NC Energy Storage Study
Stakeholder MeetingOctober 2, 2018
1
Legislative Language
2
Faculty and Staff Team Members• Joe DeCarolis, Civil, Construction & Environmental Engineering• Jeremiah Johnson, Civil, Construction & Environmental Eng.• Christopher Galik, Public Administration• Harrison Fell, Agricultural and Resource Economics• Ning Lu, Electrical and Computer Engineering• David Lubkeman, Electrical and Computer Engineering• Wenyuan Tang, Electrical and Computer Engineering• Ken Dulaney, FREEDM Center• Anderson Rodrigo de Queiroz, NC Central• Steve Kalland, NC Cleantech Center• Autumn Proudlove, NC Cleantech Center• Isaac Panzarella, NC Cleantech Center
3
Student Team Members
• Shuchi Liu, Electrical and Computer Engineering• Yao Meng, Electrical and Computer Engineering• Asmaa Alrushoud, Electrical and Computer Engineering• David Mulcahy, Electrical and Computer Engineering• Catie McEntee, Electrical and Computer Engineering• Zachary Small, Agricultural and Resource Economics• Danny Sodano, Civil, Construction, and Environmental
Engineering• Dustin Soutendijk, Civil, Construction, and Environmental
Engineering• Lisha Sun, Electrical and Computer Engineering• Chris Gambino, Public Administration
4
Hypothetical Case Studies• We don’t have a single comprehensive model to examine all
relevant grid services• Instead, we’re conducting a series of hypothetical case
studies, organized by the category of service that storage can provide
• Our approach is use a suite of different models and datasets, each adapted to address a given case study
• Under each hypothetical case study, a set of relevant scenarios are considered
• Scenario assumptions include:–Presence of renewable investment tax credit–Storage cost and performance (present and 2030)–Future scenarios driving 2030 grid mix (bulk energy storage)
5
Definition of storage
Harmonized scenario
assumptions
Cost assessment
Identify applications and services
Identify technologies
Hypothetical Case Studies:Use models and data to
assess benefit
Deployment happening?Barriers? No
How much?
Yes
Break-even cost?
Other Barriers?
[Benefit – Cost]?
+
−
Policy options to increase value to NC consumers 6
Project Status
• Today we will be presenting preliminary results, with a focus on lithium-ion battery storage
• We have not completed all planned case studies• Additional time is required to synthesize results into a
broader set of insights• Draft report and benefit-cost spreadsheet for public
review on November 1, 2018• Final report on December 1, 2018
Project website: https://energy.ncsu.edu/storage/7
Storage Technologies Considered
Mechanical • Flywheels• Pumped storage• Compressed Air
Electrochemical • Lithium-ion batteries• Lead-acid batteries• High T sodium batteries• Flow batteries
Chemical• H2 electrolysis + storage
+ fuel cells
Thermal• Chilled water• Ice storage• Phase change materials• Water heaters
Electrical • Supercapacitors• Superconducting
magnetic energy storage
8
Cost SpreadsheetWe are developing a spreadsheet to assess storage costs. It includes:• A map showing compatibility between storage
technologies and grid applications• A separate worksheet for each storage technology• Parameters include capital costs, O&M costs, roundtrip
efficiency, lifetime, and degradation (if applicable)• Embedded references to data sources from which cost
and performance data derived• Will be made publicly available once completed
9
Li-ion Battery Cost• Considered the following scales residential (10 kW), commercial
(100kW), utility (1 MW)• And the following durations:
– 0.5 hr, 2 hr, and 4 hr for utility-scale batteries– 2 hr for residential and commercial batteries
Source Current Cost ($/kWhs) Projected 2030 Cost ($/kWhs)
Schmidt et al. (2017) 1250 450
McKinsey & Co. (2018) 587 170
GTM Research (2018) 525
NYSERDA (2018) 401 (2019 estimate) 180
EIA Market Trends (2018) 399
Lazard (2017) 291 (2018 estimate excludes EPC)
10
Assume: 525 $/kWhs; and 200 $/kWhs in 2030 (~60% reduction)
Revenue Requirements
11
Topic and Speaker Lineup
• End-User Services – Isaac Panzarella• Distribution Services – David Lubkeman• Transmission Services – David Mulcahy• Frequency Regulation – Ning Lu• Bulk Energy Time Shifting / Peak Capacity Deferral –
Anderson de Queiroz• Solar Clipping – Jeremiah Johnson• Baseline Policy Review and Survey Results –
Christopher Galik
12
What to Expect in the Report• Detailed benefit-cost analysis for each service and scenario• A sense for the market size of each service (e.g., high,
medium, low)• An examination of opportunities for value stacking• A policy review that identifies potential barriers and solutions
to the deployment of energy storage
13
• Several applications show that Li-ion batteries are near parity with conventional alternatives; value stacking would likely produce positive net benefits.
• If 2030 projected battery costs realized, case for batteries is more compelling
Preliminary Insights
Process for Stakeholder Q&A• At the end of each presentation, you will have the opportunity to
write down questions or comments pertaining to the presentation• Scribes at each table will transcribe the feedback into a master
Google Sheet• Presenters will address a limited number of questions (time
permitting) as they appear on the screen• Complete compiled feedback will be reviewed and taken into
consideration as we work on the draft report
14
NC Energy Storage StudyEnd-User Services
Isaac PanzarellaNC Clean Energy Technology Center
Stakeholder MeetingOctober 2, 2018
End-User ServicesTime-of-Use/Energy Management • Shifting advantage of time-of-use Demand Charge Management• reduce monthly demand peak Backup Power• emergency backup power in the event of outagesDistributed Energy Resource Management• management of energy consumption and injection into gridPower Quality Management• Mitigate voltage fluctuation, voltage drop, and frequency
Demand Charge + DER Management• Demand Charge Management - Shifting electricity
consumption to reduce the customer’s highest peak consumption from the grid can reduce demand charges ($/kW). These are especially significant for industrial and commercial customers. This application can often be coupled with TOU rate reduction ($/kWh).
• Renewable Energy Management - For residential, commercial, industrial, and even utility distributed energy resource (DER) sites, ESS can enable the management of energy consumption and injection into grid. Depending on the objective of ESS control, DER management may respond to economic incentives for end-users or manage variability of power injection to avoid reliability issues.
