Efficient SRU Optimization via Deep Learning-driven LSTM Dynamical Models

Project Overview:

This project utilizes Long Short-Term Memory (LSTM) networks to estimate key variables in a sulfur recovery unit (SRU) of a refinery plant. The goal is to achieve accurate predictions for process optimization, even in offline analyzer scenarios.

Files Included:

• IN_Table.csv: Input dataset containing essential measurement series for predicting SO2 and H2S concentrations.
• OUT_Table.csv: Output dataset containing target variables Out1 and Out2.
• H2S_Model.ipynb: Jupyter Notebook containing the code for the H2S concentration prediction model.
• SO2_Model.ipynb: Jupyter Notebook containing the code for the SO2 concentration prediction model.
• H2S_GUI.ipynb: Jupyter Notebook containing the Streamlit code for the H2S concentration prediction GUI.

Instructions for Running:

1. Ensure you have Python and Jupyter Notebook installed on your system.
2. Clone this repository to your local machine using git clone <repository-url>.
3. Navigate to the project directory.
4. Place the IN_Table.csv and OUT_Table.csv datasets in the same directory.
5. Open the desired Jupyter Notebook (H2S_Model.ipynb or SO2_Model.ipynb) in Jupyter Notebook.
6. Follow the instructions provided in the notebook to run the code and train the LSTM model.
7. For the GUI application, open H2S_GUI.ipynb and follow the instructions to run the Streamlit app.

Dependencies:

• Python 3.x
• Jupyter Notebook
• Pandas
• NumPy
• Matplotlib
• TensorFlow
• Streamlit (for GUI)