Dieter.jl

Table of Contents

• Dieter.jl
• • Overview
  • Model code
  • Installation
  • • Example of a run script
• Model formulation
• • The optimization model
  • • Index sets
    • Variables
    • Generation and storage capacity variables
    • Objective function
    • Constraints

Overview

DIETER is the Dispatch and Investment Evaluation Tool with Endogenous Renewables, an open energy system model developed by the German Institute for Economic Research (DIW Berlin). The model is structured as a cost-minimising linear program that when solved to optimality will produce an hourly operation profile of an interconnected transmission system over a single year, including aspects such as

• generator power output,
• storage technology charge and discharge,
• transmission flows,
• variable renewable energy utilisation, and
• hydrogen electrolyser operation.

The model also decides on cost-optimal capacity expansion of relevant model technologies relative to the year under consideration.

This code repository holds a derivative version of DIETER written using the Julia language and the JuMP algebraic modelling language (a Julia package). While the original version of DIETER is implemented in GAMS, Julia/JuMP enables a free and open-source implementation. Linear programming (LP) solvers are required to compute optimal solutions, and both free/open-source and commercial solver can be freely interchanged using interfaces via JuMP (commercial solvers such as the Gurobi LP solver may required for obtaining reasonable run times when solving large-scale problems).

The Dieter.jl Julia/JuMP package has been developed at CSIRO within the Energy Systems research program. Thanks to Mario Kendziorski for an earlier version, as well as Wolf-Peter Schill.

In keeping with the original version, Dieter.jl is open source and may be freely used and modified by anyone. Contributions are welcome.

Model code

For a quick overview, the model code can be directly viewed in the model.jl file.

Note that the model is (currently) lacking some features which exist in the GAMS model such as the representation of reserves and prosumage.

General utility functions for handling DataFrames, Dicts and translations from JuMP data structures, useful independently of Dieter.jl, are to be found in the util.jl file.

Installation

This package is not registered on the Julia package system. To install this as a Julia module in a Julia REPL session, you can run the following package installation commands to install directly from the model code repository:

import Pkg
Pkg.add(Pkg.PackageSpec(url="https://github.com/csiro-energy-systems/Dieter.jl.git"; rev="main"))

The argument rev="main" selects the git branch to use; change this to use your own alternative branch as required.

Example of a run script

To run the model, one might use a script similar to the following, however this currently requires user customisation to work:

using Dieter
using JuMP

# select your favourite solver
using HiGHS
solver = HiGHS.Optimizer
# using Gurobi
# solver = Gurobi.Optimizer
# using CPLEX
# solver = CPLEX.Optimizer

current_dir = @__DIR__
datapath = joinpath(current_dir,"test","testdata")

# set the Dieter parameters
dtr = InitialiseDieterModel(DieterModel, Dict{String,Any}())
dtr.settings[:datapath] = datapath

initialise_data_file_dict!(dtr)
check_files_exist(dtr.data["files"])

dbfilepath = joinpath(datapath,"DTR_regions_test_data/DTR.db")
dfDict = parse_data_to_model!(dtr; dataname=dbfilepath)

build_model!(dtr)

JuMP.set_optimizer(dtr.model, solver)
solve_model!(dtr)

generate_results!(dtr)