RESIDENTIAL
COMMERCIAL
INDUSTRIAL
ESS for Demand Charge + DER Management
ESS for Demand Charge + DER Management• Applicable technologies
– Battery storage– Battery storage with solar PV– Thermal storage
• Scenarios presented– Large commercial/industrial (~ 1,500 kW monthly peak)– Coincident peak tariff
• Modeling tools– System Adviser from NREL– Excel based spreadsheet
Customer Peak Demand Requirement (kW)
Coincident Peak Demand Charge ($/kW/month)
Customer Peak Demand Charge ($/kW/month)
Energy Charge ($/kWh)
City of Wilson (Schedule FR-MGS-2)
>35 & <500
$23.39 $5.00 a $0.0650
City of Wilson (Schedule FR-1-1)
>500 & <10,000
$20.50 $4.10 a $0.0570
Fayetteville PWC (Pilot CP Rate) b
>1000 $20.11 $2.00 $0.04098
Greenville Utilities (Schedule MGS-CP)
>35 & <750
$17.00 $15.61 $0.03027
Greenville Utilities (Schedule LGS-CP)
>750 $22.20 $13.13 $0.02524
ESS for DCM+DER – Coincident Peak Rates
• Sample of coincident peak demand pricing from NC utilities
800
900
1,000
1,100
1,200
1,300
1,400
1,500
1,600
1,700
1,800
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Monthly Coincident Peak Demand (kW) with Solar PV & ESS
Base Solar PV 500 kW
800
900
1,000
1,100
1,200
1,300
1,400
1,500
1,600
1,700
1,800
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Monthly Coincident Peak Demand (kW) with Solar PV or ESS
Base Solar PV 500 kW 500kW 2-hr ESS 500kW 4-hr ESS
800
900
1,000
1,100
1,200
1,300
1,400
1,500
1,600
1,700
1,800
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Monthly Coincident Peak Demand (kW) with Solar PV & ESS
Base Solar PV 500 kWSolar PV 500kW + 500kW 2-hr ESS Solar PV 500kW + 500kW 4-hr ESS500kW 2-hr ESS 500kW 4-hr ESS
Solar PV 500 kW
Solar PV 500kW + 500kW 2-hr ESS
Solar PV 500kW + 500kW 4-hr ESS
500kW 2-hr ESS
500kW 4-hr ESS
Coincident Peak Demand Savings & Payback with solar PV & ESS- Wilson Energy CP FR-1-1 tariff
Net Capital Cost $889,000 $1,396,500 $1,624,100 $725,000 $1,050,000
Annual Utility Savings $71,011 $173,079 $192,934 $102,632 $121,803
Benefit $/kW/yr $204 $244 $205 $244
Simple Payback 12.5 8.1 8.4 7.1 8.6
Time of Use Rate Savings Savings & Payback with solar PV & ESS- Duke Energy Carolinas Large General Service TOU rate LGS-TOU-50
Net Capital Cost $889,000 $1,396,500 $1,624,100 $725,000 $1,050,000
Annual Utility Savings $53,309 $145,158 $163,061 $92,290 $109,632
Benefit $/kW/yr $184 $220 $185 $219
Simple Payback 16.7 9.6 10.0 7.9 9.6
Still to Come and Future Research beyond this study
This Study• Residential and commercial models• Power quality management benefits of storage• Case studies on customer-sited storage projects in NC
Future Research• Benefits of storage for energy resilience in critical
infrastructure• Incentives for energy storage• Aggregation of customer sited storage for utility dispatch
NC Energy Storage Study
Distribution ServicesDr. David Lubkeman
Analysis Contributors: Lisha Sun, Shuchi Liu
Stakeholder MeetingOctober 2, 2018
Analysis Scope
• Specific Applications Being Studied:–Distribution capacity deferral/peak shaving–Reliability enhancement–Voltage support and control (relates to distributed generation integration)
CB S
Substation
Battery energy storage
CB
VoltageRegulator
Substation
Voltage Controlled
Bus
S
Substation
Voltage Controlled
Bus
Feeder 1
Feeder 2
Feeder 4
Feeder 3.
Opt 1:Addition of Extra Capacity
Opt 2: Install Energy Storage
Capacity Deferral/Peak Shaving• Small capacity/duration
increase needed• Use energy storage instead of
significant substation/circuit upgrade
• Energy storage can also be used for monthly peak shaving
0
5,000
10,000
15,000
20,000
25,000
00:00 04:00 08:00 12:00 16:00 20:00 00:00
kW
Summer Peak Day
No ES With ES
Capacity Limit: 20 MVA
Capacity Deferral/Peak Shaving Approach
Setup Feeders
Hourly load shapes for each feeder
Load growth every year
ES Sizing and Control
Solve for added ES kW needed at each year
Calculate optimal monthly peak shaving settings for each year
Estimate Cost & Benefits
ES annual cost
Deferral benefits
Additional peak shaving benefits
Sensitivity analysis
Data and Assumptions
[1] Anderson Hoke, Randomized Hourly Load Data for use with Taxonomy Distribution Feeders, https://catalog.data.gov/dataset/randomized-hourly-load-data-for-use-with-taxonomy-distribution-feeders[2] Jim Eyer, Garth Corey, SAND2010-0815, Energy Storage for the Electricity Grid: Benefits and Market Potential Assessment Guide[3] GTM, NYSERDA and Mckinsey[4] Duke Energy Rate, Large General Services, Schedule LGS-50
Feeder Simulation
Annual Load Curve Hourly from NREL [1]
Capacity 20 MVA
Number of Feeders 4
Load Growth Rate 1%
Coincident Peak (yr.1-10) 20.3 MW – 22.2 MW
Energy Storage Added in 500kW block
Energy Storage (Li-ion 4 hr.) [3]
Cost ($/kWh) 525 (200 for 2030)
O&M Cost 10 $/kW/yr.
Life 10 Years
Round trip efficiency 85%
Self discharge 5% (24 hr.)
Loan Rate 6%
Electricity Rate [4]
Demand Charge $ 11/kW
Energy Cost for ES Operation Losses $ 55/MWh
Deferral Benefits
T&D installed Cost 150 $/kVA [2]
Upgrade Option Add Substation Capacity in 10 MVA
Fixed Charge Rate 11%
Preliminary Results
Cost Benefit Analysis Unit $/kW-yr.
