51
1 Managing the Decarbonized Grid Austin Electricity Conference 2017 April 21, 2017 Ross Baldick, Department of Electrical and Computer Engineering, University of Texas at Austin
1
Managing the
Decarbonized Grid
Austin Electricity Conference 2017
April 21, 2017
Ross Baldick,
Department of Electrical and Computer Engineering,
University of Texas at Austin
2
Outline
Temporal issues:
Load-duration and net load-duration,
Decarbonizing net load—hourly and longer
time-scales,
Shorter-timescales.
Locational issues:
Geographical aggregation,
Transmission,
Distribution system,
Distributed resources.
3
Outline
The panelists:
Gene Preston, Consultant,
Paul Wattles, ERCOT,
Brian Johnson, NREL,
Bill Muston, ONCOR.
Temporal issues. Load varies over time due to weather and
human activity.
Renewable production varies with weather.
Supply must match demand continuously.
Load-duration curve shows the implications
of matching supply to demand at hourly and
longer time-scales:
Re-orders the load levels from highest to lowest.
Net load-duration curve considers load
minus renewables. 4
Load in ERCOT in 2016.
5
0
10000
20000
30000
40000
50000
60000
70000
80000
1/1/20160:00
2/1/20160:00
3/1/20160:00
4/1/20160:00
5/1/20160:00
6/1/20160:00
7/1/20160:00
8/1/20160:00
9/1/20160:00
10/1/20160:00
11/1/20160:00
12/1/20160:00
1/1/20170:00
Source of data: ERCOT
Load-duration curve.
Re-order chronological data
from highest to lowest.
Low operating cost, or
“baseload,” generation
covers minimum load at
lowest overall cost.
Low capital cost,
“intermediate” and “peaker,”
generation required in
addition for higher load
levels.
Load, MW
Duration
Peak
Min
Baseload
Intermediate
Peaker
Decarbonizing with
renewables.
ERCOT and Texas is at forefront of
integrating wind.
Solar is also growing.
Also significant growth in renewables in
various states nationwide and in various
countries worldwide.
7
Wind in ERCOT in 2016.
8
0
1,000
2,000
3,000
4,000
5,000
6,000
7,000
8,000
9,000
10,000
11,000
12,000
13,000
14,000
15,000
16,000
17,000
Jan-01 Feb-01 Mar-01 Apr-01 May-01 Jun-01 Jul-01 Aug-01 Sep-01 Oct-01 Nov-01 Dec-01
MW
Source of figure and data: ERCOT
9
Load, wind, net load Feb 2016.
0
10000
20000
30000
40000
50000
60000
Feb-0
1…
Feb-0
1…
Feb-0
2…
Feb-0
2…
Feb-0
3…
Feb-0
3…
Feb-0
4…
Feb-0
5…
Feb-0
5…
Feb-0
6…
Feb-0
6…
Feb-0
7…
Feb-0
8…
Feb-0
8…
Feb-0
9…
Feb-0
9…
Feb-1
0…
Feb-1
0…
Feb-1
1…
Feb-1
2…
Feb-1
2…
Feb-1
3…
Feb-1
3…
Feb-1
4…
Feb-1
5…
Feb-1
5…
Feb-1
6…
Feb-1
6…
Feb-1
7…
Feb-1
7…
Feb-1
8…
Feb-1
9…
Feb-1
9…
Feb-2
0…
Feb-2
0…
Feb-2
1…
Feb-2
2…
Feb-2
2…
Feb-2
3…
Feb-2
3…
Feb-2
4…
Feb-2
4…
Feb-2
5…
Feb-2
6…
Feb-2
6…
Feb-2
7…
Feb-2
7…
Feb-2
8…
Feb-2
9…
Feb-2
9…
Load
Net load
Wind
Min of
net load
Max of net load Max of load
Min of
load
Source of data: ERCOT
Load-duration curve.
Load-duration without wind. Net Load-duration with wind.
Net load = load minus wind.
Load, MW Net load, MW
Duration Duration
Implications of load-duration
and net load-duration curves. Peak of net load nearly as large as load peak:
So meeting peak net load requires nearly as
much non-renewable production as without wind,
Typical for onshore wind; however, correlation of
coastal wind and solar with peak is stronger.
