Identification of Track Irregularity Potential Defect Based on the Track Inspection Data Feature Clustering
Research purposes:Track irregularity potential defect has the characteristics of strong concealment,and the identification of these defects with high accuracy is the premise of carrying out the preventive maintenance for railway track.Based on the dynamic inspection data,some potential defects identification indexes were constructed using track quality index(TQI)and track irregularity amplitude.Taking the identification indexes as the clustering samples,the k-means++clustering algorithm and the elbow method were combined to judge the optimal cluster number of samples to determine the potential defects and their location.The effectiveness of the proposed method was verified from the dynamic detection data of a ballasted railway.Research conclusions:(1)For the standard deviation management,two potential defects on the 16 km continuous track sections were identified effectively,with the historical maximum values of TQI of 5.5 mm and 5.3 mm,and these locations are consistent with the real site.(2)For the peak management,two potential defects and their locations within the 500 m track section were also detected effectively,with the historical maximum amplitudes of longitudinal level irregularity of 3 mm and 3.3 mm.(3)The proposed potential defects identification method can provide assistance and reference for renovation of potential defects under the preventive maintenance strategy with the machine tamping and manual interventions.
万方数据
维普
