Load Shift Working Group OCTOBER 24, 2018 10AM 3:30PM PST CPUC - - PowerPoint PPT Presentation

Load Shift Working Group

OCTOBER 24, 2018 10AM – 3:30PM PST CPUC COURTYARD ROOM

https://gridworks.org/initiatives/load-shift-working-group/

Agenda

10:00AM -10:30 AM: Intros, Updates, and Purpose

▪ Introductions ▪ DR Regulatory Updates ▪ Re-Cap on Homework Assignments

▪ RA ▪ GHG emissions ▪ Feedback on Proposals Presented to Date (key changes)

▪ Today’s Objective: Refine our thinking on considering additional products and comparing all products.

10:30AM – 11:30 AM: MIDAS Product Proposal

▪ Rick Aslin (PG&E), Michael Lee (Evolve Energy), Henry Richardson (WattTime), and Erik Woychik (Strategy

Integration).

11:30AM – 12:00 PM: Sunrun Product Proposals

▪ Steven Rymsha (Sunrun)

https://gridworks.org/initiatives/load-shift-working-group/

Agenda

12:00 - 1:00 PM: Lunch 1:00 - 2:00 PM: Pay for Load Shape Product Proposal

▪ Peter Alstone (LBNL/ Schatz Energy Research Center/Humboldt State University)

2:00 -3:00 PM: Comparing Product Proposals

▪ Matthew Tisdale (Gridworks) product comparison matrix

3:00-3:30 PM: Next Steps

▪ Final Report Timeline ▪ Update on Future Sessions

https://gridworks.org/initiatives/load-shift-working-group/

Introduction and Purpose

Introduction: Roll call

DR Regulatory Updates Today’s Objective:

▪ Refine our thinking on considering additional products and comparing all products.

https://gridworks.org/initiatives/load-shift-working-group/

Resource Adequacy Takeaways (Draft)

The following takeaways are Gridworks interpretation of facilitated conversations. They do not represent consensus positions. They will be refined through drafting of the final report.

• 1. Current Resource Adequacy construct would capture some of the capacity value possible

through load shift, but not all:

• Recognizes: the capacity value of load shed
• Does not recognize: reducing the downward ramp, raising minimum net load, flexible RA provided without also providing

System/Local RA

• 2. Recommendations for changing the Resource Adequacy Construct:
• Unbundling Flexible and System/Local RA, allowing load shift to provide Flexible RA
• Recognize a new value for downward ramp as a part of Flexible RA
• Recognize a new value for raising minimum load

https://gridworks.org/initiatives/load-shift-working-group/

https://gridworks.org/initiatives/load-shift-working-group/

RA Subgroup Takeaways (Draft)

• 3. Consider recommendations exogeneous to changing the RA construct, including:
• Recognizing the value of avoiding renewable curtailment
• Allowing customers to benefit from negative- or low-price energy
• Reduction negative energy prices and commiserate pressure on capacity prices paid to keep critical generation resources

financially viable (as necessary)

• Impact customer choices on how they adopt/use new assets
• Locational value to the distribution system (Distribution Resource Plans/Integration of Distributed Energy Resources constructs).
• 4. Energy prices may not be enough to induce load shift behavior; monetization of some of the

capacity values could be the difference between achieving load shift or not.

• 5. Performance requirements (e.g., telemetry, response time, response duration) on the

providers of load shift. Those performance requirements may vary by the capacity service being provided

https://gridworks.org/initiatives/load-shift-working-group/

https://gridworks.org/initiatives/load-shift-working-group/

GHG Emissions

Update from Anja Gilbert & Ted Ko

https://gridworks.org/initiatives/load-shift-working-group/

https://gridworks.org/initiatives/load-shift-working-group/

Product Proposals

Update from Nora Sherif (CLECA) and Jennifer Chamberlin (Cpower)

https://gridworks.org/initiatives/load-shift-working-group/

https://gridworks.org/initiatives/load-shift-working-group/

Product Proposal Comparison Preview

https://gridworks.org/initiatives/load-shift-working-group/

https://gridworks.org/initiatives/load-shift-working-group/

10

A B C D E F G H Dispatch Method Dispatch Geo- Granularity Negative Pricing? Transaction Settles at... Performance Evaluation Role of IOU Role of Third-party Aggregators 1 PDR LSR Market Sublap Yes Aggregated Resource Metered + Typical-Use Adjustment Support Rule 24; LSE Aggregate, Bid, Coordinate, Settle 2 LSR 2.0 Market Sublap Yes Aggregated Resource Meter + Typical-Use Adjustment Support Rule 24; LSE Aggregate, Bid, Coordinate, Settle 3 CCP Distribution Operator Signal Based on Market Proxy (day-ahead price) Nodal Partial Premise Baseline Program Administrator Potentially Limited 4 MIDAS 5 Sunrun Integrated 6 Sunrun Informed 7 P4LS

MIDAS

MARKET INFORMED DEMAND AUTOMATION SERVICE

PRESENTATION FOR OCTOBER 2018 LOAD SHIFT WORKING GROUP (LSWG) BY REAL TIME PRICING BREAKOUT TEAM (FROM JULY LSWG MEETING): EVOLVE ENERGY, WATTTIME, CPOWER, WILLDAN , PG&E For LSWG Discussion Purposes Only

PRESENTATION AGENDA

❖Introductions, Background and Purpose of Presentation ❖Overview of MIDAS Product ❖Examples of Current MIDAS-like Products ❖Evolve Energy/ERCOT ❖WattTime ❖Critique and Discussion ❖Suggested Readings ❖Summary and Action Items

INTRODUCTIONS, BACKGROUND AND PURPOSE OF PRESENTATION

❖ Introductions ❖ Evolve Energy – Michael Lee ❖ WattTime – Henry Richardson ❖ CPower – Jennifer Chamberlin ❖ WillDan – Eric Woychik ❖ PG&E – Richard Aslin ❖ Background – MIDAS (Market Informed Demand Automation Service) types of products was surfaced in the July LSWG ideation session. LSWG facilitator requested a presentation

• n the MIDAS product for the October LSWG meeting

❖ Purpose – There is a growing appreciation that a “market informed” product could compliment “market integrated” products in helping the State achieve its goal of a 100% non-emitting electric resource portfolio by 2050. Today’s presentation could provide a framework to pilot a market informed load shift product in future Demand Response Program funding applications.