Minimum of net load is lower than minimum
of load:
Less baseload capacity needed with higher wind
penetration,
Min net load less affected by solar. 11
Decarbonizing net load. Various options for supplying remaining
net load:
Nuclear,
Fossil with carbon capture and storage
(CCS),
Bulk storage (together with additional
renewable, nuclear, or fossil with CCS),
Demand-side management (DSM):
Displacing load temporally,
Storage at end-user,
End-use efficiency and load curtailment. 12
Nuclear. High capital cost and low
operating cost,
Most economical to supply
energy continuously
throughout year except
when under maintenance.
US nuclear fleet not
designed for dispatchability:
French fleet seems to be
more flexible.
However, increasing
nuclear capacity above min
of net load will increase
average costs of nuclear. 13
Net load, MW
Duration
Nuclear
Fossil with CCS.
CCS involves higher
capital and operating
cost (energy) than
non-CCS.
Capture and storage
potentially use off-
peak energy.
14
Net load, MW
Duration
Fossil with CCS
Storage and DSM with
temporal load displacement. Energy discharged or load
displaced must be
provided by other
resources:
Round-trip losses and
bounce back implications,
May not be possible to
utilize lowest cost
generation resources
because of limitations on
seasonal storage (Jenkins
and Thernstrom, 2017). 15
Net load, MW
Duration
Stored or
displaced
Discharged or
displaced
End-use efficiency and
curtailment.
Shift to more efficient
appliances, including
air conditioning, can
mitigate energy
consumption,
including at peak.
Price or quantity
based curtailment
typically targets peak
explicitly.
16
Net load, MW
Duration
End-use efficiency
and curtailment
Decarbonizing net load.
Fractions of each
technology depend on
economics and
adoption going
forward:
Automated appliances,
Decreasing storage
costs,
Nuclear and CCS
costs.
17
Net load, MW
Duration
Stored Fossil with CCS
Nuclear
Discharged or
displaced
End-use efficiency
and curtailment
Temporal issues over
shorter time-scales.
Time-scales:
Minutes to hours: ramping of generation to
follow ramps in net load,
Tens of seconds to minutes: response to
random fluctuations of net load and net load
forecast errors,
Sub ten second: electromechanical dynamics.
18
Minute to hour
variations.
Thermal generation fleet has historically
matched diurnal load variation:
Capability to ramp from one generating level
to another sufficient to match ramps in load.
Increase in renewables is increasing the
net load ramping:
Can require additional capacity to be on-line
to provide additional ramping capacity,
Increases operating costs. 19
20
Longer and faster ramps
in net load compared to load.
0
10000
20000
30000
40000
50000
60000
Feb-0
1…
Feb-0
1…
Feb-0
2…
Feb-0
2…
Feb-0
3…
Feb-0
3…
Feb-0
4…
Feb-0
5…
Feb-0
5…
Feb-0
6…
Feb-0
6…
Feb-0
7…
Feb-0
8…
Feb-0
8…
Feb-0
9…
Feb-0
9…
Feb-1
0…
Feb-1
0…
Feb-1
1…
Feb-1
2…
Feb-1
2…
Feb-1
3…
Feb-1
3…
Feb-1
4…
Feb-1
5…
Feb-1
5…
Feb-1
6…
Feb-1
6…
Feb-1
7…
Feb-1
7…
Feb-1
8…
Feb-1
9…
Feb-1
9…
Feb-2
0…
Feb-2
0…
Feb-2
1…
Feb-2
2…
Feb-2
2…
Feb-2
3…
Feb-2
3…
Feb-2
4…
Feb-2
4…
Feb-2
5…
Feb-2
6…
Feb-2
6…
Feb-2
7…
Feb-2
7…
Feb-2
8…
Feb-2
9…
Feb-2
9…
Load
Net load
Wind
Tens of seconds to
minutes.
Aggregating renewables over large
geographical scales:
Can reduce relative size of renewable
fluctuations in supply.
West Texas is suitable scale.
See, for example, Lee and Baldick, 2016.
Voltage control remains local:
Updates to IEEE 1547 Standard may be
helpful. 21
Sub 10 seconds. Current systems use frequency as
feedback signal to measure
supply/demand imbalance.
