Scribd
Upload
68 views

How Retail Markets Can Optimize Electricity Distribution

How Retail Markets Can Optimize Electricity Distribution. D. P. Chassin, pnnl. Introduction to real-time capacity markets Purpose, theory, basic examples. Issues Examine Olypen market design / results Objectives, implementation, results, insights.

Uploaded by

williamb285
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
68 views

How Retail Markets Can Optimize Electricity Distribution

How Retail Markets Can Optimize Electricity Distribution. D. P. Chassin, pnnl. Introduction to real-time capacity markets Purpose, theory, basic examples. Issues Examine Olypen market design / results Objectives, implementation, results, insights.

Uploaded by

williamb285
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
You are on page 1/ 28

How Retail Markets Can Optimize Electricity Distribution

D. P. Chassin Pacific Northwest National Laboratory

Overview

Introduction to real-time capacity markets


Purpose, theory, basic examples, issues

Examine Olypen market design/results


Objectives, implementation, results, insights

Preview AEP NE Columbus RTP-DA rate


Rate design and valuation process

Purpose of Retail Real-Time Pricing


Discover retail price of energy Time-varying value of (constrained) supply Incorporates time-varying value of demand response Addresses 3 major distribution issues:
Load growth, distributed resource control, demand response

Markets as optimizers
Auctions solve allocation problem
Computationally efficient (parallelizable) Equilibrium assignment of buyers and sellers Interative (either explicit or implicit)

Linear program discovers price


Maximizes total benefit (primal) Minimize local costs (dual)

Price solution is Pareto optimal

See DP Bertsekas , Linear Network Optimization: Algorithms and Codes, MIT Press, 1991

Retail Capacity Market


Energy price [$/MWh]

Buyer surplus

Cleared price

Seller surplus

Power [MW]
Cleared load

Incorporate Day-Ahead Schedule


Price ($/MWh)

RTP customers actual response

Retail price between DA and RT


Real-time price is high Cleared price Day-ahead Price is low Load (MW) Unresponsive Load Maximum Load

Scheduled Load

Some potential issues/FAQs


Should utility be allowed to own/coordinate distributed resources (analog to generation/transmission conflict)? How to ensure costs are not double-embedded? How is seller surplus from feeder congestion used? How does utility fairly compensate consumers? Are there any subsidies built into the rate scheme? How is misbid/misresponse handled? What kind of security is really needed? How is rebound managed?

Rebound peaks occur with load control

Load Shape for Single-Family (Gas) Homes on 7-18-2006


Fixed price Fixed_A

TOU_A_Group_1 Time-of-use price

Total Hourly Energy Consumption (kWh)

1200
1000 800 600

400
200 0 0 2 4 6 8 10 12 14 16 18 20 22 24

Hour of Day

Complex pricing strategies mitigate rebound

Load Shape for Single-Family (Gas) Homes on 7-18-2006


TOU_A_Group_1 Time-of-use group 1 TOU_A_Group_4 Time-of-use group 4
Total Hourly Energy Consumption (kWh)

TOU_A_Group_2 Time-of-use group 2 TOU_A_Group 55 Time-of-use group

TOU_A_Group_3 Time-of-use group 3 TOU_A_Group 66 Time-of-use group

1200 1000

800
600 400 200 0 0 2 4 6 8 10 12 14 16 18 20 22 24

Hour of Day

At some point a capacity market is easier


Load Shapes for Single-Family (Gas) Homes on 7-18-2006
Total Hourly Energy Consumption (kWh)
Fixed price
Real-time TOU/CPP price

1000
900 800 700 600

500
400 300 200 100

0
0 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 of Day

Pacific NW GridWise Testbed Projects

GridWise Testbed Participants


Bonneville Power Administration Pacificorp Portland General Electric City of Port Angeles Municipal Utility IBM Whirlpool/Sears Kenmore Clallum County Public Utility District

11

Virtual Distribution Utility Operation


IBM Invensys $
MW
Internet broadband communications

Market

Johnson Controls
12

Olympic Peninsula RTP Market

Customer participation

$35

Economy

Comfort

Economic Cooling Response


User sets: Tdesired, comfort (based on occupancy calendar)

These imply: Tmax, Tmin, k (price response parameters) Price is expressed as std. deviation from mean (over a short period, e.g., 24 hrs)

k Pbid Price Pavg Pclear

Tmin
15

Tset

Tdesired Temperature

Tcurrent

Tmax

Managing Constraints
DG required above feeder limit

Load (kW)

Market failed to cap demand for one 5-min. interval in 12 months of operation

Price ($/MWh)

Hour
16

Load Shifting RTP Customers


Winter peak load shifted by pre-heating Resulting new peak load at 3 AM is noncoincident with system peak at 7 AM Illustrates key finding that a portfolio of contract types may be preferred i.e., we dont want to just create a new peak

17

Mixing rates also manages uncertainty


It is impossible to choose a portfolio in this white region because no combination of contracts can yield such risk/return

18

Peak energy uncertainty

19

Gross margin volatility

20

Response Manages New Resources


Regulation: one or more fast-responding power plants continually throttle to match normal fluctuations in load Highest cost generation in markets (zero net energy sales, wear & tear, fuel consumption) Intermittency of wind output can exceed regulation capability and reduces cost effectiveness of wind

normal fluctuations in load

Demand management to a capacity cap with real-time prices eliminated load fluctuations for 12 hours!