Benefit 130 - Deferral 66
- Deferral Peak Shaving 4
- Added Peak Shaving 60
Cost (2018) 170 Net (2018) -39Cost (2030) 68 Net (2030) 62
0
500
1,000
1 2 3 4 5 6 7 8 9 10
ES KW Needed for Every Year
$(800,000)
$(600,000)
$(400,000)
$(200,000)
$-
$200,000
$400,000
$600,000
1 2 3 4 5 6 7 8 9 10
Cost and Benefits By Year
Energy Storage O&MDeferral Deferral Peak ShavingAdded Peak Shaving
1 311 11
15
050
100150200250300350400
1 2 3 4 5 6 7 8 9 10
Energy Storage Operation Hours
Discharge Hr (Deferral Only) Discharge Hr (Add Peak Shaving)
C S S S
Substation
400 customer 300 customer 200 customer 100 customer
OH 8 miles
Distribution Reliability
Assume no alternative back feed source
SS
Reliability Enhancement• Distribution Circuit Model
ES LocationFaults on …
Feeder Backbone Laterals
End of the feeder X
Along the feeder X
Customers X X
Compare to: Line upgrade to
reduce failure rate Backup generator
C S S S
SubstationEnergy storage
unit size 1MW,4MWh
Energy storage unit size
250kW,1000kWh
Energy storageunit size
100kW,400kWh400 customer 300 customer 200 customer 100 customer
Reliability Enhancement ApproachSize the ES
Needed
Apply load curve
Contingency analysis
Number of ES units placed
Calculate Reliability
Indices Improvement
SAIFI*
SAIDI*
Estimate Cost & Benefits
Cost for ES, line upgrade and backup generator
Avoided cost of interruptions
Sensitivity analysis
*System Average Interruption Frequency Index (SAIFI): Total number of sustained (>5 minutes)customer interruptions / Total number of customers served*System Average Interruption Duration Index (SAIDI): customer interruption duration (minutes) / Total number of customers served
Reliability Data and Assumptions
[1] Anderson Hoke, Randomized Hourly Load Data for use with Taxonomy Distribution Feeders, https://catalog.data.gov/dataset/randomized-hourly-load-data-for-use-with-taxonomy-distribution-feeders[2] Power/Forward Carolinas, Executive Technical Overview, Duke Energy, November 2017[3] Understanding the Cost of Power Interruptions to U.S. Electricity Consumers, LBNL, 2004[4] GENERAC, Home backup generator sizing calculator, http://www.generac.com/for-homeowners/home-backup-power/build-your-generator[5] Compare Commercial Generator Prices – Buyers Guide 2018, https://priceithere.com/commercial-generator-prices/
Feeder Simulation
Annual Load Curve Hourly from NREL [1]
Circuit Length 8 miles
# of Customers 1000
SAIFI Target 1.5 interrupt/customer [2]
SAIDI Target 150 minutes/customer [2]
Initial Index [SAIFI, SAIDI] = [2, 8 hrs.]
OH failure rate 0.5-0.8 /mile-yr. [2]
UG failure rate 0.03 – 0.2/mile-yr. [2]
Backup Generator Cost
Residential (10-20 kW) $300/kW [4]
Commercial(20-150 kW) $280/kW [5]
Installation $2000 – 8000 [4]
Cost of Interrupted Power [3]
Duration Residential Commercial Industrial
0 sec $2.18 $605 $1,893
1 hour $2.7 $886 $3,253
Sustained Interruption $2.99 $1067 $4,227
Feeder Conversion Cost [2]
OH to UG $400k – 500k/mile
Voltage Support and Control for PV
• Distribution Circuit Model
0.99
1
1.01
1.02
1.03
1.04
1.05
1.06
Voltage at Interconnection
Volt No ES Voltage with ES-3500-3000-2500-2000-1500-1000-500
0500
10001500
ES and PV Profile
ES PV
ANSI Limit: 1.05
Top of the Feeder Power
Power With ES Power No ES
Charge During PV
Peak Shaving
Substation
Voltage Controlled
Bus
Feeder
CircuitBreaker
PV
Voltage Support and Control Approach
Setup Model
Energy storage control + PV Case
Distribution feeder upgrades to accommodate PV
Power Flow Analysis
Analyze additional PV could be added
Quantify peak demand decrease
Quantify energy and loss reduction
Estimate Cost & Benefits
Cost for each case
Increase PV benefits
Avoided demand and energy cost
Sensitivity analysis
Voltage Support Data and Assumptions
[1] Anderson Hoke, Randomized Hourly Load Data for use with Taxonomy Distribution Feeders, https://catalog.data.gov/dataset/randomized-hourly-load-data-for-use-with-taxonomy-distribution-feeders[2] NREL, NSRDB, https://maps.nrel.gov/[3] NREL, The Cost of Distribution System Upgrades to Accommodate Increasing Penetration of Distributed Photovoltaic Systems on Real Feeders in the United States, April 2018[4] Duke Energy Rate, Large General Services, Schedule LGS-50
Feeder Simulation
Annual Load Curve Hourly from NREL [1]
PV data NREL NSRDB [2]
Feeder Length 6 miles
K factor 477 - 4/0 Conductor(%VD/kVA-mile)
0.03% - 0.055% (12.5 kV)0.009% - 0.011% (23 kV)
Load Level 5 MW/ 10 MW
Voltage Level 12.5 kV/ 23 kV
Distribution Line Re-conductor Cost [3]
Low $ 200 k/mile
Medium $ 400 k/mile
High $ 600 k/mile
Electricity Rate [4]
Demand Charge $ 11/kW
Energy Cost for ES Operation Losses $ 55/MWh
Still to Come• Improve the capacity deferral case based on feedback• Add sensitivity analysis to the capacity deferral study
• Complete the reliability enhancement analysis• Complete the voltage support analysis
• Final VBA-enabled Excel spreadsheet that can be customized for future use
NC Energy Storage Study
Transmission ServicesDavid MulcahyDanny Sodano
Stakeholder MeetingOctober 2, 2018
Analysis Scope
• Transmission services fall into two categories:
– Transmission Investment Avoidance
– Transmission Congestion Relief
Transmission Investment Deferral
Transmission Deferral/AvoidanceService:
• Energy storage can be an alternative to address reliability issues which might otherwise require transmission build out
Benefit : Avoid transmission investment for reliability or economics
ESS Utilization: •ESS can be used to supplement reliability or contingency issues
• Enabling non-spinning resources for contingency response• Avoid congestion during contingency events • Reduce element overload
•Can be alternative to economic transmission investment.