OVERVIEW OF MIDAS PRODUCT

MIDAS encompasses a variety of potential demand automation services deployed by vendors utilizing either a market or grid state informed signal that is acted upon by a controller connected to an end-use. Time granularity of market or grid state informed signals can be as low as 5 minutes but can be forecast for planning purposes for as long as 30-days. Locational granularity of market or grid state informed signals can be as low as a CAISO pricing node or a distribution feeder but can be aggregated or adjusted to meet a variety of use cases. Subscription based business model wherein customer pays vendor for providing signal, automation devices and control API in exchange for vendor providing economic and/or environmental benefits of equal or greater value to the customer.

MIDAS Product Features Technology neutral? Both the market informed signal and the control algorithm can be customized for service to be provided, location on grid, “duty cycle” of end-use technology and customer preferences Required to be Energy Neutral ? No Market Integrated/ Dispatchable by CAISO? No Grid Needs the Product Solves for Can be customized could be system GHG, local air emissions, system or local generation capacity, local T or D congestion management, renewables curtailment Dispatch Granularity (location/time) Location granularity could be as low as distribution transformer bank. Time granularity could be as low as 5 minutes.

MIDAS Business Model Customer

Customer is responsible for providing the end-use that can be controlled and for providing a set of preferences regarding how the device is controlled.

Demand Response Provider (DRP)

DRP provides or purchases the signal and provides or purchases the API working with the signal providers and the controller OEMs. The DRP develops the “program” offerings and recruits and services the customers.

Load Serving Entity (LSE)

LSEs can play the role of DRP or they can fund a third-party DRP or they can be passive. Regardless of the role played by the LSE, the LSE should be aware of the MIDAS “program” and should include the expected impacts

• f the MIDAS program in their planning and operational forecasts.

Distribution Utility (UDC)

UDC can play various roles. It could inform/modify the signal based on local capacity constraints. It could institute the program directly in an area for local needs. etc.

Independent System Operator (ISO)

ISO is essential in providing the real time price or grid state indicators at the right level of time and locational granularity that is appropriate.

ENERGY EMPOWERING TODAY’S ENERGY CONSUMER

Michael Lee, Samit Shah

EVOLVE’S MARKET VISION

Market Need

Supply has traditionally followed demand. Now demand needs to follow supply. As more inflexible renewable energy comes online, we need demand that adapts to local conditions on a real- time basis. Consumers should be compensated for providing this grid flexibility – reflected in lower prices if they load shift. This transparency is clean, simple, and easy. Complex compensation formulas constructed are almost always out-of-date as soon as they are finalized. Real-time information is the most accurate source.

Vision

3rd parties compete on UI/UX, advanced algorithms, and cost-savings vs. cost charged. The company with the most customer-focused product (usability, cost/value, carbon savings, etc.) will win in the future market. Through load shifting we can increase the value of all renewables and incentivize more to come online.

ABOUT EVOLVE

Customer-Centric Retail Energy Provider in Texas

We sell energy at cost – we don’t make money on energy. We charge $10/month and we need to save at least this amount to keep our customer relationship. We are the robo-advisor for energy.

Management of IOT Devices

We autonomously load-shift from peak prices to low prices. Our machine-learning profiles are custom to each user’s preferences.

We Respond to Wholesale Pricing Signals

• Although distribution grid can be congested, this is often highly correlated with wholesale prices
• Wholesale prices often correlated with carbon intensity
• Load shifting a customer’s usage away from $1,000+/MWh power is in everyone’s interest

THE MARKET IS BROKEN

CONSUMER PRICING IS DISCONNECTED FROM WHOLESALE

Consumer pricing is a relic from when we had mostly coal on our grid. Today’s flat pricing is not representative of costs.

$/MWh

OUR SOLUTION

Deliver power at real-time wholesale prices to consumers. Instantly save consumers 30%+ off their energy costs.

Real Time Prices For Consumers

PART 1

$/MWh

OUR SOLUTION

KW

Deliver power at real-time wholesale prices to consumers. Instantly save consumers 30%+ off their energy costs.

Real Time Prices For Consumers

PART 1

With a financial incentive to shift use during expensive times, our technology manages usage autonomously.

Shift Usage Away From Peak Hours

PART 2

$/MWh

OUR SOLUTION

Real Time Prices Only

• 30%
• vs. hedged pricing

Deliver power at real-time wholesale prices to consumers. Instantly save consumers 30%+ off their energy costs.

Real Time Prices For Consumers

PART 1

With a financial incentive to shift use during expensive times, our technology manages usage autonomously.

Shift Usage Away From Peak Hours

PART 2 Flexible Loads: IOT APIs Inflexible Loads: EVOLVE BATTERY

KW $/MWh

Real Time w/ Load Shift

• 52%
• vs. hedged pricing

CUSTOMER FOCUSED PRODUCT

Easy to access, easy to use

Built to work across platforms and allows customers an accessible way to control their energy.

Incentive to reduce use

Most energy companies have a disincentive for customers to save money. We are passing through costs, so customers are incentivized to reduce their use.

Set it and forget it

Customers can connect their IOT devices and set their preferences easily and not have to constantly manage their shifting.

Questions?

