基于轨道动检数据的轨道板的变形识别及预测

现有对高速铁路板式无砟轨道变形病害的检测效率不足,检测成本过高,而通过轨道动检数据能够一定程度上反映轨道板变形程度.因此,搜集了CRTSⅠ、Ⅱ、Ⅲ型板线路3 a内的动检数据,引入小波能量作为轨道板变形评价指标,通过建立时空数据挖掘模型实现了不同轨道板的变形定位识别和劣化预测.研究结果表明:受当地气温影响,轨道板变形程度...

Deformation Recognition and Prediction of Track Slabs Based on Track Inspection Data

Existing methods for detecting deformation of high-speed railway slab ballastless tracksis low in efficiency and high in cost, while the track dynamic geometry inspection data can reflect the deformation level of slabs in a way.In this work, the track dynamic inspection data of railway lines of CRTS Ⅰ, Ⅱ, Ⅲ slabs within three years were collected and the wavelet energy was used as the deformation evaluation index of track slab. Finally, a temporal-spatial data mining model was proposed to realize the recognition and degradation prediction of track slabs. The results show that, affected by the local air temperature, the deformationlevel of track slabs has a seasonal pattern.Type-Ⅰ and type-Ⅱ slabs display warping and arching deformation in a high-temperature environment, while type-Ⅲ slab shows frost heavingin a low-temperature environment. Of the three types of track slabs, the deformation level of type-Ⅰ slab is mild, while the type-Ⅱ slab is most severe. The residual deformation of type-Ⅱ slab will accumulate over time, which may eventually lead to surface irregularity exceeding the limit. The long-term and short-term memory network can realize short- and mid-term prediction of the track slab deformation within 15 to 30 days. The the best predictionR-square value of the type-Ⅰ platedeformation is close to 0.9, while that of the type-Ⅱ and type-Ⅲ plate deformations exceeds 0.9.