Demand Response: Fundamentals and Evolution

Demand Response: Fundamentals and EvolutionFundamentals and Path Forward
Course Outline 9:00 – 9:15am Welcome Remarks and Introductions
9:15 – 11:00 am Demand Response Fundamentals • Defining DR and DR Evolution • Wholesale Power Market Design
11:00am – 12:00pm Lunch concurrent with Interest Group meetings
12:00-3:10pm Strategies and Tactics • How Wholesale Power Markets Work • Portfolio Development • Program Design and Implementation • Evaluation, Measurement & Verification
3:10-3:40pm Technology Overview
4:30pm Adjourn 4
5
Both DR and EE reduce load, but
Energy Efficiency is a permanent change in energy consumption, generally with no decrease in service level.
Demand Response is a temporary change in energy consumption, generally with some decrease in
service level (e.g., less comfortable climate, sub-optimal lighting).
7
0
1
2
3
4
5
6
0:00 2:00 4:00 6:00 8:00 10:00 12:00 14:00 16:00 18:00 20:00 22:00
kW
8
9
meet demand • Must be available
immediately • When demand may
demand
1010
The Economic Definition of DR The “response” is based on a payment for either:
• the willingness to change behavior (capacity), or • the actual change (performance)
in the “demand” level of electric energy Payment can be based on the actual reduction controlled by:
• the electricity customer, or • programmed into the customer’s equipment
responding to either: • a grid system operator reliability request, or • a price signal, or • on the availability to be on call
11
Bombardier CRJ-700 Only Has Capacity for 70 People
80 People show up for a flight on a day of high travel demand
Airline pays 10 people to take a later flight (e.g. provide an incentive in the form of ticket vouchers)
Demand Response in the Airline Industry
Source: Skipping Stone 12
Power Grid needs 80 MW
Power Grid pays consumers to reduce 10 MW to balance supply & demand
Source: Skipping Stone 13
• Developing curtailment capability to address short- term/emergency supply shortfallsImproved Reliability
• Delaying investment in specific, localized substations and feeders using DR as a demand side resourceDeferral of T&D Upgrades
• Reduction of system operating costs through fewer starts of peaking units, reduced need for spinning reserve from generators, and economic dispatch of DR resources
Operational Cost Savings (Economic Dispatch)
• A possible alternative to new generation or a more economical way to provide ancillary services
Integration of Intermittent Renewable Resources
• Commission rulings to have ESPs fund and operate DR programs or achieve DR curtailment goalsRegulatory requirements
Source: Navigant
14
Types of DR programs • Dispatchable = call or control or bid in advance
• Wholesale market directed economic programs • Wholesale market directed reliability programs • Direct Load Control, e.g. automatic appliance shut-off • Interruptible Rates, i.e. lower rates for directed
reductions
• Non-Dispatchable / Price-Responsive Demand = pre-set • Critical Peak Pricing – scheduled • Peak Time Rebate – built into a rate • Time-Of-Use Pricing – annual schedule • Dynamic Pricing – all of the above
15
8,000
9,000
10,000
11,000
12,000
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour
16
• Manage post-event snapback to avoid new peak
• Smooth out load reduction to give steady MW that operators can count on
• Monitor indoor temperatures (for HVAC loads) to ensure customer comfort
Load Shape Impact of Residential Direct Load Control
0
1
2
3
4
5
6
0:00 2:00 4:00 6:00 8:00 10:00 12:00 14:00 16:00 18:00 20:00 22:00
kW
Source: Navigant
17
• Smooth load shape during control event
• Like a bull pen, different loads can be called when needed
• Its all about the load impact, not what is controlled
Peak-Day Load Shape Before and After Load Control
6,000
7,000
8,000
9,000
10,000
11,000
12,000
13,000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour
Source: Navigant 18
Demand Response Evolution
Planning & Emergencies
Visibility
• Distribution & Transmission Relief
20
• Master Station
Introductory Questions • Which of the following devices are DERs?
• A. Load Control Switch • B. Smart Thermostat • C. Smart Plug and Switches • D. All of the above • E. None of the above
• Which communication technology is fully backward compatible throughout its history
• A. AMI Mesh • B. Cellular • C. Wi-Fi / Broadband • D. None of the above
23
Introductory Questions • Which of the following devices are DERs?
• A. Load Control Switch • B. Smart Thermostat • C. Smart Plug and Switches • D. All of the above • E. None of the above – The items are devices that monitor
and control DERs.
• Which communication technology is fully backward compatible throughout its history
• A. AMI Mesh • B. Cellular • C. Wi-Fi / Broadband • D. None of the above – All technologies evolve. Suppliers try
to make things backward compatible, but at some point can’t. 24
Broadband Internet
• Control Strategies • End Point Controllers • Commercial DR • Master Station Evolution
25
DR Technology Evolution • Communication
26
Utility 1 Utility 2 Utility 3 Utility 4 Utility 5
Emerging Establishing
Martin Cooper 1973
2007
AMI Mesh Gen 1
Broadband Internet
• Current Entry Point • Home Gateway
• Multiple Communication Radios • Home Intelligence
• Cloud Services • Security • Lifestyle • Energy Markets
• Home Automation Communication • Bluetooth & Zigbee -> dominate • Wi-Fi, Wave, X10, et. al. ->
• Devices • Customer Convenience • Security • Energy
29Cellular Wi-Fi/ IPBluetoothZigbee
Digital Paging Mesh Cellular Broadband + Wi-Fi Lora Home EMS Communication
Type One-Way Two-Way Two-Way Two-Way Two-Way Two-Way Bandwidth Low Medium to High High High Medium to Low
Effective Network Latency Minutes 15 - 30 Seconds < 1 Seond < 1 Second 1-5 minutes Owner Utility &
Commercial Utility Commercial Customer Commercial Customer
Comm Cost Maintenance Cents/message
DR Solutions Emergency
Frequency Following DR Event Attributes
Event Delivery (System) < 10 Min Seconds - Mins Seconds - Mins Seconds - Mins Seconds - Mins Seconds - Mins Device Acknowledge Field Monitors Device Device Device Device Gateway
Acknowledge Quantity Small Medium Large Large Large Large Event Performance Substation Meters Home Meter
Device Data Home Meters Device Data
Home Meters Device Data
Home Meters Device Data
Home Meter Gateway
M&V Device Performance Field Visit Network Network Network Network Network - Home
Quantity Small All All All All All Program Evaluation Sample All All All All All
System Maintenance Individual Device Status No Yes Yes Yes Yes No - Home
DR Technology Evolution • Communication
31
• AC Units • Cycling • Temperature Set Point & Ramp
• Heating • Dual Fuel, Storage • Off Heat pump Aux • Temperature Set Point & Ramp • On/Off Baseboard
• Water Heater • On/Off • Just in Time/ Comfort bump • Frequency Following
• Pool Pump / Hot Tub • On/Off & Scheduling • TOU Optimization
• Solar PV • EV • Battery
Utility 1 Utility 2 Utility 3 Utility 4 Utility 5
AC Cycling Control Strategies • Standard Cycling
• Control for % of time • 50% control = 15m
run, 15 m off
unit runtime • 50% control = unit
typical run 10m / 30 min. Let run 5m
Issue: How do we maximize?
