PORTFOLIO OPTIMIZATION FOR OPEN ACCESS CONSUMERS/DISCOMS By Dr. PARUL MATHURIA POST DOCTORAL FELLOW DEPARTMENT OF INDUSTRIAL AND MANAGEMENT ENGINEERING INDIAN INSTITUTE OF TECHNOLOGY KANPUR 2017 15-05-2017 1
BID AREAS IN INDIA 1 N1 North Region Jammu and Kashmir, Himachal Pradesh, Chandigarh, Haryana 2 N2 North Region Uttar Pradesh, Uttaranchal, Rajasthan, Delhi 3 N3 North Region Punjab 4 E1 East Region West Bengal, Sikkim, Bihar, Jharkhand 5 E2 East Region Orissa 6 W1 West Region Madhaya Pradesh 7 W2 West Region Maharashtra, Gujarat, Daman and Diu, Dadar and Nagar Haveli, North Goa 8 W3 West Region Chhattisgarh 9 S1 South Region Andhra Pradesh, Telangana, Karnataka, Pondicherry (Yanam), South Goa 10 S2 South Region Tamil Nadu, Pondicherry (Puducherry), Pondicherry (Karaikal), Pondicherry (Mahe) 11 S3 South Region Kerala Source: IEX 12 A1 North East Tripura, Manipur, Mizoram, Nagaland Region 13 A2 North East Assam, Arunachal Pradesh, Meghalaya Region
Source: IEX VOLATILITY
HISTORICAL DATA OF MCP APRIL 2013-1017 Source: IEX
WHY ELECTRICITY PRICES REPRESENTS HIGH VOLATILITY? ISSUES Demand supply balance Non-storable nature of electricity Trading decisions are made well in advance Prices depends upon the real time conditions REASONS Uncertain demand Availability of production units & network components Power production of non-dispatchable generators Availability of generation resources Energy prices of other markets such as fuel, emission Legal reasons (market rules & structure) Others 15-05-2017 5
MARKET TIMEFRAME Timeline of Participation Long term Medium term more than a year Week to year more than one day ahead Short term one day ahead real-time operation Construction & Investment Planning Hedging against the price risk & optimizing the financial part of the power portfolio Optimizing physical part of the power portfolio Balancing Market Long Term Power Purchase Agreements Forward Market Day Ahead Market Intraday Market Scheduling own generation for real-time 15-05-2017 6
PERFECT MARKET Many Buyers many eligible consumers/retailers with the willingness & ability to buy the product ata certain price Many Sellers with the willingness & ability to supply the product at A certain price No Market Power due to competition no seller can abuse his position & control prices Sufficient Liquidity sufficient traders so that planned trading is achievable Price Taker firms aim to sell where marginal costs meet marginal revenue Regular Market Updates for both consumers & producers Homogeneous Products the products of the different firms are similar 15-05-2017 7
RISK & UNCERTAINTY RISK A chance that future value of considered parameter would be different than expected Viewed as A negative Possibility of suffering harm or loss Costs of future uncertainty REASONS No information about future events at the time of planning Exact estimation is not possible UNCERTAINTY SOURCES INCLUDES Technical, Institutional & Legal issues 15-05-2017 8
RISK MANAGEMENT METHODOLOGY THAT MAKE BEST USE OF AVAILABLE RESOURCES THREE STEPS PROCESS RISK IDENTIFICATION RISK ASSESSMENT RISK CONTROL Risky v/s Risk Free trading options MANY POSSIBLE OBJECTIVES: To minimize exposure to risk To maximize profit for A controlled level of risk Optimum selection of risk-return trade-off 15-05-2017 9
RISK CONTROL V/S RISK MITIGATION MANAGEMENT Diversification Risk sharing Uncertain outcomes are correlated to reduce certain variability Interdependency AVOIDANCE Hedging Contingent claims Contractual arrangement as insurance Controlling financial consequences
DIVERSIFICATION RISK CONTROL BY DIVERSIFICATION Diversification is about diversifying the investment in multiple trading options, so that exposure to risk associated with any particular asset is limited This concept is applied through portfolio construction by investing energy in available different trading options.
