Download - MT and ST PASA
Agenda
2
1. Recap Section 3.16 of the WEM Rules
2. Understanding Key Componentsa) ST and MT PASA Market Model Concept
b) Inputs and Pre-processing
c) ST and MT Scenarios Modelling
d) Outputs & Post -Processing
3. Next Steps
Section 3.16A of the WEM Rules – Projected Assessment of System Adequacy (PASA)
• AEMO will be required to conduct periodic PASA assessments and publish a rolling MT PASA and ST PASA as set out in clause 3.16.1• At least each week of the 36 month period from the starting date of the assessment (MT PASA)
• At least each day of the seven day period from the starting date of the assessment (ST PASA)
• Clause 3.16.2 set out the new PASA objective linking it to Power System Security and Power System Reliability framework.
• Clause 3.16.3 set out the obligation on Rule Participants to provide the information for AEMO to conduct and prepare MT PASA and ST PASA. Details of the information required will be set out in the WEM Procedure.
• Clause 3.16.7 deals with forecasting, specifically the 36-month forecast as the week ahead forecast is covered in the SCED rules under 7.3 (Forecast Operational Demand).
• Clause 3.16.8 sets out the core requirements of the MT and ST PASA report which are linked to the objectives set out in clause 3.16.2.
• Clause 3.16.9 enables AEMO to issue an updated MT PASA or ST PASA where there has been a material change.
• Clause 3.16.10 provides head of power for AEMO to develop WEM Procedure which contains the details in respect of the PASA framework.
ST PASA Base Market Model
Current Bids and Offers
WEMDE
Realtime Constraints
Future Outage Constraints
Demand Forecast
Dispatch Targets
Binding Constraints
Violating Constraints
Dispatch Targets
Binding Constraints
Violating Constraints
Repeat for each interval
SCADA ParametersRegistrationSystem Normal
Constraints
Variable input data for ST
PASA
Static (or derived) input data
Dispatch Targets used as a proxy for
SCADA for subsequent intervals
A collection of WEMDE output data for each interval of
the horizon
Reflects non-System Normal
constraints applied by controllers
Reflects when outage constraints
will be applied/removed in the
future
Potential parameters could be:• DFCM (RoCoF)• Load relief • Engine parameters (max
iterations, etc)
Overview ST PASA Conceptual Diagram
Current Bids and Offers
Current Realtime Constraints
Known Future Outage Constraints
Expected Demand Forecast
Publish the following:
• Full data report of all intervals for each scenario
Identify the following:
• Not enough online capacity to meet demand (violating energy constraint)
• Shortages of ESS (violating ESS constraints)
• Other security issues (other violating constraints)
• Occurrence of network constraints binding
• Occurrence of DSP dispatch
And determine likelihood based on:
• Number of occurrences• Scenario combinations that
occurrences appear in
WEMDE
Varied Bids and Offers
Bids and Offers
Varied Realtime Constraints
Varied Outage Constraints
Varied Bids and Offers
Demand Forecast
Constraints
Outages
Demand
Repeat for each interval
Dispatch Targets
Binding Constraints
Violating Constraints
Dispatch Targets
Binding Constraints
Violating Constraints
Dispatch Targets
Binding Constraints
Violating Constraints
Repeat for each combination of input scenarios
• Identify and issue Low Reserve Notifications
• Identify and manage any dispatch interventions
Varied Bids and Offers
SCADA ParametersRegistrationSystem Normal
Constraints
Different NSG/SSG Forecast Quantities
Unplanned Outages
Variations on in/out of service quantities
Changes to outage start/completion times
Different demand scenarios
Combinations of input data to form
input scenario s
Different variations of input data (see below for examples)
A collection of output data for
each interval in the horizon, from each
of the different scenario s
Run output processing on the data to identify risks
Highlight type of LRC and likelihood of occurrence
Variations on DSP available quantities
Static (or derived) input data
Inputs and Pre-processing
Current Bids and Offers
Current Realtime Constraints
Known Future Outage Constraints
Expected Demand Forecast
Varied Bids and Offers
Varied Realtime Constraints
Varied Outage Constraints
Varied Bids and Offers
Demand Forecast
Varied Bids and Offers
Different NSG/SSG Forecast Quantities
Unplanned Outages
Variations on in/out of service quantities
Changes to outage start/completion times
Different demand scenarios
Variations on DSP available quantities
ST PASA Input VariationsInput variations considerations
Intermittent (NSG/SSG) variations• I1 - Base (per offers)• I2 - Base + statistical error margin (e.g. per Facility, per time of year/
time of day)• I3 - High output (e.g. 95% capacity)• I4 - Low output (e.g. 5% capacity)• I5 - Reserve capacity levels
Availability (In/Out of Service Quantity) variations• A1 - Base (per offers)• A2 - Base + adjusting in service quantities for large/key Facilities to zero
Realtime Constraint variations• RT1 - Base (currently applied realtime constraints)• RT2 - Base + random forced outages (e.g. based on statistical outage
rates)• RT3 - Base + largest remaining contingencies
Future Outage variations• O1 - Base (current planned outage start/finish times)• O2 - Early commencement/late finish variations• O3 - Base + largest network contingency
DSP available quantity variations• DSP1 - Base (per offers)• DSP2 - Base + statistical error margin (e.g. per
DSP, per time of year/time of day)• DSP3 – low availability (5% capacity)
Demand variations• D1 - Expected demand• D2 - Expected demand + statistical error margin• D3 - Expected demand + low PV output• D4 - Expected demand + high PV output• D5 - Expected demand + highly variable PV
output• D6 - High demand• D7 - Low demand• D8 - POE10 demand
We will define valid combinations of these input variations to form the overall list of input scenario s for ST PASA
E.g.