• Cloud based analytics • Analyze Customer
Clusters • Control by individual
customer attributes
• Rate based control • Pre-Cool during off-peak • Manage peak rate • Full Release after peak
rate time
• Allow x degrees per hour temperature increase to max temp
Issue: Reduction not flat for full control period & Free Riders
Issue: New Rate Design
• Rate based control • Pre-Cool during off-peak • Manage peak rate • Full Release after peak
rate time
Clusters • Control by individual
WH Control Strategies
during high prices • Ancillary Services
• Add to Unit Commitment as system reserves
• Frequency Dispatch • Add to AGC as a frequency
dispatch resource • Regulation
34
Control Signal every 4-8 seconds Dispatch every few seconds
Automatic Dispatch. Requires < 1 minute Response. 25-60 events
Automatic Dispatch. Requires < 10 minute Response. 25-60 events
Daily. Curtail during day, charge WH at night
Rate Profile Sent to WH which manages unit temperature and usage
0.8 kWh
0.4 kWh
• Maximize Customer’s solar output • Issue: Voltage Violations
• Smart Inverter • Keeps Solar from creating voltage violations
• Direct Control • On/Off • Adjustable ???? Available (See Smart Inverter)
• Home EMS • Home Optimization Energy Storage (Battery, Water Heater,
Pool Pump, Pre-Cool, Pre-Heat)
• Control Strategies • End Point Controllers • Commercial DR • Master Station • Standards
36
• Switches
• Thermostats
• Home EMS / Gateway • Smart Home • Home Resource Optimization • Solar & Battery Integration
• Behavioral • Media Communication • Home Energy Monitor • Display (Beat the Peak, Energy Orb) • Home Energy Reports
Switch T-Stat Home
Baseboard Heating Heat Pump, Dual Fuel
Water Heater Pool Pump
Program Design BYOD
Communications Wi-Fi
Cellular ZibBee
• 110 Loads • Window AC • De-Humidifiers
• 21st Century Residential • Loads
38
• Medium / Large Commercial • BMS/EMS • Business Type Tailored Solution • Load specific Solutions
• Industrial • Generally Site Specific Solution • Sophistication / Complexity varies • Requires Engineering Analysis
• Large loads worth the effort
• Agriculture • Irrigation, Crop Dryers, Dairy Hot Water
Utility Segment Average Customer Count per Utility Average Annual Revenue/ Customer Utilities Total Residential Commercial Industrial Residential Commercial Industrial
IOUs Large 24 2,118,443 1,869,053 244,388 5,003 1,181$ 7,636$ 112,738$ IOUs Medium 104 420,003 365,545 52,163 2,295 1,151$ 6,658$ 88,973$ Municipals Large 103 131,946 117,166 14,385 395 1,351$ 4,741$ 121,418$ Cooperatives Large 66 97,222 87,714 9,274 234 1,521$ 4,471$ 90,491$
Small and Medium Business • Requires 12-18 month payback • Standard Solution • Mismatch between market size & energy intensity
39
Energy Intensity kWh/ft^2
Food Store 4% 40.99 Restaurant 8% 40.2 Unref Wharehouse 7% 20.2 Hotel 1% 19.61 Large Office 1% 17.7 Retail Store 15% 14.06 Small Office 35% 13.1 Health Care 1% 12.26 Misc 25% 12.13 Refr Warehouse 0% 9.84 College 1% 7.46 School 2% 4.45
DR Technology Options
EE Synergy
Interior Lighting 28.7% Cooling 14.9% Refrigeration 13.4% Ventilation 11.9% Office Equipment 7.1% Exterior Lighting 5.8% Miscellaneous 5.8% Cooking 4.2% Motors 4.2% Heating 1.6% Air Compressors 1.0% Water Heating 0.9% Process 0.3%
100% 88% 17% 40% 35% 65%
Industrial and Large Commercial
Dispatch Utility
Aggregator ISO
LonTalk, Profibus, SDS, SRTP, et. al. BACnet, Modbus, DALI, X10, et. al.
AC
2030.5 + Legacy
Dispatch Utility
Aggregator ISO
LonTalk, Profibus, SDS, SRTP, et. al. BACnet, Modbus, DALI, X10, et. al.
AC
2030.5
Comms
Comms
42
• User Integration
DER Lifecycle Software Environment
Behind the
meter Storage
DERMS TelemetryIT Architecture StandardsCo-
Customer Engagement Field Services
Back office/IT DRMS > DERMS
Customer Systems Battery Systems
Demand Charge Reduction TOU Bill Management
T&D Congestion Relief
Operating Reserve (10-min spinning reserve
Area & Distribution Operations
Avoided T&D
Unknown: “Fast DR, PV + Battery Storage, Grid-Interactive Loads
DR Management Systems
Dispatch Strategy Management
Program Design & Implementation
• Broadband Internet, Cellular, AMI Mesh
• DERs are resources: • HVAC, WH, Hot Tub, Appliances, Pumps, Motors, etc.
• Controllers make resources available for Demand Response: • Thermostats, Switches, Smart Plugs, Home EMS, BMS, Controller, etc.
• Home Automation has a mix of radio technologies: • Wi-Fi, Bluetooth, Zigbee, etc.
• Utility IT Infrastructure (EMS, BMS, DRMS, DERMS, CIS) is complex
50
52
Market • Maintains Bulk Power System
Reliability
• Supplier • Owns & operates generation facilities • Provides Energy • May also provide Ancillary Services
• Distribution Company • Owns & operates the distribution grid • Generally a regulated entity • May be a municipality or cooperative
• Energy Retailer • Competitive Provider (areas with
competition) • Local Utility (areas without
competition) • Procures and delivers energy to
customers
58
Wholesale Desks
• Provide Open Transmission Access • Administer Wholesale Electricity Market • Maintains Bulk Power System Reliability
RTO = Regional Transmission Organization • Adds Regional Coordination • Works across multiple states
60
• Energy • Capacity • Ancillary Services
• Financial Hedges • Bi-Lateral Trades • Financial Transmission Rights (FTRs)
Timeframes • Capacity (Typically 3-5 Years
Ahead) • Forward/Long-Term (Years &
Similarities • Electric energy is treated as a
commodity • Markets are set up for trading
electricity • Associated commodities
(ancillary services) also traded • Traded forward and in real-time • Traded financially and physically
Differences • Cannot be stored in large
quantities • Delivery is influenced by
transmission constraints
At 7:00 am ISO/RTO Publishes Load Forecast Operating Reserve Requirements
By 11:00 am Market Participants Submit Day-Ahead Demand Bids Day-Ahead Resource Offers Outage Notifications Bi-Lateral & Physical Transactions
At 4:00 pm ISO Clears & Publishes Results For Clears Day-Ahead Market
Each market has different timelines
Typical Day-Ahead Market Timeline
0%
1%
2%
3%
4%
5%
6%
7%
8%
9%
10%
64
Generation – Wholesale Process • Suppliers offer to sell energy to the market (often erroneously called
“bidding” into the market) • Offers can be made by blocks (as many as 10 generally) each consisting of a
price and a quantity. • The sum of all offer quantities must be less than the generation maximum
capacity • The block prices can be any amount; perfect competition would force all
offers to “marginal cost” • Offers are generally daily and are split into “trading intervals” such as an
hour-long block
PSEG SEC Filing Form 10-K (2013)
66
“Standard” Market Design Pricing Model
• Forward prices derived from offer curve and bid curve • Real-Time prices derived from offer curves and actual demand
Locational Component • System Lambda is based on pure economics • LMP “separation” stems from impacts from:
• Transmission losses • Transmission constraints
• Thousands of Locational Marginal Prices (LMP) per region • Hundreds of LMPs grouped into Zones
69
Regulation • Regulation service supplies the grid with small increases and decreases
in supply every few seconds • The makes up the difference between a dispatch “block” of generation
and the instantaneous actual load which moves faster than every 5 minutes.
70
71
Grid-Connected Renewables Output, 10/20/2020 72
California ISO: Things that Make You Go Quack
Aggregate Demand Minus Renewables Output, 10/20/2020 73
FERC Order 2000: RTO Features (1999) Minimum Characteristics
1. Independence 2. Scope and Regional
Configuration 3. Operational Authority 4. Short-term Reliability
Minimum Functions 1. Tariff Administration and
Design 2. Congestion Management 3. Parallel Path Flow 4. Ancillary Services 5. OASIS and Transmission
Capability 6. Market Monitoring 7. Planning and Expansion 8. Interregional Coordination
74
FERC Order 888: ISO Features (1996) • Fair and non-discriminatory governance • No financial interest in power markets
• Provide open access to the transmission system • Ensure short-term reliability • Control over interconnected transmission • Identify & relieve constraints • Efficient management and administration
• Pricing that promotes the efficient use and investment • Publish transmission system information • Coordinate with neighboring control areas • Establish an dispute resolution process
75
Federal Energy Regulatory Commission
“A demand response resource participating in an organized wholesale energy market administered by an RTO or ISO has the capability to balance supply and demand as an alternative to a generation resource and when dispatch of that demand response resource is cost-effective as determined by the net benefits test described in this rule, that demand response
resource must be compensated for the service it provides to the energy market at the market price for energy, referred to as the locational marginal price (LMP). This approach for compensating demand response resources helps to ensure the competitiveness of organized wholesale energy markets and remove barriers to the participation of demand response resources, thus ensuring just and reasonable wholesale rates.” [FERC, March 15, 2011]
7676
System Operator Administers the demand response process
Service Provider Coordinates resources to deliver electricity products and services to a market or distribution operator
Transmission / Distribution Service Provider
Scheduling Entity Submits bids/offers and receives Schedules and Awards
Designated Dispatch Entity
Receives and processes demand resource dispatch instructions or market information and provides response feedback
Metering Authority Provides data necessary to determine the performance of a resource.
