Echo-Guard: Industrial Predictive Maintenance System 🔊

Echo-Guard is an end-to-end Deep Learning application designed to predict machinery failure before it happens. It processes raw vibration sensor data from the NASA Bearing Dataset, converts signals into Mel-Spectrograms, and uses a Convolutional Neural Network (CNN) to classify equipment health in real-time.

🚀 Key Features

• Signal Processing: Automated pipeline to convert Time-Domain vibration data to Frequency-Domain Spectrograms using Librosa.
• Deep Learning: Custom 2D-CNN architecture built in TensorFlow/Keras achieving >95% accuracy.
• Real-Time Dashboard: Interactive User Interface built with Streamlit for live sensor monitoring.
• Fault Detection: Distinguishes between "Healthy" operation and "Critical" bearing degradation.

🛠 Tech Stack

• Python 3.9
• TensorFlow/Keras (CNN Implementation)
• Librosa (Audio Feature Extraction)
• Streamlit (Frontend)
• Pandas/NumPy (Data Engineering)

📸 How It Works

1. Input: System accepts raw NASA sensor data (or .wav files).
2. Preprocessing: Applies Short-Time Fourier Transform (STFT) to generate a spectrogram.
3. Inference: The CNN analyzes the visual pattern of the spectrogram.
4. Output: Returns a confidence score and a Go/No-Go maintenance alert.

💻 How to Run

# 1. Install Dependencies
pip install -r requirements.txt

# 2. Run the Dashboard
streamlit run app.py