- Benchmark_Plots
- Step test plots
- Optimal steady state input dynamics
- ControlLoops
- Decoupled system drawings
-
FIS: Contextual Bandit Fault Identification System
-
FTC_Case1: MeOH priority.
-
FTC_Case2: Water priority.
-
FTC_Case3: Both considered.
- Velocity_Case4_initial: Initial results
- Velocity_Case4_Ucost: Reward contains deltaU cost
- Case1_Case2_using_MISO: Case 1 and case 2 using the MISO RL Code
- SMDP: SMDP Case studies of case 3
- MATLAB_Files:
- Decoupler: Decoupled Wood-Berry distillation tower
- wood_berry_distillation: Raw Wood-Berry distillation tower m file. Has transfer function and conversion from tf2ss.
- Paper_Plots
- Bad system dynamic plots (inv response, overshoot, oscillations)
- Case study plots:
Case 1: Top product priority
Case 2: Bottom product priority
Case 3: Top & bottom priority
Case 4: Adaptability Study
-
Decoupled_Distillation: Decoupled Wood-Berry distillation tower simulation with PI regulatory controllers.
-
Woodberry_Distillation: Initial Wood-Berry distillation tower simulation with PI regulatory controllers.
-
Woodberry_Distillation_RL: Initial Wood-Berry distillation tower simulation with RL in supervisory layer.
-
RL_Module_Position: Position reinforcement learning module. Takes absolute states and actions.
-
RL_Module_Velocity: Velocity reinforcement learning module. Takes relative actions given error.
-
RL_Module_Velocity_MIMO_SMDP: MIMO SMDP reinforcement learning module. Considers semi-Markov processes.
-
RL_Module_Velocity_SMDP: SMDP reinforcement learning module.