Load (kW)

Hour

21

AEP NE Columbus Project


Many tariffs are planned
Fixed Rate (standard) Interruptible Tariff (direct load control)

2-Tier Time of Use (2-TOU)


3-Tier Time of Use (3-TOU) Real Time Price Double Auction (RTPDA) Each tariff enable a difference kind of response

RTP Rate Design


Determine RTP-DA pricing method PJM DA Hourly LMP 5-minute RTP LMP Customer bids (Heating, AC, hotwater) Feeder constraints (physical limits) System limits not expressed in LMP Residential (exc. RR1), small commercial May include special terms (e.g., 1 yr harmless) May also include other resources TBD PUCO approval required

System requirements
Advanced Metering Infrastructure (AMI) Home Energy Manager (HEM) Advanced equipment controls
Heating systems (electric only) Air-conditioning system Hotwater heaters (electric only)

Resource control (e.g., CES strategies) Smart Grid Dispatch engine

RTP-DA Valuation

Determine costs/benefits of RTP-DA


Values included Wholesale energy production Generation capacity Ancillary services (regulation and reserves) Transmission congestion Distribution congestion Values excluded Scarcity pricing Subtrans. constraints Environment constraints Wind/bundling/firming Reactive power Emergency/reliability Financial transmission rights

How Does RTPDA work?


Invoke Day Ahead OPC Scheduler (Mainframe) DA

AEP OHIO BATTELLE RTP PROJECT


Enterprise Integration (EI)
Source Services EI Broker Target Services Day Ahead
d Da hea yA d

CSP Web Services

DA

Retrieve Day Ahead (Projected) LMP Prices

Day Ahead

Da

yA hea

Day Ahead 5 Min RT 5 Min RT

Day Ahead

ccs
Deliver AEP Zone LMPs

eMarket (XML API)


RT

RT

Retrieve Real Time LMP Prices

Real Time

Real Time (Both 5 Min & Hourly)

Real Time

5 Min RT Hourly RT

5 Min RT Hourly RT Hourly RT

Integ eDataFeed (XML API)


S

Hourly RT

Internet
Key: DA=Day Ahead, RT=Real Time, S=Settled

AEP Firewall

Retrieve Daily (Settled) LMP Prices

Daily Settled

ttle Se

Guaranteed Delivery

Se ttle

Settled

Settled Daily Settled

FTP Server

ftp://ftp.pjm.com/pub/account/lmp
OPC Scheduler (Mainframe)

Invoke Settled

upply

5m Input
JM (14 P

inute

RTPs

n statio

s) node

Common Solution Platform (CSP)

Circuit loads(80)

Usage Summarized

AEP - DAS

Demand Input

BATTELLE Application

Send RTP Prices Interval Usage

Data Store

CALCULATION WATCHDOG ENGINE Interval Amount

PJM LMP Transm. Node


(14 Nodes)

me usto

(~ 10

00 C

d5 tere r Me
5

Meter

R ute Min

P TP

s rice
oa eL ds

Daily Set-Up File Data

DIST LMP (80 Nodes) D Nodes

ustom

ers)

ta e Da inut

Interval Rate

Detail View for each 5 minute interval

Home Area Network

Cu

sto

me

rM

re ete

dA

ppl

ian

Interval Usage Send Bill Trigger Data & Retrieve Summary Level Changes

Interval Amount

Summary View Marginal Energy Cost

MACSS MAINFRAME

Cirucuit Loads

Circuit Load View

AMI Head-END

Appliance Load

Send Register Reads

Appliance Loads

Appliance Load View

Interval Data
Dynamic Prices Repository

Interval Data

Dynamic Prices Repository

MDM
RTP Display Data

RTP Display Data


2/23/2010
Graph Summary Detail

Customer

AEP. COM

Conclusions
Retail capacity markets
Energy price of Pareto-optimal allocation

Olypen project a simple/full example


Demonstrated basic concept Showed important of enabling technology

AEP NE Columbus project


Significant scaling up of implementation Stronger integration into wholesale operations

Questions/Comments

Contact: David P. Chassin Pacific Northwest National Laboratory [email protected]

You might also like