Design Reference

Sygration Generation Market Simulator



Simulator Optimization and Constraint Reference

This section explains how the Generation Market Simulator arrives at the final dispatch solution and determines the settlements quantities. Where appropriate, a reference to the Ontario Electricity Market Rules has been provided.

Minimum & Maximum Energy Constraints (Ramp Rates)

Energy ramp rates are used to determine how much generators can change their energy output, under normal conditions, from one interval to the next. This may result in imposing upper and lower limits against the energy output and reserve for the simulator optimization run. 

While the Constrained Optimization run uses the normal ramp rates, the Unconstrained Optimization run uses a multiplier factor of 12 applied to all energy rates.  [Reference material refers to the use of a 60-minute period and not a 5-minute period – giving the same effect.  The market rules do not specify the value but actually state the use of a “variable” TradingPeriodLength for this purpose, however, it is currently configured for 60 minutes (multiplier of 12)]

 

References:

MR App 7.5 6.5
MR App 7.5 8.2


MR App 7.5 2.11.2
MR 7.5 4.13.1

10-Minute Reserve Ramp Rate Constraint

The amount of 10-Minute reserve considered for optimization is limited by the OR Ramp Rate X 10 minutes. This is applied to the sum of the Spinning and Non-Spinning 10-Minute OR.

Any quantity of 10S or 10N offers that cannot be scheduled for either market will be ignored. 

This constraint is applied equally to both the Constrained and Unconstrained optimization runs.

 

 

MR App 7.5 6.3.5A.1

10-Minute Spinning Reserve Loading Point

The amount of 10-Minute Spinning reserve that can be scheduled may be further limited by its Reserve Loading Point, when generation is below this point. The RLP value must be provided as part of each OR Offer and must be at least 0.1 (MW) for the 10-Minute Spinning offer. It is applied equally to both Constrained and Unconstrained runs.

The DSO uses the RLP as part of the final optimization process, using the current generation dispatch, while the Simulator pre-processes the RLP using the previous generation dispatch prior to enterring the optimizing phase; resulting in a very small difference if the generator is ramping up/down while below the RLP. Any quantity of OR10S offers that cannot be scheduled due to this limitation will be ignored.

 

 


MR App 7.5 6.3.5.1

30-Minute Reserve Loading Point

The amount of 30-Minute reserve that can be scheduled may be further limited by its Reserve Loading Point, when generation is below this point. The RLP value must be provided as part of each OR Offer and must be at least 0.0 (MW) for the 30-Minute offer. It is applied equally to both Constrained and Unconstrained runs.

The DSO uses the RLP as part of the final optimization process, using the current generation dispatch, while the Simulator pre-processes the RLP using the previous generation dispatch prior to enterring the optimizing phase; resulting in a very small difference if the generator is ramping up/down while below the RLP. Any quantity of OR30 offers that cannot be scheduled due to this limitation will be ignored.

 

 

MR App 7.5 6.3.5.1

30-Minute Ramp Rate Constraint

The amount of 30-minute reserve considered is limited by the OR Ramp Rate X 30 minutes.

Any quantity of 30-minute offers that cannot be considered as a result will be ignored.

This constraint is applied equally to both the Constrained and Unconstrained optimization runs.

 

 

MR App 7.5 6.3.5A.2

Maximum MW of Energy and Reserve

The total scheduled quantity of all offered products (Energy + OR 10S + OR10N + OR 30) cannot exceed the maximum energy quantity offered for that hour.

The simulator may be run in Historical Generation Mode, where its dispatch is compared against actual generator data captured in the Sygration data warehouse. In this situation, the recorded Generator Capability for the generator is used in addition to the above limit to further limit the total scheduled quantity.

 

 

MR App 7.5 6.1.4
MR App 7.5 6.3.4

Marginal Offers

A marginal offer is one where the price offered equals the market or locational price, and has a zero marginal economic surplus to the generator.  In real life, only a portion of the quantity of Energy or Reserve that is needed will be scheduled. 

As the simulator has no means to know how much would actually be needed, it compromises by selecting ½ of the lamination quantity with the zero surplus.

 

 

Tied Surplus

In the context of the simulator, a tied surplus occurs when the two or more products (Energy, or any Reserve) have the same marginal economic surplus, and are positive.  If the quantity of tied laminations cannot be scheduled, due to any constraint above, the simulator treats the products fairly by scheduling a proportion of all tied surplus laminations.

 

 

Negative Marginal Surplus

Negative marginal surplus refers to bid quantities where the offered price is higher than the market (or locational) price.  These prices have therefore not “cleared the market”. With the exception of the affect of the Minimum Generation Constraint, no quantities of any products with negative marginal surplus will be scheduled as it results in a reduction in economic benefit to the participant.

 

 

MR 4.3.3

RD Filter

The resource dispatch filter applies to the energy dispatch portion of the constrained optimization run. Where the change from the previous dispatch is less than 2% of the maximum generation offer, to a maximum of 10MW, the previous value will be used.

The RD filter is not applied on the 1st and 7th interval of each hour.

 

 

MM 4.3 1.8.1

Nodal Price Adjustments

The Simulator can output settlements figures based on nodal prices for those users interested in the impact of Ontario potentially moving to a locational priced system (currently it is a uniform market price for energy and operating reserve). 

For the purpose of nodal settlements; after the constrained optimization run, all nodal prices are adjusted to within the minimum and maximum market clearing prices:

·         LMP Energy Prices below -$2000 are adjusted to -$2000

·         LMP Energy and Operating Reserve Prices above $2000 are adjusted to $2000

·         Operating Reserve Prices below $0 are adjusted to $0

 

 

 

 

MR 4.4.6

Simulator Limitations

The Simulator is not a transmission congestion model and does not have the full set of bids and offers for supply and demand to determine exactly what the dispatch would have been.  Given that limitation, the simulated output would be accurate with the following exceptions:

1.       Simulations assume full market liquidity: bids and generation dispatches are assumed to have no impact on prices. In reality, a change of an existing bid may change the overall outcome of the market, while new generation that receives a dispatch by the IESO will have an impact on the market.

2.       If a bid is marginal, there is no way to know how much quantity would have been scheduled. The simulator dispatch will be pro-rated within the bid lamination by 50% of the optional portion (i.e. for energy, the portion that is within the ramp-rate capabilities).

3.       If multiple products (i.e. energy, reserve) have the same marginal benefit and cannot be entirely scheduled, a proportion of each product will be scheduled.

4.       The bid simulator assumes that dispatched energy quantities will be followed exactly.  This impacts two areas:

·         IESO settlements are based in part on the actual quantities of energy measured by the revenue meters. This energy may differ slightly from the dispatch instruction depending on how well the generator is able to comply with the dispatch.

·         The previous Simulator dispatch is used to determine the ramping trajectory for the current interval for both constrained and unconstrained runs. The IESO uses the actual generation at the start of the interval for the constrained run, which may differ from the dispatch.

5.       Actual IESO dispatches may result from manual operator intervention. The bid simulator has no way to predict this.

6.       The simulator does not do multi-interval optimization, while the IESO DSO does.  MIO affects the market and shadow prices, which are reflected in the outcome of the simulator.  However, change-of-hour resource pre-ramping, which occurs because of MIO, has not (yet) been incorporated into the simulator. 

7.       Additional facility registration data used by the IESO in its dispatch is not used in the simulator. Examples are Minimum Loading Point and Generator Forbidden Regions.

 Contact Sygration if you would like to see these or additional capabilities added to the simulator.

Copyright © 2005 Sygration