Continuous Stirred Tank Reactor (CSTR) sample

Project Bonsai code sample demonstrating chemical process optimization in a continuous stirred tank reactor (CSTR). Efficient control of an exothermic, non-linear chemical reaction with a CSTR is a benchmark in which to compare PID, MPC and Bonsai brains.

For an overview of this sample, watch Chemical Reactor Optimization with Microsoft Project Bonsai.

The chemical process here considers a transition from low to high conversion rate (high to low residual concentration). Because the chemical reaction is exothermic (produces heat), the reactor temperature must be controlled to prevent a thermal runaway. The control task is complicated by the fact that the process dynamics are nonlinear and transition from stable to unstable and back to stable as the conversion rate increases. The reactor dynamics are modeled in Simulink. The controlled variables (states) are the residual concentration and the reactor temperature, and the manipulated variable (action) is the temperature of the coolant circulating in the reactor's cooling jacket.

This example shows how to use Project Bonsai's Machine Teaching strategies to learn a controller for a chemical reactor transitioning from low to high conversion rate. For background, see Seborg, D.E. et al., "Process Dynamics and Control", 2nd Ed., 2004, Wiley, pp. 34-36. This sample is largely adapted from the MathWorks': Gain scheduled control of a chemical reactor.

Adapative PI Control

Bonsai Brain Control

Actions (Manipulated Variables)

Bare minimum for the sim (all units are continuous):

Final set for Bonsai training:

Note: Performance improved when making the brain learn the per-timestep adjustment to apply to previous dTc. Control is maintained with the addition of dTc_adjust, and an accumulator was added on the simulator side.

Note: given an additional rule that requires keeping dTc changes at no more than 10 Kelvins/min, we forced dTc_adjust to be on the [-5, 5] range (for Ts=0.5min)

States (Control Variables)

Which matches the set of Observable States used for bonsai training

Note, the .ink file defines ranges higher than the ones shown here. That is made on purpose since the brain will try to explore, and thus will hit extreme limits in doing so.

Tref was removed as observable state since brain to simplify brain training. With Bonsai's solution, Tref is not needed to be able to drive the concentration linearly between points.

Constraints

• Tc < 10 degrees / min
• Tr < 400 to prevent thermal runaway

Model Overview and Instructions

Please open the MATLAB livescript, Chemical_Process_Optimization.mlx, for further descriptions and instructions for getting started with this sample.

Project

This repo has been populated by an initial template to help get you started. Please make sure to update the content to build a great experience for community-building.

Contributing

This project welcomes contributions and suggestions. Most contributions require you to agree to a Contributor License Agreement (CLA) declaring that you have the right to, and actually do, grant us the rights to use your contribution. For details, visit https://cla.opensource.microsoft.com.

When you submit a pull request, a CLA bot will automatically determine whether you need to provide a CLA and decorate the PR appropriately (e.g., status check, comment). Simply follow the instructions provided by the bot. You will only need to do this once across all repos using our CLA.

This project has adopted the Microsoft Open Source Code of Conduct. For more information see the Code of Conduct FAQ or contact opencode@microsoft.com with any additional questions or comments.

Trademarks

This project may contain trademarks or logos for projects, products, or services. Authorized use of Microsoft trademarks or logos is subject to and must follow Microsoft's Trademark & Brand Guidelines. Use of Microsoft trademarks or logos in modified versions of this project must not cause confusion or imply Microsoft sponsorship. Any use of third-party trademarks or logos are subject to those third-party's policies.