RUL Prediction using LSTM and Hyperparameter Optimization

This project implements Remaining Useful Life (RUL) prediction for NASA CMAPSS Turbofan Jet Engines using Long Short-Term Memory (LSTM) neural networks based off engine variables such as cycle times, operational settings, and sensor readings. The solution utilizes sliding windows for time-series data processing and Optuna for hyperparameter optimization across multiple datasets.

---

Project Overview

This project:

• Processes CMAPSS time-series datasets using sliding windows with padding and masking.
• Builds LSTM models for RUL prediction.
• Performs hyperparameter optimization using Optuna for window size, batch size, learning rate, and epochs.
• Ensures robust model evaluation with RMSE metrics.

---

Setup Instructions

1. Clone the Repository:

   git clone https://github.com/your_username/JetEngineRUL.git
   cd JetEngineRUL
   

2. Create a Virtual Environment:

   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate
   
   or, using conda,
   
   conda create --name JetEngineRUL python=3.12.2 -y
   conda activate JetEngineRUL
   

3. Install Dependencies:

   pip install -r requirements.txt
   
   or, using conda,
   
   conda install requirements.txt
   

4. Download Engine Data: https://data.nasa.gov/Aerospace/CMAPSS-Jet-Engine-Simulated-Data/ff5v-kuh6/about_data

Dependencies

The following dependencies are required for this project:

• Python (3.12.2)
• TensorFlow (>=2.9.0)
• NumPy
• Pandas
• Scikit-learn
• Optuna
• Matplotlib

File Structure

   JetEngineRUL/
     ├── data/
     │   ├── raw/
     │   │   ├── raw_data_instructions.txt   
     │   │   └── CMAPSSData/                        # Place C-MAPSS dataset here
     │   └── processed/                             # Contains processed dataset
     │       └── processed_data_instructions.txt
     ├── notebooks/
     │   ├── 01_eda.ipynb                           # Exploratory Data Analysis
     │   ├── 02_preprocessing.ipynb                 # Data Preprocessing
     │   └── 03_modeling_and_evaluation.ipynb       # Modeling and Evaluation
     ├── results/
     │   ├── figures/                               # Figures from results
     │   ├── metrics/                               # Performance Metrics
     │   └── models/                                # Saved Models
     ├── requirements.txt                           # Python dependencies
     ├── LICENSE                                    # MIT License
     └── README.md                                  # Project description and instructions