Day-Ahead Market Evolution Detailed Overview Presentation to - - PowerPoint PPT Presentation

Day-Ahead Market Evolution Detailed Overview

Presentation to Technical Panel July 8, 2008

2

Focus of Presentation

• Provide some detail (“nuts and bolts”) of the

“common elements” of the day‐ahead design

• Currently working on a detail design document

for the common elements which is expected to be presented to you through market rule amendments throughout the coming months

3

Reminder of Day-Ahead Design Work

• Identify day‐ahead mechanism improvements that

would result in net benefits to the province as a whole relative to the existing Day‐Ahead Commitment Process (DACP)

• Elements which provide the majority of benefits to

Ontario, specifically:

• 1. Day‐ahead unit commitment improvements (operational

and integration of the changing supply mix expected to emerge in the next few years)

• 2. Better scheduling incentives (includes imports/exports),

reduced transaction costs

• 3. Day Ahead Demand Response Efficiencies (providing
• pportunities to better manage demand response

are common to all options

4

Common Elements

• Common design elements enhancing

constrained algorithm includes:

– 3‐part bids/offers – 24‐hour optimized unit commitment process – multi‐passes of the constrained schedule using peak and average – adding day‐ahead exports – review and change the existing guarantees

• Discuss the design of each item above and

including how design differs from today algorithm

5

Why Three Part Offers/Bids

• Multi‐part offers/bids requires market participants to

submit not only the incremental cost of energy but also

• ther financial costs and physical limitations in advance
• f operations
• Allows all energy offer/bid components (total cost) to be

used in determining the most economic dispatch for entire system ‐ more accurately reflects the complex cost structure of some generators and loads

• Drives correct outcomes as these offers/bids accurately

reflect: – The cost of being available for dispatch – The physical restrictions of the load or generator

6

Typical Three-part Offer

• Minimum generation block
• Ramp rates
• Minimum run time
• Minimum down time
• Maximum stops per day
• Maximum daily energy limit

Incremental Energy Costs to reach/maintain minimum output Operational Restrictions

+ +

• Incremental energy offer
• Start‐up cost (cost to get to minimum loading

point)

• Minimum generation cost

7

Start-up Cost

• Start‐up cost is the cost to bring an off‐line

resource up to the minimum generation level

• Includes all unit specific procedures such as

ramp up and synchronisation

• Submission options:

– Hourly Start‐up cost – A single cost for each hour of the day – Time Since Last Stoppage – Cost is submitted as a function of the time the unit was last on line – No Start‐up Cost – A $0 startup cost

8

Ongoing Minimum Generation Cost

• Minimum generation cost is the ongoing cost to

maintain resource output at the minimum generation level

• Can be changed hourly

9

Minimum Level of Generation

• The Minimum Generation (or “Min Gen”) level

is the lowest output a facility is physically capable of sustaining while remaining on‐line

• Only for a short period during start‐up or shut‐

down can a resource be below this level

• Initially the Min Gen limit indicated during

Facility Registration

10

(MW)

40

($)

30 20 10 50 60

150 100 50

Min Gen.

Multi-Part Offer Example

Incremental Energy

Generator A offer:

• 50MW minimum: $2,500/hr
• 50MW to 100MW: $20/MWh
• 100MW to 150MW: $30/MWh

Generator A must operate at least to MW minimum

11

(MW)

40

($)

30 20 10 50 60

150 100 50

Multi-Part Offer Example

Generator A offer:

• 50MW minimum: $2,500
• 50MW to 100MW: $20
• 100MW to 150MW: $30

Dispatching Generator A to meet a 110MW demand results in a cost of $3,800 = $2,500 + 50MWx$20 + 10MWx$30

12

Multi-Part Offers Compared

$30 $50 $20 $31 110MW 110MW

Generator A Generator B

A market that incorporates 3‐part offers would schedule Generator B. Today’s market does not see the min‐gen offer and would schedule Generator A based on its lower incremental offer.

