/Transformer_MPC

Primary LanguageJupyter NotebookMIT LicenseMIT

Transformer_MPC

Hybrid Machine Learning and Physics-based model for Process Control and Optimization

pinn_off

pinn_on

results_0.mp4

Folders:

01_Emulation: LSTM MPC Emulation. Additional details are available with PID emulation and MPC emulation.

02_Surrogate: All case studies in the paper referenced below including the FOPDT, fluidized bed roaster for gold ore refinement, and Temperature Control Lab hardware. The fluidized is a case study in the paper to emulate MPC on a complex nonlinear process. For additional information see:

Park, J., Babaei, M.R., Arce Munoz, S., Venkat, A.N., Hedengren, J.D., Simultaneous Multistep Transformer Architecture for Model Predictive Control, Computers & Chemical Engineering, 2023, DOI 10.1016/j.compchemeng.2023.108396 Article | Preprint

03_HybridModel: Hybrid physics-based and machine learned (Transformer) case study with the Temperature Control Lab. For additional details, see the PhD Dissertation of Junho Park.