77
Wholesale Market Demand Response Active Response
Paid To Change Energy Consumption • Real-Time Energy Market aka “Spot Market” • Day-Ahead Energy Market • Regulation Market (small changes every few seconds)
Stand-By Response Paid To Be Available To Change Consumption • Emergency Response • 30-Minute Reserve Market aka “Non-Spinning Reserve” • 10-Minute Reserve Market aka “Spinning Reserve” • Potentially ~2-Minute Reserve aka “Fast Reserve” • Capacity Market (as an obligation for Capacity Award)
78
79
“[d]emand response can provide
• competitive pressure to reduce wholesale power prices; • increases awareness of energy usage; • provides for more efficient operation of markets; • mitigates market power; • enhances reliability; • and in combination with certain new technologies, can support the use of
renewable energy resources, distributed generation, and advanced metering.”
[FERC, October 28, 2008] 80
Federal Energy Regulatory Commission
"...enabling DER aggregators to compete in all regional organized wholesale electric markets....empowers new technologies to come online and participate in a level playing field, further enhancing competition, encouraging innovation and driving down costs for customers."
[FERC, September 17, 2020] 81
How the Markets View DR and EE Demand-Side Management
Demand Response
Energy Efficiency
Dispatchable Non-Dispatchable
Reliability Economic
Load as a Capacity Resource
EmergencySpinning Reserves
Non-Spinning Reserves
Energy- Price
(4CP Response)
Demand-Side Management
Demand Response
Energy Efficiency
Energy- Price
(4CP Response)
Demand-Side Management
Demand Response
Energy Efficiency
Energy- Price
(4CP Response)
Energy Efficiency
• Customer saves costs without changes in service with energy efficiency measures
• Incentive programs improve return on investment of measures installed
• EE is “Invisible to Markets” • EE can qualify as “capacity” in Capacity Markets with some
exceptions • Not a commodity that can be easily brokered
83
cost of electricity • Requires premises technology to support automated, frequent
response such as TOU controls • “Set-And-Forget” mode
• Again, Invisible to Markets • Impact to Markets leads to more intricate forecasting tools as
rates change over time
Economic Demand Response • “Active” Energy price response - dispatchable • Triggered by Market Price, not grid problems • Ancillary Services provide additional revenue options • Demand response aggregators can simplify response process for groups
of customers • Retailers can develop custom programs
• “Demand-as-a-Hedge”
• Generally 50-200 Hours/Year based on scarcity or congestion of local or system energy prices
85
Emergency Demand Response • “Stand-By” Energy dispatch opportunity • Implemented at Bulk Power (Transmission) Level or
Distribution Level • Paid a service fee even if no emergencies are called in a given
year – reservation based on opportunity • Often connected to capacity markets as delivery condition for
capacity awards • Generally <50 Hours Year • The first form of demand response (1970’s)
86
Flat -Tiered Time of Use Critical Peak Pricing Real Time Pricing
A B
2
3
88
89
Energy Basics for DR • DR resources are generally treated as pseudo-generation
• Modeled as generator • Use most wholesale market interfaces • Most market rules apply without change
• Costs are generally allocated to loads, often zonal based • Unique differences
• Registration/qualifications rules • Payments subject to “Net Benefits” test • Performance is substantially more complicated
90
Demand Response Bids • Historically the “Load is the Load” meaning there is usually no
elasticity in demand based on price • Elasticity means one might reduce load when prices are high or
even consume more if prices are low • For DR in wholesale markets, MPs may bid in their demand in
accordance with wholesale tariffs. • The markets now need to develop the optimization platforms to
integrate the demand side offers • The market clears, and the MPs then manage the loads to
optimize to the market outcomes
91
92
Role of DR Aggregators in the Market • Demand-Side Aggregators collect small resources (residential and small
commercial) and take a share of the wholesale market proceeds • Without the Aggregator, small resources cannot participate as they do not
make the minimum reduction size requirements • Aggregators simplify the very complex market rule into understandable
“pay-for-performance” offerings • But, Aggregators cannot operate a stand-alone program, since they have no
costs to offset.
Demand Response Load Sources
Data Source: State of the Market Report for PJM, 2020-Q2, Monitoring Analytics, LLC 94
Manufacturing: 3,978.5 MW, 42%
HVAC 3,120.8 MW, 33%
Lighting: 862.6 MW, 9%
Refrigeration: 215.4 MW, 2%
Batteries: 49.3 MW
Day-Ahead vs. Real-Time Participation
Data Source: State of the Market Report for PJM, 2020-Q2, Monitoring Analytics, LLC 95
Real-Time Revenue: $345k, 35%
Day-Ahead Revenue: $634k, 65%
Demand Response Event Timing
Market Settlements for DR performance
What is baseline? • Would-be load consumption during DR event if DR were not
called • “Normal” load is estimated (counterfactual)
Baseline calculation method • Important process for DR performance payments • Based on large numbers of studies • Balance between accuracy (complexity) and easy-of-use
(simplicity) • Performance is verified via this process
98
0
5
10
15
20
25
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
D em
an d
[k W
Hour Ending
DAY 1 DAY 2 DAY 3 DAY 4 DAY 5 BASELINE
99
Raw baseline is a simple average of ten days Date Type Event HE 16 Meter Value
5/25/2010 Weekday N 14.75 5/26/2010 Weekday N 15.50 5/27/2010 Weekday N 15.25 5/28/2010 Weekday N 15.75
6/1/2010 Weekday N 13.75 6/2/2010 Weekday N 14.00 6/3/2010 Weekday N 14.75 6/7/2010 Weekday N 13.75 6/8/2010 Weekday N 15.50
6/10/2010 Weekday N 14.50
Trade Date (TD) = 6/15/10 Hour Ending (HE) = 16
Raw customer baseline = 14.75
Performance Evaluation Models Baseline Type-I:
A Baseline performance evaluation methodology based on a Demand Resource’s historical interval meter data which may also include other variables such as weather and calendar data.
Baseline Type-II: A Baseline performance evaluation methodology that uses statistical sampling to estimate the electricity consumption of an Aggregated Demand Resource where interval metering is not available on the entire population.
(based on terminology from NAESB M&V Standards) 101
Baseline: Baseline vs. Event Day
102
Demand Response Energy Measurement
106
According to wholesale market settlement calculations and methodology, the utility’s load reduction shows that the utility ended up making a payment to the wholesale market (instead of receiving payment)
Baseline Method Determines Payment
What is DR Potential?
• Estimates how many MWs of load reduction (or other load modification) are feasible over time
• Represents what could be achieved given certain conditions are met (e.g., customers accept the program, cost effectiveness, etc.)
Why is it Important?
• Helps to prioritize program focus and design efforts • Important caveat – DR
potential studies are not program designs
• Serves as regulatory and business due diligence tool
• A tool in the IRP process • Can identify the “game
changer” large customers
Technical Potential
Economic Potential
Achievable Potential
• DR Potential represents what could potentially be achieved given certain conditions are met, e.g.:
• Cost-effectiveness requirements are met
• Customers accept the programs
• DR potential is not intended to advocate what programs/measures should be adopted/implemented; it only sets forth estimates, should certain things occur.
FERC DR Potential Estimates – Entire US
• National, regional and utility- specific studies indicate a 5-10% reduction in load is within the range of achievable potential
• FERC National DR Potential Study (2009) projected a range of potential at 4-20%
• The various scenarios reflect estimates of DR potential, should certain things occur
0
50
100
150
200
Other DR Interruptible Tariffs DLC Pricing w/o Tech Pricing w/Tech
38 GW, 4% of peak
82 GW, 9% of peak
138 GW, 14% of peak
188 GW, 20% of peak
Source: FERC National Assessment of DR Potential, 2009 http://www.ferc.gov/legal/staff-reports/06-09-demand-response.pdf Reflects peak demand reductions for capacity – does not reflect DR potential for ancillary services, etc. 112
Key Considerations for DR Potential Analysis •Sectors- residential, commercial, industrial, agricultural? •Sizes– small, medium, large, extra-large? •Business types – retail, office, grocery, education, etc.?
Customer Classifications to
Level of Geographic Coverage
•All manually enabled? •All automated (PCTs, Auto-DR)? •Combination of both?Technology Choices
•Will DR help meet reliability or economic needs? •Summer, winter or both? •Contingency, regulation, load following?
Seasonality and Time-Sensitivity
Source: Guidehouse 114
Present potential estimates by DR option, customer class, season, region, etc., along with annual and levelized costs.
