Power grid failure is very costly to any modern society, and preventing upheavals like the blackout in eastern US and Canada in the summer of 2003 is extremely important. Complete power grid failure may be triggered by the failure of a transformer, underground cable, overhead line insulator or any other component of the power grid. While close monitoring of expensive, centrally located assets like transformers, generators and circuit breakers is feasible and economically justified, it is extremely difficult to continuously monitor assets that are spread over long distances, and in some cases very difficult to reach, like overhead lines accessories and outdoor insulators. Condition monitoring of outdoor insulators is prohibitively costly, time consuming and unsafe. To overcome these problems, the use of machine learning (ML) in outdoor insulators condition monitoring and diagnostics could be a viable solution.

中文翻译：

---

机器学习在户外绝缘子状态监测与诊断中的应用


电网故障对于任何现代社会来说都是非常昂贵的，因此防止像 2003 年夏天美国东部和加拿大停电这样的剧变极为重要。变压器、地下电缆、架空线路绝缘子或电网任何其他组件的故障可能会触发整个电网故障。虽然对变压器、发电机和断路器等昂贵的集中资产进行密切监控是可行的，并且在经济上也是合理的，但持续监控分布在很远距离的资产却极其困难，在某些情况下很难到达，例如架空线路配件和户外绝缘子。室外绝缘子的状态监测成本高昂、耗时且不安全。为了克服这些问题，在室外绝缘子状态监测和诊断中使用机器学习 (ML) 可能是一个可行的解决方案。