that on an average day it would consume 7.3x10 6 mwh of renewable energy At the optimistic cost of $150/kWh the capacity to store one day of renewable energy production would cost $1.095 trillion. At Elon Musk s retail of $350/kWh it costs $2.555 trillion. 14

15 Cost Offsets - Opex Installing solar/wind, is prepaying electric power for years One kwh requires short tons of coal or mcf of natural gas. Taking prices to be $40/short ton $2.75/mmBTU and assuming a 50/50 mix of coal and gas, zero marginal costs of renewables saves fuel of $ bn per year once renewables have fully replaced fossil energy sources. Assuming renewables replace fossil linearly over 30 years average saving is half of this, over thirty years a total of $ tn. 15

16 Cost Offsets - CapX Cost of replacing fossil plants that come to the ends of their lives over the next three decades Most coal plants are already at least 41 years old, against an expected life of years. The rest were built before 1990, making them at least 26 years old and due for retirement by % of all gas generators were over 10 years old as of 2010, making them candidates for replacement by 2050 Offsets $1.06tn 16

17 Costs Cost category Best case ($ trillion) Worst case ($ trillion) Capacity $1 $1.79 Transmission $0.1 $0.2 Storage $2.2 $4.00 Gross total $3.3 $5.99 Fuel savings $0.96 $0.96 Plant replacement offset $1.06 $1.06 Net total $1.28 $3.97 Worst case corresponds to average cost over last four years. Best case is 10% lower than 2016 costs.

18 Over 34 Years.. Best case $37.6bn/yr In 2015 US CapX on electricity capacity was $42bn. Worst Case $135bn/yr Roughly 3X current spend Note the big unknown is the amount of storage we will need 18

19 Storage again Requirement depends on Covariance between different energy sources spatial diversity and wind/solar breakdown Grid connectivity Extent of demand response Exact fraction of renewable vs dispatchable energy Could probably get to 40/50% renewable without storage 19

20 Transportation Light duty vehicles account for 63% of US transport-related GHGs (cars 34% and light trucks 28%): heavy-duty vehicles account for 21% All major carmakers offer BEVs: how fast will they replace ICEs? Obstacles: Range Cost Charge time First 2 now overcome, or will be within 2 years Working on charge times 20

21 Transportation Autos last about 15 years so we have two+ auto generations between now and 2050 Need most cars sold in 2035 to be electric (BEV, FCEV,.) to phase out ICEs in cars by 2050 McKinsey s by 2030 BEVs will be 50% of light vehicles in US Seems challenging to get 100% Evs by 2050 but could get > 50% Biofuels are another route 21

22 Conclusion Goal - to reduce emissions 80% from 2005 by 2050 Already 9% down, leaving 71% Decarbonizing electricity is possible and reduces emissions 30%: requires net investment of $37-135bn/yr (39%) Electrifying 75% of light vehicles reduces another 13.5% (52.5%) Replacing 50% of residential/industrial uses of fossil fuels (space, water, process heating) by electricity reduces another 16.5% (69%) Otherwise boost LULUCF sinks, reduce HFCs, tackle heavy vehicles, biofuels, efficiency. 22

23 Energy Storage And the alternatives for dealing with intermittency 23

24 Why Energy Storage May Be The Most Important Technology In The World Right Now Forbes April /17/2017

25

26 Three options for Intermittency Store energy from times of surplus to times of shortage Rely on spatial diversification, lack of correlation between renewable outputs in different locations Demand-side management cut demand when output falls via negawatt markets

27 Storage Conceptually simple. Options are Hydro/pumped hydro Compressed air energy storage Batteries Li-ion or Redox Flow or Mg-ion... 27

28 Storage Currently expensive to store energy on a large scale Cheap storage could allow us to smooth output of wind, solar farms Storage is important if non-fossil fuels are to take over from fossil fuels completely Also to flatten load curve and reduce peak generating capacity Tremendous potential for arbitrage between off-peak and peak hour prices 28

29 Storage Currently most grid-scale energy storage in the US takes the form of pumped hydro power stations Battery storage capacities are typically measured in megawatt hours (mwh) when used in the grid, or kilowatt hours in cars. (A Tesla model S battery has a capacity of kwh depending on the options chosen.) Battery storage has historically in the region of $400-$500 per kwh 29

30 Storage Consider a wind turbine with capacity of 2mW, a typical turbine. Assume a capacity factor of 32.5%: then on average it produces 24x0.65mWh daily, 15.6mWh/day. At a capacity cost of $1,700/kW it will cost $3.4m. At $300/kWh a battery large enough to store one average day s output will cost $4.68m A 10 mw solar installation would cost $19.1m, produce on average 24x2.86 mwh daily At $300/kWh a battery to store this would cost $20.6m 30

31 Storage At the promised price of redox flow batteries, $150/kWh, the costs to store a day s output from a 2mW wind turbine or a 10mW solar farm are respectively $2.3m and $

32 Storage The US consumes 4 bn mwh/year, of which in our scenario two thirds would be from renewables. This means that on an average day it would consume 7.3x10 6 mwh of renewable energy At the optimistic cost of $150/kWh the capacity to store one day of renewable energy production would cost $1.095 trillion. At Elon Musk s retail of $350/kWh it costs $2.555 trillion. 32

33 33

34 S = storage C= total daily consumption 34

35 35

36 36

37 Distribution of total output of all US wind farms 37

38 Pose problem as statistical decision theory How much capacity do we need, and where, to be 95% (or 99%) certain of meeting load? 38

39 39

40 40

41 Y axis storage in number of hours of average load, X axis capacity as % of average load 41

42 Stock of stored energy Consumption of energy Output of renewable energy 42

43 43