rtc-tools-heat-network 0.4.3

Heat network models for RTC-Tools 2.

rtc-tools-heat-network

Rtc-tools-heat-network is an optimization application for optimal planning, design and operation of Energy Systems with the current main focus on District Heating Systems (DHS). The current application focuses on a Mixed Integer Linear Problem (MILP) approach, with multiple linearization strategies to conservatively approximate the steady-state physics and financial models. All physics are combined in the HeatMixin class. When inherited this class can be combined with objective functions (that typically incorporate the financial aspects) and interface methods to create an optimization workflow (see also running an example).

The main supported method for defining your Energy system is ESDL (Energy System Description Language), which is a modelling language for energy systems. See also:https://github.com/EnergyTransition/ESDL. With ESDL you can define assets like demands, sources, pipes, etc. and fill in their attributes. The ESDLMixin class will parse the ESDL file and utilize the attributes to build up the model representation.

This optimization package was originally developed for operational optimization and hosts two optimization approaches 1) A MILP approach and 2) Nonlinear Problem (NLP) approach. These two approaches were developed to run sequentially for operational optimization. the MILP would fix the integer decision for the NLP problem, such that only the continuous variables need to be solved. The NLP problem would then find the optimized solution with the steady-state non-linear physics included. The existing outdated (still to be updated / update in progress) documentation can be found on: http://warmingup.pages.ci.tno.nl/rtc-tools-heat-network/

Installation

Installation of the RTC-Tools Heat Network library is as simple as::

# 1a. Use pip to install directly
pip install rtc-tools-heat-network

If you are going to develop and change the source code, you probably want to do something like::

# 1b. Use git clone and pip to make an editable/developer installation
git clone https://ci.tno.nl/gitlab/warmingup/rtc-tools-heat-network
pip install -e rtc-tools-heat-network

RTC-Tools Heat Network depends on RTC-Tools <https://gitlab.com/deltares/rtc-tools.git>_, which is automatically installed as one of its dependencies.

Running an example

To make sure that everything is set-up correctly, you can run one of the example cases. These do not come with the installation, and need to be downloaded separately::

# 1. Clone the repository
git clone https://github.com/Nieuwe-Warmte-Nu/rtc-tools-heat-network.git

# 2. Change directory to the example folder
cd rtc-tools-heat-network/examples/pipe_diameter_sizing/src

# 3. Run the example
python example.py

You will see the progress of RTC-Tools in your shell. If all is well, you should see something like the following output:

In this example.py file you can see a small workflow being set-up. The PipeDiameterSizingProblem class inherits from (Note only the *classes are defined in rtc-tools-heat-network the others come from rtc-tools package):

• CollocatedIntegratedOptimizationProblem: This class does all the discretization of the state variables in your problem.
• *ESDLMixin: This class does the parsing and setting up of a model based on an ESDL file.
• GoalProgrammingMixin: This class allows you to add Goals (objective functions) with different priorities.
• LinearizedOrderGoalProgrammingMixin: This class allows you to add higher order goals (e.g. order=2) for MILP problems.
• *HeatMixin: This class adds all the heat network physics for MILP problems.

Within the PipeDiameterSizingProblem class you can see that the path_goals() function is overwritten and that a path_goal with priority one is added to meet the heat demands. The definition path_goal is used to define a goal that is applied to a state variable at every time step. Furthermore, the goals() method is also overwritten in this case where an objective with priority two is added to minimize length*diameter. The goals() method is used here for global variables that do not change over time. The priorities indicate the sequential order in which the optimizer would be applied to the goals. In this example the heat demand is matched first, after which priority 2 length*diameter is minimized. In this example the objective of the priority one goal constraints the priority two goal optimization, which ensures that the optimization of the priority two goal does not have impact on the optimal result of the priority one goal.

Contribute

You can contribute to this code through Pull Request on GitHub. Please, make sure that your code is coming with unit tests to ensure full coverage and continuous integration in the API.

GitHub: https://github.com/Nieuwe-Warmte-Nu/rtc-tools-heat-network

Release

This package is released on pypi here whenever a new tag is pushed. In order to release this package:

1. Make sure that all relevant merge requests and commits have been merged to the master and/or poc-release branch.
2. Run git checkout master or git checkout poc-release to switch to the release branch.
3. Run git pull origin master or git pull origin poc-release to pull all latest changes.
4. Run git tag <new_version> where <new_version> is the new version number.
5. Run git push origin <new_version> to push the tag to Github.
6. Check Github to confirm the release is processed without errors.
7. Once the release has finished, confirm the new version is available on pypi.