Public

Main module for EnergyModelsRecedingHorizon, a package that extends EnergyModelsX with the implementation of a receding horizon framework for stress testing the invested energy system in an operational framework.

struct RecHorOperationalModel <: RecHorEnergyModel

Operational energy model without investments, receding horizon implementation.

Fields

• emission_limit::Dict{<:ResourceEmit, <:TimeProfile} is a dictionary with individual emission limits as TimeProfile for each emission resource ResourceEmit.

• emission_price::Dict{<:ResourceEmit, <:TimeProfile} are the prices for the different emissions types considered.

• co2_instance is a ResourceEmit and corresponds to the type used for CO₂.

run_model_rh(case::AbstractCase, modeltype::RecHorEnergyModel, optimizer; check_timeprofiles::Bool=true, use_op_per_strat::Bool=false, optimizer_param::Dict=Dict(MOI.Silent() => true))

Create and run the model utilizing a receding horizon framework. The results are returned as a dataframe indexed by the model variables.

Arguments

• case::Case - The case type represents the chosen time structure, the included Resources and the 𝒳 and potential coupling between the 𝒳. It is explained in more detail in its docstring.

  Required input

  While the Case type is flexible, we have to follow certain structures:

  • The case type requires as additional input in the dictionary field misc the entry

  :horizons corresponding to an AbstractHorizons type.

  • The order of the individual elements vector in the field elements cannot be arbitrary at the moment due to the structure of the code. You must use the following order:

    1. Vector{<:EMB.Node}
    2. Vector{<:Link}
    3. Vector{<:Area}
    4. Vector{<:Transmission}

  If you do not use this structure, the model will not run.

• modeltype::RecHorEnergyModel - Used modeltype, that is a subtype of the type EnergyModel.

• optimizer - A MathOptInteface optimizer.

Keyword arguments

• check_timeprofiles::Bool = true - A boolean indicator whether the time profiles should be checked or not. It is advised to not deactivate the check, except if you are testing new components. It may lead to unexpected behaviour and potential inconsistencies in the input data, if the time profiles are not checked.
• use_op_per_strat::Bool = false - A boolean indicator whether the provided parameters indexed over strategic periods should be based on the parameter op_per_strat of the time structure or not. The former could correspond to utilizing, e.g., providing annual values for the fixed operating expenses and the CO₂ limit. The parameters are otherwise relative to a duration of 1 of an operational period If the value is specified as false,
• optimizer_param::Dict = Dict(MOI.Silent() => true) - A dictionary in which the key represents the parameter and the value the chosen value for the parameter.
  Parameter names

  The name of some parameters is optimizer specific. You must hence be careful when changing the optimizer

DurationHorizons{T} <: AbstractHorizons{T}

Type used for specifiying the optimization and implementation horizon of a receding horizon model as the duration of the horizons. This implies that the number of operational periods in the different iterations can vary.

Iterating a DurationHorizons results in SingleHorizon which includes all information required for a single run.

Fields

• len::Int64 is the total length of the investigated timeframe as a multiple of the duration of 1 of an operational period.
• dur::Vector{T} is a vector of the duration of each individual operational period.
• optim::Int64 is the sum of the duration of the operational periods that are used in the optimization horizon.
• impl::Int64 is the sum of the duration of the operational periods that are used in the implementation horizon.

PeriodHorizons{T} <: AbstractHorizons{T}

Type used for specifiying the optimization and implementation horizon of a receding horizon model as multiples of the number of periods.

Iterating a PeriodHorizons results in SingleHorizon which includes all information required for a single run.

