disruptive electricity market innovation: motivation ... · disruptive electricity market...

LABORATORY FOR INTELLIGENTINTEGRATED NETWORKSOF ENGINEERING SYSTEMSEMPOWERING YOUR NETWORK






1

Prof. Amro M. Farid1,2,3

1-Thayer School of Engineering at Dartmouth2-Department of Computer Science at

Dartmouth3-Research Affiliate, MIT Mechanical

Engineering

Disruptive Electricity Market Innovation: Motivation & Pathways

http://engineering.dartmouth.edu/liines




Invited PresentationWhat is Next for Our Electricity

Markets? Towards a New Market Regulatory Framework

IX Electricity Security Advisory Panel Workshop

International Energy AgencyParis, FR

June 17, 2018







2

Today’s restructured electricity markets are based on economic principles developed more than twodecades ago, allowing competition in the power sector with the objective of improving economicefficiency, shifting risks to investors and, in relevant emerging economies, to reduce the burden on thetreasury. Since then, many markets have successfully followed a pragmatic and constant learning-by-doing process, resulting in innovations and changes that allow them to improve power systemperformance while delivering electricity in a secure and affordable manner. Recently, however, a numberof unprecedented developments are threatening to disrupt power systems around the world. Theseinclude the need to quickly and affordably decarbonize the economy, the introduction of new supply-and demand-side technologies, and expanded opportunities for consumer engagement. Thispresentation seeks to provoke a discussion towards disruptive innovation in electricity market designdrawing on several recent publications including the ISO New England System Operational Analysis andRenewable Energy Integration Study and the New England Energy-Water Nexus Study. It provides sevenkey insights: 1.) The reliance on VRE curtailment is unsustainable leading to a 40% erosion of the VREcapacity factor. 2.) The reliance on virtual power plant demand response is also sustainable leading to acascade of system imbalances and costs relative to a transactive energy model. 3.) Instead, eIoT-enabledtransactive energy retail markets are required to activate the grid with a proliferation of decentralized“uber-like” grid services. 4.) Deep decarbonization will impose a “Beyond-Energy-Mindset” thatfacilitates an expanded portfolio of incentivized ancillary grid services. 5.) Demand-side retailaggregators will provide net social benefits while eroding the incumbent utility business model. 6.)Similarly, demand-side wholesale aggregators will do the same. 7.) Finally, energy storage will emerge asas a capable resource as the LCOE of Storage+VRE systems continues to decline in an energy excess-future.

Presentation Abstract







3

Goal: To provoke discussion towards disruptive innovation in electricity market designs§ Insight 1: Reliance on VRE Curtailment is Unsustainable

- VRE curtailment leads to a 40% erosion in the VRE capacity factor.§ Insight 2: Reliance on Virtual Power Plant Demand Response is Unsustainable

- The presence of demand baseline errors – present only in the VPP implementation – leads to a cascade of system imbalances and costs as compared to a Transactive Energy model.

§ Insight 3: eIoT-enabled Transactive Energy Retail Markets are Required- The integration of network-enabled physical devices can activate the grid periphery with a

proliferation of decentralized “uber-like” grid services. § Insight 4: Beyond Energy -- the Proliferation of Ancillary Services

- Deep decarbonization imposes tremendous requirements for an expanded portfolio of incentivized ancillary grid services.

- Deep decarbonization requires an expanded portfolio of incentivized ancillary grid services.

§ Insight 5: The Emergence of the Demand Side Retail Aggregator- Self-aggregation of retail customers provides net social benefits while eroding the incumbent utility

business model. § Insight 6: The Emergence of the Demand Side Wholesale Aggregator

- The entrance of demand-side aggregators into wholesale electricity markets further provides net social benefits while further eroding the incumbent utility business model.

§ Insight 7: The Emergence of Energy Storage- The LCOE of Storage +VRE systems continues to decline in an energy-excess future.

Presentation Outline







4

Drivers for the Evolution of the Electric Power Grid

∴ We need holistic techniques that integrate multiple layers of control andsimultaneously manage technical and economic objectives.

Several drivers have emerged to challenge this assumption:

§ Decarbonization§ Growing electricity demand§ Deregulation of electricity markets§ Active end-user participation§ Digital innovations in energy

technologiesGeneration Transmission

& DistributionDemand

Physical

Cyber

Dispatc

hable

Stochas

tic

Control Objectives

Economic

Technical

Balancing &Frequency Control

Line CongestionControlVoltage Control

Investment Cost

Operating Cost

Traditional power systems were built upon the assumption that generation was controlled by a few centralized generation facilities that were designed to serve fairly passive loads.







5

Electric Power Enterprise Control System Simulation (EPECS)

∴ EPECS Simulation has been used to study techno-economic systemperformance in the presence of variable renewable energy resources.