46
Energy Storage as Alternative to Transmission
• Planning reliability requirements drive transmission investment
• Transmission expansion is system specific and difficult to compare from other studies beyond simple avoided cost factors
• Therefore, ESS’s value comes as alternativeto otherwise required transmission builds.
47
Compared to Proposed Projects
• Proposed projects from NC Transmission Planning Collaborative (NCTPC) and NC IRPs– NCTPC plan outlines why transmission is required– Can storage help mitigate any of the current issues?
• Without detailed power flow models or load forecasts, limited to simple calculations and determining if energy storage has any potential to mitigate reliability issue
• Examined current proposed projects and evaluate applicability of storage at high level and include more qualitative analysis
48
Criteria for Applicability
Potentially Applicable:• Line Capacity Violations• Low voltage• Double circuit
Not Applicable:• Reactive Support• Generator interconnection
Transmission projects are in response to contingency events in forecasted operating
conditions
Preliminary Results
Applicability of ESS to current transmission projects• Total potential as high as $283m
• Scale of investment is significant
• ESS should be considered as alternative to transmission for reliability
Note: Only shows upper bound to value. Does not show benefit nor sizing of ESS to projects
ESS Alternative Potential
Number of Projects Cost ($M)
No 4 142
Potential 11 283
Total 15 425
Next Steps and Future Work
Next steps:• Finalize classification and description of projects
Potential future work:• Include ESS in power flow model of contingencies
• Examine use cases of storage related to frequency and duration of contingency events
Transmission Congestion Relief
Transmission Congestion ReliefService:
• Energy storage can reduce transmission congestion during constrained periods
Benefits: •Allows cheaper generators to operate more•Lowers risk of line overloads•Potential alternative to economic transmission investment
ESS Utilization:•Charges when marginal generator is cheap•Discharges during congestion period •Allows cheaper generators to have higher capacity factor 53
Congestion Approach
Develop Supply Curve
• Use unit-level generator data from public sources
• Assume merit order for supply curve
Identify Congestion
• Compare hourly generation to merit order supply curve
• Identify hours with significant deviation in commitment from merit order
Calculate value
• Calculate avoided operational cost from storage
• Identify generator locations where storage has value
Method Assumptions
Avoided Cost of Congestion ($)
For hours with congestion:
• Calculate value of storage comes from the operational cost saved by dispatching cheaper generators
• Compare to cost of energy storage system located near constrained generators
Publicly Available Data Sources
Generator Data:
Source: EIA-860 Form Data
Includes: Operational units, nameplate capacities, and locations
Hourly Operation:
Source: U.S. EPA Air Markets Program Data
Operational data limited to thermal generators subject to emission reporting requirements
Method for Calculation1. Rank merit order by annual Capacity Factor (CF)
2. Determine committed generators at each hours
3. Find marginal generator (highest rank by CF)
4. Determine which generators are out of order
Preliminary Results
• Developed supply curve based on generator operation
• Many generators are operating frequently out of merit order based on annual CF
• Need to examine causes, frequency and verify results for these generators
Next Steps
• Establish marginal/operational costs of generators
• Examine if reason other than congestion is reason for out of merit order
• Use data longer timeframe (beyond 2017 data)
• Estimate sizing of ESS to determine cost for improvements
NC Energy Storage Study
Generation Peak Capacity Deferral & Bulk Energy Time-shifting
Anderson R. de QueirozJoseph F. DeCarolisJeremiah X. Johnson
Dustin SoutendijkDanny Sodano
Stakeholder MeetingOctober 2, 2018
Introduction
• The overall goal is to evaluate how storage can contribute to Generation/Resource Adequacy, more specifically:
–Peak Capacity Deferral • How storage can contribute to postpone investments in generation
–Bulk Energy Time Shifting• How storage can contribute to better economic generation resources
Storage Technologies to be Considered
Mechanical • Flywheels• Pumped storage• Compressed Air
Electrochemical • Lithium-ion batteries• Lead-acid batteries• High T sodium batteries• Flow batteries
Chemical• H2 electrolysis + storage
+ fuel cells
Thermal• Chilled water• Ice storage• Phase change materials• Water heaters
Electrical • Supercapacitors• Superconducting
magnetic energy storage
Peak Capacity Deferral
• A spreadsheet calculation is developed first to estimate the break-even cost associated with a Li-ion battery
• Some simple assumptions: –1 MW, Natural Gas CT plant cost: $900/kW–Variable Operations & Maintenance (VOM) costs: $3/MWh–Natural Gas price: $4/MMBtu–Heat Rate: 6.5-11 MMBtu/MWh–1 MW, 4 MWh Li-ion Cost: $2100/kW–WACC: 10%–Technology Lifetime: 20 year lifetime for both
• Compare Cost of New Entry (CONE) and operating costs of both technologies at various capacity factors
Peak Capacity Deferral
• A battery needs a CF of 0.31 or more to be more cost-effective
• When used only for this application, the CONE of the Li-ion battery needs to be $96.6/kW-yr lower to be more cost-effective at CF of 0.1*CF proportional to battery MWh throughput
More rigorous results to come from additional modeling
• Benefit from charging with more efficient electricity (HR 6.5 vs 11)• Off peak → 3.5 ¢/kWh and Peak → 4.4 ¢/kWh
Analysis Scenarios
1. Base case
2. Duke IRP
3. Expanded RPS
4. Clean Energy Standard
5. Carbon Cap
6. Natural Gas Prices
7. Deployment of Residential Solar PV
8. Deployment of Plug-in Electric Vehicles
• 2017 Carolinas Power generation system• HB589 solar PV deployments• Fixed representation of the exchanges
• RPS expanded to 2030 with a target of 40% for renewables (solar, wind, biomass, small hydro)
• 60% target of clean energy sources by 2030
