Settings

Settings are used to support settings for the models that are not typical use cases. The settings is a field name in the InfrastructureModels data structure and the most common way to populate or add information to settings is to include them in the kwargs of the solve routines. Example:

    data = GasModels.parse_file("../test/data/matgas/case-6.m")
    settings = Dict("output" => Dict("duals" => true))
    solve_ogf(data, WPGasModel, nlp_solver, setting=settings)

Output

The output key is used to specify settings for uncommon outputs for solutions to a model

dual: is the key used to report dual variable values in the output. Use the boolean true to report dual variables in the solution.

Config

The config key is used to specify settings for for configuration of a model

networks: is the key used to identify the network keys to be used in populating an optimization model. This is primarly used for specifying a network in the multi network structure to model with optimization. While multiple network ids can be specified, this is generally ill advised as this will produce a single large decooupled optimization problem that is better solved individually.
use_nominal: is the key used to switch load/production/storage variables from physical engineering bounds to nominal usage bounds (set automatically by solve_ogf_nominal).
pipe_zero_length_tolerance: is the key used to override the tolerance below which a pipe's length is treated as zero, forcing p_i == p_j instead of applying the Weymouth equation. Defaults to 1e-6.
pipe_zero_lambda_tolerance: is the key used to override the tolerance below which a pipe's friction factor is treated as zero, forcing p_i == p_j instead of applying the Weymouth equation. Defaults to 1e-6.
inclined_pipe_threshold: is the key used to override the incline angle (in degrees) above which a pipe switches from the horizontal Weymouth model to the inclined pressure-drop model. Defaults to 5.0.
num_flow_breakpoints: is the key used to override the number of breakpoints used to build the piecewise-linear relaxations of f*abs(f)/f^2 in LRWPGasModel/LRDWPGasModel (more breakpoints tighten the relaxation at the cost of added binary variables). The value may be a plain Int applied to every pipe/resistor/ne_pipe, or a Dict for finer control, with lookup order id-specific -> type-wide -> global default -> the built-in default (1 if the flow range straddles zero, else 0):

    settings = Dict("config" => Dict("num_flow_breakpoints" => Dict(
        "default" => 2,             # applied to any component/type not listed below
        "pipe" => Dict(
            "default" => 5,         # applied to every pipe not listed below
            1 => 10,                # pipe with index 1 gets 10 breakpoints
        ),
        "resistor" => 1,             # every resistor gets 1 breakpoint
    )))
    solve_ogf(data, LRWPGasModel, lp_solver, setting=settings)

The following is an example of solving all networks in a multi network in a single optimization problem

    data = parse_multinetwork("../test/data/matgas/case-6.m", "../test/data/transient/time-series-case-6a.csv", time_step=864.0)

    settings = Dict("config" => Dict("networks" => parse.(Int, keys(mn_data["nw"]))))
    solve_ogf(data, WPGasModel, nlp_solver, setting=settings)

The following is an example of solving all networks in a multi network as a sequence of optimizations

    data = parse_multinetwork("../test/data/matgas/case-6.m", "../test/data/transient/time-series-case-6a.csv", time_step=864.0)
    
    result                                   = Dict{String, Any}()
    result["solution"]                       = Dict{String, Any}()
    result["solution"]["nw"]                 = Dict{String, Any}()
 
    for nw in keys(data["nw"])
        settings = Dict("config" => Dict("networks" => [parse(Int,nw)]))
        solution = GasModels.solve_ogf(mn_data, WPGasModel, nlp_solver, setting=settings)
        @test solution["termination_status"] == LOCALLY_SOLVED
        result["solution"]["nw"][nw] = solution["solution"]["nw"][nw]

        # storing some individual solve information
        result["solution"]["nw"][nw]["termination_status"] = solution["termination_status"]
        result["solution"]["nw"][nw]["dual_status"]        = solution["dual_status"] 
        result["solution"]["nw"][nw]["solve_time"]         = solution["solve_time"]
        result["solution"]["nw"][nw]["primal_status"]      = solution["primal_status"] 
        result["solution"]["nw"][nw]["objective"]          = solution["objective"] 
        result["solution"]["nw"][nw]["objective_lb"]       = solution["objective_lb"] 

        # aggregate information
        result["solve_time"]   = get(result, "solve_time", 0) + solution["solve_time"]
        result["objective"]    = get(result, "objective", 0) + solution["objective"]
        result["objective_lb"] = get(result, "objective_lb", 0) + solution["objective_lb"]
        result["optimizer"]    = solution["optimizer"]
        result["model_type"]   = solution["model_type"]
        result["model_name"]   = solution["model_name"]

        #top level information
        result["solution"]["base_density"]         = solution["solution"]["base_density"] 
        result["solution"]["multinetwork"]         = solution["solution"]["multinetwork"] 
        result["solution"]["base_volume"]          = solution["solution"]["base_volume"] 
        result["solution"]["base_length"]          = solution["solution"]["base_length"] 
        result["solution"]["base_mass"]            = solution["solution"]["base_mass"] 
        result["solution"]["per_unit"]             = solution["solution"]["per_unit"] 
        result["solution"]["base_time"]            = solution["solution"]["base_time"] 
        result["solution"]["base_flow"]            = solution["solution"]["base_flow"] 
        result["solution"]["base_pressure"]        = solution["solution"]["base_pressure"] 
    end