optimizerConfig.xml

The optimizerConfig.xml file is the template for any configuration file for the optimization routines. The actual configuration file should be located in MiGRIDSProjects/<myProjectName>/InputData/Setup/ and called optimizerConfig<myProjectName>.xml. Without such a file optimization fails.

The following describes the various tags, their function and default values where warranted.

optimizerConfig Tag

Root tag wrapping all other tags

optimizationObjective Tag

This contains the objective of the optimization. The value given determines which fitness function is actually called in getFitness.py.

Default value: "maxREContribution"

Available choices:

• maxREContribution: looks for the EES size which maximizes renewable energy contribution
• minFuelUtilization: looks for the EES size which minimizes diesel fuel utilization of the generator fleet

If additional fitness methods are made available they need to be registered in the choices attribute of the optimizationObjective tag.

optimizationConstraints Tag

A container for various possible constraints on the optimization. Simulations failing to meet these constraints are discarded.

maxAnnualGenConfigChanges Tag

NOTE: this tag currently is not used in the optimization.

Default value: 2000

Default setting: active="False"

minESSP Tag

Lowest ESS power capacity to be considered. Units: kW. If 'active' is true, enforce this constraint. Default value is 0. NOTE: the setting in this tag may be overridden by the selection in optimizationEnvelopeEstimator

Default value: 0

Default setting: active="True"

maxESSP Tag

Highest ESS power capacity to be considered. Units: kW. If 'active' is true, enforce this constraint. Default value is 10000. NOTE: the setting in this tag may be overridden by the selection in optimizationEnvelopeEstimator

Default value: 10000

Default setting: active="True"

minESSE Tag

Lowest ESS energy capacity to be considered. Units: kWh. If 'active' is true, enforce this constraint. Default value is 0. NOTE: the setting in this tag may be overridden by the selection in optimizationEnvelopeEstimator

Default value: 0

Default setting: active="True"

maxESSE Tag

Highest ESS energy capacity to be considered. Units: kWh. If 'active' is true, enforce this constraint. Default value is 10000. NOTE: the setting in this tag may be overridden by the selection in optimizationEnvelopeEstimator

Default value: 10000

Default setting: active="True"

optimizationEnvelopeEstimator Tag

Configuration for getOptimizationBoundaries.py interface. Based on the value given here the interface passes on to a specific boundary value estimation. Any new method to estimate boundaries needs to be registered with the choices attribute here.

Default value: variableSRC

Available choices:

• variableSRC: looks at the estimated SRC requirements to narrow down the ESS range of power and energy capacities , implemented in calcSRCMethodBoundaries.py. Additional configuration data for this method is provided in variableSRCCongig.
• noBoundaries: skips boundary calculation and merely uses the minimum and maximum values configured in the respective children of optimizationContstraints.

variableSRCConfig Tag

Additional configuration passed through to calcSRCMethodBoundaries.

mode Tag The attribute 'mode' describes how to reconcile the estimated envelope with the sizing constraints (minESSP/E, maxESSP/E); options are 'tight', which uses the narrower band of values, or 'relaxed' which uses the wider band of values. NOTE: this is currently not used.

minDynSRC Tag Fraction of total load that always should be kept as SRC. Default value is 0.15.

minDuration Tag Time in seconds that the energy storage has to cover SRC. Units: seconds. Default value is 60 seconds

minPercent Tag Minimum percentile as a fraction that the energy storage has to be able to cover SRC. Units: none. Default value is 0.7.

maxPercent Tag Maximum percentile as a fraction that the energy storage has to be able to cover SRC. Units: none. Default value is 0.99.

maxMargin Tag The margin put on the maximum power calculated as a fraction. Units: none. Default value: 1.10.

dataReductionMethod Tag

Choices of methods used to reduce the long time series, in order to reduce algorithm run time. Only one method is implemented thus far. If additional methods are implemented they will have to be registered in the choices attribute.

Default value: RE-load-one-week

Availble choices:

• RE-load-one-week: calls the reduction algorithm in getDataSubsetsRELoadOneWeek.py
• noReduction: skips data reduction and runs on the entire data set. Not recommended.

optimizationMethod Tag

Selects from the available optimization algorithms. NOTE: the only functioning algorithm at this time is the hillClimber choice.

Default value: hillClimber

Availble choices:

• hillClimber: selects an adaptive hill climbing algorithm. See hillClimberConfig for additional configuration.
• geneticAlgorithm: Not implemented due to total compute time requirements. Here for illustration only.

hillClimberConfig Tag

Configuration parameters for the hillClimber algorithm. Note: some configuration values are currently hardcoded, if these are desired to be configured they should be added here as child tags.

minRunNumber Tag Minimum number of iterations to dispatch. Default value is 3.

maxRunNumber Tag Maximum number of iterations to dispatch. Default value is 50. Note that 50 iterations might create considerable amounts of data and may require significant (>days) compute time.

**convergenceVariance Tag ** Fraction of variance from best so far considered convergent. Default value is 0.1. Note: this is not used currently. Convergence is only achieved if the exact best fitness value is found repeatedly.

convergenceRepeatNum Tag Number of consecutive repeats at convergence variance to terminate. Default value is 3.

dispatchFullRunAtConvergence Tag If true, the best found configuration is dispatched for a full length simulation. Default value is False. Note: due to the total time requirement for dispatch of a full simulation, this is currently not implemented.

geneticAlgorithmConfig Tag

Place holder only. Note implemented.