• Base scenario (all base input data) = I1, DSP1, A1, RT1, O1, D1• High capacity risk scenario = I4, DSP3, A2, RT3, O2, D7• Medium capacity risk scenario = I2, DSP2, A1, RT3, O2, D5Etc
Scenario’s Modelling
WEMDE
Bids and Offers
Constraints
Outages
Demand
Repeat for each interval
Dispatch Targets
Binding Constraints
Violating Constraints
Dispatch Targets
Binding Constraints
Violating Constraints
Dispatch Targets
Binding Constraints
Violating Constraints
Repeat for each combination of input scenarios
SCADA ParametersRegistrationSystem Normal
Constraints
Combinations of input data to form
input scenario s
A collection of output data for
each interval in the horizon, from each
of the different scenario s
Static (or derived) input data
Outputs Post Processing
Publish the following:
• Full data report of all intervals for each scenario
Identify the following:
• Not enough online capacity to meet demand (violating energy constraint)
• Shortages of ESS (violating ESS constraints)
• Other security issues (other violating constraints)
• Occurrence of network constraints binding
• Occurrence of DSP dispatch
And determine likelihood based on:
• Number of occurrences• Scenario combinations that
occurrences appear in
Run output processing on the data to identify risks
• Identify and issue Low Reserve Notifications
• Identify and manage any dispatch interventions
Highlight type of LRC and likelihood of occurrence
ST PASA: Data Requirements
• Given shorter timeframes, most information is expected to come through offers
• Will need modelling data from participants to support estimating different output scenarios (e.g. wind turbine power curves)
• Variability estimated from historical data
• Risks based on including contingency events (e.g. loss of next largest generator, major transmission outage, loss of major ESS provider, etc)
MT PASA Market Model: Key Requirements
• Approximation of single node wholesale electricity market model
• Mixed integer linear programming (MILP) solver to co-optimise energy and essential system services provision for every 30-min interval in the MT PASA horizon (time-sequential)
• Handles linear constraint equations (security-constrained dispatch)
• Determines estimates of unit commitment (time-sequential over specified commitment horizon)
• Randomised generator forced outages
• Uses generator planned outage schedules
• Assesses Power System Security/Reliability risks based on different input scenarios
MT PASA Base Market Model
Market model (e.g. Plexos) based on :• co-optimised (energy & ESS), • gross pool,• single node with linear network constraints provided from external source
Market Model manages the modelling inputs such as:Unit CommitmentFacility bid creation (Economic parameters)Synthetic maintenance outages (monte carlo)Randomised forced outages (monte carlo)ESS Requirements (CR and RoCof)
Future Network Outage /ESS Constraints
Demand Forecast
Dispatch Targets
Binding Constraints
Unserved Energy
Repeat for cost variability(internal in
model)
Intermittent Forecast
Dispatch Targets
Binding Constraints
Unserved Energy
Loss FactorsFacility and
Standing DataGenerator
Planned OutagesSystem Normal
Constraints
Future generator/network
augmentation
Other Rule Participant Data
DSP Forecast Available Quantities
LOLP
Variable input data for MT PASA
Static (or derived) input data
A collection of output data for each interval of the horizon allowing for price sensitivity
Reflects when outage/ESS
constraints will be applied/removed
in the future
A collection of output data for each interval
of the horizon
Assumptions around known changes to the power system (e.g. new generators, network changes)
Information such as heat rates, fixed/variable O&M, fuel costs/contracts, energy constraints, unit modelling data/limits)
Overview MT PASA Conceptual Diagram
Publish the following:
• Full data report of all intervals for each scenario
Identify the following:
• Not enough online capacity to meet demand (violating energy constraint)
• Shortages of ESS (violating ESS constraints)
• Other security issues (other violating constraints)
• Occurrence of network constraints binding
• Occurrence of DSP dispatch
And determine likelihood based on:
• Number of occurrences• Scenario combinations that
occurrences appear in
Market model (e.g. Plexos) based on :co-optimised (energy & ESS), gross pool,single node with linear network constraints provided
Market Model manages the modelling inputs such as:Unit CommitmentFacility bid creation (Economic parameters)Synthetic maintenance outages (monte carlo)Randomised forced outages (monte carlo)ESS Requirements (CR and RoCof)
Intermittent Forecasts
Outages
Demand
Repeat for cost variability(internal in
model)
Repeat for each combination of input scenarios
Dispatch Targets
Binding Constraints
Unserved Energy
Dispatch Targets
Binding Constraints
Unserved Energy
Dispatch Targets
Binding Constraints
Unserved Energy
Dispatch Targets
Binding Constraints
Unserved Energy
• Identify and issue Low Reserve Notifications
• Identify and manage any dispatch interventions
Loss FactorsFacility and
Standing Data
Generator Planned Outages
System Normal Constraints
Future generator/
network augmentation