Michael Lee mlee@evolveenergy.co

AUTOMATED EMISSIONS REDUCTION

October 2018

27

WattTime

Mission

To give people the power to choose cleaner electricity WattTime unlocks the power to choose the cleanest energy by accurately measuring the moment-to-moment changes in carbon emissions on our electricity system—presenting a host of new opportunities to mitigate the climate impact of

• ur electricity use through more informed renewables procurement and by

automatically and effortlessly shifting load.

Approach

COMBINING AER AND DEMAND RESPONSE

29

LMP Price Frequency

Price of electricity (In dollars per megawatt-hour, Northern California, 2017)

Price [$/MWh] 5-Minute Segment, Sorted by Price

30

Combining AER & Demand Response

Demand response target Automated emissions reduction opportunity

Price [$/MWh]

Automated Emissions Reduction

32

Summary

What is AER?

WattTime, a technology nonprofit, has developed a fundamentally new approach to significantly reduce emissions from power plants using software known as Automated Emissions Reduction (AER).

Demands & Technology Create an Opportunity AER value across multiple sectors

AER enables internet-enabled, electricity consuming devices to seamlessly reduce emissions by combining:

• real-time grid data on power plant emissions, and
• internet-enabled control of electricity-consuming devices using new comfort and cost

algorithms

• With 23 billion “smart” devices expected worldwide by 2020, a rapidly growing share
• f electricity consumption is capable of supporting AER
• Current-generation AER has the capability to reduce CO2 emissions by the equivalent
• f 1 million cars
• As technology matures, impacts per device will grow
• AER offers institutional and residential energy users a new source of rapid, low-cost

emissions reductions

• AER also offers ancillary benefits to numerous other energy sector actors
• Strong potential for new entrants and business models

33

Sub-Regional Signal Availability

Geographic Specificity

• Signal aligns with ISO control areas
• Marginal emissions of plants serving the

region

34

The Problem: Grid Emissions Vary By Time

The marginal power plant that reacts when you flip a switch is always changing. Power customers don’t know in real time how dirty their power is.

A dirty time on the

• grid. Using electricity

at this time causes more carbon emissions. A clean time on the

• grid. Using electricity

at this time causes fewer carbon emissions.

35

Emissions Reductions through Timing

• Much electricity use is at least

partially flexible in time

• E.g. devices with compressor cycles

can sync cycles to cleaner moments

Normal operation Emissions-optimized Example: fridge cycles

36

AER & Demand Response Program Design

Program Sketch:

• Participants enroll wifi-enabled, smart devices in program

AER/DR Program

• During infrequent peak events (defined by the LSE), devices

automatically shift to a low energy operation

• Load shifts out of peak event period
• If program enrolled in market mechanism, savings can be

shared with consumers, subsidize devices, pay program costs

• In remaining non-peak periods, device continuously optimize for

emissions reductions

37

Technical Underpinnings

• Marginal emissions rates (lbs/MWh)
• Real-time
• 5-minute frequency
• Environmental index
• Percent (0-100)
• Rating (0-5)
• Switch (0-1)
• Balancing Authority (lat, long)
• Forecast (December 2018)
• Device Specific Optimization (November 2018)

API (api.watttime.org)

END USER VALUE

39

Individual consumers are expecting more environmentally friendly options, and are willing to pay for them

Consumers in America want and expect more sustainable solutions

• A survey of 1,500 customers conducted by

SmartEnergy IP found that 32% expect their utility to adopt automation technologies to save energy[1]

• A 2016 Gallup poll revealed that 73% of

Americans want to emphasize alternative energy instead of oil and gas production[2]

Consumers are increasingly willing to pay for environmentally conscious brands [3]

Source: [1] Navigant Research; [2] Gallup; [3] Nielsen

40

Demand Response Participation with AER

Study (WattTime):

• 300 randomly selected individuals across 30 U.S.

states were asked if they would sign up for a hypothetical ADR program. Unbeknownst to these individuals, they were randomly assigned to different ADR program descriptions: a regular program, one that offered an unusually large financial incentive (half that users’ annual electricity bill), or one with AER. Result:

• As expected, adding environmental impact to a DR

program (by adding AER to it) increased signups. Contrary to researcher expectations, AER increased signups even more than financial gain did.

41

AER Adoption with a Price Premium

Study (EMotorWerks):

• A company sold 700 electric vehicle charging

stations (EVSEs) with and without WattTime’s AER feature side-by-side in its online marketplace. To receive the AER-equipped version, consumers had to check a box agreeing to pay $50 more than the regular price. Result:

• 82% of customers choose to voluntarily pay $50

extra to receive the WattTime feature.

AER POTENTIAL & IMPACTS

43

Customers are also increasingly demanding communicating, controllable, and “smart” devices and control systems

Smart devices, appliances, and controls are growing in availability and popularity

• The smart thermostat market is projected to

quadruple in size, reaching a $4.4 billion dollar industry by 2025.`

• Large consumer technology companies are now

competing for market share in the growing “smart home” space.

• In institutional, commercial, and industrial

facilities, business priorities are driving customers to demand connected, intelligent control systems to manage loads.

Some 30 billion devices may be connected to the Internet of Things (IoT) by 2020[2]

Source: [1] Navigant Consulting; [2] McKinsey, December 2014

2013: 7–10 billion devices 2020: 26–30 billion devices

45

As the IoT expands, greater connectivity offers new

• pportunities to capture value from connected devices
• Reduce peak demand by shifting the timing of electricity usage to non-peak hours.

Existing programs in the United States are already capable of reducing peak loads by up to 32 GW.

• Lower energy costs by scheduling load to take advantage of relatively low-cost

electricity at different times of day. U.S. utilities currently have over 7.5 million customers enrolled in some form of dynamic pricing program, which directly incentivize this temporal flexibility.

• Reduce emissions by shifting load to coincide with renewable energy production, or

cleaner, more-efficient conventional generators.

Source: EIA

Existing capabilities Emerging Opportunity Using current technology, it is possible to stack the value of these use cases, achieving both cost reductions for capacity and energy, as well as emissions reductions.