DERIVATIVE TRADING/ HEDGING Having a Position In Security Using Derivatives Trading with Financial Instruments or Contracts (Agreements) such as Forward, Future, Option, Swap, CfD, FTR Or TCC Limitations Market Of Hedging Contracts Is Limited Requires Additional Payment Restricts Opportunities For Higher Profit 15-05-2017 12
PORTFOLIO OPTIMIZATION PORTFOLIO Energy combination of available trading approaches Aiming to maximizing participants benefits (profits/ returns/ cost) & minimizing the corresponding risk Substantially reduces the variability of returns without an equivalent reduction in expected returns There is a reward for bearing risk 13
MARKET MECHANISM Two Types Of Markets Mandatory Transaction Notification PHYSICAL MARKET Spot Market (Exchange) Bilateral Contracts (OTC) GenCos Bilateral/OTC Transactions Pool Trading Loads FINANCIAL MARKET Forward, Future, Option, Transactions are scheduled by MO+SO Swap, CfD, FTR Or TCC Power Pool Derivative Instruments FORWARD FUTURE OPTION SWAP OTC (Over the counter ) trading Exchange Traded Derivatives Pool Day-ahead Market Adjustment Market Balancing Market 15-05-2017 14
MARKET PRODUCT PORTFOLIO INDIA Source: PPT 2016, Mr. Prasanna Rao, IEX
BUYERS IN ELECTRICITY MARKET Buyers Open Access Consumers Captive Consumers DisComs Retailers/ Aggregators ALLOWED TO TRADE IN POWER EXCHANGE With Higher Voltage Grade Larger Power Consumption Procures Electricity for Forecasted Demand Risk of price Risk of availability of transmission corridor Risk of getting cleared in market OBJECTIVE Minimize Total Purchasing Cost Minimize Risk
RETAILERS V/S DISCOMS Retailers are subsidiary of a DISCOM Manage two sets of contracts, on supply & demand side Supply Side: Electricity procurement from various contracts and pool for fulfilling customer demand Demand Side : Obliged to serve varying customer demand Retailer s RM problem is basically bi-level optimization problem Purchase Cost Minimization Selling Price Determination with consideration of elastic nature of demand 15-05-2017 17
POWER PROCUREMENT PROBLEM Participate in wholesale trading Procures electricity for its known demand Optimally decide its mix of electricity purchase from Pool, (day ahead ) Bilateral contracts (local and non-local) Supplier 2 Supplier 3 Self Generation Self production Free to purchase from any supplier, irrespective of its location Supplier 1 Large Consumer Spot Market Prices are correlated with each other 15-05-2017 18
PROCUREMENT COST Bilateral contract, with home location supplier T B B B 1 1, t 1, t t1 C P for i 1 Bilateral contract with supplier of non-home location would be T B B S S B i i, t 1, t i, t i, t t1 Spot market T C 1, tpt Self-generation Facility Total electricity procurement cost 2~ C P for i n S S S t1 T C c u b ( P ) a ( P ) c G G G 2 su t t t t t1 n S G B i i1 CP C C C 15-05-2017 19
PURCHASE PORTFOLIO SELECTION Expected Procurement Cost Risk of Cost Minimize Risk Weighted Cost N P 0 0 i i i1 E C w C w E C 2 min Z E CP P N wi 1 w i 0 i0 E C C C E C E C i2 n G B S B P 1 i N N N N N w w CovC, C w Var C w w Cov C, C 2 2 P i j i j i i i j i j i1 j1 i1 i1 j1 S B S B B B n n n n Var( C ) Var C Var C 2 Cov C, C Cov C, C 2 P P i i i j i2 i2 i2 j2 i j i j 15-05-2017 20
