Research on the Evolution Characteristics and Collaborative Detection Strategies for Bottom Oil-Spill Fires in Oil-Immersed Transformers

Conventional single-type detectors generally face difficulties in achieving both early warning and fire-source localization for rapidly developing bottom oil-spill fires occurring in large oil-immersed transformers. In this study, a 500 kV transformer bottom oil-spill fire model is established to analyze the evolution of temperature field, smoke density, CO and CO2 concentrations, and thermal radiation, and to compare the response characteristics of various detectors. A stable high-temperature zone forms within 30 s after ignition, exhibiting a central temperature of 950–1000 °C and a distinct temperature gradient. CO2 and CO concentrations stabilize at 60,000–90,000 ppm and 600–1300 ppm, respectively, while thermal radiation intensity attenuates sharply with distance. Aspirating smoke detectors achieve early response but suffer from poor positioning and anti-interference performance. Distributed optical fiber detectors realize thermal anomaly identification and fire localization. Infrared and ultraviolet detectors respond within 10 s, and ultraviolet detection is less susceptible to thermal interference. This paper proposes a collaborative detection strategy of fiber temperature sensing and infrared–ultraviolet composite flame detection, forming a hierarchical fire identification framework. The findings inform the optimal design of transformer fire detection systems.