Overhead Assessment

Sigmoid and tanh implementations with error detection using Maclaurin series.

Compile & Run

gcc -o activation Overhead_Assessment/C/main.c -lm
./activation

Functions

• sigmoid(x) - Sigmoid activation
• tanh(x) - Tanh activation
• sigmoid_error_detection(x, &err) - With fault detection
• tanh_error_detection(x, &err) - With fault detection

Notes

• Input range: -10 to 10 (auto-clipped)
• 30 Maclaurin terms
• Error tolerance: 0.000001

Training

NASA Jet Engine RUL Prediction Enabled with Error Detection

CNN model to predict Remaining Useful Life of aircraft jet engines using NASA C-MAPSS dataset.

Quick Start

1. Setup Environment

cd Training
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate
pip install -r requirements.txt

2. Download Data

Download NASA C-MAPSS dataset and place in archive/CMaps/:

• train_FD001.txt, test_FD001.txt, RUL_FD001.txt
• train_FD002.txt, test_FD002.txt, RUL_FD002.txt
• train_FD003.txt, test_FD003.txt, RUL_FD003.txt
• train_FD004.txt, test_FD004.txt, RUL_FD004.txt

3. Run

python training.py

Project Structure

nasa-rul-prediction/
├── README.md
├── requirements.txt
├── training.py
├── models/              # Output
└── archive/CMaps/       # Dataset folder

Requirements

• Python 3.12.3
• See requirements.txt

Results

Models saved to models/ folder with metrics (Accuracy, F1-score, Recall) and average inference time.