An automation tool for Typhoon HIL simulations
Created by the Center for Sustainable Electric Energy Systems at the University of Wisconsin - Milwaukee.
This project is a work in progress and is still in its early phases. See the Contributing section below!
Intended uses of this tool include:
- Real-time Controller Hardware in the Loop (CHIL) testing
- Generating data for Machine Learning and Artificial Intelligece (ML/AI)
- Online training of Reinforcement Learning (RL) and similar learning methods
- Development of Digital Twins (DT)
The goal of this project is to create a tool which can automate the interaction with Typhoon HIL simuluations. Its primary features should include:
- Reading and writing SCADA values and model variables while the simulation is running
- Recording high resolution sensor and probe data to files
- Repeating a large number of simulations with random variations to each run
Users should have the ability to write a script which will automate the interaction with an HIL simulation. Model values can be read and written in near real-time while the simulation is running. Each run of the simulation may be randomized, e.g. a fault could be induced randomly with a normal distribution around a pre-scheduled simulation time. Sensor and probe data from each run of the simuation can be captured and stored. The real-time capability of Typhoon hardware will allow a large volume of rich simulation and test data to be rapidly generated.
Users will need a Typhoon license suitable for running their models on the intended simulation platform.
Use of this tool should be as simple as providing a Typhoon schematic and an automation script to the tool. This feature, like many other features, is not yet complete. Contributions to this project are welcome and encouraged! Feel free to post questions or comments on the Issues page or to fork the repository and improve the project.
Active work should merged into the dev branch, preferably through a pull request with appropriate reviewers. Adding unit testing, CI/CD, and documentation would be fantastic. The main branch should be reserved for release-ready code.