Collaborative Research: Risk Informed Decision Making for Maintenance of Deteriorating Distribution Poles Under Extreme Wind Hazards

美国的电力分配依靠木杆支撑的架空电线。该配电系统在大风事件中受损，导致企业和居民的电力供应中断。木杆受到老化和风害压力增加的综合影响，导致配电基础设施不可靠和潜在不安全。作为关键的能源输送基础设施的一部分，1亿多根配电杆在暴风雨期间遭受巨大损失。目前用于预测一组通用电力系统部件的故障时间分布的方法不能从整体上解决老化和危险的同时影响。关于电杆失效年龄的研究一直是稀少、不完整和不一致的。该项目提出了结合时变衰减模型的配电杆随机有限元模型，以模拟配电杆在不同气候条件下的当前和未来响应状态。认知性不确定性将减少，产能和需求建模中任意不确定性的各种来源将被捕获。一种利用n维脆弱性表面与区域飓风风险模型相结合的新策略将被合并来估计故障概率，开发配电杆故障时间的预测模型，并量化预测中的置信度水平。将在配电网随机优化的基础上制定基于优先级的诊断和风险信息维护框架，该框架既考虑了物理部件在地理上对应激源的暴露，也考虑了它们对网络可靠性的重要性。这种方法所取得的成功并不能保证故障的避免，但指出在某些地区的网络中，使用维护良好的基础设施可能会降低风暴期间的故障风险，并为选择在正常运行情况下进行预防性更换的电杆提供指导。

Distribution of electricity in the United States relies on overhead lines supported by wood poles. This distribution system suffers damage in high wind events resulting in interruption in electrical power to businesses and residencies. The wood poles suffer from the combined effect of aging and increasing stress from wind hazards leading to an unreliable and potentially unsafe distribution infrastructure. Over 100 million distribution poles, which are part of the critical energy delivery infrastructure, suffer immensely during wind storms. Current methods for predicting the distribution of times-to-failure of a general set of power system components do not address simultaneous effects of aging and hazards with a holistic view. Research on age-at-failure of poles has been sparse, incomplete, and inconsistent. This results in increased uncertainties in predictions of pole lifetimes, thus obstructing informed and accurate decision-making on effective pole management strategies.The project proposes the development of stochastic Finite Element models for distribution poles integrated with time-dependent decay models to simulate the current and future state of the response of distribution poles to various climate conditions. Epistemic uncertainties will be reduced and various sources of aleatoric uncertainties in capacity and demand modeling will be captured. A novel strategy utilizing n-dimensional fragility surfaces integrated with regional hurricane hazard models will be merged to estimate failure probabilities, develop predictive models for the time-to-failure of distribution poles, and quantify the confidence levels in the predictions. A priority-based diagnostics and risk-informed maintenance framework will be developed on the basis of stochastic optimization for distribution networks that takes into account both the geographic exposure of the physical components to stressors and their importance for the reliability of the networks. The success afforded by this approach does not guarantee failure avoidance, but points to the regions in networks where use of well-maintained infrastructure may lower the risk of failures during storms and also provides guidance in selecting the poles for preventive replacement under normal operating circumstances.