Connectivity and control allow energy-using devices to be optimized against several

• criteria. Devices can be programmed to:

46

• Choice in how they consume

electricity

• Alignment with personal

motivations through emissions reductions

• Potential cost savings through

shared revenue from peak pricing events

Participant Value Streams

• Reduced energy costs during

peak periods

• Higher utilization of

renewable resources during curtailment periods (increasing generation of RPS eligible RECs)

• Load flexibility to bring on more

variable renewable generators

Consumer/End-user Load Serving Entity ISO/Grid Resource

48

• Using current technology and

data about marginal emissions, individual customers are empowered to make informed decisions about their next unit of energy consumption.

• These immediate emissions

savings are verifiable, easily demonstrated, and simple to quantify.

The impact of small changes on the margin today can add up to major emissions reductions over time

Source: RMI analysis

• As more customers make

incremental changes to their usage, there will be an emerging opportunity to adjust the control signals and directly impact power plant operational decisions (i.e., unit commitment).

• While harder to quantify,

these savings can be much greater (e.g., targeted shifting to eliminate the need for coal plant

• peration).
• As these operational impacts

are reflected in system

• perations, spot prices, and

forward capacity prices, emissions-aware load shifting can drive emissions-reducing investment decisions.

• These impacts are difficult to

forecast, but could materially increase investment in renewable energy resources.

Planning for next kilowatt-hour… ... leads to grid

• perational changes …

… and eventually impacts resource investment

Thank You

Henry Richardson

Environmental Analyst henry@WattTime.org 415.300.7475

ADDITIONAL SUGGESTED READINGS ON MIDAS-LIKE PRODUCTS

CAISO -- White Paper Proposal Wholesale Grid State Indicator to Enable Price Responsive Demand EPRI -- Transactive Incentive-signals to Manage Energy-consumption (TIME) ComED RTP Pilot -- [add link to evaluation report here]

Critique/Discussion Opportunities for “merchant layers” to augment markets, add value, increase investment incentives, sharpen operational incentives – all good!

• “How to” enable products (e.g., WattTime, Evolve Energy) suggests we ask

how to monetize average pricing in retail tariffs and CAISO settlements: ○ Can new products provide equivalent of “contract-for-differences (CFD)

• ver (beyond) retail tariff?

○ Can new products provide CFD over (beyond) average subLAP CAISO prices, or prompt use of more granular settlement for customers (to Pnode)? ○ Whether WattTime/Evolve Energy prompts wholesale (generation) to monetize/arbitrage markets (e.g., with bilateral contracts)?

Critique/Discussion

• How IFOM and BHM DERs will directly benefit from new merchant layers to

enable even more value?

• Increased complexity: two-way value creation from 1) customers-to-grid, 2)

from distribution-to-customers, 3) from wholesale-to-customers, and 4) from distribution-to-wholesale – a lot to work out

• Tradeoffs exist between 1) energy (PDR in ISO terms), 2) capacity (Supply in ISO

terms and distribution) ○ GHG is largely “energy” related, but storage and preheating/cooling trade

• ff directly with capacity

○ A new GHG price may need to reflect both LMP (plant dispatch at the “margin”) and GHG from Local/Flexible ramping as this component becomes larger -- a combined “marginal price”

SUMMARY AND ACTION ITEMS

Fill this out “live” based on the discussion that results.

Sunrun

Product Proposals for Coordinated Distribution and Bulk System MUA Benefit

https://gridworks.org/initiatives/load-shift-working-group/

Steven Rymsha 10/24/2018

LSWG Product Proposals for Coordinated Distribution and Bulk System MUA Benefit

Actual neighborhood of Sunrun customer homes 61

Integrated System Operations Planning

Integrated System Operations Planning

A Distribution System Operator must be more than a “gatekeeper” preventing interconnection & market participation.

• PG&E’s presentation indicate operational challenges including

PDR-LSR load consumption misalignment with distribution needs.

• Load consumption capacity service provides capacity expansion of

bulk power, transmission, distribution systems, and service nodes.

• Customers can

■ Build load ■ Reduce DER export

Load Shift/Load Consumption products incorporating distribution system operational coordination and grid modernization planning is a much more powerful and valuable solution to meet the grid needs. The LSWG must not overlook planning and operational

• pportunities within the various domains

63

Market Informed Load Consumption MUA Capacity Product

Market Informed Load Consumption MUA Capacity Product

This market informed product will not directly bid capacity or respond to real time markets, but operate in a specific programmatic manner based on California's excess energy planning needs.

• DERs are scheduled to deliver load consumption capacity

during specified hours based on market informed planning needs.

IOU/DSO

• Offers programs to customers as rider tariff
• Within the midday excess energy time period

■ IOU/DSO coordinate load consumption response

• Coordinate operations and planning for capacity

expansion benefits across service node, distribution feeders, and Sub-Lap.

65

Market Informed Load Shift MUA Capacity Product

Market Informed Load Shift MUA Capacity Product

These market informed resources will not directly bid capacity or respond to real time markets, but operate in a specific programmatic manner based on California's load shift planning needs.

• DERs are scheduled to deliver load consumption capacity and

load curtailment capacity within specified time periods based

• n market informed planning needs.

IOU/DSO

• Offers programs to customers as rider tariff
• Within the midday excess energy time period

■ IOU/DSO coordinate load consumptions response

• Coordinates operations and planning for capacity

expansion benefits acrosses specific service node, distribution feeders, and Sub-Lap.

• Within the peak load time period

■ IOU/DSO coordinates load curtailment/export response

67

Market Integrated Load Consumption MUA Capacity Product

Market Integrated Load Consumption MUA Capacity Product

PDR-LSR does not require load curtailment participation unless resources have must offer obligations to do so.

• Thus the working group should encourage enabling resources

that have no ability or requirements to provide peaking capacity an opportunity to addresses excess energy planning and

• perational needs.