Fields

• len::Int64 is the total length of the investigated timeframe as a multiple of the duration of 1 of an operational period.
• dur::Vector{T} is a vector of the duration of each individual operational period.
• optim::Int64 is the number of operational periods that are used in the optimization horizon.
• impl::Int64 is the number of operational periods that are used in the implementation horizon.

durations(single_hor::SingleHorizon)

Extracts the fiels dur from the SingleHorizon single_hor corresponding to a vector of the durations of the individual time periods.

indices_optimization(single_hor::SingleHorizon)

Extracts the fiels idx_optim from the SingleHorizon single_hor corresponding to a vector of the indices of the optimization horizon.

indices_implementation(single_hor::SingleHorizon)

Extracts the fiels idx_impl from the SingleHorizon single_hor corresponding to a vector of the indices of the implementation horizon.

abstract type AbstractInitData <: EMB.Data

Abstract type for initialization data.

struct InitData{T} <: AbstractInitData

Initialization data type for the inclusion of initial data before the first horizon. The standard initialization data is using a single value for a given variable. Multiple variables can be initialized simultaneously.

Fields

• init_val_dict::Dict{Symbol,T} is a dictionary with the variable symbol as key and the value in the beginning of the analysis as value.

StorageInitData(val::Real)

Constructor for providing initialization for Storage nodes. A StorageInitData must be provided for each Storage node.

Arguments

• val::Real: initial value for storage level.

has_init(n::EMB.Node)
has_init(l::Link)

Checks whether the node n or link l has initialization data.

is_init_data(data)

Checks whether the argument data has initialization settings.

data_init(n::EMB.Node)
data_init(l::Link)

Returns initialization data for node n or link l.

abstract type  FutureValue <: AbstractElement

Declaration of the abstract type for future value descriptions in the optimization.

struct ElementValue{T<:Union{TimeProfile, Real}}

An ElementValue represents an instance of a given AbstractElement with an assigned value. It replaces dictionaries in which an AbstractElement is used as key value so that it is possible to reset the AbstractElement

Fields

• element::N is the instance of the element.
• value::T is the used value.

struct StorageValueCuts <: FutureValue

A collection of multiple StorageValueCut that constructs a piecewise linear upper bound on the future value of the stored resource.

Fields

• id::Any is the name/identifier of the StorageValueCuts.
• time::Int is the time where the cut is valid relative to the start of the operational period.
• weight::Real is the weighting of the StorageValueCuts in the objective function. For example used if the end time of the optimization arrives between two different StorageValueCuts.
• time_weight::Real is the weighting of the StorageValueCuts in the objective function due to the elapsed time.
• cuts::Vector{StorageValueCut} is a vector of all the cuts that are included in the future value description.

struct StorageValueCut

A StorageValueCut represents a cutting hyperplane that puts an upper bound on the value of the stored resource at the end of the optimization horizon.

Fields

• id::Any is the name/identifier of the StorageValueCut.
• coeffs::Vector{<:ElementValue} are the cut coefficients associated with the level of the given Storage nodes. They can also be provided as Dict{<:Storage{<:Accumulating}, <:Real}.
• rhs::Real is the cut right hand side constant.

struct TypeFutureValue <: FutureValue
TypeFutureValue(element_type::Type{<:AbstractElement}, key::Symbol, val::Real)

A future value for a given nodal type and model key. It utilizes only the final value and directly adds it to the cost function for all instances of the given type.

Fields

• element::Type{<:AbstractElement} is the nodal type for which the future value applies.
• val_dict::Dict{Symbol, Real} is a dictionary for including the variables provided as keys with a given value for calculating the future value. It is important to consider the unit of the value.

get_future_value(𝒳ᵛᵉᶜ::Vector{Vector})
get_future_value(case::Case)
get_future_value(𝒰::UpdateCase)

Returns the vector of FutureValue of the Case case or the vector of elements vectors 𝒳ᵛᵉᶜ.

If the input is an UpdateCase, it returns the new FutureValuess of the individual FutureValueSub types of UpdateCase 𝒰.

has_cuts(v::FutureValue)
has_cuts(v::StorageValueCuts)

Returns true if the FutureValue is of type StorageValueCuts.

struct RecedingAccumulating <: Accumulating

StorageBehavior which accumulates all inflow witin a strategic period. RecedingAccumulating allows for initializing level values for the storages.