OVERALL OPTIMIZATION PROBLEM OBJECTIVE FUNCTION POWER BALANCE CONSTRAINT STARTUP COST GENERATION LIMITS RAMP UP LIMIT RAMP DOWN LIMIT LIMITS ON BILATERAL CONTRACTS min Z E CP. w c u i t su i, t, t, t,, n S G B t t t i, t i1 PD P P P t c c u u t su su t t t1 P u P P u t G G G min t t max t P P R u t G G up t t1 t P P R u t G G dw t1 t t1 P v P P v t B B B i min i, t i, t i max i, t 2 P VARIABLE DECLARATION CONSTRAINT S su P, c 0 t t i, t 15-05-2017 21 t t u, v 0,1 t
CASE STUDY Large Consumer located at APS Case Study Of PJM Electricity Market Planning Period Is 120 Hours, With Each Hour As A Trading Interval TABLE I Specifications for Bilateral Contracts Contract Index Location Contracted Price Minimum Limit per hour Maximum Limit per hour Contract 1 APS 52 $/ MWh 60 MW 400 MW Contract 2 PECO 56.5 $/ MWh 20 MW 200 MW Contract 3 DOM 58.5 $/ MWh 50 MW 500 MW TABLE II Specifications for Self-Generation Facility Total capacity 120 MW Minimum power output 20 MW Ramp rate 80 MW/h Quadratic cost 0.01 $/MW 2 h Linear cost 42 $/ MWh No-load cost $ 600 Start-up cost $ 200 TABLE III Variance-Covariance Matrix between Uncertain Costs at 0.0001 Spot Market Contract 2 Contract 3 Spot Market 1207830135-96097141.9-382901659.8 Contract 2-96097141.9 71896411.19 40024748.14 Contract 3-382901659.8 40024748.14 708777796.3 15-05-2017 22
Price $/MWh Price in $/MWh Demand in MW CASE STUDY 110 100 90 80 70 60 50 40 30 20 APS DOM PECO 1 11 21 31 41 51 61 71 81 91 101 111 Hours Day ahead LMPs of three different locations 56 54 52 50 48 46 44 42 40 38 570 550 530 510 490 470 450 430 410 390 Contract 1 Contract 2 Contract 3 1 11 21 31 41 51 61 71 81 91 101 111 Hours Demand data 1 11 21 31 41 51 61 71 81 91 101 111 Hourly expected procurement price from risky bilateral contracts 15-05-2017 23 Hours
RESULTS Expected Portfolio Cost $X10 6 3 2.9 2.8 2.7 2.6 2.5 2.4 2.3 SCENARIO I WITH CORRELATION SCENARIO II WITHOUT CORRELATION α=0.1 2.2 0.5 1.5 2.5 3.5 4.5 5.5 6.5 15-05-2017 Standard Deviation $X10 4 Efficient Frontier Scenario I Scenario II α=0 Traded Power in MWh X 10 3 Traded Power in MWhX10 3 30 25 20 15 10 5 0 40 35 30 25 20 15 10 5 0 Spot Market Contract 2 Contract 3 Spot Market Contract 2 Contract 3 0 0.002 0.004 0.006 0.008 0.01 Risk weighing factor α (a) Risky Procurement Options Self-generation Contract 1 Self-generation II Contract 1 II 0 0.002 0.004 0.006 0.008 0.01 Risk weighing factor α (b) Risk-free Procurement Options Electricity purchase from different contracts for various values of α 24
Energy in MW Energy in MW RESULTS 600 500 Spot Market Self Generation Contract 1 Contract 2 Contract 3 Demand 400 300 200 100 600 500 0 1 21 41 61 81 101 Hours Mix of electricity purchase for each trading interval at α =0 Spot Market Self Generation Contract 1 Contract 2 Contract 3 PD 400 300 200 100 0 1 21 41 61 81 101 Hours Mix of electricity purchase for each trading interval at α =0.01 15-05-2017 25
OPEN ACCESS CONSUMER: INDIAN CONTEXT. Open Access Consumer Price uncertainty and demand flexibility Short term power trading Unscheduled Interchange (UI) Mechanism Renewable purchase obligations (RPO) UI Charge for deviation from Scheduled withdrawal Frequency linked UI charge FiT REC RPO 5/15/2017 26/16