IOU/DSO

• Offers programs to customers as rider tariff
• Within the midday excess energy period

■ IOU/DSO coordinates load consumption capacity dispatch via 3rd party aggregator.

• Abnormal distribution circuit configuration
• Limit distribution backfeed
• IOU/DSO bid additional available load

consumption capacity within CAISO PDR-LSR

69

Market Integrated Load Shift MUA Capacity Product

Market Integrated Load Shift MUA Capacity Product

IOU/DSO

• Offers programs to customers as rider tariff
• Within the midday excess energy period

■ IOU/DSO coordinates load consumption capacity dispatch via 3rd party aggregator.

• Abnormal distribution circuit configuration
• Limit distribution backfeed
• IOU/DSO bid and dispatch additional incremental

load consumption capacity into CAISO PDR-LSR

• Within the peak load period

■ RA resources participate per existing market rules and contracts ■ For non-RA resources IOU/DSO to have extraordinary incremental dispatch rights via 3rd party aggregator under program rules to utilize aggregated incremental capacity for greater power system needs.

71

Integrated Load Shift Operational Planning Benefits

Integrated Load Shift Operational Planning Benefits

MUA benefits across multiple power system domains

• Greater value and power system benefits can be achieved by

coordinating within distribution domain planning and

• perations.

Load consumption/shift products can enable broader utilization of resources by enabling participations paths, which do not exist today for the majority DER’s to capture value for capacity services align within distribution planning,

• perations, and grid modernization needs.

Load shift programs if coordinated offer an opportunity to expand the available capacity as specified within ICA maps. Potential for significantly more capacity to be aligned with planning and operational needs across the power system.

73

Thank You

74

Lunch Break

Pay for a Load Shape

Peter Alstone Schatz Energy Research Center / Lawrence Berkeley National Laboratory Load Shift Working Group October 24, 2018

What is the time and spatial granularity?

Time step of dispatch signal Hourly time steps, Weekday-Weekend Day Advance notice time to customer 1-month advance notice, persistent for 1-3 months Scale of geographic detail Range of detail possible: circuit-feeder-PNode-SubLAP-DLAP-ISO Expected frequency of dispatch Seasonal dispatch, 1-3 months What are the driving factors?

• Reduce cost of serving load (energy prices),
• reduce curtailment,
• reduce capacity needs.

The Net Load could be an organizing principle to define desired dispatch. This example is from 2017 operations data released by CAISO through Renewables Watch.

The Prices could also be an organizing principle to define desired dispatch. This example is from 2017 operations data synthesized by LGC Consulting.

A straw-proposal process concept

1. Every 3 months*, LSE’s work to establish an updated target load shape that supports grid needs (possibly different targets for different customer classes and different geographic areas). These are published publicly after any needed modification by distribution system operators to ensure reliability, a month in advance of changes from

• ne target shape to another.

2. Participating customers and/or aggregators work to match loads to the target using automated, structural, and behavioral approaches . 3. The total savings from the reduced cost of serving customer loads is estimated and returned to participants through incentives and performance payments.

1. Incentive pool could include energy market operations savings, avoided generation capacity cost, avoided T&D, avoided curtailment based on evaluation of the performance of the portfolio. 2. Performance based (or mixed fixed+performance based) incentives provide nudges for participating customers / sites to continue improving compared to the average participating customer.

*the period is a design choice and could in principle be anywhere from days to months. On a supporting slide, show a preliminary analysis of market prices to show the order of magnitude different in the cost to serve load. $60M/y if the 5% of the

These example target load shapes are based on “anti-duck” and “anti-price” for 2017.

Customer

Participate, Invest in enabling technology, Respond, Get Incentives.

Third Party / Aggregator

Design and deliver retail programs / products for aggregations of small and medium customers. Incorporate into existing DR portfolios and business models.

Load Serving Entity (CCA or IOU or DA provider)

Define SubLAP or DLAP level target load shape that minimizes cost of service. Publish target load shape (if appropriate)

Distribution utility / service territory LSE (IOU )

Refine target based on distribution system constraints. Provide incentives for modified targets. Provide AMI meter data access to support settlement. Participate in publication of target load shapes and ensuring cybersecurity.

CAISO

Support forecasts of net load / price and market data access. Support program evaluation and valuation of response with CAISO analysis.

CPUC (new addition to table)

Provide regulatory oversight for target load shape definition, publication, and verification processes? Editorial Note: Defining the target load shape for the system-scale (before any local modifications) would weigh costs, pollution, and customer experience. How to balance public oversight need with the need for maintaining nimble response to changes in grid needs? How will non-IOU LSE’s be treated in any regulatory oversight?

Organizational Roles

For Illustrative Purposes – Look at 24 grocery store load shapes to see variability within a sector – “good” match here (a “structural winner”)

Setting the “target”: 1) Find the “average” shape in the period. 2) Identify the baseload (the minimum of the average shape). 3) The target is the baseload plus a “new” variable load that is rescaled to match the system target shape. The total load is the same in target vs. actual.

Poorly performing sites have worse match between actual and target loads

A correlation-based performance metric (comparing target to actual load) is strongly related to the cost to serve loads at the sites.

Better-performance (higher scoring) sites are less costly to serve.

If 5% of daily load participated in this program with a 3- month target loadshape and hit the target, the total value in energy market cost savings is $~60M/year annually based on a simplified estimate. This only includes the avoided cost of energy, which are ~20% lower for sites that match the target shape compared to the average net load. Additional value streams from reduced curtailment, GHG, local air pollution, capacity could be incorporated into product design.

The duration of target load shape persistence changes the estimated energy market savings, but not by much.

This plot shows different “duration” target load shapes based on the daily net load, weekly average, monthly average, and three-month average.