• Duke’s 2017 Climate Report to Shareholders: 40% reduction in 2005 CO2 emissions levels by 2030
• High and Low Projections from EIA AEO 2018
• Scenario matches the build-outs proposed by Duke’s 2018 IRP
Analysis scenarios … TEMOA run
(Capacity Expansion) …
Determine build out plans for the system
…*No energy storage is
considered in the initial operational dispatch runs
• hourly operations• operational marginal
prices [$/MWh]TEMOA run
(Operational Dispatch)
…
1 3 5 7 9 11 13 15 17 19 21 23
MW
Hour of Day
Nuke Coal Gas RE
0
10
20
30
40
50
1 3 5 7 9 11 13 15 17 19 21 23Mar
gina
l Pric
e ($
/MW
h)
Hour of the Day
Introduce Energy Storage
*Assume a fixed duration (energy to power ratio) and efficiency
…
• Run different storage-size configurations
• Determine cost-optimal build plan for storage
• Minimize operational costs with storage
• Calculate change in production costs [$/year]
…
Run sensitivities- Load- Storage costs- Storage efficiency
Approach, Data and Assumptions
Carolinas
25,606 MW actual demand
Data & Assumptions• System representation:
• Existing power generators represented as individual power plants• Future generators grouped by their respective generation class
• Sources:
Power Interchanges
EIA Annual Electric Generator data, form EIA-860EIA electric utility data survey, form EIA-923EIA's U.S. Electric System Operating Data Tool NREL Annual Technology Baseline - ATBNREL Solar and Wind Energy Resource Assessment - SWERA
19412 MW avg demand (2017)
05
10152025303540455055606570
2017 2020 2025 2030
Inst
alle
d C
apac
ity [G
W]
Base Case - Installed Capacity [GW]
Energy EfficiencySolar PVBiomass - STLandfill Gas - Int. G.Landfill Gas - GTDieselNatural Gas - CCNatural Gas - CTHydro - PSHHydroCoal - STNuclear
Solar PV to reach 16.6 GW from 3.6 GW in the existing system
2.1 [GW]
05
1015202530354045505560657075
BaseCase
DukeIRP
E-REPS CESTD CO2cap NG-L NG-H
Inst
alle
d C
apac
ity [G
W]
Base Case – 2030 Installed Capacity [GW]
Energy EfficiencySolar PVBiomass - STLandfill Gas - Int. G.Landfill Gas - GTDieselNatural Gas - CCNatural Gas - CTHydro - PSHHydroCoal - STNuclear
0.36 [GW]
6.8 [GW]
4.24 [GW]
2.03 [GW]
5.5 [GW]
0.29 [GW]
2.5 [GW]
Next Steps
• Finish the Operational model runs–We already a working model for the operational runs–Runtime: About 30 minutes of computing time in a cluster to run one year for 8760 hours
• Include energy storage and re-run Operational Models–Obtain hourly operations–Operational Marginal costs
• Compute benefits–Peak generation deferral–Bulk energy time shifting
Future Research• This is the first comprehensive open source modeling effort
to develop projections for the Carolinas power system• It can be used to assess economic, technical, and policy
futures and provide valuable insights to decision makers• Model and analyze other scenarios, e.g.:
–Bidirectional capabilities for EVs–100% of clean energy–Wider range of future fuel prices–Policies under consideration
• Analyze storage deployment directly in the capacity expansion model
NC Energy Storage Study:Frequency Regulation Services
Professor Ning Lu and Yao Meng
Electrical and Computer Engineering Department
North Carolina State University
Regulation Service• Regulation services: balances generation and load in real-time to maintain
system frequency and tie-line power flows at the scheduled values.
• Inputs: Area Control Error(ACE) and Tie-line Flow Deviations.
• Signal resolution: 2-10 seconds
• Characteristics: mostly energy neutral, random in magnitude (very hard to forecast)
0 50 100
Time of the day(hr)
-400
-200
0
200
400
AC
E(M
W)
-400 -200 0 200 400
ACE Signal(MW)
0
0.02
0.04
0.06
0.08
0.1
Pro
babili
ty o
f O
ccurr
ence
NY-ISO ACE signal of June,2017 and its probability density function
Energy Storage for Regulation Service
Advantages
Energy storage systems are controlled by power electronics.
Excellent controllability allows them to follow regulation signals precisely.
1. Reduce the wear-and-tear of the traditional generators2. Reduce the amount of required regulation capacity3. Improve the quality of regulation services
• Energy storage systems have energy limits. When regulation signals have significant DC components, energy storage devices will soon be fully charged/discharged
• Three approaches to deal with this issue
─ Design energy-neutral frequency regulation signal─ Design operation strategy to maintain the state-of-charge(SOC) levels─ Allow storage to adjust its committed regulation services in a shorter interval• The first method has been implemented by PJM and ISO-NE. Fast regulation signal: Applying a
high-pass filter to the AGC signal. • Signals with a fast ramping rate but energy neutral.
Technical Challenges
Enabling Factors
0 10 20 30 40 50 60
Time(mins)
-1
-0.5
0
0.5
1
Regu
lation
Sign
al
One hour PJM RegD signal
0 10 20 30 40 50 60
Time(mins)
-1
-0.5
0
0.5
1
Reug
lation
Sign
al
One hour NY-ISO regulation signal
Design Considerations: FERC Order 784 requires the improvement of signal design considering the state of charge constraint of energy storage system
Monetary Incentives: FERC Order 755 requires the implementation of pay-for-performance regulation market
PDF of PJM RegD
-1 -0.5 0 0.5 1
Hourly Energy Bias(MWh)
0
0.02
0.04
0.06
0.08
0.1
0.12
Prob
abilit
y of
Occ
urre
nce
PDF of PJM RegA
-1 -0.5 0 0.5 1
Hourly Energy Bias(MWh)
0
0.1
0.2
0.3
0.4
Prob
abilit
y of
Occ
urre
nce energy neutral
Pay-for-performance Market Mechanism
Regulation capacity: participating resource will be rewarded by the bidding capacity 𝑃𝑃𝑏𝑏𝑏𝑏𝑏𝑏
𝑟𝑟𝑟𝑟𝑟𝑟, unit: $/MWh. Regulation-up and regulation down signals have the same power limit except in the CASIO control area
Regulation mileage 𝑴𝑴: the sum of the absolute values of the regulation control signal movements, unit $/∆MW, 𝑃𝑃𝑡𝑡
𝑟𝑟𝑟𝑟𝑟𝑟 is the power output of a regulation unit at 𝑡𝑡
Performance factor 𝝀𝝀: A value between 0 and 1, represent the response accuracy with respect to the regulation instructions. A general penalization format is as follows:
Payment = 𝑃𝑃𝑏𝑏𝑏𝑏𝑏𝑏𝑟𝑟𝑟𝑟𝑟𝑟(ρ𝑐𝑐 + 𝛌𝛌Mρ𝑀𝑀)
where ρ𝑐𝑐, ρ𝑀𝑀 are capacity clearing price and mileage clearing price, respectively. In this analysis, we assume 𝜆𝜆 = 1.