Unit Commitment

Day-AheadResource Scheduling

RegulationService

Cyber Layer of Controls

Real-Time Balancing

Physical Power Grid Layer

P̂ST (t)

P̂DA(t)∆PST (t)

PREG

∆PST (t)

PLOAD

RRAMP

P LOADREQ

RRAMPREQ

P (t)

P REGREQ

∆PRT (t)Regulation LevelImbalanceMeasurement

ImbalanceMeasurement

I (t)

P (t) - Actual net load;

P̂DA(t)

P̂ST (t)

P LOADREQ

RRAMPREQ

P REGREQ

PLOAD

RRAMP

PREG

∆PDA(t)

∆PST (t)

∆PRT (t)

- Net load day-ahead forecast;

- Net load short-term forecast;

- Load following reserve requirement;

- Ramping reserve requirement;

- Regulation reserve requirement;

- Actually scheduled load following reserves;

- Actually scheduled ramping reserves;

- Actually scheduled regulation reserves;

- Imbalances at the day-ahead scheduling output;

- Imbalances at the real-time balancing output;

- Imbalances at the regulation service output;

- Residual imbalance at the system output;

I (t)

Reserve Scheduling

Storage Commitment Regulation

Economic Dispatch

Storage Dispatch







6



















7

Real-Time Energy Market Dispatch: 2025 Conventional2025-4-0423a Week 4: Real Time Schedules

Jan 23 Jan 24 Jan 25 Jan 26 Jan 27 Jan 28 Jan 29Time (hours) 2025

0

2000

4000

6000

8000

10000

12000

14000

16000

18000

Powe

r (M

W)

NuclearNatural GasHydroWoodMunicipal Solid WasteCoalLandfill GasKeroseneDistillate Fuel OilLiquid Natural GasJet FuelResidual Fuel OilNo. 6 Fuel OilSolarWind







8

Real-Time Energy Market Dispatch: 2030 High VREs Plus2030-3-0423a Week 16: Real Time Schedules

Apr 17 Apr 18 Apr 19 Apr 20 Apr 21 Apr 22 Apr 23Time (hours) 2030

0

2000

4000

6000

8000

10000

12000

14000Po

wer (

MW

)

NuclearNatural GasWoodHydroMunicipal Solid WasteLandfill GasKeroseneCoalDistillate Fuel OilJet FuelNo. 6 Fuel OilResidual Fuel OilSolarWind







9

The Emergence of Variable Renewable Energy (VRE)

∴ The emergence of VRE necessitates active demand side resources.

Past:

Future: Generation/Supply Load/Demand

Well-Controlled &Dispatchable

Thermal Units: (Potential erosion of

capacity factor)

Demand Side Management:(Requires new control

& market design)

Stochastic/Forecasted

Solar & Wind Generation:(Can cause unmanaged

grid imbalances)

Conventional Loads:(Growing &

Needs Curtailment)

Generation/Supply Load/Demand

Thermal Units: Few, Well-Controlled,

Dispatchable, In Steady-State

Conventional Loads:Slow Moving, Highly

Predictable, Always Served







10

VRE Curtailment: 2030 Scenarios

∴ Relative to the conventional scenarios, high VREs require significant curtailment.







11



















12

Reliance on Virtual Power Plant DR is Unsustainable

∴FERC Order 745 vs Transactive Energy: A Techno-Economic Policy Case

Tota

l ope

ratio

n co

st, $

Milli

on

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

Socialwelfare

SCUC baselineerror = 0.05

SCUC baselineerror = 0.1

50.71% 52.56% 54.35%

37.01% 37.55% 38.08%

5.85%8.77% 11.69%6.43%6.43%

6.43%+5.31%+10.55%

Generation Virtual generation Load following Regulation

Tota

l ope

ratio

n co

st, $

Milli

on0

0.20.40.60.8

11.21.41.61.8

22.2

Socialwelfare

SCED baselineerror = 0.05

SCED baselineerror = 0.1

50.73% 52.48% 54.27%

37% 37.55% 38.07%

5.84% 5.84% 5.84%6.43%21.43%

37.5%+17.3%

+35.69%

Generation Virtual generation Load following Regulation







13



















14

Enabling DSRs with the energy Internet of Things

∴ The emergence of demand side resources necessitates the distributedtransactive control through the energy Internet of Things.







15

New England Power Plants and Rivers

Most power plants are located near a water source. This indicates the strong coupling between the water supply and energy supply systems.







