🏭 Industrial Digital Twin & Predictive AI

📋 Executive Summary

This project is a full-stack Industrial Digital Twin application designed to simulate and predict machinery failures in real-time.

Unlike static analysis scripts, this software features a GUI (Graphical User Interface) that allows engineers to interact with the Digital Twin, run auto-pilot simulations, and visualize decision boundaries dynamically. It acts as the "Brain" (Layer 2) of my IoT Portfolio, processing telemetry from the C-Firmware Layer.

---

🖥️ Application Screenshot

🛠️ Key Features

• 🧠 Real-Time Inference: Uses a Random Forest Classifier to predict failure probability instantly based on user input or simulation.
• 🎮 Interactive Simulation (Auto-Pilot): Features a "Live Mode" that generates synthetic fluctuations in temperature and vibration to test the model's reaction speeds.
• 📊 SCADA-like Dashboard: A professional GUI built with Tkinter and Matplotlib, featuring:
  • Control Sliders for manual testing.
  • Live Event Logging (Audit Trail).
  • Dynamic Scatter Plot with "Glow" effects for active sensors.
• ⚠️ Visual Alerts: Immediate color-coded feedback (Green/Red) based on risk thresholds (>85°C or >4.5mm vibration).

---

🏗️ Technical Architecture

The system follows a Model-View-Controller (MVC) pattern:

1. Backend (Model): * Ingests raw CSV data (Robust ETL).
   • Trains the Random Forest model on startup (n_estimators=100).
   • Exposes an API-like method predict_status(temp, vib).
2. Frontend (View):
   • Tkinter window with responsive layout.
   • Embedded Matplotlib canvas for data visualization.
3. Controller:
   • Handles events (Button clicks, Slider moves, Timer loops for auto-pilot).

---

⚙️ How to Run

Prerequisites: Python 3.8+ and the following libraries:

pip install pandas numpy scikit-learn matplotlip