Home

Gridifix Wiki

Gridifix is an end-to-end fault detection and localization engine for medium-voltage (MV) power distribution networks. It pairs a Deep Neural Network (DNN) for normal-state prediction with Random Forest classifiers to detect and locate single-bus faults with high precision.

---

Navigation

• Power Grid Basics — CIGRE MV Network, pandapower, and distribution grid fundamentals
• Pipeline Overview — End-to-end system architecture and workflow explanation
• Dataset & Features — Dataset structure, feature engineering, and label definitions
• Dataflow — Complete data pipeline flow from raw inputs to model outputs
• Getting Started — Installation, setup, and quick-start guide

---

Quick Summary

Component	Technology	Purpose
DLPF Solver	Custom Python	Distribution Linear Power Flow — replaces Newton-Raphson for 10x+ speedup
DNN Baseline	Keras / TensorFlow	Learns normal (healthy) bus voltages & line power flows
Residual Engine	NumPy	Computes spatial deviation between live measurements and NN predictions
Fault Detection	YDF (Random Forest)	Binary classifier: healthy vs. faulted
Fault Localization	YDF (Random Forest)	Multi-class classifier: predicts the faulted line/bus ID
Data Synthesizer	pandapower	Generates 21k+ synthetic healthy & faulted grid states
MLOps	MLflow	Tracks experiments, metrics, and model artifacts

---

Repository

GitHub: Divyesh-Kamalanaban/gridifix