Characterize market for DR potential estimation (no. of customers and coincident peak load estimates) by customer
class (residential, small/med/large C&I), end use, season, etc.
Define and characterize DR options and map applicable options to relevant customer classes.
Develop participation and unit load reduction assumptions under each DR option; also develop itemized cost
assumptions.
Develop baseline forecasts for customer count and load over the study period.
Step 4: Estimate Potential and Costs
Step 4: Develop Key Assumptions for Potential & Costs
Step 3: Define DR Options and Characterize
Step 2: Develop Baseline Forecasts
Step 1: Market Characterization
Level 1: By Sector • Residential • Commercial and Industrial (C&I)
Level 2: By Size Customer Class
• Residential
• C&I customers by size, based on maximum demand values, following rate schedules: » Small C&I: less than 50 kW maximum demand » Medium C&I: 50-200 kW maximum demand » Large C&I: greater than 200 kW maximum demand
Level 3: By Building Type (under each C&I class)
C&I segments/building types- align with segments/sub-segments considered by Eversource.
Typical segments include the following: Retail, office, grocery, lodging, schools, colleges/universities, restaurant, healthcare, hospital, warehouse, industrial.
Level 4: By End-use (for each building type)
Disaggregation by end-use. • Residential: space cooling, electric water heating, electric vehicle, battery. • C&I: HVAC, electric water heating, lighting, refrigeration, industrial, battery.
Step 1: Market Characterization
Baseline Peak Demand Projections
116
• Bottom-up peak demand projections using 8760 hourly system load data, load profiles by segment and end-use loadshapes provide a solid foundation for developing DR potential estimates.
Source: Guidehouse
Baseline Peak Demand Projections Example
117
Projections by Customer Class Projections by Customer Segments
Typical DR Options Step 3: Define DR Options
DR Options Characteristics of DR Options Eligible Customer Classes Targeted/Controlled End uses
Direct Load Control (DLC)
Load Control Switch
Control of electric loads by a thermostat and/or load control switch. Typically customers receive fixed monthly, seasonal, or annual incentives.
Residential,
Small C&I, Medium C&I Space cooling, space heating, electric water heating.
C&I Curtailment
Auto-DR enabled
Manual and/or Auto-DR enabled load reduction. Typically customers have a firm load reduction commitment.
$/kW payment based on contracted capacity, plus $/kWh payment based on energy reduction during an event.
Could be voluntary reductions with an energy payment only.
Large C&I
Various load types, e.g.- HVAC, lighting, water heating, refrigeration, pumping, industrial process loads, refrigerated warehouses.
May also include load shifting to back up generators (BUGs).
Pricing Options Time varying rate offers (e.g., critical peak pricing), or discounts on electricity use reduction during critical peak periods (Peak Time Rebate); Can be offered either as opt-in or as default with opt-out.
All All
Behavioral DR Behavioral modifications in energy use during peak periods through targeted messaging and outreach efforts.
Residential Any
Behind The Meter (BTM) Battery Use of BTM batteries for load shifting and/or dispatch during peak demand periods.
All BTM batteries.
Electric Vehicle Charging Control
Auto-DR enabled Managed charging of electric vehicles for peak demand reduction.
All Electric vehicles.
Key Variables Description
Participation Rates
• Percentage of eligible customers enrolled in DR programs • Participation and ramp
Unit Impacts • kW reduction per device • Percent of enrolled load
Costs
and variable costs)
• Program Lifetime • Discount Rate
Calculate DR Achievable Potential
Number of customers; sq. ft. of floorspace; sector peak demand
Eligibility and other program-specific factors
Customer acceptance – varies over time as programs ramp up
Potential peak load reduction from DR
measures (either manual or technology-enabled)
120
121
Potential Results by Customer Class & Segment Examples
122
Step 5: Calculate Potential and Costs
Cost-Effectiveness Assessment Example
Behavioral DR 5.52 5.78 0.96 0.96
BTM Battery Control 19.17 22.91 0.84 0.73
EV Charging Control 3.05 11.08 0.28 0.26
Cost and Benefit Items TRC PAC
Program Development Cost Cost Cost
Program Administrative Cost Cost Cost
Program Delivery Cost Cost Cost
Marketing & Recruitment Cost Cost Cost
Technology Enablement Cost Cost Cost
O&M Cost Cost Cost
Incentives Transfer Cost
Avoided T&D Capacity Costs Benefit Benefit
Avoided Energy Purchases Benefit Benefit
Participant Cost Cost N/A
Annual Itemized Program Costs
al $
) Program Development Cost Program Delivery Cost Program Administrative Cost O&M Cost Incentives Marketing & Recruitment Cost Technology Enablement Cost
Supply Curves for DR Options
125
• Supply curves show how DR options stack in terms of increasing levelized costs vs. their contribution to the potential.
• Levelized costs include all programmatic costs, technology enablement costs, customer incentives, and operations & maintenance (O&M).
• Supply curves are useful for prioritizing DR options (least cost, highest contribution option).
• These feed into cost-effectiveness calculations.
Source: Guidehouse (formerly Navigant)
The Mechanics of Doing DR Potentials • Deciding on the level of rigor: How much is too much?
• DR potential study without primary data collection might take 6-12 weeks to complete (consultant cost might range from $50,000 to $150,000)
• DR potential study that includes primary data collection might require 4- 6 months to complete (consultant cost would range from $200,000 to $350,000)
• Models and tools • There are many tools and resources to draw from, ranging from simple
Excel spreadsheets to those built on platforms such as Analytica
126
• Characterize technologies that provide joint EE-DR benefits (e.g., smart thermostats, EMS).
• Consider EE-DR interactions in baseline load (DR baseline load after EE effects).
• Consider customer adoption of technologies from a joint EE-DR perspective.
• Establish a model and identify data needs to represent EE-DR interactions and jointly estimate EE and DR potential.
Integration with DER Technologies • DR potential assessment commonly includes DER technologies, such as EVs, behind-the-meter
(BTM) battery storage, and thermal storage.
• EV inputs for DR potential assessment include: EV adoption forecast Charging profiles Unit impacts from EV charging control DR enablement costs (charging control)
• Storage (BTM batteries and thermal storage) inputs for DR potential assessment include: Storage adoption forecast Charge and discharge profiles % load that could be shifted to storage technologies DR enablement costs (Balance of System Costs)
128
Source: Guidehouse (formerly Navigant), 2017 Distributed Energy Resources (DER) Potential Study, Prepared for Consolidated Edison Company of New York, Inc., http://documents.dps.ny.gov/public/Common/ViewDoc.aspx?DocRefId=%7B59359020-BA53-4C6D-A79F-C7B5E2BF4BAE%7D
• A set of programs that: o Serve various customer
segments o Achieve various business
objectives o Provide various operational
needs
• Enables holistic planning to meet multiple objectives
• Identifies mix of programs prior to detailed design of any one program
• Ensures programs do not overlap/cannibalize participants
131 131
Key Considerations in Deciding on a Mix of DR Programs
Regulatory drivers
system upgrades
metering
• Can one program address multiple goals?
• How many programs do you need? • Other considerations
• Interest/buy-in program internal stakeholders
• Interest in DR among customers • Existing systems and processes • Customer characteristics
eg, large industrial; high residential A/C usage
132
Steps to DR Portfolio Development Step 1: Identify needs (DR objectives) • Do you have a specific MW target? • What types of resource requirements do you need DR to address?
Step 2: Map needs to specific program attributes • What do programs need to provide to operations and planning? • How does this translate to program rules/capabilities?
Step 3: Develop high-level program concepts • Use identified DR “considerations” to shape program ideas • Target markets, major technology components, incentive type • Much less detailed than a “program design”
133
Step 1: Identify Needs
Defer T&D investments/ Address local network constraints
Reduce energy supply costs (via economic dispatch)
Provision of ancillary services (e.g. spinning reserves)
Support expansion of intermittent renewables
Reduce load during grid emergencies
Non-Operational Drivers Provide customer value and market
engagement
• “Reliability” DR programs are larger & more common than “economic” DR
• Fewer programs with lower MW have attributes that are able to provide energy or ancillary services
• Smallest subset of all might have attributes needed to provide regulation services
Step 2: Map Needs to Program Attributes
135
Required Characteristics of DR Programs
* Recent capacity needs in non-summer periods have led to increased interest in annual DR availability
DR Benefit Stream
DR Operational Characteristics
Notification/ Speed of Response
Up to 10 or more desirable 3-6 hours Typically hours
Deferral of T&D Upgrades
May be similar to Capacity Deferral; Specific to utility’s unique situation and the location on the system
Economic Dispatch
~30 minutes to several hours Minutes to hours
Balance Renewables All Year Potentially dozens or
even > 100/Yr < 1 hour < 10 minutes
Step 2: Map Needs to Program Attributes
Program Type Residential/Mass Market Large Commercial & Industrial
Direct Load Control (DLC) Seasonal or annual incentives to control A/C or other appliances Not applicable
C&I Curtailment/ Interruptible Rates Not applicable
Manual and/or Auto-DR enabled curtailment of Large C&I loads; can be either voluntary or require firm capacity reduction; Interruptible Rider: Rate discount to provide capacity.
