关键词: 多尺度视角, 变压器油, 热解机制, 故障预警

Abstract: Under the context of constructing new power systems, transformers are subjected to prolonged high loads and complex stresses, leading to a marked increase in the frequency of thermal faults in oil–paper insulation systems. Although existing methods such as dissolved gas analysis (DGA) can identify fault types, they remain limited in capturing microscopic evolution mechanisms and providing early failure warnings. This paper reviews recent advances in the thermal decomposition mechanisms of transformer oils from a multiscale perspective, establishing the linkage between microscopic reactions and macroscopic characteristic parameters. It compares conventional mineral oils with emerging environmentally friendly ester-based insulating oils in terms of thermodynamic equilibrium and characteristic gas generation pathways, and reveals a synergistic degradation mechanism in which copper lowers the activation energy while organic acids initiate free-radical chain reactions. Finally, in view of the insufficient coupling between molecular dynamics simulations and real operating conditions, future research directions are proposed, including high-resolution free radical detection, evaluation of solvation effects, and mechanism-oriented molecular design, providing a theoretical basis for early warning of transformer thermal faults.

Key words: multiscale perspective, transformer oil, pyrolysis mechanism, fault early warning