Research on multidimensional perception fusion early warning method based on fiber optic energy supply

To fully leverage the advantages of fiber optic energy supply technology, such as safety, reliability, efficiency, and maintenance free, and to address the high rates of missed and false alarms in existing cable fire warning methods. This article proposes a multi-dimensional perception fusion warning method based on fiber optic power supply. Firstly, the overall architecture of the system was designed, including fiber optic power supply modules and various information acquisition sensors such as temperature, smoke, CO concentration, etc. And explained its working principle. Secondly, multiple key models were constructed, including a local decision processing model, a multidimensional perception fusion model considering feature coefficients, and a BP neural network model. On this basis, a two-level detection and early warning process for cable fires combining local decision-making and BP neural network is proposed. Finally, through experiments in three scenarios: no fire, smoldering fire, and open fire, this method will be compared and analyzed with other methods in terms of warning accuracy. The results validated the effectiveness of the method proposed in this paper.