MBDoE and MO analysis for co-current and counter-current RPB

Authors: Jialu Wang, Ryan Hughes, Debangsu Bhattacharrya, Alexander Dowling

Installation instructions

The following instructions assume you have anaconda installed. We suggest create an environment with the following commands to run code:

Step 1: create a new environment

• create new environment, called for e.g. rotary, with conda

conda create --name rotary python==3.8

conda activate rotary

Step 2: install IDAES-PSE

• this step is necessary for running homotopy solve for the model

pip install idaes-pse

idaes get-extensions

Step 3: install Pyomo from specified branches

• install from the following branch for a modified version of Pyomo:

pip install git+https://github.com/jialuw96/pyomo.git@RPB_MBDOE

Step 4: install jupyter notebook

• this is needed only for the draw_figure.ipynb notebook

  conda install jupyter notebook

Software versions we use for the results

Python: 3.8

IDAES-PSE: 2.2.0

Pyomo: 6.7.0 dev 0

Code content

Co-current flow model

• RPB_model_cocurrent.py: Co-current flow RPB model provided by Ryan, modified for running k_aug

• Cocurrent_flow_MBDoE.ipynb: Apply Pyomo.DoE to the RPB model with compute_FIM and run_grid_search. stochastic_program is also tried but not debugged

Counter-current flow model (Results are represented in Jialu's thesis chapter 4)

• RPB_model_countercurrent_kaug.py: Conter-current flow RPB model provided by Ryan, modified for running k_aug

• Countercurrent_MBDoE.ipynb: MBDoE analysis applying Pyomo.DoE and k_aug to the RPB model with compute_FIM and run_grid_search. stochastic_program is also tried but not debugged

• Counterflow_MO_data_process.ipynb: Process the sensitivity data for MO analysis

• Counterflow_Jacobian_process.ipynb: Reformulate the data structure for MO analysis

• countercurrent_MO.py: Measurement optimization analysis to the RPB model

• counterflow_finite_difference_analysis.ipynb: MBDoE analysis to the RPB model with finite difference method, not using Pyomo.DoE

• draw_figure.ipynb: Generates all results figures in the manuscript

Results content

Cocurrent flow model

• Cocurrent_results/MO_results: folder, contains the MO results at different budgets. For e.g., Sep17_2000_a means budget=2000, optimized with A-optimality

• Cocurrent_results/MBDoE_results: folder, contains the process models ran with different variables. For e.g., Nov25_368_388 means ads.T = 368K, des.T=388K

• Countercurrent_results/MO_MILP_solution: folder, contains the MO results at different budgets. For e.g., updated_2000 means budget=2000, optimized with A-optimality

• Countercurrent_results/MBDoE_results/MBDOE_sens: folder, contained the process models ran with different variables. For e.g., .Tgin_des_391/Tgin_ads_368 means ads.T = 368K, des.T=391K

Run Models

Model code and how to run them are provided by Ryan Hughes in Ryan-Hughes-8/fixed_bed_adsorption repository:

Co-current flow model: https://github.com/Ryan-Hughes-8/fixed_bed_adsorption

Counter-current flow model: https://github.com/Ryan-Hughes-8/fixed_bed_adsorption/tree/counter-current-configuration

Run measurement optimization results

The rerun instructions are for counter-current flow model, which is the model we use for the paper / thesis results.

Step 1: Achieve sensitivity information (Jacobian) and error variance matrix

• Run Counterflow_MO_data_process.ipynb to generate all the sensitivity data

Step 2: Process the data for MO analysis

• Run Counterflow_Jacobian_process.ipynb to organize the sensitivity data to .csv files that MO code use

Step 3: Run MO analysis

• In Countercurrent_MO.py, find the toggles in line 17 - 27
• Choose mixed-integer option by mip_option
• Choose objective by objective
• Choose budget set. It is default to be the budget ranges we use for the paper, from 2000 to 10000

Run MBDoE analysis

Run MBDoE with k_aug

• Countercurrent_MBDoE.ipynb: MBDoE analysis applying Pyomo.DoE and k_aug to the RPB model with compute_FIM and run_grid_search

Run MBDoE with finite difference

• Counterflow-finite-difference-analysis.ipynb: MBDoE analysis applying finite difference method to the RPB model. It runs the grid search, and generates the thesis results.

Reproduce figures

• draw_figures.ipynb: include the functions and the path of results to reproduce Fig. 4.2, 4.3, 4.4, 4.5 in the Chapter 4 of Jialu's thesis.