Pricing Programs:
High-priced events to encourage voluntary reductions (may be technology enabled) offset by lower rates other times
Peak Time Rebate (PTR) Bill discounts based on voluntary reductions Not applicable
Demand Bidding/Buyback Not applicable Customer bids curtailment supply curve or responds to utility pricing
Behavioral Programs: Behavioral DR
Voluntary reductions in response to a utility/ISO request; often enabled via a mobile app Nascent programs starting
Specialty Programs: Ag pumping, EVs, Storage
These and similar programs could be considered part of the above program types, but are sometimes called out to focus on specific technologies or customer segments
DR Program Types
137
Example DR Program Portfolio
• Customer segments: Res, small C&I (Mass Market); medium & large C&I. • Incentive types: Curtailment incentive, pricing
Step 3: Program Concepts
138 Source: Guidehouse
Year 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027
Legacy Program, Continued
Manual DR Pilot
TOU Default CPP Opt In CPP Default
Legacy Program, Phase Out
TOU TOU Pilot Modified TOU Expanded TOU & CPP*
Legacy Program, Continued
* Expanded TOU & CPP contingent on advanced metering infrastructure (AMI) deployment for mass market customers Curtailment Program
Pricing Program
Other Electric Vehicle (EV) Interruptible Rate
Mass Market (MM)
Pool Pumps, Window A/C
Interruptible Rider
phased in over several years
• Some test concepts via pilots
• May phase out legacy programs
Source: Guidehouse
* Examples of “Other DR” programs include electric vehicles, industry- specific (e.g., agriculture), etc.
Growth Trajectory of the DR Portfolio • Existing programs
continue to grow, or phase out
• Some programs may reach maturity/ saturation quickly
• Pricing is contingent on AMI
500 MW portfolio
Step 3: Program Concepts
141 Source: Guidehouse
Xcel Energy Demand Response Program Portfolio
Residential Saver’s Switch Business Saver’s Switch AC Rewards Program (smart thermostat)
Peak Partner Rewards (C&I Curtailment) Critical Peak Pricing Interruptible (ISOC)
Residential Battery Demand Response Pilot Geo-Targeting Pilot Future considerations: EV Smart charging; Behavioral DR
Mass Market Direct Load Control Programs
Medium and Large C&I Programs
Demand Response Pilots
Map needs to program attributes with internal stakeholder input
Choose the number of programs and high-level program types considering all relevant factors
Plan and implement pilots if needed and as time allows
Proceed with detailed program design for entire portfolio
Integrated Distributed Energy Resources (IDER)
A program design type that delivers the benefits of multiple DERs to both customers and the grid using similar technology intervention and/or a linked incentive while leveraging complementary program delivery resources and infrastructure.
Electric Vehicles
Demand Response
Key Takeaways – DR Portfolio Development • The DR resource is more than the peak MW it provides; when and how those
curtailments are used are also important • The DR portfolio is a mix of programs that serves multiple customer segments
and utilizes several incentive mechanisms • Business and regulatory drivers of DR determine what type of DR programs and
portfolio is appropriate • Most DR portfolios are geared toward few, multi-hour capacity/reliability
events, with some economic dispatch • Use of DR for ancillary service (including integration of renewables) requires a
different mix of program characteristics—more, shorter events with fast response • The DR resource is itself part of a broader portfolio that include iDSM (energy
efficiency) and other DERs
Input from Outside
DR Portfolio Considerations
Source: Navigant 148
DR Pilot Program Development Process
Market Assessment • Start with knowledge of
customer types, size and type of load, load shapes by type and end-use.
• Identify the most attractive DR candidates by commercial building/industrial activity type
Technical Assessment • Conduct detailed site survey • Determine potential shed
amount and cost of shed • Determine load shed strategies • Conduct detailed assessment
for technology requirements
Pilot Operations • Set up and test functioning of
all technical components • Call DR events • Measure load shed amount • Evaluate pilot performance
149
Source: Navigant/PGE 150
Elements of the DR Program Design 1: Assess Target Market, Eligibility and Risk Factors
2: Develop Marketing/Recruitment Strategy
4: Develop Customer Incentive Strategies
5: Establish EM&V Plan
6: Develop Implementation Logistics
151
Element 1: Assess Target Market, Eligibility and Risk • Segmenting the market
• Matching customer load characteristics to system needs • Targeting end-uses and segments that show the most DR potential
• Setting goals • Targeted number of customers for each program and each program
component • Translating the expected numbers of participants into MW reductions
• Developing criteria for customer eligibility • For DLC programs, eligibility is based on equipment type (usually central A/C
units) • For Interruptible programs, customer size is typically a criteria and whether
there is an interval meter in place (usually customers > 200 kW)
• Identify the risks of deployment and evaluation; examples might include:
• Back-office and customer technologies • Communication path and cost • Customer adoption and fatigue • Legal
1: Assess Target Market, Eligibility and Risk Factors
Advanced Technology
(Critical peak pricing)
switch technology
~7000 Participants
~1000 Participants
~2000 customers on new rates
Target: ~10,000 total participants 153
Element 2: Develop Marketing/Recruitment Strategy
• Different marketing approaches should be considered for each DR program and even tailored for different customer types within a program
• Residential DR programs are typically marketed through bill inserts, print and social media, and advertising via television and internet
• Commercial and industrial customers typically require a more tailored marketing message, where one-on-one contact is essential
• DR customer messaging is a challenge as customers don’t typically distinguish DR from other aspects of efficiency
some industrial)
light Industrial, most agriculture)
Highly Specialized Target Marketing
155
Element 3: Specify Technology Standards & Choices • The decision about the level of customer
technology is very much a function of the program objectives
• DR programs where there is the predictability of the need for DR may not necessitate significant amounts of technology… so-called demand-side DR
• DR programs that necessitate dispatchability will invariably require some level of customer technology… so-called supply-side DR
• Interoperability standards such as OpenADR establish the criteria that can be used in DR program designs
Two Way Communication via Internet and/or AMI Network
157
Example 2: Technology Architecture of an OpenADR Program OpenADR provides a non-proprietary, open, standardized and secure demand response interface that allows communication of DR signals.
Source: OpenADR Alliance 158
Element 4: Develop Customer Incentive Strategies • Strategy must balance these factors:
• Cost-effectiveness • Regulatory compliance (including customer equity
concerns) • Customer acceptance
• Incentive approaches for mass market programs • One-time payment vs. recurring payment • Payments can be structured in different ways, depending
on the scope of services provided • Integration with other non-DR program incentives
• Incentive approaches for programs such as curtailable/interruptible or load aggregator programs
• One-time incentives/rebates ($/kW) for DR enablement equipment
• Capacity payment ($/kW-year) for committed and reserved kW (sometimes referred to as a reservation payment)
• Energy payment ($/kWh reduced) for measured energy reductions
Target: ~10,000 total participants
Noon to 5p summer 4p to 9p winter
• Free load switch • $5/mo
Incentive Approaches:
(Critical peak pricing)
switch technology
Step 1: Define Scope
• Determine budget • Assess measurement data infrastructure
Step 2: Specify Methodologies
Step 3: Plan for Data
• Identify data sources needed for evaluation • Develop plan to collect data at beginning stages of
implementation
Timeline
Element 6: Develop Implementation Logistics • Develop timeline for implementation • Set budget targets – either through top-
down (e.g., percent of revenue) or built from the bottom-up (e.g., number of FTEs, contractors, etc.)
• Budget components should include: • Program development (potential, program
design) • Program operations (marketing, recruitment,
sign-up, implementation, validation, EM&V)
163
Portfolio Design
Program Design
Program Implementation
• Objectives of the DR resources for the utility • T-stats role within the portfolio • Cannibalization of existing programs
• Technology equipment choices • Technology services • Eligibility • Incentives • Dynamic rates • Load control temp offset strategies
Example 1 (cont.)
switch program?
