Avoided Emissions from Renewables: Timing is Everything NREL Energy Analysis Seminar Series

1
Avoided Emissionsfrom RenewablesTiming is
EverythingNREL Energy Analysis Seminar Series
LFEE MIT Laboratory for Energy and the
Environment

• Stephen ConnorsMike Berlinski, Kate Martin, Mike
  Adams
• Analysis Group for Regional Electricity
  AlternativesLaboratory for Energy and the
  Environment Massachusetts Institute of
  Technology
• One Amherst St., Cambridge, MA 02139-4307, USA

2
Timing Isnt Everything

• But will be for this talk
• Avoided Emissions from Renewables is Really a
  Story About Multiple Resource Dynamics
• Solar and Wind Hydro and Biomass
• Electricity Demand Fossil (and other) Dispatch
• Energy Market Dynamics (both electric and fuel)
• Which Fossil Generation Displaced?
• Load Shape Following (LSF) Fossil
• Powerpool-Specific Hourly LSF Avoided Emission
  Rates
• Avoided Emissions from Solar PV (EPA)
• Offshore Wind in the Northeast (MA-RET/GE/DOE)
• Implications for Future Research (All in an
  hour!)
• Statistics vs. Operational Modes and
  Thresholds
• Technology Cost, Performance and Complexity

3
Fossil Unit Dispatch/Duty Cycle and Avoided
Emissions

• Real-life Data (Historical, Hourly)
• Fossil Unit Operations (EPA Clean Air Markets)
• Regional Power Pool Loads Prices (NERC
  Subregions)
• Actual or Simulated Renewable Generation (using
  hourly weather data - insolation/windspeed)

• Some interesting (e.g. challenging) dynamics and
  implications
• Renewable power is state specific
• Short-time duration solar, wind, others
• Medium long time scales hydro and biomass
• Fossil dispatch is also very state specific
• Available generation (maintenance, hydro
  nuclear availability)
• Level of Demand (capacity margin)
• Factors affecting dispatch order (fuel costs,
  reserve markets)
• How do they interact?

4
Fossil Unit Operation Modes

• Generation is generally thought of as
  baseload, intermediate, peaking, and/or
  spinning.
• In reality, unit dispatch is much more complex
• However, several Operation Modes have
  been observed by looking at hourly operations.

• Even these modes do not, in an of
  themselves, tell us which units respond to
  increases in renewable generation, or changes in
  demand
• These units we call Load Shape Following

5
How Do They Really Run?

• Fourteen Days in January 2002

6
How Do They Really Run?

• Fourteen Days in July 2002

7
How Do Renewables Run?

• Fourteen Days in January July

8
Example Peak Demand

• RepresentativeNortheastGenerators

9
Example Overnight/Off-Peak

• RepresentativeNortheastGenerators

10
A Year At a Glance
Total Load Profiles are normalized to peak
subregion demand in 2002
Fraction of 2002 Peak
365 Days of Year
eGrid Fossil Generation Profiles are normalized
to peak subregion eGrid generation in 2002.
24 Hours of Day
11
LSF Avoided Emissions Rates
Load Shape Following Emission Rate
Profiles represent emissions from 1MWh of load
shape following generation in each hour.
These are the emissions offset by 1MWh of
non-emitting generation applied in each hour of
the year. Units are kg/MWh in each hour.
SO2 Offset (kg/MWh)
12
Generation from Photovoltaics
PV Generation Profiles Show hourly PV generation
as a fraction of monitored capacity. Units are
kWh/kWcap in each hour.
kWh/kWcap
13
Avoided Emissions from PVs
Emissions Offset from PV Offsets are the product
of hourly PV generation as a fraction of
monitored capacity and the hourly load
shape following emission rates in the
subregion. Units are g/kWcap in each hour.
g/kWcap
14
Some Select Subregions

• The East
• New England (NEWE)
• Ohio Valley (ECOV)
• Texas (ERCT)

