Hydraulic Digital Twin

Problem

Real industrial and laboratory datasets are often sensitive, incomplete or difficult to share publicly. This project demonstrates the shape of a digital-twin monitoring workflow without exposing confidential facility data.

Approach

• Create synthetic hydraulic time-series data with realistic operating states.
• Run validation checks to identify missing, saturated or suspicious sensor behaviour.
• Extract features for anomaly detection and operating-state classification.
• Generate Markdown/HTML reports that summarise diagnostics and results.

What it demonstrates

The repository shows how to structure an applied-AI engineering demo: reproducible scripts, synthetic data, documented assumptions, tests and visual outputs. It is designed to communicate the workflow, not to claim that the synthetic data reproduces a specific facility.