The electrical grid is the backbone of our modern life, delivering power to homes, business and industries. However, with the increasing demand for electricity, renewable energy integration, and the rise of decentralized energy sources, maintaining grid stability has become a complex challenge. This is where Artificial Intelligence (AI) comes into play, offering innovative solutions to optimize grid performance and resilience.
By leveraging optimized AI algorithms and advanced data analytics, grid operators can gain real-time insights into grid operations, monitor critical parameters and automatically detect anomalies for early identification of potential issues enabling prompt intervention to prevent disruption and ensure a stable power supply.

Approach

This use case is based on the "Electrical Grid Stability" dataset from the UCI Machine Learning Repository.
The goal was to determine if an electrical grid is stable or not, only using 12 parameters such as production, consumption, time reaction, etc.
Each sample was measured every 2 seconds, 10 000 real signals were gathered.
We then used NanoEdge AI Studio to create an N-class classification project based on these inputs to try to predict whether the power grid was stable or unstable.

Sensor

Generic sensors.

Data

2 classes of classification Stable, Unstable
Signal length 12 (multi-sensors)
Data rate 0.5 Hz

Results

N-class classification:
92.35% accuracy, 0.3 Kbytes of RAM, 1.6 Kbytes of Flash memory

uc-neais-electrical-grid-stability-e1669384252433 uc-neais-electrical-grid-stability-e1669384252433 uc-neais-electrical-grid-stability-e1669384252433

Green points represent well classified signals. Red points represent misclassified signals. The classes are on the abscissa and the confidence of the prediction is on the ordinate 

Model created with
NanoEdge AI Studio
NanoEdge AI Studio
Compatible with
Any STM32 MCU
Any STM32 MCU