Point of reference for scale:

The unit cost savings ($/MWh) from shifting estimated on the previous slide are $25-30 per Shifted MWh. These are roughly consistent with E3 RESOLVE model estimates used in the DR Study (to the right) A relatively low value in the market is consistent with DR potential study results for current-day grid

• perations. As curtailment (and

average price differentials) increase

• ver time, this value should go up.

FIGURE FROM THE DR Potential Study

Questions and Discussion Points

• Can this kind of program be deployed alongside traditional Shed DR?
• … or alongside dispatchable Shift?
• Could aggregators find a way to incorporate this into their business

models?

• Should different customer classes have different targets?
• Is this kind of “slow changing” shift better or worse for managing

distribution system constraints?

• Is there policymaking usefulness from a having a “synthesis of value”

product like this that incorporates cost evaluation across many grid needs, instead of trying to capture discrete value streams?

This plot shows the effective value in terms of reduced unit cost to serve load for different “duration” target load shapes based

• n the daily net load, weekly average, monthly average, and

three-month average.

Gridworks

Report Outline & Comparing Product Proposals

https://gridworks.org/initiatives/load-shift-working-group/

Report Outline

https://gridworks.org/initiatives/load-shift-working-group/

• 1. Intro
• a. Load Shift defined
• b. CPUC orders identified
• c. Working Group introduced
• 2. Why Shift Load?
• a. Potential Study findings
• b. IRP findings
• c. Negative price trends
• d. Other
• 3. Evaluation Criteria:
• a. Evaluation criteria introduced and defined
• b. Justification for criteria
• c. How the criteria are being used
• 4. Products
• a. Enhanced PDR
• b. CCP
• c. MIDAS
• d. Pay for Load Shape
• e. Sunrun product proposals
• 5. Product Evaluation
• a. Compare, Contrast, Draw Insights
• b. Identify relative product strengths compared to

evaluation criteria

• c. Identify barriers
• 6. Considering Questions from the Ruling
• a. RA
• b. Data Access
• c. GHG
• d. Coordination with CAISO
• 7. Recommendations
• a. Findings (e.g., market integration, technology

neutral, not energy neutral)

• b. Recommendations on further research and

consideration

• c. Next steps

Begin to tackle 5.a today; 5.b and 5.c in November.

Preliminary Product Comparison Notes

https://gridworks.org/initiatives/load-shift-working-group/

• Full comparison matrix available for those who would like to

use it on Load Shift workshop website.

• The following summary is preliminary and in need of review

and feedback.

• Purpose is to see where we stand and begin the process of

completing the Working Group’s product evaluation

What the products have in common

https://gridworks.org/initiatives/load-shift-working-group/

• All products are technologically neutral
• No product is energy neutral – take and shed are distinct
• Performance evaluation of all products are a “work in

progress”

• All products anticipate the ability to dual-participate (aka

Multi-Use Application)

• All products need further consideration of potential

ratepayer costs

94

A B C D E F G H Dispatch Method Dispatch Geo- Granularity Negative Pricing? Transaction Settles at... Performance Evaluation Role of IOU Role of Third-party Aggregators 1 PDR LSR Market Sublap Yes Aggregated Resource Metered + Typical-Use Adjustment Support Rule 24; LSE Aggregate, Bid, Coordinate, Settle 2 LSR 2.0 Market Sublap Yes Aggregated Resource Meter + Typical-Use Adjustment Support Rule 24; LSE Aggregate, Bid, Coordinate, Settle 3 CCP Distribution Operator Signal Based on Market Proxy (day-ahead price) Nodal Partial Premise Baseline Program Administrator Potentially Limited 4 MIDAS Distribution Operator Signal Based on Grid Conditions Distributed Partial Premise ? Program Administrator Aggregate, Coordinate 5 Sunrun Integrated Distribution Operator Signal Based on Grid Conditions Scalable Partial Device Baseline Active DSO Aggregate, Coordinate 6 Sunrun Informed Scheduled Scalable Partial Device TBD Passive DSO Aggregate, Coordinate 7 P4LS Scheduled Scalable Partial Premise or Aggregated TBD Determine Schedule Potentially Limited

95

A B C D E F G H

Dispatch Method

Dispatch Geo- Granularity Negative Pricing? Transaction Settles at... Performance Evaluation Role of IOU Role of Third-party Aggregators 1 PDR LSR

Market

Sublap Yes Aggregated Resource Metered + Typical-Use Adjustment Support Rule 24; LSE Aggregate, Bid, Coordinate, Settle 2 LSR 2.0

Market

Sublap Yes Aggregated Resource Meter + Typical-Use Adjustment Support Rule 24; LSE Aggregate, Bid, Coordinate, Settle 3 CCP

Distribution Operator Signal Based on Market Proxy (day- ahead price)

Nodal Partial Premise Baseline Program AdministratorPotentially Limited 4 MIDAS

Distribution Operator Signal Based on Grid Conditions

Distributed Partial Premise ? Program AdministratorAggregate, Coordinate 5 Sunrun Integrated

Distribution Operator Signal Based on Grid Conditions

Scalable Partial Device Baseline Active DSO Aggregate, Coordinate 6 Sunrun Informed

Scheduled

Scalable Partial Device TBD Passive DSO Aggregate, Coordinate 7 P4LS

Scheduled

Scalable Partial Premise or Aggregated TBD Determine Schedule Potentially Limited

• Product dispatch method vary widely
• “Scheduled” implies pre-determined target load curve the customer is incentivized to match; may be based on system,

local, or distributed conditions

• Range reflects varying sensitivities to ability of customer to quickly market prices and perceived value of solving wholesale
• vs. distribution level grid challenges

96

A B

C

D E F G H Dispatch Method

Dispatch Geo- Granularity

Negative Pricing? Transaction Settles at... Performance Evaluation Role of IOU Role of Third-party Aggregators 1 PDR LSR Market