Other Rule Participant Data
DSP availability
Known Future Network Outage
Constraints
Expected Demand Forecast
Varied Network Outage Constraints
Demand Forecast
Intermittent Forecast
DSP Available Quantities
Other Rule Participant
Data
Future generator/
network augmentation
System Normal
Constraints
Generator Planned Outages
Loss Factors
Facility and Standing Data
Combinations of input data to form input scenario s
A collection of output data for each interval in the horizon, from each of the different
scenario s
Run output processing on the data to identify risks
A collection of output data for each interval of
the horizon
Highlight type of LRC and likelihood of occurrence
Static (or derived) input data
Assumptions around known changes to the power system (e.g. new generators, network changes)
Information such as heat rates, fixed/variable O&M, fuel costs/contracts, energy constraints, unit modelling data/limits)
Changes to outage start/completion times
Different demand scenarios
Different variations of input data (see below for examples)
Different intermittent generation forecasts
Different DSP availability assumptions
Static (or derived) input data
Inputs and Pre-processing
Known Future Network Outage
Constraints
Expected Demand Forecast
Varied Network Outage Constraints
Demand Forecast
Intermittent Forecast
DSP Available Quantities
Other Rule Participant
Data
Future generator/
network augmentation
System Normal
Constraints
Generator Planned Outages
Loss Factors
Facility and Standing Data
Changes to outage start/completion times
Different demand scenarios
Different variations of input data (see below for examples)
Different intermittent generation forecasts
Different DSP availability assumptions
Static (or derived) input data
MT PASA Input Variations
Input variations considerations:
Intermittent (NSG/SSG) variations:• I1 – Statistical time of year (e.g. per Facility, per time of year/time of day)• I2 - High output (e.g. 95% capacity)• I3 - Low output (e.g. 5% capacity)• I4 - Reserve capacity levels
DSP available quantity variations:• DSP1 - Base (per offers)• DSP2 - Base + statistical error margin (e.g. per DSP, per time of year/time of
day)• DSP3 – low availability (5% capacity)
Future Outage variations:• O1 - Base (current planned outage start/finish times)• O2 - Early commencement/late finish variations• O3 - Base + largest network contingency
Demand variations:• D1 - Expected demand• D2 - High demand• D3 - Low demand• D4 - POE10 demand• D5 – POE50 demand
We will define valid combinations of these input variations to form the overall list of input scenario s for MT PASA
E.g.
• Base scenario = I1, DSP1, O1, D1• High capacity risk scenario = I3, DSP3, O3, D4• Medium capacity risk scenario = I4, DSP2, O2, D2Etc
Scenario’s Modelling
Market model (e.g. Plexos) based on :• co-optimised (energy & ESS), • gross pool,• single node with linear network constraints provided
Market Model manages the modelling inputs such as:• Unit Commitment• Facility bid creation (Economic parameters)• Synthetic maintenance outages (monte carlo)• Randomised forced outages (monte carlo)• ESS Requirements (CR and RoCof)
Intermittent Forecasts
Outages
Demand
Repeat for cost variability(internal in
model)
Repeat for each combination of input scenarios
Dispatch Targets
Binding Constraints
Unserved Energy
Dispatch Targets
Binding Constraints
Unserved Energy
Dispatch Targets
Binding Constraints
Unserved Energy
Dispatch Targets
Binding Constraints
Unserved Energy
Loss FactorsFacility and
Standing Data
Generator Planned Outages
System Normal Constraints
Future generator/
network augmentation
Other Rule Participant Data
DSP availability
Combinations of input data to form input scenario s
A collection of output data for each interval in the horizon, from each of the different
scenario s
A collection of output data for each interval of
the horizon
Static (or derived) input data
Assumptions around known changes to the power system (e.g. new generators, network changes)Information such as heat rates,
fixed/variable O&M, fuel costs/contracts, energy constraints, unit modelling data/limits)
Outputs Post Processing
Publish the following:
• Full data report of all intervals for each scenario
Identify the following:
• Not enough online capacity to meet demand (violating energy constraint)
• Shortages of ESS (violating ESS constraints)
• Other security issues (other violating constraints)
• Occurrence of network constraints binding
• Occurrence of DSP dispatch
And determine likelihood based on:
• Number of occurrences• Scenario combinations that
occurrences appear in
Run output processing on the data to identify risks
• Identify and issue Low Reserve Notifications
• Identify and manage any dispatch interventions
Highlight type of LRC and likelihood of occurrence
MT PASA: Data Requirements
• Will need modelling data from participants to support estimating output (e.g. wind turbine power curves)
• May require information from Demand Side Programmes on future demand estimates (e.g. future demand availability reductions)
• Information on future network augmentations and decommissioning
• Information on future generation connections
• Information on future generation retirements