Optimal Experimental Design and Uncertainty Propagation for Process Scale-Up Paper

Optimal experimental design for the isolated lab-scale batch reactor (BR) model is performed in the batch_reactor_exp_design.py file.

Parameter estimation and uncertainty quantification for the kinetic parameters of the isolated BR model is performed in the batch_reactor_parameter_est.ipynb file.

The uncertainty in the estimated kinetic parameters of the isolated BR model is propagated to the design of the Williams-Otto [1] process in the williams_otto_process.ipynb file.

The batch_experiment.py file holds the Experiment class and the mathematical model of the isolated BR.

Data and results are stored in the Data and Figures folder, respectively.

Running these files require the Pyomo and IDAES-PSE packages. The following section provides guidance on how to install Pyomo and IDAES-PSE.

Making an IDAES-PSE environment

We recommend using a Conda environment.

1. Create a new Conda environment (replace my-idaes-env with your preferred name)

conda create --yes --name my-idaes-env python=3.10
conda activate my-idaes-env

2. Install IDAES

conda install --yes -c conda-forge idaes-pse

3. Install the IDAES extensions

idaes get-extensions

The IDAES extensions include the compiled solver binaries and function libraries required by many IDAES examples.

4. Install Pyomo with your preferred package manager

pip install pyomo

5. Install NumPy and Pandas

pip install numpy pandas

6. Install Matplotlib and SciPy

pip install scipy matplotlib

[1] Biegler LT. Nonlinear Programming: Concepts, Algorithms, and Applications to Chemical Processes. Society for Industrial and Applied Mathematics (2010). ISBN 978-0-898717-02-0