The Research on Transformer Fire Alarm Based on Multi-Sensor Data Analysis

Artificial intelligence (AI) neural network algorithms have been applied to fire alarm systems by some researchers due to their high fault tolerance and successful results in engineering. However, existing research lacks neural network data fusion analysis specifically for transformer fires. In this study, BP neural network and RBF neural network algorithms are selected, a multi-sensor test platform for transformer fires is developed, and both transformer fire and interference tests are conducted. Multi-sensor data are collected, and neural network data fusion training is performed. The results demonstrate that both the BP and RBF neural network data fusion algorithms can reduce false alarms and shorten the alarm time compared to single sensors, with the RBF neural network proving to be more effective. This research provides a foundation for the application of neural network data fusion algorithms in transformer fire detection scenarios.