16

New England Energy Water Nexus Study: Key Findings

In 6 New England Scenarios, flexible energy-water resources are shown to enhance:

§ Average levels (~150MW) and minimum levels (~150MW) of upward load following reserves

§ Average levels (~900MW) and minimum levels (~1.2GW) of downward load following reserves

§ Water withdrawals (~26,000 m3/min)

§ Day Ahead Energy Market Production Costs (~3600$/hr)

§ Real Time Energy Market Production Costs (~500/hr)







17



















18

Regulation Reserves: 2030 Scenarios

∴ Relative to the conventional scenarios, high VREs saturate regulation reserves.







19

Load Following Reserves: 2030 High VREs Plus Scenario

∴ In Spring & Autumn, the ability to track low net load conditions is constrained.







20

Ramping Reserves: 2025 Conventional Scenario







21

Ramping Reserves: 2030 High VREs Plus Scenario

∴ In Spring & Autumn, the ability to track low net load conditions is constrained.







22



















23

The Emergence of the Demand Side Retail Aggregator

Conventional Case: Utility Revenue = (2000kWh) ∗ (0.1$/kWh)

− (1200kWh) ∗ (0.08$/kWh) = $200 − $96 = $104

Consumer Cost = $200

Solar PV Revenue = $96

Transactive Energy Case: Utility Revenue = (800 kWh) ∗ (0.1$/kWh)

= $80Consumer Cost = (800kWh) ∗ (0.1$/kWh)

+ (1200kWh) ∗ (0.09$/kWh)= $188

Solar PV Revenue = $108

102 4 Transactive Energy Applications of eIoT

Transactive Energy Enabled

Apt 1:Prosumer

Apt 2:Prosumer

Apt 4:Prosumer

Conventional

Apt 1:Consumer

Apt 2:Consumer

Apt 3:Consumer

Apt 4:Consumer

Utility

Apt 3:Consumer

Utility

Fig. 4.1: A Use Case Comparison between a Conventional and an eIoT Trans-active Energy Enabled Apartment Building

Tenant Payment = (1200kWh) ⇤ (0.09$/kWh) (4.4)+ (800kWh) ⇤ (0.1$/kWh) = $188 (4.5)

Finally, the tenants with rooftop solar now receive $108 as opposed to $96.

Solar Tenant Credit = (1200kWh) ⇤ (0.09$/kWh) = $108. (4.6)

While this specific case may appear ideal, it is illustrative. In the transactiveenergy case, the presence of solar generation provides an incentive for greatercompetition which ultimately benefits all the participating prosumers whilesimultaneously eroding the utility’s billable energy. Because the tenantshave collectively agreed to interact with the electric utility through a singlecommercial meter, the utility simply sees a decrease in the total amount ofelectricity purchased.

The eIoT transactive energy aggregation use case above shows net socialbenefits due to several enabling factors.

1. The presence of prosumers with local solar generation that is, at times,inadequately compensated by utilities encourages the emergence of atransactive energy marketplace.

∴ Self-aggregation of retail customers provides net social benefits while erodingthe incumbent utility business model.







24

Emergence of the Demand Side Retail Aggregator

The eIoT transactive energy aggregation use case above shows net social benefits due to several enabling factors:

§ The presence of prosumers with local solar generation that is, at times, inadequately compensated by utilities encourages the emergence of a transactive energy marketplace.

§ The solar generator’s value proposition leaves local consumers at times over-billed by utilities.

§ The transactive energy marketplace is likely to be strengthened if there is a strong sense of community within the apartment building.

§ There exists nearly ubiquitous measurement, communication, and decision-making capabilities within the building to support the transactions. It provides price and quantity information for rational decision-making. The user-friendliness of these information technologies encourages greater adoption.

§ There exists a sparsity of measurement, communication, and decision- making capabilities between the building and the utility.







25



















26

Emergence of the Demand Side Wholesale Aggregator

Conventional Retail Rate Case: Consumer Cost = $200

Wholesale Real-Time Pricing Case:Consumer Cost = $162Utility Revenue = $0Wholesale Real-Time Pricing w/ Load Shifting Case:Consumer Cost = $134Utility Revenue = $0

∴ The entrance of demand-sideaggregators into wholesale electricitymarkets further provides net socialbenefits while further eroding theincumbent utility business model.

104 4 Transactive Energy Applications of eIoT

4.2.2 An eIoT Economic Demand Response in WholesaleElectricity Markets Use Case

The second eIoT use case is based upon economic demand response (DR) asit is currently implemented in wholesale electricity markets. Consider Figure4.2. On the left is the same conventional apartment building. On the rightis the same transactive energy enabled building which now acts as a singleeconomic demand response participant.