The West Pacific Northwest (NWPN) Southwest
(WSSW) California (CALI)
15
The East Operating Hours
(eGrid Only)
16
New Englandby the hour
New England(NEWE-2002)
365 days x 24 Hours Total Load/Elec. Demand
Fossil Generation (Normalized to 2002 Peak
Load) Fossil emissions rates (kg/MWh) are
for weighted Load Shape Following
generating for that hour.
17
Ohio Valleyby the hour
Load
Fossil Gen
SO2/MWh
NOx/MWh
CO2/MWh
Ohio Valley(ECOV-2002)
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
365 days x 24 Hours Total Load/Elec. Demand
Fossil Generation (Normalized to 2002 Peak
Load) Fossil emissions rates (kg/MWh) are
for weighted Load Shape Following
generating for that hour.
18
Texasby the hour
Load
Fossil Gen
SO2/MWh
NOx/MWh
CO2/MWh
Texas(ERCT-2002)
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
365 days x 24 Hours Total Load/Elec. Demand
Fossil Generation (Normalized to 2002 Peak
Load) Fossil emissions rates (kg/MWh) are
for weighted Load Shape Following
generating for that hour.
19
New EnglandPV by the hour
New England(NEWE-2002)
365 days x 24 Hours Total Load/Elec. Demand
PV Generation (Monitored Systems, normalized
to 1 kW of installed PV capacity) Avoided
LSF fossil emissions from PV generation
(kg per hour)
20
Ohio ValleyPV by the hour
Load
PV Gen
SO2 kg
NOx kg
CO2 kg
Ohio Valley(ECOV-2002)
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
365 days x 24 Hours Total Load/Elec. Demand
PV Generation (Monitored Systems, normalized
to 1 kW of installed PV capacity) Avoided
LSF fossil emissions from PV generation
(kg per hour)
21
TexasPV by the hour
Load
PV Gen
SO2 kg
NOx kg
CO2 kg
Texas(ERCT-2002)
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
365 days x 24 Hours Total Load/Elec. Demand
PV Generation (Monitored Systems, normalized
to 1 kW of installed PV capacity) Avoided
LSF fossil emissions from PV generation
(kg per hour)
22
The West Hours and Energy
(eGrid Only)
23
Pacific NWby the hour
Load
Fossil Gen
SO2/MWh
NOx/MWh
CO2/MWh
Pacific NW(NWPN-2002)
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
365 days x 24 Hours Total Load/Elec. Demand
Fossil Generation (Normalized to 2002 Peak
Load) Fossil emissions rates (kg/MWh) are
for weighted Load Shape Following
generating for that hour.
24
Southwestby the hour
Load
Fossil Gen
SO2/MWh
NOx/MWh
CO2/MWh
Southwest (WSSW-2002)
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
365 days x 24 Hours Total Load/Elec. Demand
Fossil Generation (Normalized to 2002 Peak
Load) Fossil emissions rates (kg/MWh) are
for weighted Load Shape Following
generating for that hour.
25
Californiaby the hour
Load
Fossil Gen
SO2/MWh
NOx/MWh
CO2/MWh
California(CALI-2002)
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
365 days x 24 Hours Total Load/Elec. Demand
Fossil Generation (Normalized to 2002 Peak
Load) Fossil emissions rates (kg/MWh) are
for weighted Load Shape Following
generating for that hour.
26
Across NERC Subregions

• PV Generation Avoided Emissions

LSF Load Shape Following // SOF Slice of
Fossil (eGrid)
27
Mid-Atlantic Best Time?
(MAAC-2002 Load Shape Following)
Load
Fossil Gen
LSF SO2
LSF NOx
LSF CO2
Mid-Atlantic (PJM) (MAAC-2002)
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
365 days x 24 Hours Total Load/Elec. Demand
Fossil Generation (Normalized to 2002 Peak
Load) Fossil emissions rates (kg/MWh) are
for weighted Load Shape Following
generating for that hour.
28
Implications/Opportunities

• What about other Renewables?
• Windpower (Onshore/Offshore)
• Hydropower (Integrated/Deferred)
• Sustainable Biomass
• What about Energy Conservation?
• Can We Target DSM for Emissions Reductions?
• Should Certain Appliances Be Pushed? (Heat
  Pumps over Air Conditioners?)
• What Role Distributed Generation?
• What Role TD Energy Storage?