PROBLEM DESCRIPTION UI Mechanism Part of Availability Based Tariff Penalty for deviation from schedule (against grid frequency) Incentivizes to support grid frequency Real time balancing mechanism Maintain grid frequency in narrow band Post transaction charges Deviation Settlement Mechanism and Regulations (DSM, 2014) RPO Fixed percentage renewable energy purchase FiT contracts as long term PPAs FiT near to cost of production of renewable energy RECs as environmental attributes 1 REC = 1 MW h of electricity injected into grid. RECs traded in PXs 5/15/2017 27/16
PROBLEM DESCRIPTION Mean- Variance Open Access Large Consumer UI Charge Frequency Indian Grid System Demand Generation Self generation DA Contracts RPO Bilateral Contracts IEX DA PXIL DA 5/15/2017 7 REC FiT
PROBLEM DESCRIPTION Grid frequency is calculated from [12] f t = Grid frequency L t =System Load G t =System generation PFR= Power deficit- frequency fall ratio f t Lt [ Gt UIt ] 50 PFR * L t Mean Variance approach for Indian case study. Demand shifting using flexibility in projected mind accounting UI deviations 5/15/2017 29/16
OBJECTIVE To develop a planning model for short term power procurement of a large Indian electricity consumer considering uncertainties (DAM price) and renewable promotional policies while addressing real time grid frequency imbalances using demand flexibility. 5/15/2017 6/16
MODELLING Objective Cost and Risk Minimization Cost = Cost of power purchase from (bilateral contracts + Spot Markets + FiT Contacts+ Self Generation) + UI Penalty/Revenue. Risk = Uncertainty of spot market prices Constraints Demand Balance Base Demand + Demand Fluctuations = Shifted Demand + UI deviations Expected Demand = Scheduled Demand RPO Purchasing a percentage from FiT contracts 5/15/2017 6/16
PLANNING MODEL Minimum and Maximum Purchase Constraints Bilateral contracts, Spot market Self Generation Constraints Quadratic Cost Function Minimum and Maximum Generation Ramp up and Ramp down UI Charge Calculated from grid frequency Deviation limitations according to DSM 2014. 5/15/2017 6/16
b1 Average DAM Price (Rs./MW) Actual Demand (MW) CASE STUDY Generation Unit. 5000 4000 PXIL avg price IEX avg price Capacity Minimum power output Ramping limit (up/down) Quadratic Cost 120 MW 20 MW 80 MW 0.6 Rs./(MW) 2 h 3000 2000 Linear Cost No-load Cost Startup Cost 2700 Rs./MWh 2000 Rs. 1000 Rs. 0 50 100 150 Time (Hours) Other Data Values Bilateral contract price 3000 Rs./MW h Min./Max bilateral vol. 30 MW/800MW 1110 1060 1010 960 0 50 100 150 limit Demand Flexibility 12% Min./Max limit on SI 900 MW/ 1100 MW Min./Max limit on flexible -40MW/40 MW load RPO, PFR 10 %, 4% Trading intervals 168 hours Time (hours) RPO purchase price 5000 Rs./MW 5/15/2017 System demand/gen. 100 GW 33/16
UI Allocation (MW) Frequency (Hz) Frequency (Hz) Cost ( 10 6 Rs.) Power Procurement (MW) 617 612 607 602 597 592 Efficient Frontier 587 1.4 3.4 5.4 Standard deviation ( 10 6 Rs.) RESULTS 150000 100000 50000 Self Generation IEX DAM 0 Bilateral PXIL DAM 0 0.000005 0.00001 α UI Allocation Grid Frequency Improved Frequency Grid Frequency 5/15/2017 0-20 -40-60 -80-100 -120-140 0 100 Time (Hours) 50.06 50.04 50.02 50 49.98 49.96 49.94 50.04 50.02 50 49.98 49.96 0 100 Time (Hours) 12/16
Demand (MW) Demand (MW) RESULTS SCHEDULED DEMAND EXPECTED DEMAND 1250 1200 1150 1100 1050 1000 950 900 850 1 49 97 145 Time (Hours) 1250 Scheduled Demand Expected Demand 1150 1050 950 5/15/2017 850 0 10 20 30 40 Time (Hours) 12/16
THANK YOU 15-05-2017 36