Inertia of machines electromechanically
coupled to system slows the change in
frequency.
Increasing renewables implies less
coupled inertia.
Could synthesize inertia.
Could reduce reliance on inertia to
stabilize frequency.
22
Locational issues.
Geographical aggregation,
Transmission,
Distribution system,
Distributed resources.
23
Geographical
aggregation.
Already touched on geographical
aggregation being effective for averaging
out relatively fast variations of renewable
fluctuations.
Longer term fluctuations require
aggregation at potentially vast scales:
Globally for solar.
24
Transmission.
Large-scale wind and solar resources can
be far from load centers.
Large-scale aggregation requires
connections over long distances.
Necessitates increased transmission.
Steady-state and transient stability issues
occur with low inertia at end of long
system.
25
Distribution system.
Essential link for access between
distributed resources and wider market.
Various distribution system issues add
complexity to integration of distributed
resources:
System protection,
Voltage management,
Safety considerations, including anti-islanding
under faulted conditions.
26
Distributed resources.
Recent large growth rate for distributed
solar:
Distribution wires ambivalent about direction
of flow, but
Added complexity of distribution system
issues.
IEEE Standard 1547 modifications may
help.
27
Gene Preston http://egpreston.com
CEO of Transmission Adequacy
Consulting.
Performs Solar & Wind Transmission
Studies and Loss of Load Probability
Studies.
Holds Ph.D. in electrical and computer
engineering from The University of Texas
at Austin.
28
29
http://egpreston.com/cases.htm http://egpreston.com/case6a.txt
30
New 345 kV Transmission Lines Are Needed
Red=existing lines
Green=new 345 kV lines
12 345 kV lines to
SPP’s Great Plains
10 345 kV lines to
Western Deserts
Many New 345 kV
Lines to West Texas
20 345 kV lines to
Serve Coastal Wind
32 GW West Tex Wind
12 GW SPP Plains Wind
24 GW Coastal Wind
44 GW Cen Tx Solar
10 GW Western Solar
22 GW West Tx Solar
31
Without Gas Backups Load Areas Are Unreliable
44 GW Cen Tx Solar
32 GW Remote Solar
68 GW Remote Wind
Green areas can retire their internal
backup gas generators only when the
the transmission system is strong or
there is long term storage in the area.
32
Nuclear Is Also An Option
Eliminates much of the transmission and storage
Long term fuel supply using IFR waste burner
Ultra fast load follower using off the shelf parts
Thorium fuel designs are proliferation resistant
New designs operate at low 1 ATM pressure
Spent fuel we currently have is worth trillions
Nuclear plants give jobs and local tax revenue
Nuclear can be made both safe and low cost
Don’t let fear rule; Trust engineering solutions
Paul Wattles Senior Analyst, Market Design and
Development, at ERCOT.
Market Design team leads initiatives to
enhance long-term efficiency of
ERCOT’s markets.
With ERCOT since May 2004, with
experience in market design, advanced
metering, distributed energy, demand
response, and governmental affairs.
Graduate of the University of Arizona in
Tucson. 33
PUBLIC
Evolving resource mix
• Ancillary Services are market-procured reliability services (operating reserves)
• Current AS framework reflects grid reality of late ‘90s, not the grid of the future with high renewable penetration
– Designed around capabilities of steam boiler units
– These requirements are barriers to entry for new resource types that could efficiently meet the technical needs of the system
• The future grid: – Will require changes to both the need for AS and the capabilities of resources
providing AS
– Will be more efficient with an AS framework that attracts a more flexible resource mix
Austin Electricity Conference 2017 34
Gas-Steam
50%
Coal
25%
Gas-CT/CC
5%
Nuclear
8%
Cogen
11%
Other
0.9%
Renewables
0.008%
Gas-CT/CC
36.5%
Gas-Steam
14.8%
Coal
22.4%
Renew-
ables
13.8%
Other
1.4% Nuclear
5.7%
Cogen
5.3%
Late 1990s 2015
PUBLIC
Wind Generation Capacity
35
The data presented here is based upon the latest registration data provided to ERCOT by the resource owners and can change without notice. Any capacity changes will be reflected in current and subsequent
years' totals. Scheduling delays will also be reflected in the planned projects as that information is received. This chart reflects planned units in the calendar year of submission rather than installations by peak of
year shown.