Will the utility end up running two systems to keep legacy switches?
Choose the t-stat vendor Multiple t-stat
models
Topic Switches Thermostats Comments
Installation If BYOT Need to get into house for thermostats; wiring incompatibilities
Cost If BYOT Thermostats more expensive and higher installation cost
Comfort/Opt-outs Comfort control and temperature thresholds with thermostats; switch customers do not notice or opt out
Customer Engagement
Energy Efficiency/ DR Convergence
Some energy efficiency savings due to programming; two-way and self-learning can increase and ensure savings
Reliability— Short Term
Similar if by paging; WiFi t-stats may go offline, but utilities know, and self-correcting; Switch failures over time
Reliability— Long Term
Switch inspections needed; paging networks in decline; will customers want a shiny new thermostat in five years? Wi-Fi?
(Source: Various, Compiled by Navigant)
Advantage to… Example 1 (cont.)
165
Program Design Example 2 – Fast DR for Large C&I Customers
Program Concept
• Large C&I customers required to respond to DR events within less than 10 minutes of notification time.
• Essential for participation in the Ancillary Services Market. • Participants need to have real-time data measurement and monitoring systems to
measure load reduction, say at 4 second intervals. • Extremely high level of availability required for participating loads
Targeted Customer Segments
Examples of controlled end-uses
• Any loads that are programmed to shed • Examples- shut-off of roof-top HVAC units; increasing temperature set points by 2
degrees or more; turning off process-related load; Lighting system switching/dimming.
166
• Participant has control over end-use equipment operation during DR event. • Auto-DR enables pre-programming of load reduction/shifting strategies in
response to DR events.
Technology Components
• OpenADR Virtual End Node (VEN) interface device • Facility EMCS • Two-way real-time IP-based data communication method • Telemetry Built into OpenADR2.0b communication for transmitting energy use
data • 15-minute interval meter for settlement
DR Event Notification Time
• Notification time is less than 10 minutes • Resources need to be made available for a certain specified duration (say 2
hours)
Customer Payment • Capacity payment for committed load (based on "available" hours) • Payment for the actual energy reduced • Strict non-compliance penalties
167
Event is transmitted to commercial or industrial customer
Non-critical loads are shed automatically
Controls are pre- programmed and automated to react quickly to event
Source: Global Energy Partners for Korea Electrotechnology Research Institute (KERI) 2010 168
DR Program Lifecycle Overview
Costs and Benefits of Demand Response TRC PAC RIM Participant
Administrative costs COST COST COST Avoided costs of supplying electricity BENEFIT BENEFIT BENEFIT Bill Increases COST Bill Reductions BENEFIT CAISO Market Participation Revenue BENEFIT BENEFIT BENEFIT Capital costs to LSE COST COST COST Capital costs to participant COST COST Environmental benefits BENEFIT Incentives paid COST COST BENEFIT Increased supply costs COST COST COST Market benefits BENEFIT BENEFIT BENEFIT Non-energy/monetary benefits BENEFIT BENEFIT Revenue gain from increased sales BENEFIT Revenue loss from reduced sales COST Tax Credits BENEFIT BENEFIT Transaction costs to participant COST COST Value of service lost COST COST
Source: California Public Utilities Commission, 2010 Demand Response Cost Effectiveness Protocols. Shaded rows indicate those costs and benefits which are not included in the CA Standard Practice Manual but have been added to these Demand Response protocols.
Primary DR Benefit
Source: Navigant/PGE 171
Constructing the DR Organization
• Single DR lead runs organization and is liaison with other departments • Multiple program managers • Support team reports either to DR lead or to program managers • Other positions and reporting structures may also be effective
Source: Navigant
Demand Response Manager
operations, policy development. • Responsible for coordinating across all
functional groups.
Operations Manager
• Coordination with all operational stakeholders (e.g., system operators, account managers) on day before, day of and day after events.
• DR requires seamless implementation of many steps across diverse functions
• DR operations requires focused oversight of complex processes
Program Manager
stakeholder management, Quality Assurance, and risk management
• DR resources are a result of multi-faceted programs that run from customer engagement to control center operations
• Success requires the focused accountability for each program in the portfolio
Source: Navigant
Marketing Manager
• Market research and develop a marketing plan to engage customers in DR participation
• DR requires changes in operations or behavior for a limited number of hours and events per year.
• An effective marketing plan describing DR’s unique attributes as well as customer benefits requires the dedicated activities of a marketing specialist.
• May oversee outsourced marketing functions.
Business Analyst
• Maintain and update database for notification of customers who enroll in programs.
• Produce regular and ad hoc reports for program results, budget management, key metrics and performance indicators.
• Program management requires extensive data reporting, from customer enrollments and drop outs, to ongoing MW reductions, to budget and quality metrics.
M & V Analyst
• Design, manage and coordinate complex evaluations of the DR portfolio
• Estimate load impacts • May oversee outside evaluators
• All stakeholders must have confidence that DR programs are delivering real, measurable demand reductions.
• Programs require an expert analyst to independently verify peak load reductions using widely accepted, industry standard methods.
Position Descriptions and Purpose
Example of FTE Growth to Develop a 500 MW DR Portfolio
Reports
ownership of various program components
Turnkey, Full Service Utility fully outsources all components of
DR program implementation 177 177
Guaranteed Load
Source: EnerNOC 178
Expansion into higher-value ancillary services markets
Diversification into other energy services (including renewable integration)
Increased value of DR due to carbon pricing & renewables
Downward pressure on provider’s margin of curtailment payments
Continued depression of the price for capacity
Uncertain regulatory policies
Impact of New Technologies
High Medium Low Low Medium High Risk of Failure Opportunity for Success
DR Business Risks and Mitigations » Risks and Opportunities of Strategic Outsourcing
Utility control over the dispatch flexibility
179
•Prequalify vendors through RFQ process − Vendor roster(s) could be segmented based on DR program type
• Issue RFPs to qualified vendors − Indicate very specific tasks and expectations − Consider outlining acceptable business models − Define key performance parameters and specify how they will be
measured − Specify preferred method for load impact settlement
•Select vendor(s) •Seek necessary regulatory approvals •Launch DR program(s)
Outsourcing Process
Evaluation
• To assess the merits of the program • Ex Post: To uncover the principles
underlying successful program performance
• Ex Ante: To forecast the future effectiveness under different scenarios
Program Process
• To understand whether and how well objectives are being fulfilled
• To determine the reasons for successes and failures and find gaps and develop recommendations for program design, implementation and optimization
Program Evaluation
Ex Ante
Ex Post
184
• What is the value for capacity markets?
• Is there value for ratepayer investment?
• Are we getting what we planned?
Program Evaluation
Cost Effectiveness
Forecasting
185
• Regression is the better approach, where feasible (data and time limitations)
• Settlement estimate is typically unbiased, but highly variable at an individual customer level
Ex ante versus Ex post Load Impact Evaluation
186
Ex Post Evaluation • Estimation of load impacts after events occurred • Tells us “What was the load impact during a specific
event?” • May be performed using any of the baseline
methods
Ex Ante Evaluation • Estimation of load impacts prior to calling events • Tells us “What do we think the impact to be when
we call an event?” • Best done using regression models with predictive
parameters • Ex-post regression coefficients tell us how load
impact is affected by temperature, hour, month, cycling strategy, etc.
• We select values of interest for these inputs • The output is a forecast of the load impact
Ex Post Evaluation Methodology In most instances, ex post impacts are estimated by comparing the reference level energy use in each hour with actual energy use in the hour on each event day.
Load Impact = Estimated
Ex Ante Evaluation Methodology • Forward-looking estimate under
specific conditions • Reflects the load reduction capability
of a DR resource under a standard set of conditions that match the market and system conditions that drive the need for additional capacity: 1-in-2 and 1-in-10 system peaking conditions.
• Analysis of historical program data is leveraged to produce ex-ante load impact estimates that are subsequently employed for resource adequacy, cost-effectiveness assessment and, by connection, resource planning.
