NOTE The rtc-tools repository has been migrated from gitlab to here; see migration from gitlab.
RTC-Tools is an open-source Python package for simulation and optimization of cyber-physical systems. It can be used for model-predictive control and operational decision making of water and energy systems. It is developed and maintained by Deltares, in collaboration with partners.
RTC-Tools offers the following functionalities:
-
Model building using extensible libraries: Build complex system models using extensible libraries of model components. Modelers can also implement their custom modeling components, for instance, using the Modelica language or Python script.
RTC-Tools can be used with the following libraries and packages for specific applications:
- rtc-tools-channel-flow: water system models
- rtc-tools-hydraulic-structures: hydraulic assets, such as weirs and pumps
- rtc-tools-heat-network: heat networks
Please note that this list is not exhaustive. Users can also create libraries for other types of applications.
-
Running simulations: Simulate a given model.
-
Specifying and solving optimization problems: Define optimization goals, constraints and decision variables to specify the optimization problem for a given model. RTC-Tools supports both open-source solvers (CBC, HiGHS, Ipopt) and commercial solvers (Gurobi, CPLEX) for solving several types of optimization problems:
-
Linear, non-linear: RTC-Tools supports both linear and non-linear optimization problems.
-
Continuous and discrete: RTC-Tools can handle both continuous and discrete decision variables. This makes it suitable for optimizing systems with a mix of continuous controls (such as pump speeds or gate positions) and discrete decisions (such as on/off states of equipment).
-
Goal programming: When multiple, and perhaps conflicting, objectives are to be considered (e.g., minimize operational costs while minimizing deviations of water levels from a given range), RTC-Tools offers two approaches to multi-objective optimization: The Weighting Method, which assigns weights to each objective and optimizes them simultaneously, and the Lexicographic Goal Programming method, which optimizes different objectives sequentially.
-
Optimization under uncertainty: RTC-Tools supports the use of ensemble forecasts for optimization under uncertainty. To reduce the size of the optimization problem, ensemble members can be automatically aggregated using a scenario tree reduction method.
-
To streamline the integration with user interfaces and data management systems (such as Delft-FEWS), RTC-Tools supports CSV and XML file formats for reading/writing timeseries and other model parameters. Support for other formats can be implemented using mixins.
RTC-Tools uses CasADi as a symbolic framework for algorithmic differentiation, as well as for interfacing with numerical optimization solvers.
pip install rtc-toolsDocumentation and examples can be found on readthedocs.
The rtc-tools repository has been migrated from gitlab (https://gitlab.com/deltares/rtc-tools) to here. To change the git remote url, run
git remote set-url origin https://github.com/Deltares/rtc-tools.git
and
git remote set-url --push origin https://github.com/Deltares/rtc-tools.git.
RTC-Tools is licensed under the GNU Lesser General Public License v3.0, and can be used free of charge. Deltares offers support packages for users who require assistance.
If you use RTC-Tools in your work, please acknowledge it in any resulting publications. You can do this by citing RTC-Tools and providing a link to our website or Gitlab repository.