𝑀𝑀 = �0
𝑇𝑇𝑃𝑃𝑡𝑡𝑟𝑟𝑟𝑟𝑟𝑟 − 𝑃𝑃𝑡𝑡−1
𝑟𝑟𝑟𝑟𝑟𝑟
𝑃𝑃𝑏𝑏𝑏𝑏𝑏𝑏𝑟𝑟𝑟𝑟𝑟𝑟
Regulation Mileage
0 100 200 300 400 500 600 700 800 900 1000
Time Step
-5
-101
5
10
20
MW
regulation mileagepower outputreg-up capacity
reg-down capacity
Payment = 𝑷𝑷𝒃𝒃𝒃𝒃𝒃𝒃𝒓𝒓𝒓𝒓𝒓𝒓(ρ𝑐𝑐 + 𝛌𝛌𝑴𝑴ρ𝑀𝑀)
𝑃𝑃𝑏𝑏𝑏𝑏𝑏𝑏𝑟𝑟𝑟𝑟𝑟𝑟 = ±1 𝑀𝑀𝑀𝑀
Mileage 𝑴𝑴
𝑀𝑀 = �0
𝑇𝑇𝑃𝑃𝑡𝑡𝑟𝑟𝑟𝑟𝑟𝑟 − 𝑃𝑃𝑡𝑡−1
𝑟𝑟𝑟𝑟𝑟𝑟
𝑃𝑃𝑏𝑏𝑏𝑏𝑏𝑏𝑟𝑟𝑟𝑟𝑟𝑟
Modeling of Energy Storage Devices
Energy Storage Models
Modeling Parameters
𝐸𝐸𝑡𝑡 − 𝐸𝐸𝑡𝑡−1 = ∆𝑡𝑡𝜂𝜂𝑐𝑐𝑃𝑃𝑡𝑡𝑅𝑅𝑟𝑟𝑟𝑟𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅 − ∆𝑡𝑡𝜂𝜂𝑏𝑏𝑃𝑃𝑡𝑡
𝑅𝑅𝑟𝑟𝑟𝑟𝑅𝑅𝑅𝑅 − ∆𝑡𝑡𝑃𝑃𝑡𝑡𝑆𝑆𝑟𝑟𝑆𝑆𝑆𝑆𝑅𝑅𝑏𝑏𝑆𝑆𝑐𝑐
Discharged energy
Charging energy
Self-discharged energy
discharging efficiencycharging efficiency
self-discharging rate
0 ≤ 𝑃𝑃𝑡𝑡𝑅𝑅𝑟𝑟𝑟𝑟𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅 ≤ 𝑃𝑃𝑏𝑏𝑏𝑏𝑏𝑏
𝑟𝑟𝑟𝑟𝑟𝑟
0 ≤ −𝑃𝑃𝑡𝑡𝑅𝑅𝑟𝑟𝑟𝑟𝑅𝑅𝑅𝑅≤ 𝑃𝑃𝑏𝑏𝑏𝑏𝑏𝑏
𝑟𝑟𝑟𝑟𝑟𝑟
𝐸𝐸𝐿𝐿𝑅𝑅𝑅𝑅𝑟𝑟𝑟𝑟𝑆𝑆𝑏𝑏𝐿𝐿 ≤ 𝐸𝐸𝑡𝑡 ≤ 𝐸𝐸𝑅𝑅𝑅𝑅𝑅𝑅𝑟𝑟𝑟𝑟𝑆𝑆𝑏𝑏𝐿𝐿
• Start-up nor shut-down costs are not considered• Actual annual revenue for year 2017 is calculated and we assume that the same revenue is
received over the entire lifetime.• Revenue includes two payments: mileages and capacity• Cost includes installation and O&M cost• NPV (Net Present Value) is calculated assuming the discount rate is 10%
Cost-benefit Study Models
𝑅𝑅 = 𝑅𝑅𝐿𝐿𝑏𝑏𝑆𝑆𝑟𝑟𝑚𝑚𝑟𝑟𝑟𝑟 + 𝑅𝑅𝑐𝑐𝑚𝑚𝑅𝑅𝑚𝑚𝑐𝑐𝑏𝑏𝑡𝑡𝑐𝑐
𝐶𝐶 = 𝐶𝐶𝑏𝑏𝑅𝑅𝑆𝑆𝑡𝑡𝑚𝑚𝑆𝑆𝑆𝑆 + 𝐶𝐶𝑂𝑂&𝑀𝑀
𝑁𝑁𝑃𝑃𝑁𝑁 = �𝑏𝑏=1
𝑁𝑁𝑁𝑁 𝑖𝑖
1 + 𝑟𝑟 𝑏𝑏
𝑷𝑷𝒓𝒓𝑷𝑷𝑷𝑷𝒃𝒃𝑷𝑷 = 𝑁𝑁𝑃𝑃𝑁𝑁𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑅𝑅𝑟𝑟𝑟𝑟 − 𝑁𝑁𝑃𝑃𝑁𝑁𝑐𝑐𝑅𝑅𝑆𝑆𝑡𝑡
Revenue
Cost-of-service
Net Present Value
• Lifetime of a battery storage system can be estimated based on how many charging/discharging cycles it has completed at different depth of discharge(DOD)
• Rain-flow algorithm is used for estimating battery lifetime depreciation• The flywheel lifetime is assumed to be constant
Battery Lifetime Estimation Methods
20 30 40 50 60 70 80 90 100
Depth of discharge(%)
0
0.5
1
1.5
2
Num
ber o
f cyc
les
10 4
FlywheelBattery High
Battery Low
Two Types of Services MechanismOne directional service: Energy storage system only takes “up” signal when discharging, while only taking “down” signal when charging
Two Types of Services MechanismTwo directional service: Energy storage system can take both “up” and “down” signal when possible.