188
Ex Ante Load Impact Projections by Weather Scenario for Peak Day
Source: Opinion Dynamics
What Are the Similarities? Ex Post and Ex Ante • Ex Post estimates what happened in the past • Ex ante forecasts the future effects • Both are informed by actual past performance and estimates of future
conditions • Both can use similar statistical techniques to calculate the parameter
estimates, e.g., econometric time-series analysis
189
»There are 3 overarching approaches
Estimating Load Impacts of DR
191
•Must reduce demand to agreed upon level
•Primarily for interruptible rates Rarely used for new DR programs Does not matter what load was prior to event
“X of Y days” Approach
•Compare to prior days, same hours •For “settlement” (incentive payment):
•Easily understandable/transparent •Can be quickly calculated
Econometric Regression Modeling
•For grid management planning and regulatory reporting
•Uses event and non-event day load data and parameters including: •Time of day, Day of week, Month
Temperature •May instead use DR participants as a
control by not calling all participants •Adjustment for same-day load prior to
the event
» Interval consumption data › At least hourly › More granular helpful for spinning
reserves/ancillary services and snapback assessment
› Many programs provide run-time from device telemetry data
› Device telemetry data collection will vary based on type of device – one-way, two-way, or smart
» Weather—temperature and humidity at similar intervals as consumption or run-time data
» Event start/end and control strategy
Data Needs for Load Impact Estimation
192Source: Opinion Dynamics
193
AMI/Interval Data
Produces load impacts net of takeback (e.g., running fans or increased run time for refrigeration, process cooling, etc.) Is essential for non-dispatchable programs (e.g., dynamic pricing, CPP, etc.)
Typically a census (all participants)
Telemetry Data
Produces load impacts gross of takeback
Residential A/C control typically has required data logging of A/C compressor
Careful! This data will require a conversion based on full load of HVAC units to get to kW
Data collection dependent on type (e.g., one-way/two-way)
Can be a sample (if one-way)
194 Source: Opinion Dynamics
(“adaptive”?) •Season/end-uses •Outdoor temperature •Hours of the day
Consistent Approach
•Insulation of homes •Customer Behavior
Example: Ex Post Impacts from C&I • C&I evaluation specific to
individual customers • A few large customers often
account for majority of curtailment
Source: Opinion Dynamics 195195
80%
% Savings
45%
196
• Regression is the better approach, where feasible (data and time limitations)
• Settlement estimate is typically unbiased, but highly variable at an individual customer level
• Performance as compared to nominated capacity can be highly variable
Program
197
RCTs vs. Proxy Days for DR Baseline Development • Focus on DR impacts using two approaches:
• Compare results by approach to test accuracy and bias of assessing impacts • Randomized Controlled Trial (RCT) approach for DR events by randomly assigning participants to
treatment and control status for each event • An RCT eliminates bias from self-selection by participants and from different comparison days • Distinct from typical approach that uses non-event weather days to serve as reference load • Results:
• Matching biases impact estimates downward • Well-matched weather day results are closer to RCT results than poorly matched days
Research Design Matching Approach Modeling Approach Comparison
Experimental Random Assignment Difference Event Day
Quasi-Experimental Mahalanobis Distance Day Matching Linear Fixed Effects Regression Similar Day
A Closer Look at Reference Load Average RCT Reference vs Modeled Reference Event Day Usage
Weather Matters! Average Summer Ex Post Demand Response Event Impacts (Well Matched)
Cool Days Underestimate Reference Load Average Summer Ex Post Demand Response Event Impacts (Poorly Matched)
Events are typically called on hottest days, so non-event days will be cooler leaving cooler days with lower demand as comparison days
in times of COVID
202
Hourly Load Differences Case Study: Impact of COVID on average hourly load between 2019 and 2020 (excluding event days)
These graphs represent a subset of 43 customers who participated in both 2019 and 2020 event seasons.
Hourly load is generally lower on weekdays in
2020
Baseline Accuracy Differences Case Study: Baseline sensitivity analysis using proxy days to test baseline prediction accuracy for C&I Aggregator Program
Ideal performance is 100%, where the baseline succeeds in predicting the actual
-100%
0%
100%
200%
300%
400%
500%
600%
700%
-900% -800% -700% -600% -500% -400% -300% -200% -100% 0% 100% 200% 300%
20 20
B as
el in
e Ac
cu ra
Baseline 5/10 Baseline 4/5 Baseline 10/10 Linear (Baseline 5/10) Linear (Baseline 4/5) Linear (Baseline 10/10)
This study was conducted on a subset of 64 customers for whom we had both 2019 and 2020 data. All three types of baselines have a symmetric same-day adjustment. The percent in the graph is the baseline divided by actual demand on proxy days.
• Baseline accuracy is far more variable for 2020 than 2019.
• 10/10 baseline does a slightly better job than the others
Case Study: Baselines in an Integrated World
205
Integrated DR 3.0 Requires an Evolution in Baseline Considerations • With integrated technologies,
increasing adoption of DERs, and more DR value streams, baselines are increasingly complex
• Value Streams: Shape, Shift, Shed, Shimmy
• Technology Applications: GEBs, NWS, Electrification, Solar + Storage, Home EV Charging
• Program Interventions: DR, EV, DG, EE, rates
• Reflect continuous or overlapping systems
206Source: Opinion Dynamics & LBNL
Case Study: Smart Thermostat with EE+DR Customer Engagement Strategy
Case Study: Smart Thermostat with EE+DR Baseline Development Strategy
Case Study: CPP + TOU • Rate-based DR programs may produce
spillover effects, which provoke challenges in developing appropriate baselines for load impact estimates
• Price signals may lead participants to lower consumption not just on event days, but also on non-event days due to risk aversion, habituation or other factors
• Spillover can artificially depress reference loads for impact analysis, potentially leading to lower impact estimates
Update program theory, document program activities, and compare program operation with theory.
Assess and make recommendations on program documentation and processes needed to support demand savings estimation.
Investigate barriers to program participation and other obstacles to program implementation.
Evaluate how the program meets the portfolio needs.
Recommend improvements and solicit implementer feedback on usefulness of the recommendations.
Case Study: C&I Program Evaluation Task Description
Program Materials and Database Review
Analyzed the DR program databases to characterize and understand the participant population
Program Staff Interviews Conducted telephone or in-person interviews with program staff from each of the three IOUs
Program Stakeholder Interviews Conducted telephone interviews with program stakeholders. These stakeholders include key account representatives, vendors, program verification engineers and DR program staff
Participant Survey Fielded telephone survey to program participants with contact information, resulting in maximum completes
Participant Interviews Conducted telephone interviews with participants
Industry Manufacturing (5 of 10) Manufacturing (10 of 10)
Customers Represented by Sites
Single Site (7 of 8) Multiple Sites (7 of 10)
Dually Enrolled BIP (10 of 10) BIP (10 of 10)
Auto-DR Enabled 1 of 10 are enabled 4 of 10 are enabled Size Large Large Avg. # of Events Bid 5 of 9 events 12 of 13 events
Avg. % of Baseline Reduced (Hourly kWh)
56% (range 28% and 99.9%)
71% (range 36% and 99%)
Avg. Incentive Earned per event
$16,600 (range $300 and $69,000)
$32,000 (range $11,000 and $62,000)
The Majority of Load Reduction is Concentrated Among a Small Number of Participants
Utility 1
80%
% Savings
45%
7% 14% 7%
Multiple Sites 38% (n=85)
Yes 86%
No 14%
for Bid
Average Bid 518
Average Reduction 630
58% 57%
29% 7%
during Demand Bidding Event
Have kW Reduction Goal
Large Customer
Top 2 Barriers
6.4
5.5
Knowledge
The process for participating in Demand Bidding events
overall (n=39)
How to submit, adjust or withdraw a bid for an
upcoming event (n=39)
How to view results of your participation after an event
(n=38)
Arrows indicate differences between active and dormant participants. *Based a review of program databases
214
Correlated barriers responses to likelihood to participate in future events
Reduce these barriers to
events
PG&E barriers to participating in events: Active and dormant respondents
Shutting down or reducing your production and/or service schedule ,
6.4
Loss of revenue due to shutting down equipment, 4.9
Your facility’s operating hours, 4.5
Your facility’s ability to adjust production or service schedules, 5.4
Your facility’s product or service, 5.1
Employee comfort during events, 3.1
The time required to participate in events, 4.6
The amount of manual effort
required to participate in events, 3.8
Lack of support from utility
staff/customer relationship
managers, 1.7
My company is often unaware of Demand Bidding Program events ,
1.6
events, 0.7
Your facility’s product or service, 6.6
The current state of the economy, 3.9
The amount of load reduction needed to meet bid is difficult to
understand, 4.5
managers, 2.7
Large Barrier, Strong Correlation with Interest in Participating
Large Barrier, Weak Correlation with Interest in Participating
Small Barrier, Strong Correlation with Interest in Participating
Small Barrier, Weak Correlation with Interest in Participating
Average size of the barrier on a scale from 0 to 10, where 10 is a big obstacle"
Co rre
la tio
n Be
tw ee
n Ba
rri er
a nd
In te
re st
in P
ar tic
ip at
in g
in F
ut ur
e Ev
en ts
e)
Barriers highlighted in blue indicate that they may be adressed through program design changes
B
C
A
D
217
2% 27%
Reactions to Event Duration
% Overrode Temperature Settings 30% (n=2,348)
Reactions to Change in Indoor Temperature During Events
17%
35%
43%
Post-Event Season Survey • Sources of program awareness and motivation for
participation;
• Experiences with program enrollment and device registration;
• Satisfaction with various program processes, the program overall, and utility in general;
• Demand-response event participation (awareness, experience, and satisfaction);
• Participation in energy optimization component (awareness, experience, and satisfaction); and
• Sociodemographic and household characteristics
Case Study: Why do customers opt-out? • Participant Comfort
• Participants who prefer cooler homes tend to opt-out • Participants in homes with less thermal integrity tend to opt-out
• Thermostat Engagement • Frequent engagement with the thermostat is correlated with opt-out behavior
• Household Occupancy Patterns • Participants who always opt-out are more likely to be at home • Participants who opt-out blame other household members who were home at
the time • Underscores importance of customer engagement and understanding
program participation motivations (process evaluation)
Source: Opinion Dynamics 218
219
What Have We Learned? • What is Demand Response? Why is it needed? • How to communicate a common, descriptive language for describing
the evolution of demand response as seen through the eyes of industry leaders
• The current state of the fundamental market principles and technology enabling demand response initiatives today
• Examples of portfolio design strategies and implementation tactics to address the challenges of a fading boundary between bulk power exchange, retail electricity usage, and new market entrants
• How to compare and contrast demand response program performance • Demand response is complicated, innovative, but exciting!!
