Solid mechanics and AI hybrid approach to mitigation of climate change driven railway track buckling

铁路网的一个重大问题是防止在炎热天气下的轨道弯曲，特别是在气候变化导致极端和可变条件的情况下。失灵风险意味着列车运行速度缓慢，降低了网络容量，并给客户带来了糟糕的体验。穿越弯曲的轨道会导致脱轨，造成严重的安全后果。数据显示，在特定的赛道条件下，屈曲更为普遍。单独的微不足道的因素组合在一起，导致没有明显原因的屈曲。影响因素可能通过部件、安装、年龄、加载历史或其他因素的可变性而具有随机性。铁路网上的大规模数据收集为人工智能(AI)方法在理解铁路基础设施方面伴随传统力学提供了可能性。谢菲尔德将轨道系统视为受约束的梯形结构，建立了轨道屈曲的力学分析模型。在这项研究中，预计将用有限元模型来扩展这一点，以更充分地捕捉现实(和随机)行为。此外，还将利用铁路网数据开发一种人工智能屈曲模型。一种特殊的基于模糊集的方法已被证明可以提供快速可靠的预测(在另一种情况下，准确率为95-97%)，估计显著影响结果的因素。学术监督者和业界对这项研究的共同支持将提供一个难得的机会，利用具有潜力的最新想法来解决研究问题，从而带来改善铁路系统性能的真正好处。

A significant problem for rail networks is prevention of track buckling in hot weather, particularly with climate change leading to extreme and variable conditions. Buckling risk means trains run at slow speeds, reducing network capacity and giving poor customer experience. Traversing buckled track can lead to derailment with severe safety consequences. Data shows buckles are more prevalent for specific track conditions. Individually insignificant factors occur in combination leading to a buckle without an obvious cause. Contributory factors may have stochastic nature through variability in components, installation, their age, loading history or other factors.Large-scale data collection on the railway network is opening the possibility of Artificial Intelligence (AI) approaches to accompany conventional mechanics in understanding rail infrastructure. An analytical mechanics model of rail buckling considering the track system as a restrained ladder structure is available in Sheffield. In this research it is anticipated this will be extended with a finite element model to more fully capture realistic (and stochastic) behaviour. Alongside this an AI model of buckling will be developed with rail network data. A particular fuzzy-set based methodology has proven to offer fast reliable predictions (in alternative cases 95-97% accuracy) estimating factors influencing results significantly. The joint support of the research by academic supervisors and industry will provide an exceptional opportunity to address the research problem using the latest ideas with potential to deliver real benefits of improved rail system performance.