Sublap

Yes Aggregated Resource Metered + Typical-Use Adjustment Support Rule 24; LSE Aggregate, Bid, Coordinate, Settle 2 LSR 2.0 Market

Sublap

Yes Aggregated Resource Meter + Typical-Use Adjustment Support Rule 24; LSE Aggregate, Bid, Coordinate, Settle 3 CCP Distribution Operator Signal Based on Market Proxy (day-ahead price)

Nodal

Partial Premise Baseline Program AdministratorPotentially Limited 4 MIDAS Distribution Operator Signal Based on Grid Conditions

Distributed

Partial Premise ? Program AdministratorAggregate, Coordinate 5 Sunrun Integrated Distribution Operator Signal Based on Grid Conditions

Scalable

Partial Device Baseline Active DSO Aggregate, Coordinate 6 Sunrun Informed Scheduled

Scalable

Partial Device TBD Passive DSO Aggregate, Coordinate 7 P4LS Scheduled

Scalable

Partial Premise or Aggregated TBD Determine Schedule Potentially Limited

• Product dispatch geo-granularity range from sublap, to nodal to distribution level
• Like Dispatch Method, reflects varying perceived value of solving wholesale vs. distribution level grid challenges
• Q: If dispatch isn’t a response to distribution conditions, how are potential negative impacts on distribution operations avoided?

97

A B C D E F G H Dispatch Method Dispatch Geo- Granulari ty

Negative Pricing?

Transaction Settles at... Performance Evaluation Role of IOU Role of Third-party Aggregators 1 PDR LSR Market Sublap

Yes

Aggregated Resource Metered + Typical-Use Adjustment Support Rule 24; LSE Aggregate, Bid, Coordinate, Settle 2 LSR 2.0 Market Sublap

Yes

Aggregated Resource Meter + Typical-Use Adjustment Support Rule 24; LSE Aggregate, Bid, Coordinate, Settle 3 CCP Distribution Operator Signal Based on Market Proxy (day-ahead price) Nodal

Partial

Premise Baseline Program Administrator Potentially Limited 4 MIDAS Distribution Operator Signal Based on Grid Conditions Distribute d

Partial

Premise ? Program Administrator Aggregate, Coordinate 5 Sunrun Integrated Distribution Operator Signal Based on Grid Conditions Scalable Partial Device Baseline Active DSO Aggregate, Coordinate 6 Sunrun Informed Scheduled Scalable Partial Device TBD Passive DSO Aggregate, Coordinate 7 P4LS Scheduled Scalable Partial Premise or Aggregated TBD Determine Schedule Potentially Limited

• Product proposals differ in how strongly their dispatch correlates to negative market pricing
• CCP, Sunrun Informed, and P4LS aim for negative price period, but are based on some forecast and therefore may not be a
• match. MIDAS and Sunrun integrated give priority to distribution-level grid needs, which may not match negative prices.
• Are negative prices our strongest indicator whether shift avoids renewable curtailment?

98

A B C D

E

F G H Dispatch Method Dispatch Geo-Granularity Negative Pricing?

Transaction Settles at...

Performance Evaluation Role of IOU Role of Third-party Aggregators 1 PDR LSR Market Sublap Yes

Aggregated Resource

Metered + Typical-Use Adjustment Support Rule 24; LSE Aggregate, Bid, Coordinate, Settle 2 LSR 2.0 Market Sublap Yes

Aggregated Resource

Meter + Typical-Use Adjustment Support Rule 24; LSE Aggregate, Bid, Coordinate, Settle 3 CCP Distribution Operator Signal Based on Market Proxy (day-ahead price) Nodal Partial

Premise

Baseline Program AdministratorPotentially Limited 4 MIDAS Distribution Operator Signal Based on Grid Conditions Distributed Partial

Premise

? Program AdministratorAggregate, Coordinate 5 Sunrun Integrated Distribution Operator Signal Based on Grid Conditions Scalable Partial

Device

Baseline Active DSO Aggregate, Coordinate 6 Sunrun Informed Scheduled Scalable Partial

Device

TBD Passive DSO Aggregate, Coordinate 7 P4LS Scheduled Scalable Partial

Premise or Aggregated

TBD Determine Schedule Potentially Limited

• Product settlement level ranges from aggregated resource, to premise, to device.
• “Premise” and “Device” levels can always be aggregated up, but PDR-LSR can’t be disaggregated down, right?
• Device level settlement introduces metering requirements and/or acceptance of third-party meter data

99

A B C D E F G H Dispatch Method Dispatch Geo- Granularity Negative Pricing? Transaction Settles at...

Performance Evaluation

Role of IOU Role of Third-party Aggregators 1 PDR LSR Market Sublap Yes Aggregated Resource

Metered + Typical-Use Adjustment

Support Rule 24; LSE Aggregate, Bid, Coordinate, Settle 2 LSR 2.0 Market Sublap Yes Aggregated Resource

Meter + Typical- Use Adjustment Support Rule 24; LSE

Aggregate, Bid, Coordinate, Settle 3 CCP Distribution Operator Signal Based on Market Proxy (day-ahead price) Nodal Partial Premise

Baseline

Program AdministratorPotentially Limited 4 MIDAS Distribution Operator Signal Based on Grid Conditions Distributed Partial Premise

?

Program AdministratorAggregate, Coordinate 5 Sunrun Integrated Distribution Operator Signal Based on Grid Conditions Scalable Partial Device

Baseline

Active DSO Aggregate, Coordinate 6 Sunrun Informed Scheduled Scalable Partial Device

TBD

Passive DSO Aggregate, Coordinate 7 P4LS Scheduled Scalable Partial Premise or Aggregated

TBD

Determine Schedule Potentially Limited

• No silver bullets; Lots of questions remain regarding evaluation with increased frequency, how device level evaluation can work,

and the desired period of measurement (from 5 minute to one-year).

• Lines 6 and 7 likely do not require ex post evaluation; however, the challenge of determining the ex ante target schedule will

have its own challenges.