Performance CriterionTo evaluate the accuracy of following regulation signals, we calculated response rate as:
𝑅𝑅𝑅𝑅 =𝑛𝑛𝑆𝑆𝑟𝑟𝑆𝑆𝑆𝑆𝑏𝑏𝑆𝑆𝑆𝑆𝑟𝑟𝑏𝑏𝑛𝑛𝑡𝑡𝑅𝑅𝑡𝑡𝑚𝑚𝑆𝑆
× 100%
where 𝑛𝑛𝑆𝑆𝑟𝑟𝑆𝑆𝑆𝑆𝑏𝑏𝑆𝑆𝑆𝑆𝑟𝑟𝑏𝑏 is the number of regulation signals fully following by the ESS and 𝑛𝑛𝑡𝑡𝑅𝑅𝑡𝑡𝑚𝑚𝑆𝑆 is the total number of regulation signals.
To evaluate the lifetime depreciation when providing regulation services, we calculated the aging ratioas:
𝐴𝐴 =𝐿𝐿𝑏𝑏𝑟𝑟𝑆𝑆𝑚𝑚𝑟𝑟𝑆𝑆𝑡𝑡 − 𝐿𝐿𝑟𝑟𝑟𝑟𝐿𝐿𝑚𝑚𝑏𝑏𝑅𝑅
𝐿𝐿𝑏𝑏𝑟𝑟𝑆𝑆𝑚𝑚𝑟𝑟𝑆𝑆𝑡𝑡× 100%
where 𝐿𝐿𝑏𝑏𝑟𝑟𝑆𝑆𝑚𝑚𝑟𝑟𝑆𝑆𝑡𝑡 is the default lifetime of battery, 𝐿𝐿𝑟𝑟𝑟𝑟𝐿𝐿𝑚𝑚𝑏𝑏𝑅𝑅 is the remaining lifetime after certain period of service estimated by rain-flow algorithm.
Simulation Setup• Regulation signals and the corresponding price data were
downloaded from PJM and NY-ISO website, the data was collected from January 1, 2017 to December 31, 2017
• Designed lifetime of Li-ion battery is 10 years, while the designed lifetime of flywheel is 21 years
• The power and energy rating of Li-ion battery and flywheel is 1MW and 0.5 MWh, respectively
• Cost Parameters
Results
Results Summary: Service Quality & Lifetime
1. Regulation signal design makes a significant difference.2. When providing regulation services, battery lifetimes are shortened. 3. When providing RegD services, battery lifetimes can be further shortened but not by much. 4. When providing 1-directional services, battery lifetimes can be prolonged.5. As the flywheel can cycle as many times at low DOD as at high DODs, its lifetime is not affected by providing
the regulation services.
NY RegD RegA RegD+RegA
Regulation Signal
0
1000
2000
3000
Rev
enue
($)
Comparison of daily revenue
NY RegD RegA RegD+RegA
Regulation Signal
40
60
80
100
Res
pons
e R
ate(
%)
Comparison of daily response rate
NY RegD RegA RegD+RegA
Regulation Signal
0.5
1
1.5
2
2.5
Agi
ng C
ost(
%)
Comparison of lifetime depreciate
NY RegA RegD
Regulation Signal
-100
0
100
200
300
Annu
al P
rofit
($/k
Wyr
)
0.5hr
2hr
4hr
Profits comparison of different ESS technologies
Profits comparison of different battery sizesA larger size battery has a longer service life. When supplying RegD, the service life are 3.8, 5.5, 13.5 years for 0.5, 2 and 4 hours battery, respectively.
1-directional 2-directional
PJM RegD
0
200
400
600
800
Annu
al P
rofit
($/k
Wyr
)
Battery
Battery-2030
Flywheel
Now
Future
Flywheel
Now
Future
Flywheel
0.5
2 4
0.52 4
0.52 4
What to come
• We have finished the following comparisons–Regular regulation signals v.s. storage-friendly signals–1-directional v.s. 2-directional services–Regional differences (PJM v.s. NYISO)–Different battery sizes–energy storage technologies (Li-ion Battery v.s. Flywheel; lifetime sensitive to DOD v.s. lifetime not sensitive to DOD)
• What to come–Market-based v.s. non-market based regulation services
• Need signals from non-market based systems–different energy storage control algorithms
• Optimize energy storage operation • Stack the regulation service with other type of services
NC Energy Storage Study
Energy Storage to Reduce Solar ClippingJeremiah Johnson
Stakeholder MeetingOctober 2, 2018
Background• Solar photovoltaics use invertors to convert DC generation to AC
generation.• You can reduce inverter costs by “undersizing” to yield reductions in
delivered energy cost. This results in “clipping.”• Energy storage can reduce clipping and allow for higher invertor
utilization.
02468
101214
0:00
2:00
4:00
6:00
8:00
10:0
012
:00
14:0
016
:00
18:0
020
:00
22:0
0
Sola
r G
ener
atio
n M
W
Unclipped GenerationClipped
02468
101214
0:00
2:00
4:00
6:00
8:00
10:0
012
:00
14:0
016
:00
18:0
020
:00
22:0
0
Sola
r Gen
erat
ion
MW
ClippingDischarged from ESUnclipped generation
Approach, Data and Assumptions
• In this analysis, we will ultimately use solar resource data at a 1-minute resolution provided by Strata (under NDA).
• We assume crystalline silicon modules and test two configurations: south-facing fixed tilt at 20 degrees and single axis tracking (with the axis running north-south).
• The PV array DC output is calculated using the California Energy Commission performance model in NREL’s System Advisor Model with the Nominal Operating Cell Temperature (NOCT) method to estimate temperature impacts. We model inverter performance is modeling using an empirical method from Sandia National Laboratory.
Solar Array Model Inverter Model
Energy Storage Operations
Potential DC Output
System AC Output
Clipping Losses
Solar Resource Data
Planned Scenarios/SensitivitiesParameter Values Considered DescriptionLocation Multiple, throughout state • Minute-level data for one year
Array Design 10 MWDC fixed tilt10 MWDC single axis tracking
• Fixed tilt at 20 degrees, south-facing• Single axis (north-south) tracking with
backtracking, +60 degree rotationDC/AC Ratio 1.2 to 2.0, in 0.1 increments • Rated inverter capacities range from 5
MWAC to 8.3 MWAC
Energy Storage Rated Capacity
0% to 100% of the difference between inverter rating and module nameplate rating, in 10% increments
• Ten rated power capacities will be considered for each DC/AC ratio
Energy Storage Duration
1 hour to 4 hours at rated power, in 1 hour increments
• The rated MWh of the energy storage system will be varied between one and four hours of discharge at the battery’s rated power
0%5%
10%15%20%25%30%35%
1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2
DC/AC Ratio
• For now, this uses 1-minute resolution data from Oak Ridge National Lab.