Financial security posted for funding interconnection facilities does not include CREZ security deposits, which are refunded to the Interconnecting Entity when an IA is signed.
18,589 18,589 18,589 18,589
4,2915,383 5,664 5,664
733
4,2454,245 4,403
116816 977 1,173 1,385
1,854
2,875
4,785
8,005
8,9169,400 9,604
10,40711,065
12,470
15,764
17,604
23,613
28,247 28,528 28,686
0 MW
5,000 MW
10,000 MW
15,000 MW
20,000 MW
25,000 MW
30,000 MW
35,000 MW
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020
Cumulative MW Installed IA Signed-Financial Security Posted IA Signed-No Financial Security
Notes:- The data presented here is based upon the latest information provided to ERCOT by resource owners and developers and can change without notice.- Installed capacities for the current year account for changes reported by the facility owners during the reporting month, and will be reflected in subsequent years' totals.- Installed capacities include only wind facilities that have registered with ERCOT (Those larger than one megawatt and supply power to the ERCOT system.)- This chart reports annual planned units with projected Commercial Operations Dates throughout the calendar year. In contrast, ERCOT's Capacity, Demand and Reserves (CDR) report shows planned capacity projected to be commercially available on or before the start of the Summer and Winter Peak Load seasons.
- Financial security posted for funding interconnection facilities does not include CREZ security deposits, which are refunded to the Interconnecting Entity
ERCOT Wind Additions by Year (as of April 1, 2017)
• Texas leads the U.S.A. in wind capacity. If Texas were a
separate country, it would rank 6th globally in wind capacity
• Wind Generation record: 16,141 MW (March 31, 2017)
• Wind Penetration record: 50 percent (March 23, 2017)
Austin Electricity Conference 2017
PUBLIC
Utility Scale Solar Generation Capacity
36
The data presented here is based upon the latest registration data provided to ERCOT by the resource owners and can change without notice. Any capacity changes will be reflected in current and subsequent
years' totals. Scheduling delays will also be reflected in the planned projects as that information is received. This chart reflects planned units in the calendar year of submission rather than installations by peak
of year shown.
420
Austin Electricity Conference 2017
PUBLIC
Proposed Future Ancillary Services
Austin Electricity Conference 2017 37
Regulation Up
Fast-Responding Regulation Up
Current Proposed
Fast Frequency Response 1
Primary Frequency Response
Contingency Reserves 1
Synchronous Inertial Response
Supplemental Reserves 1
Mostly unchanged
59.8 Hz, Limited duration
59.7 Hz, Longer duration Fast Frequency Response 2
Contingency Reserves 2
SCED-dispatched
Manually dispatched
Supplemental Reserves 2
SCED-dispatched
Manually dispatched
Deferred development
Non-Spin
Responsive
Regulation Down
Fast-Responding Regulation Down
Regulation Up
Regulation Down
Fast-Responding Regulation Up
Fast-Responding Regulation Down
PUBLIC
Substitutability = >liquidity
Austin Electricity Conference 2017 38
Regulation Up
Fast-Responding Regulation Up
Current Proposed
Fast Frequency Response 1
Primary Frequency Response
Contingency Reserves 1
Supplemental Reserves 1
Mostly unchanged
59.8 Hz, Limited duration
59.7 Hz, Longer duration Fast Frequency Response 2
Contingency Reserves 2
SCED-dispatched
Manually dispatched
Supplemental Reserves 2
SCED-dispatched
Manually dispatched Non-Spin
Responsive
Regulation Down
Fast-Responding Regulation Down
Regulation Up
Regulation Down
Fast-Responding Regulation Up
Fast-Responding Regulation Down
Brian Johnson
Electrical Engineer with the National
Renewable Energy Laboratory.
Conducts research in power electronics,
power systems, renewable energy
systems, and control systems.
Holds M.S. and Ph.D. in electrical and
computer engineering from the University
of Illinois at Urbana-Champaign.