100

A B C D E F

G

H Dispatch Method Dispatch Geo- Granularity Negative Pricing? Transaction Settles at... Performance Evaluation

Role of IOU

Role of Third-party Aggregators 1 PDR LSR Market Sublap Yes Aggregated Resource Metered + Typical-Use Adjustment

Support Rule 24; LSE

Aggregate, Bid, Coordinate, Settle 2 LSR 2.0 Market Sublap Yes Aggregated Resource Meter + Typical- Use Adjustment

Support Rule 24; LSE

Aggregate, Bid, Coordinate, Settle 3 CCP Distribution Operator Signal Based on Market Proxy (day-ahead price) Nodal Partial Premise Baseline

Program Administrator

Potentially Limited 4 MIDAS Distribution Operator Signal Based on Grid Conditions Distributed Partial Premise ?

Program Administrator

Aggregate, Coordinate 5 Sunrun Integrated Distribution Operator Signal Based on Grid Conditions Scalable Partial Device Baseline

Active DSO

Aggregate, Coordinate 6 Sunrun Informed Scheduled Scalable Partial Device TBD

Passive DSO

Aggregate, Coordinate 7 P4LS Scheduled Scalable Partial Premise or Aggregated TBD

Program Administrator

Potentially Limited

• Proposals offer wide range of perspective on the potential role of the IOU
• Evaluation of the advantages and disadvantages of these different roles relatively incomplete so far
• Evaluation of role of CCAs (and other non-IOU LSEs) relatively incomplete

101

A B C D E F G

H

Dispatch Method Dispatch Geo- Granularity Negative Pricing? Transaction Settles at... Performance Evaluation Role of IOU

Role of Third-party Aggregators

1 PDR LSR Market Sublap Yes Aggregated Resource Metered + Typical-Use Adjustment Support Rule 24; LSE

Aggregate, Bid, Coordinate, Settle

2 LSR 2.0 Market Sublap Yes Aggregated Resource Meter + Typical-Use Adjustment Support Rule 24; LSE

Aggregate, Bid, Coordinate, Settle

3 CCP Distribution Operator Signal Based on Market Proxy (day-ahead price) Nodal Partial Premise Baseline Program AdministratorPotentially Limited 4 MIDAS Distribution Operator Signal Based on Grid Conditions Distributed Partial Premise ? Program Administrator

Aggregate, Coordinate

5 Sunrun Integrated Distribution Operator Signal Based on Grid Conditions Scalable Partial Device Baseline Active DSO

Aggregate, Coordinate

6 Sunrun Informed Scheduled Scalable Partial Device TBD Passive DSO

Aggregate, Coordinate

7 P4LS Scheduled Scalable Partial Premise or Aggregated TBD Determine Schedule

Potentially Limited

• All proposal imagine the roll of the aggregator may include recruiting customers and service customers in coordinating the

response to dispatch method

• Role of third-party aggregators differ primarily on whether they are bidding a price for their service or “simply” respond to a

signal and receive an administratively determined price for their service

Conclusions

• The Working Group has developed a wide variety of choices for the

Commission to consider.

• Some assessment of “viability” seems in order. How to accomplish?
• What product is “the best” depends what criteria you give the most

weight.

• The working group can suggest weighting as a part of its

evaluation or frame the choice of weighting for the Commission. Gridworks recommends the latter.

https://gridworks.org/initiatives/load-shift-working-group/

Next Steps

▪ Final Report Timeline ▪ Final Report Expectations ▪ Update on Future Sessions

https://gridworks.org/initiatives/load-shift-working-group/

https://gridworks.org/initiatives/load-shift-working-group/ Date Milestone Topics Notes

• Oct. 26

Stakeholder Submissions WG comments due on LS products Deadline for Stakeholders to provide written feedback on Pay for load shape, MIDAS, and Sun Run Proposals

• Nov. 1

Report outline to WG Gridwork to WG Monday, Nov. 8 Stakeholder Submissions WG comments due on draft outline of Report Deadline for stakeholders to provide written feedback on report outline Wednesday, Nov. 14: CPUC’s Golden Gate Room Working Group Meeting Review feedback on draft report outline Working group to compare and contrast products, identify pluses and minuses, make recommendations for final report sections. Review finalized RA proposals Finalize product proposals (as necessary) Friday, Nov. 30 Report Draft to WG Gridwork to WG Monday, Dec. 10 Stakeholder Submissions WG comments on first draft of report Deadline for stakeholders to provide written feedback on report draft Wednesday, Dec. 12 CPUC’s Courtyard Room Working Group Meeting Review draft 1 and stakeholder comments on it Consider if a phone or in-person meeting is needed in January

• n the report
• Dec. 14

Report Draft V2 to WG Gridwork to WG Friday, Dec. 21 Stakeholder Submissions WG comments due on report draft V2 Deadline for stakeholders to provide written feedback on report V2 draft Monday, Dec. 31 Final report to working group Gridwork to WG Redline + clean draft for transparency Jan, 2019 Admin time to submit report Possible phone or in-person meeting to discuss the

Final Report Expectations: Shared?

https://gridworks.org/initiatives/load-shift-working-group/

• Final report will be broadly accessible: 15 pages or less and supported by professional design
• Final Report will be reasonably reflective of the Working Group’s perspective: highlight perspectives, not every

perspective on everything possible.

• Gridworks suggests we say “not everything is agreed to by everyone” at the outset of the Report -- make this a

collective expression. Leave party positions for the many comment opportunities that will follow.

• Final Report will be ~90% insights and ~10% caveats about those insights. We share the challenge of being

accurate without reversing the 90/10 ratio.

• Final Report will be largely complete by December 14, leaving 6 weeks for design, clean-up and submission.
• Review heavily weighted toward the input your provide on November 14, December 10, and December 12;

by December 14 “new material” is too late