• Fixed tilt (20°), due south• The AC Capacity Factor represents the utilization based on the
invertor’s rated capacity
Clipped solar
Preliminary Results: Solar Clipping vs. DC/AC RatioAC
Cap
acity
Fac
tor
• 10 MWdc solar PV, 5 MWac system 2.0 DC/AC ratio• Experiences 2,400 MWh of clipping annually (16% of potential
generation)
Preliminary Results: Reducing Clipping Using Li-ion Batteries
-
400
800
1,200
1,600
2,000
1 2 3 4 5 6
Clip
ping
Red
uctio
n fr
om
Ener
gy S
tora
ge (M
Wh/
yr)
Energy Storage Duration (hr)
5 MW4 MW3 MW2 MW1 MW
Reduction in Solar Clipping
- 100 200 300 400 500 600 700 800
1 2 3 4 5 6
Reve
nue
Requ
irem
ent t
o Su
ppor
t Sol
ar C
lippi
ng
($/M
Wh)
Energy Storage Duration (hr)
5 MW4 MW3 MW2 MW1 MW
1 2 3 4 5 6
Energy Storage Duration (hr)
5 MW4 MW3 MW2 MW1 MW
Preliminary Results: Revenue Requirement for Li-ion Batteries to Reduce Solar Clipping
2.0 DC/AC ratio, Solar data from Oak Ridge, TN; linear interpolation to estimate Li-ion costs for all capacities
Breakeven Value of “Saved” Solar2018
Breakeven Value of “Saved” Solar2030
Min cost = $320/MWh4 hour durationRecovers ~one-third of clipped solar
Min cost = $180/MWh4 hour durationRecovers ~one-third of clipped solar
In the coming months…• Update analysis using 1-minute resolution solar resource
data for North Carolina• Conduct analysis for single-axis tracking solar• Synthesize the findings
• Consider varying value of time of day of generation
Potential future work…
NC Energy Storage Study
Current Policy and Potential OptionsChristopher Galik
Christopher GambinoAutumn Proudlove
Stakeholder MeetingOctober 2, 2018
Analysis Scope
• What is the regulatory structure that governs storage in NC?
• What changes are possible to increase value of energy storage to North Carolina Consumers?
• Among the list of possible interventions, which are the most feasible/practicable, and what are the associated implications for storage?
Approach and Data• Analysis, R&D, and Market Support: R&D and storage deployment,
analysis like this, worker training programs, etc.• Planning and Access: Define, reform, or refine utility planning processes
and/or the rules affecting access to, e.g., data, state/wholesale markets. • Business Models and Rate Reform: Changing how utilities are regulated
or operate, targeted changes to rates (e.g., time-of-use, demand charges).• Mandates: Policies establishing minimum deployment targets or
performance standards.• Process and Approvals: Policies that govern the process for storage
deployment (e.g., interconnection standards, compensation rules).• Incentives and Financing: Provide funding to or defray the cost of the
deployment of storage (e.g., loans, tax credits, rebates, exemptions).• Utility-Driven Demonstrations and Deployment Programs: Utility-led
programs to purchase, fund, or deploy storage.
Approach and Data• Review of existing regulatory context consisted of:
–NCUC Rules, energy-related legislation (e.g., Senate Bill 3, House Bill 589), as well as the regulations, rate cases, settlements, rulemakings, and/or orders that emerged as a result. Also included are policies or provisions explicitly referencing storage issued by PJM regulators or Tennessee Valley Authority (TVA);
–Third-party news alerts, policy briefs, issue analyses, and summary reports to identify provisions that our initial scan may have missed;
–Anonymous survey distributed to the energy storage project stakeholder group.
Preliminary Results
• Analysis, R&D, and Market Support:–Possible relevance of $1M REPS cost-recovery provision for R&D.
• Planning and Access:–Under PJM, FERC Orders 841, 845, 784, 819, and 890 have direct to storage wholesale markets. Storage already eligible to participate in PJM ancillary service (Reg D) market.
Preliminary Results
• Planning and Access:–Rule R08-41 (Emergency Load Reduction Plans and Emergency Procedures) requires demonstration of black start capabilities, which could conceivably include storage.
–Rule R08-60 (Integrated Resource Plannings and Filings) Requires identification of load requirements and resource options, conceivably including storage. Also requires identification of implications of smart grid deployment on planning, within which storage is mentioned.
Preliminary Results
• Business Models and Rate Reform:–In its 2017 avoided cost ruling (Docket No. E-100, SUB 148), NCUC signaled an intention to refine avoided cost calculations so as to provide clearer signals to qualifying facilities to further facilitate the deployment of advanced solar or storage applications.
• Mandates:–No storage-specific mandates, but unclear if storage qualify for muni/co-op REPS compliance (i.e., DSM)
Preliminary Results
• Process and Approvals:–May 2015 interconnection standards (Docket No. E-100 Sub 101; storage may be connected under same process as other small generating facilities). Revisions to interconnection standards are ongoing, with multiple provisions related to storage being debated under the auspices of the proposed Competitive Procurement of Renewable Energy (CPRE) program;
–At the county and municipal level, decisions regarding contracting, zoning, compliance with fire codes, and decommissioning requirements are likely
Preliminary Results• Incentives and Financing:
–NA• Utility-Driven Demonstrations and Deployment Programs: In a
June 2018 settlement agreement that was ultimately rejected, DEC pledged to deploy at least 300 MW of storage by 5/26, with 200 MW of that by 5/23. By the 2021 IRP, DEC proposed to include methodology for including DERs and NWAs in integrated system operations planning (ISOP) processes.
• Initial filings under the CPRE program discussed a modified power purchase agreement (PPA) to allow for pre-inverter storage. Storage must be included as part of a renewable project (i.e., no apparent provision for stand-alone storage).
Still to Come
• Currently reviewing a database of policies, programs, regulations, proceedings, etc., related to storage as deployed in other states, RTOs (October-November).
• Pairing of policy options to analytical findings will be, by design, conducted near the conclusion of the study (November).