29
Dynamics in Space and Time
Source Mass Renewable Energy Trust TrueWind
Solutions
30
Where When Is It Windy?

• Look Near Shore (Data from two NOAA Buoys)
• Near Hull Wind Turbine
• Near Cape Wind Site
• How Much?
• How Predictable?

Slightly Higher Winds Yield Much Higher kWh
Production!
(Source GE Wind)
31
Multi (2)-Year Performance

• Highly Variable in Production, Revenues and
  Avoided Emissions

32
Demand, Price Revenues 01
Source Data ISO-NE, NOAA
Mass Bay 01
Buzz Bay 01
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
(MWh)
(MWh)
33
Demand, Price Revenues 02
Source Data ISO-NE, NOAA
Load 2002
MCP 2002
Mass Bay 02
Buzz Bay 02
Mass Bay02
Buzz Bay 02
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
(GW)
(/MWh)
(MWh)
(MWh)
()
()
34
Looking Longer Further Out

• 20 NOAA stations selected for analysis starting
  with 6 select data sites (3 Near, 3 Far)

Hub Height _at_ 75m
(Calm seas)
35
Wind Energy by Hour Month
Buzzards Bay Nantucket
Total Generation by month andhour of day.
Far offshore
Near shore
Boston Hotel
Source Data NOAA ( Wind Speeds Scaled to 75m )
For One MW of Installed Wind Capacity(At 75m
using GE Wind 3.6 MW Power Curve)
36
Average Annual Windspeed _at_ 75m

• Average windspeeds vary considerably from year to
  year.

Source Data NOAA
Preliminary results Please do not cite without
permission MIT/UMass-Amherst Pilot Project
Offshore Wind Collaborative Sponsors Mass.
Renewable Energy Trust/GE-Wind/U.S. DOE
37
Annual Capacity Factors

• Differences in wind resources across data sites
  are magnified when we go from windspeeds to
  energy production.

Source Data NOAA, GE
Preliminary results Please do not cite without
permission MIT/UMass-Amherst Pilot Project
Offshore Wind Collaborative Sponsors Mass.
Renewable Energy Trust/GE-Wind/U.S. DOE
Power production at a 75m hub height using a GE
Wind 3.6 MW Power Curve (104m dia.)
38
Average Windspeeds _at_ 75m
Source Data NOAA
Preliminary results Please do not cite without
permission MIT/UMass-Amherst Pilot Project
Offshore Wind Collaborative Sponsors Mass.
Renewable Energy Trust/GE-Wind/U.S. DOE
39
Average Windspeeds _at_ 75m
A Closer Look
The three far offshore data sites and
Buzzards Bay
Source Data NOAA
Preliminary results Please do not cite without
permission MIT/UMass-Amherst Pilot Project
Offshore Wind Collaborative Sponsors Mass.
Renewable Energy Trust/GE-Wind/U.S. DOE
The Buzzards Bay data site is closer to shore
than Nantucket, though it is windier than
Nantucket for many (but not all) of the last 15
years - though this may be due to height
scaling issue.
40
Average Capacity Factors
The three far offshore data sites and
Buzzards Bay
Source Data NOAA, GE
Preliminary results Please do not cite without
permission MIT/UMass-Amherst Pilot Project
Offshore Wind Collaborative Sponsors Mass.
Renewable Energy Trust/GE-Wind/U.S. DOE
Power production at a 75m hub height using a GE
Wind 3.6 MW Power Curve (104m dia.)
41
Volatile Capacity Factors
Percent Change from Previous Years Capacity
Factor
Source Data NOAA, GE
Preliminary results Please do not cite without
permission MIT/UMass-Amherst Pilot Project
Offshore Wind Collaborative Sponsors Mass.
Renewable Energy Trust/GE-Wind/U.S. DOE
Power production at a 75m hub height using a GE
Wind 3.6 MW Power Curve (104m dia.)
42
Variability Matters!