Economic Demand Response

Apt 1:Prosumer

Apt 2:Prosumer

Apt 4:Prosumer

Conventional

Apt 1:Consumer

Apt 2:Consumer

Apt 3:Consumer

Apt 4:Consumer

Utility

Apt 3:Consumer

Utility

ISO

Fig. 4.2: A Use Case Comparison between a Conventional and an eIoT Eco-nomic Demand Response Apartment Building

The building’s conventional load profile is shown in Figure 4.3a. Forsimplicity, assume that the building is relatively small compared to the peakload of the wholesale electricity market. Consequently, the building acts as aprice taker because its bids have little e↵ect on the locational marginal prices(LMPs) that clear the wholesale electricity market. Figure 4.3b shows thehourly LMPs for the full day. They are assumed to closely follow the trend ofthe “duck curve” mentioned earlier in Section 1.2.

The financial benefits for the transactive energy-enabled building can becalculated. As stated previously, the building’s tenants pay $200 when ex-posed to the retail rate. However, simply by entering the wholesale electricitymarket, they would pay $162 without shifting their behavior. This is because,

4.2 Potential eIoT Energy Management Use Cases 105

00:00 04:00 08:00 12:00 16:00 20:00

Hour of the day Jun 23, 2016

31

32

33

34

35

36

37

38

Lo

ad

Pro

file

, kW

Daily Load Profile

00:00 06:00 12:00 18:00 00:00

Hour of the day Jun 23, 2016

-5

0

5

10

15

20

25

Price

, ce

nts

/kW

h

Hourly Prices

Fig. 4.3: eIoT Economic Demand Response in Wholesale Electricity MarketsUse Case Data: (a) On left, the daily net load profile of the prior to demand re-sponse incentives. (b) On right, the hourly locational marginal prices (LMPs)experienced within the wholesale electricity market.

on average, wholesale electricity rates are lower than retail rates. In such acase, the tenants have saved $38 but the utility naturally has lost all $200because the transactive energy-enabled building has e↵ectively “cut out themiddle-man”. Now imagine that the transactive energy enabled building isable to shift its loads so that it is no longer exposed to evening peak pricingand, more importantly, it makes use of negative LMPs during peak sunlighthours. A perfectly flat load curve would mean that the tenants now pay$134 for a savings of $66. In this case, as well, the utility has no access tothe associated revenue. A flattened load curve could be achieved in multi-ple ways. Significantly sized loads, like a fleet of electric vehicles or factoryproduction, may have the required flexibility. In residences, eIoT-enabledhome appliances (e.g. dishwashers, washers, and dryers) can be timed to shiftload during the day. In commercial buildings, HVAC units and hot waterheaters can be controlled to curtail energy consumption during peak hours.Residential and commercial applications may be relatively small-scale, butthey have the intended impact with load aggregation. Industrial loads maynot need aggregation, and examples include water pumping, desalination,and factory production. In all cases, eIoT devices and infrastructure enablethe transactive energy applications.

Again, this specific case is illustrative although it may appear ideal. Theability to aggregate so as to have access to wholesale electricity rates pro-vides a financial benefit to the building’s end consumers. Furthermore, theability to participate in that market through economic demand responseallows the building to fill the troughs and shave the peaks of the duck curve.In both cases, this is financially beneficial [550]. Filling the troughs of the







27



















28

§ In May 2017, SolarCity (a subsidiary of Tesla) partnered with the Kauai Island Utility Cooperative (KIUC) to install a 15 MW solar array paired with an 11 MW battery storage system. The project was financed using a 20-year PPA at a price of $139/MWh.

§ In early 2017, KIUC signed a PPA with AES corporation at $110/MWh to finance a 28 MW solar array paired with a 20 MW/100 MWh battery system that is slated to come online by the end of 2018.

§ In May 2017, NextEra Energy entered into a 20-year PPA with Tucson Electric Power to finance a 100 MW solar array paired with a 30 MW/120 MWh energy storage system—the agreed-upon price was $45/MWh.

§ In December 2017, Xcel Energy’s Colorado utility subsidiary announced the results of a recent solicitation where the median bid price for solar-plus-storage projects was $36/MWh and the median bid price for wind-plus-storage projects was $21/MWh.

§ FERC Order 841 is a game-changer for energy-storage assets§ Prediction: Driven by a need to capitalize on solar projects, energy storage adoption will

follow a physics-constrained supply curve 1.) fluidic-hydro 2.) thermal-building 3.) chemical-battery 4.) chemical-hydrogen

The Emergence of Energy Storage

Source EIA U.S. Battery Storage Market Trends 2018

∴ The LCOE of Storage + VRE systems continues to decline in an energy-excessfuture.







29



















30

Thank You

http://engineering.dartmouth.edu/liines




Prof. Amro M. Farid1,2

1-Thayer School of Engineering at Dartmouth2-Department of Computer Science at

Dartmouth

Email: amfarid_at_dartmouth_dot_edu

disruptive electricity market innovation: motivation ... · disruptive electricity market...

Documents