220
www.peakload.org
Leadership White Paper • Sign in at www.peakload.org/evolution-alumni for recording and
downloads. Continue the conversation at DR Evolution Training Alumni Conversation Circle
• Note: Access is limited to individuals who have taken or instructed this course
• Introduction to Demand Response Fundamentals, online
Visit www.peakload.org/demand-response-training
• Please participate in the PLMA interest groups, Dialogues, additional training, conferences, and reach out to PLMA members!
• Visit the PLMA Resource Directory at www.peakload.org/page/ResourceDirectory
Publisher Information: Demand Response: Fundamentals and Evolution
8th Edition (October 2020) Copyright © 2020 by Peak Load Management Alliance
All rights reserved. No part of this presentation may be reprinted or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording or by any information storage or retrieval system, without prior written permission of
the copyright owner and publisher.
Trademarked names may be used in this book. Rather than use a trademark symbol with every occurrence of a trademarked name, we use the names only in editorial fashion and to the benefit of the trademark owner with no intention of infringement of the trademark.
The information in this book is distributed on an “as is” basis, without warranty. Although every precaution has been taken in the preparation of this work, neither the authors nor the publisher shall have any liability to any person or entity with respect to any loss
or damage caused or alleged to be caused either directly or indirectly by the information contained herein.
Acknowledgements PLMA would like to recognize its Executive Director, the Officers and Executive Committee, and the Education Planning Group,
Training Partners, and Training Volunteers for oversight and editorial review of the Demand Response Training Series.
225
Reductions in load could allow generators time to ramp up
Increase in load could allow generators time to ramp down
226
Load met by Wind
Surplus Wind
Generation
Demand response and energy storage are sources of power system flexibility that increase the alignment between renewable energy generation and demand.
• Demand Response provides a means to shift demand to times of relatively high wind generation and low load
• Storage Technologies can store excess wind generation for use in times of relatively low wind generation and high load
This figure also shows how more flexible generation could accommodate increased RE penetration and can provide an alternative or supplement to DR and storage.
Source: GreeningtheGrid.org/ National Renewable Energy Laboratory Sept. 2015 227
DSM
DR is Part of Integrated Demand Side Management (IDSM)
A program design type that delivers the benefits of EE to customers and DR to the grid using the same technology intervention and/or a linked incentive while leveraging the same program delivery resources and infrastructure.
228
Evolution of Demand Response to Distributed Energy Resources: Fundamentals and Path Forward
Welcome, Safety Moment
What is Demand Response?
“Simple” DR & EE Definitions
Why Utilities Need Demand Response
The Economic Definition of DR
Demand Response in the Airline Industry
Demand Response in the Utility Industry
Slide Number 14
DR Technology Overview
DR Technology Evolution
Communication Standards
CIM Domains
Cellular Network Standards
IEEE Wireless Standards
Federal and State Regulators
Slide Number 63
Generation – Wholesale Process
Equilibrium Pricing
Price “Inversion”
FERC Order 2000: RTO Features (1999)
FERC Order 888: ISO Features (1996)
Slide Number 76
Wholesale Market “Actors”
Slide Number 80
Slide Number 81
Slide Number 83
Slide Number 84
Slide Number 85
Slide Number 86
Slide Number 87
Who Can Manage Demand or Provide Services?
Energy Basics for DR
Demand Response Load Sources
Day-Ahead vs. Real-Time Participation
Demand Response Event Timing
Baseline: Historical Data Analysis
Performance Evaluation Models
Baseline: Morning Adjustment
Baseline: Calculated Reduction
Baseline Method Determines Payment
DR Potential & Portfolio Development
Representative DR Potential Analysis Framework
Typical Market Segmentation
Typical DR Options
Key Assumptions Development
Potential Results by Customer Class & Segment Examples
Cost-Effectiveness Assessment Example
The Mechanics of Doing DR Potentials
EE-DR Potential Integration Areas
Integration with DER Technologies
DR Program Portfolio Development
Key Considerations in Deciding on a Mix of DR Programs
Steps to DR Portfolio Development
What Do You Want Your Programs to Achieve?
Most DR Programs Address Capacity and Little More
Required Characteristics of DR Programs
DR Program Types
Growth Trajectory of the DR Portfolio
Example of Utility Program Portfolio
Trajectory of a DR Portfolio Design
Slide Number 143
ProgramDesign
Program Design Example: C&I DR Curtailment
Elements of the DR Program Design
Element 1: Assess Target Market, Eligibility and Risk
Customer Segments and Offerings: Residential Example
Element 2: Develop Marketing/Recruitment Strategy
Target Marketing Approach
Example 1:Technology Simple Architecture of a DR Program
Example 2: Technology Architecture of an OpenADR Program
Element 4: Develop Customer Incentive Strategies
Residential Example with Incentives
Element 6: Develop Implementation Logistics
Program Design Example 1 – Smart Thermostats
Example 1 (cont.)
Example 1 (cont.)
Program Design Example 2 – Fast DR for Large C&I Customers
Example 2 (cont.)
Example 2 (cont.)
Evaluation Feeds Program Redesign: C&I DR Curtailment
ProgramImplementation
Constructing the DR Organization
Example of FTE Growth to Develop a 500 MW DR Portfolio
Range of Outsourcing Services
DR Business Risks and Mitigations »
Outsourcing Process
Ex Post Evaluation Methodology
Ex Ante Evaluation Methodology
What Are the Similarities?
Load Impact Evaluation Methodology
Data Needs for Load Impact Estimation
Your Data Source Can Measure Different Results
Example: Ex Post Impacts from Residential Control
Example: Ex Post Impacts from C&I
Regression vs. Settlement Baselines
Case Study: Ex Post Baseline Development for Smart Thermostat Program
RCTs vs. Proxy Days for DR Baseline Development
A Closer Look at Reference Load
Weather Matters!
Cool Days Underestimate Reference Load
Case Study: Ex Post Baseline Development for Aggregator Program in times of COVID
Hourly Load Differences
Baseline Accuracy Differences
Integrated DR 3.0 Requires an Evolution in Baseline Considerations
Case Study: Smart Thermostat with EE+DR
Case Study: Smart Thermostat with EE+DR
Case Study: CPP + TOU
Case Study: C&I Program
The Majority of Load Reduction is Concentrated Among a Small Number of Participants
Slide Number 214
Correlated barriers responses to likelihood to participate in future events
PG&E barriers to participating in events: Active and dormant respondents
Case Study: Smart Thermostat Event Experience
Case Study: Why do customers opt-out?
Closing Remarks and Where to Learn More
What Have We Learned?
For more info:
Renewables Growth Creates DR Opportunities
DR, Storage and Renewables Integration
Slide Number 228

Documents

Demand Response: Fundamentals and Evolution