• (Renewable) Resource Variability
• Wind and Sun (Magnitude and Timing)
• Rainfall (Hydropower, Biomass)
• Fuel Markets
• Fuel Prices/Price Differentials (esp. Natural
  Gas)
• Infrastructure Investments (Pipelines/Storage/LNG)
  
• Conventional Generation
• Nuclear Availability, Hydro Potential
• Power Market Structure (Capacity Markets, Bid
  Rules)
• Power Grid Operations (Reliability/Contingency
  Practices)
• Energy Demands
• Demand Growth Relative to Supply Growth
• Heating Degree Days/Cooling Degree Days

43
Scenario Resolution Models

• Scenario Modeling Lots of Flavors
• Bottom-Up, Top-Down
• Optimization vs. Simulation
• System-Dynamics, Agent-Based
• The Finer the Resolution, the More Prescriptive
  Results Will Be
• Computers are only getting more powerful

44
Using History in Modeling..

• How to Best Use High Resolution Data in Future
  Scenarios/Pathways Analysis?

• Multiple/Coupled Renewable Resources

45
Using History in Modeling..

• How to Best Use High Resolution Data in Future
  Scenarios/Pathways Analysis

• Multiple/Coupled Renewable Resources

46
Using History in Modeling..

• How to Best Use High Resolution Data in Future
  Scenarios/Pathways Analysis

• Multiple/Coupled Renewable Resources

47
Using History in Modeling..

• How to Best Use High Resolution Data in Future
  Scenarios/Pathways Analysis

• Multiple/Coupled Renewable Resources

48
Using History in Modeling..

• How to Best Use High Resolution Data in Future
  Scenarios/Pathways Analysis

• Multiple/Coupled Renewable Resources

49
Using History in Modeling..

• How to Best Use High Resolution Data in Future
  Scenarios/Pathways Analysis

• Multiple/Coupled Renewable Resources

• Supply, Demand and Markets Prices

50
Using History in Modeling..

• How to Best Use High Resolution Data in Future
  Scenarios/Pathways Analysis

• Multiple/Coupled Renewable Resources

• Supply, Demand and Markets Prices

51
Using History in Modeling..

• How to Best Use High Resolution Data in Future
  Scenarios/Pathways Analysis

• Multiple/Coupled Renewable Resources

• Generation and Fuel Market Dynamics

52
Operating Modes and Resource Portfolios

• Renewable Resource Variability
• Wind and Sun (Magnitude and Timing)
• Rainfall (Hydropower, Biomass)
• Fuel Markets
• Fuel Prices/Price Differentials (esp. Natural
  Gas)
• Infrastructure Investments (Pipelines/Storage/LNG)
  
• Conventional Generation
• Nuclear Availability, Hydro Potential
• Power Market Structure (Capacity Markets, Bid
  Rules)
• Power Grid Operations (Reliability/Contingency
  Practices)
• Energy Demands
• Demand Growth Relative to Supply Growth
• Heating Degree Days/Cooling Degree Days

53
(More) Outstanding Questions

• Daily, Seasonal, and Inter-Annual Dynamics
  Matter!
• Are there Climate Change implications?
• On/Off Generation Better than Up/Down
  Generation?
• Are a bank of diesels better than one
  largefossil steam unit running many hours in
  standby?
• Environmental Benefits from Power Grid
  Modernization?
• Reducing transmission bottlenecks may have
  (operational) environmental benefits.
• How Beneficial Might Storage Be?
• Can you come out ahead by reducing the number of
  hours in Standby generation mode?
• Too Much of a Good Thing?
• What about greater than marginal Renewable
  kWhs?
• What if more fossil gets pushed into Standby
  mode?

54
Have We Had It Too Easy?

• We need to develop tools and information for a
  more distributed, less dispatchable energy
  sector.