Real-Time Monitoring of Air Discharge in a Switchgear by an Intelligent NO₂ Sensor Module.

Area of Science:

• Materials Science and Engineering
• Electrical Engineering
• Environmental Monitoring

Background:

• Air-insulated power equipment relies on air as an insulator, but discharge faults produce nitrogen dioxide (NO2).
• Existing NO2 sensors face interference from other gases and humidity, limiting their reliability.
• Indium oxide (In2O3) based sensors show promise for NO2 detection due to their advantages.

Purpose of the Study:

• To develop a robust sensor array for accurate NO2 detection in power systems.
• To mitigate interference from common gases and humidity affecting sensor performance.
• To establish a practical platform for real-time detection of electrical discharge faults.

Main Methods:

• Integration of four In2O3 composite materials into a sensor array for NO2 detection.
• Utilizing a convolutional neural network (CNN) and long short-term memory (LSTM) model with attention mechanism for NO2 concentration evaluation.
• Development of a microgas sensor-based platform for fault detection in switchgear.

Main Results:

• The In2O3 sensor array demonstrated detection of 250 ppb NO2 with excellent selectivity against CO.
• The CNN-LSTM model achieved NO2 concentration evaluation within 1 ppm with a detection error of 63.69 ppb, reducing humidity impact.
• The platform detected an average of 726.58 ppb NO2 from 10 discharge faults at 15 kV, indicating severe issues.

Conclusions:

• The developed In2O3 sensor array and advanced AI models offer a reliable solution for NO2 detection in power equipment.
• The microgas sensor platform effectively identifies electrical discharge faults in switchgear.
• This technology holds significant potential for widespread deployment in low- and medium-voltage switchgear monitoring.