39
% V
ari
ab
le R
en
ew
ab
le E
ne
rgy
(of
annual energ
y)
System Size (GW)
80
5
23
Alaska
n
Village
Ireland Cont. USA
Actual Operating System
35
Maui
14 CA*
Relatively
Easy
Much
Harder
WWSIS
CA 50%
Lanai Modeled System
ERGIS
REF
54
DOE 2050 Goals
35% Wind (404 GW)
19% PV (632 GW)
Deep
Decarbonizatio
n 1400 GW wind
900 GW Solar
78
* Part of a larger synchronous AC power system
Germany*
Transforming the Grid at a Scale That Matters
Extremel
y Difficult
12 Texa
s
29
The New Interface
41
The Core Component
Synchronous machines form the basis for all system
operations & control
The impact of decreased inertia:
underfrequency tripping
limit
The Need for Next-generation Inverter
Controls
Grid-following units use a “phase-
locked loop” to synchronize to an
existing grid and inject a controlled
current
Grid-forming inverters can establish
a grid and synchronize to other
inverters + machines without
communication
Bill Muston
Manager of R&D at Oncor, a regulated
electric utility in Texas, and in Oncor’s
corporate strategy & technology group
Focus on role of energy storage in utility
distribution systems, future growth of
distributed resources, and adaptations to
accommodate smooth grid operation with
all the above.
Holds B.S. in Electrical Engineering and
M.S. in Engineering from UT, Austin. 44
The old grid
45
• Distribution feeders deliver energy from substations – predictable one-way flow, maintain stable voltage & power quality
• Any interconnected distributed generation is mainly related to Diesel gensets for customer emergency
• Energy flows from generators to
transmission to substations to
distribution to customers
• Customer loads vary predictably
and smoothly – by hour, by day of
week, by season
• Feeder management and control
based on customer load profiles
• Demand-side provides a degree of
support
The future grid
46
•Distribution feeders still deliver energy from substations, maintain stable voltage & power quality
•Accommodate distributed generation
Today in Oncor Over 3200 feeders total
Over half have an inter-
connected distributed generator,
yet most are small rooftop solar
with limited penetration
• Variations from distributed solar
• Intra-hour, intra-15-minute
• Days not consistent
• Impact also varies by the purpose the
solar serves
• Customer solar simply offsetting grid
energy use
• Customer with solar plus storage to
firm solar and to supply services
upstream to ISO or sell into market
• 3rd-party solar selling into the market
or providing auxiliary services
• 3rd-party, non-renewable sources?
• Growth of electric vehicle charging?
Technical factors for incorporating DER
47
Interconnection Classifications
Interconnection study detail and requirements vary by type of interconnection
•Certified systems – inverter-based systems certified to IEEE 1547 or UL 1741 standards
•Paralleling with the grid, >100 msec
•Paralleling with the grid, < 100 msec
•Small induction generators & non-certified generators
•Downtown networks
• Voltage management –
electromechanical devices
• System protection coordination
• Trip transfer requirements
• Feeders that may be reconfigured
thru distribution automation
• Imposed overvoltage on
transmission grid during fault
• Bulk power grid considerations
with DER
• Emergency curtailment with
rotating outages
• Black start and adding load
• NERC
• Reverse flow through transformers
• Downtown urban networks – no
reverse flow
Solar output at single site – how does more solar
geographically diverse along a feeder affect voltage
management?
48
Beneficial for integration of DER
• Grid edge sensing and control
• Advanced meter systems – energy, demand, voltage
measurements, outage management
• Distribution automation devices with sensing &
autonomous operation
• Communications – high-speed & dedicated
• Advanced inverters – IEEE 1547 update
• Energy storage & microgrids
49
References
Jesse D. Jenkins and Samuel Thernstrom,
“Deep decarbonization of the electric
power sector: Insights from recent
literature,” Energy Information Reform
Project, Available from
innovationreform.org, March 2017.
Duehee Lee and Ross Baldick, “Wind
variability and impact on markets,” Invited
presentation at Wind Farms, Dallas, TX,
May 23-25, 2016. 50
References
ERCOT 2017, “Market Information”
Available from www.ercot.com, Accessed
April 10, 2017.
51
