Linear Electricity Spot Market Constraints for Managing Post-Separation Frequency Deviations

Transcription

1 Linear Electricity Spot Market Constraints for Managing Post-Separation Frequency Deviations Stuart Thorncraft, IEEE PES GM, Tampa Florida, USA, June Outline consider motivation for work events conceptual view of security management in a restructured electricity industry post- dynamic power system model derivation of linear constraint sets from model security management & electricity market interface illustrative example conclusions and further work 2

2 NEM Model 14 Sept 03 NEM QLD Scope Freq: 50.3Hz X NEM characteristics long transmission lines between major load centres ~4000km Southern Island Freq: > 49.85Hz (within standards) increases risk of? (TAS not connected until Apr 06) ~2000km Linear Operation Source: Electricity of NEMMCO Ancillary Spot Service Market Constraints Markets for for Large Managing Amounts Post-Separation of Wind EnergyFrequency Deviations 3 8 March 04 NEM Scope NEM characteristics ~4000km long transmission lines between major load centres increases risk of? X SA Freq: 47.6Hz ~2000km Linear Operation Source: Electricity of NEMMCO Ancillary Spot Service Market Constraints Markets for for Large Managing Amounts Post-Separation of Wind EnergyFrequency Deviations 4

3 Jan NEM characteristics 15:03 ~4000km NEM Scope 3 islands long transmission lines between major load centres increases risk of? X X VIC Freq: Hz CBD blacked out ~2000km Linear Operation Source: Electricity of NEMMCO Ancillary Spot Service Market Constraints Markets for for Large Managing Amounts Post-Separation of Wind EnergyFrequency Deviations 5 Jan NEM characteristics 15:45 ~4000km NEM Scope => 2 islands long transmission lines between major load centres increases risk of? X commence load restoration ~2000km Linear Operation Source: Electricity of NEMMCO Ancillary Spot Service Market Constraints Markets for for Large Managing Amounts Post-Separation of Wind EnergyFrequency Deviations 6

4 Security & Commercial Concept unreachable or unacceptable outcomes present state 5 min security envelope uncertainty increases emergency control time suboptimal but feasible economically efficient set of future states ( optimal ) Source: H. Outhred, CEEM commercial regime possible futures managed by decentralised decisions security regime (futures managed by centralised decisions) 7 Security & Market Operation conflicting objectives between system operations & market participants power system operations concerned with keeping core system intact (has cost implications) identifying & preventing rare events (ly high impact ) uncertain physical outcomes => uncertain control actions market & industry participants profit maximisation => push system to boundary commercial transactions can t proceed if system fails consistent & objective decision-making framework reduces uncertainty security concept: system operators compute and apply a secure envelope secure envelope is based on objective criteria: system standards can the security space for managing post- frequency deviations be defined in this way? 8

5 Basic Definitions: Consider an under-frequency island post- frequency deviation (Hz) X load shedding? event initial rate of change time steady state frequency deviation commence restoration (or sooner) maximum frequency deviation 9 Post-Separation Dynamic Power System Model low-order frequency response model for each island (deviations): standard LTI system for each possible island 10

6 Steady-State Frequency Deviations set derivatives to zero & solve power output of generator j, island i power lost or gained in island 11 Initial Rate of Change of Frequency put t = 0 & consider initial conditions initial state (freq & power outputs) (not necessary) put initial conditions = 0 12

7 Maximum Frequency Deviation (1) use standard analytical expression solve for t to find extreme values of states find t = t max 0 to make zero find t = t j 0 to make zero 13 Maximum Frequency Deviation (2) Find t max & t j to get maximum deviations can also compute maximum output deviations 14

8 Automatic Constraint Generation gen. offers demand forecasts, etc. system operator input other triggers online linear post freq. constraint generator technical data spot market optimization (5-minute) contingency constraints status data prices dispatch measurements SCADA / EMS state measurements data collection database of technical parameters update parameters parameter or model estimation 15 Spot Market + (ex-ante) Post-Separation Frequency Constraints LP Optimization (1) basic LP market formulation (lossless model) post- steady-state linear frequency constraints 16

9 Spot Market + (ex-ante) Post-Separation Frequency Constraints LP Optimization (2) post- rate-of-change of frequency linear constraints post- frequency deviation linear constraints 17 Simple example: hypothetical scenario L1 & L2 (double outage) becomes credible => convey to system operator demand L1 zone 2 demand zone 1 L2 F1+F2 200MW Gens: G21, G22, Gens: G11, G12, post-contingency frequency standard: max freq dev: 0.3 Hz L1 + L2 outage will lead to & present conditions (e.g. bush fire or lightning) are of concern system operator decision: use tool to construct & invoke post- frequency constraints? 18

10 Simple example: without PSF constraints Gen MW Gmax G G G G G MW zone 2 550MW 200MW $15.50/MWh $14/MWh zone 1 Gens: G21, G22, Gens: G11, G12, dispatch cost: $14,950 Gen MW Gmax G G G G Simple example: post- outcome in zone 1 without PSF constraints violated postcontingency frequency standards 20

11 Simple example: post- outcome in zone 2 without PSF constraints violated postcontingency frequency standards some generators exceed limits 21 Simple example: with PSF constraints Gen MW Gmax G G G G G MW zone 2 550MW 61.22MW $16.50/MWh $13/MWh zone 1 Gens: G21, G22, Gens: G11, G12, flow backed-off & increased dispatch costs dispatch cost: $15, Gen MW Gmax G G G G out-of merit dispatch: reduced generation: head-room for Linear ensures Electricity G14 Spot >= Market Gmin Constraints (0MW) for Managing Post-Separation Frequency responding Deviations to event 22

12 Simple example: post- outcome in zone 1 with PSF constraints postcontingency frequency is OK: freq dev < Simple example: post- outcome in zone 2 with PSF constraints postcontingency frequency is OK: freq dev 0.3 generators within limits 24

13 Variation change in freq standards say the post-contingency frequency standard is modified to be: maximum post-contingency frequency excursion 0.3 Hz; and steady-state frequency deviations within 0.2 Hz Invoke 2 sets of post- frequency constraints: one set to ensure max frequency deviation is 0.3 Hz; and one set to ensure steady-state frequency deviations within 0.2 Hz 25 Simple example: with alternative PSF constraints Gen MW Gmax G G G G G MW zone 2 550MW 41.67MW $16.50/MWh $13/MWh zone 1 Gens: G21, G22, Gens: G11, G12, flow backed-off even more & dispatch costs a bit higher dispatch cost: $15, Gen MW Gmax G G G G slight variation to similar to previously gens Linear dispatch Electricity pattern Spot Market Constraints for Managing Post-Separation Frequency backed Deviations off (not as much) 26

14 Simple example: post- outcome in zone 1 with alternative PSF constraints postcontingency freuency OK 27 Simple example: post- outcome in zone 2 with alternative PSF constraints postcontingency frequency is OK deviations hit - 0.2Hz generators within limits 28

15 Conclusions events are rare but are high impact therefore warrants investigation into mitigation important to have a consistent & well-defined interface between security processes & electricity market that: enables system operators to protect the core system can still enable market to proceed shown a simple way of linking the following: dynamic power system model post-contingency frequency standards security envelope & system operator decision-making interface to electricity market while the process may only be used infrequently, it could prevent highcosts of a post- frequency collapse more research to be done though! 29 S. R. Thorncraft, H. R. Outhred, D. J. Clements, Linear Electricity Spot Market Constraints for Managing Post-, to be published in the Proceedings of the IEEE PES General Meeting Tampa Florida, USA, June , available: Many other CEEM publications are available at: 30