Generator B offer: 110MW at $31, no min‐gen or start‐up cost Cost of Generator A = $3,800 Cost of Generator B = $3,410

13

Difference From Today

• Currently single part (incremental) offer/bid

structure

• No consideration of the total cost of being

available for dispatch or the physical restrictions

• f the load or generator
• Physical limits and costs to reach/maintain

minimum output are used after the fact to constrain on units in multiple hours to respect minimum run times and settle production cost guarantees

14

Incorporating 3 part bids

• Initially costs to reach/maintain minimum
• utput or the physical restrictions to be

included as a static submission in registration data

• Software program will “scrape” information

into optimization process

• Timing of these submission offer/bid data

updates under review

• Upgrade to MIM in 2009/2010 will allow for

data to be submitted as part of offer/bid process

15

Integrating Three-Part

Offers/Bids in Real-Time

• “Guts” of real time (RT) algorithms will not

change from their current form ‐ incremental energy only

• Evaluation process for changes to unit

commitments after close of the DA process is completed will be addressed by a manual process or aided by an offline tool ‐ limited use

• f the three‐part offers/bids in real‐time

16

What is 24 Hour Optimization?

• Dispatch model that determines the optimal

commitment (minimizing total of all as offered costs) and scheduling of generation and load response to meet energy and operating reserve requirements over the entire day

• Uses three‐part offer/bid structure
• In contrast current model optimizes hour‐ by‐

hour with no regard to future hours

17

Multi-Hour Optimization Example

• Assume the following multi‐part offers :

Start‐up Cost Minimum

• Gen. Cost

Incremental Energy Offer Minimum Load Point Minimum Run‐Time Generator A (400MW) $5000 $5500 $30/MWh 300MW 4 hours Generator B (500MW) $6000 $7000 $40/MWh 100 MW 2 hours Generator C (100MW) N/A N/A $50/MWh 0 MW N/A

18

Multi-Hour Optimization

• Today’s market would commit Generators A and B based on

incremental energy offers only

• Generator B would be constrained ON in Hour 5 to respect its

minimum run‐time, constraining down Generator A

Hour 1 400 MW Demand Hour 2 400 MW Demand Hour 3 400MW Demand Hour 4 500 MW Demand Hour 5 400 MW Demand Generator A schedule 400 MW 400 MW 400 MW 400 MW 300 MW Generator B schedule 100 MW 100 MW Generator C schedule

19

Multi-Hour Optimization

• 24 Hour optimization enables a comparison of

total costs, including start‐up for the dispatch day to arrive at least cost reliability commitment

500 MW 400 MW 300 MW 200 MW 100 MW $5,000 Generator C Cost $29,000 $25,000 $21,000 $17,000 $13,000 Generator B Cost $13,500 $10,500 Generator A Cost

20

Multi-Hour Optimization

Hour 1 400MW Demand Hour 2 400MW Demand Hour 3 400MW Demand Hour 4 500MW Demand Hour 5 400MW Demand Generator A schedule 400MW 400MW 400MW 400MW 400MW Generator B schedule Generator C schedule 100MW

With 24 hour optimization, a more expensive incremental energy offer may be selected to save the cost of start‐up and minimum run for the extra 100MW Also impacts constrained on/off payments

21

Difference From Today

• Constraints are not considered by the

current DACP when it determines whether to commit a resource, causing the resulting commitment to be inefficient at times

• CBA analysis illustrated 3 part offers/bids

along with 24 hour optimization will resulted in approx. savings of $5M/yr through unit commitment efficiencies

22

Multi-pass Constrained Algorithm

• Calculation engine design based on multiple

passes:

– First pass commits resources to forecasted average demand – Second pass adds an additional resources required to meet peak – Lastly pass use commitments from second pass to calculate hourly advisories based on average

• Using today’s signal pass method based on peak
• verstates requirements impacting efficient

market operation of both IESO and participants

23

Multi-Pass Runs

• f Constrained Algorithm

Pass 1 Constrained Run Based on Average Forecast to Obtain Unit Commitments Ontario demand forecast less dispatchable load bids Dispatchable load bids Import offers Generation

• ffers

Export bids Pass 2 Constrained Run Based on Peak Forecast to Ensure Sufficient Unit Commitment to Meet Expected Peak Demand – Uses a bias to solve with internal bids and offers

Blocked on from pass 1:

• Non‐quick starts (cannot go below minimum)
• Import and export schedules

Pass 3 Constrained Run Based on Average Forecast to Obtain Advisory Schedules

This should be a very quick pass of the algorithm as there is no additional unit commitment than what was determined in Pass 2 The schedules produced by Pass 3 are the day‐ahead constrained schedules (DACS)

24

Rule Amendments

• Additions of 3 part offers/bids, 24 hour
• ptimization and multiple security constrained

passes rules to reflect changes to “calculation engine”, (dispatch scheduling and optimization engine) will be reflected in Appendix 7.5 Market Clearing and Pricing Process

25

Why Production Cost Guarantee for Non-quick Starts

• Possible that the total amount a generator is paid

for scheduled energy will be less than the sum of as‐offered costs for scheduled start‐ups, minimum generation, and incremental energy

26

Quantity (MW) 40 50 Price ($/MW)

$50

30

On‐going cost to maintain resource at minimum output level

50 MW Generator:

• Start up cost to minimum

generation is $3,000

• Minimum output is 30 MW

– 0 to 30 MW: $70 ($2,100/hr)

• Minimum runtime is 4 hours
• The incremental price per MW

above minimum is:

• 30 to 40 MW: $60
• 40 to 50 MW: $65

PCG – Generator

$100

27

PCG – Generator

• Schedule is for 4 hours at 40 MW
• Cost to generator to start up and run for 4 hours is

$3000 + 4 x $2,100 + 4 x 10 MW x $60 = $13,800

• If the real‐time clearing price is $60, revenue will

be: $60 x 40 MW x 4 hours = $9,600

• This schedule would leave the generator out of

pocket – some sort of guarantee is required.

28

Production Cost Guarantee for Non-quick Starts

• Offer a guarantee that the participant will be

kept whole to the costs associated with the day ahead advisory schedule

• But challenge is to provide an appropriate

guarantee incentive to operate to day‐ahead advisory schedules but also to make efficient changes based on RT market conditions

29

Difference From Today

• Under the proposed method for calculating PCG

payments for made to an internal generator would consist of three elements:

– Any shortfalls in payment of RT delivery of DA constrained advisory minus the real‐time revenue received for that amount

• f energy will based on day ahead offer PCG = max(0, (DAO‐

RTP)) – For the portion of DA constrained advisory that is not implemented in RT (dispatch schedule), the PCG will guarantee the cost of arranging the delivery PCG = max(0, (DAO‐RTO)) – Any income from RT CMSC included in a generator’s DA constrained advisory will be used to reduce the PCG payment

• Guarantee will be based on advisory schedule versus

today’s minimum load point

30

Rule Changes

• Chapter 9 rules used to settle guarantees
• For example Chapter 9 4.7D to settle guarantees

against minimum loading point calculations will need to be amended

31

Export Excluded From DACP

• Working group unable to identify a means of

including exports, that did not either:

– “Require significant administrative measures to ensure exports included and/or committed day‐ahead did not result in over committing supply with the cost of the corresponding reliability guarantees borne by Ontario customers; or – Require a complex market mechanism to allow committed exports to buy out of their day‐ahead position when real‐time conditions warranted”

32

Day-Ahead Exports

• Goal ‐ inclusion of export will not degrade unit

commitment efficiencies

• Assumption now that with an appropriate

export failure charge and offsetting guarantee claw backs, design will not:

– penalize exports financially for being in the DACP rather than the RTM, – provide any incentive to over schedule exports in the day‐ahead

• Inclusion under review at TSG

33

Rule Changes

• Removing exclusion of exports from DACP
• Day‐ahead export failure charges similar to

today’s market for imports

• Automatically applied to day‐ahead advisory

transactions that fail in whole or in part in real‐ time

– Does not apply to failures for reasons outside of exporter’s control

34

Rule Changes Cont

• Day‐ahead and real‐time viewed as separate

markets for the purposes of offer cost guarantee and offsets

• Scheduled exports receive a day‐ahead

guarantee, based on a break‐even assessment of the transaction profits

• Chapter 9 settlement and billing amendments

anticipated

35

Other Guarantee Consideration

• Looking at the need to keep SGOL in current

form when changing day‐ahead guarantee policy

• Ensuring day‐ahead guarantee for imports

36

• Questions