Abnormal data identification method for track smoothness evaluation

Abstract

The invention discloses an abnormal data identification method for track smoothness evaluation, which comprises the following steps of: acquiring track monitoring index data, and segmenting the monitoring index data according to a set track length to form a data set; the method comprises the following steps of: preprocessing a data set, constructing L isolation trees by adopting an isolated forest algorithm, then dividing the L isolation trees into n groups by applying a system sampling method, and constructing n sub-forest anomaly detectors; generating a base forest anomaly detector; through a base forest anomaly detector, performing anomaly judgment on each piece of data reaching the sliding window, and updating the data set to obtain an updated data set; and calculating an anomaly rate difference value between each sub-forest anomaly detector and the base forest anomaly detector based on the updated data set to form a new base forest anomaly detector, and performing abnormal data identification on the data set through the new base forest anomaly detector.

Description

technical field [0001] The invention relates to the field of rail transit, in particular to an abnormal data identification method used for rail ride comfort evaluation. Background technique [0002] As an important part of urban transportation, subway rail transit is the top priority of subway companies to ensure its healthy and safe operation. Track smoothness is the main source of train vibration and increased wheel-rail force. Safety, stability, comfort, service life and environmental noise all have important influences. Therefore, it is necessary to obtain the smoothness results of the subway track from the evaluation of various indicators of the subway track, which can be used for the maintenance decision analysis of the subway operation company. [0003] However, in practice, the massive raw data collected by the subway track index collection system is affected by the diversity, uncertainty and complexity of the environment, making the actual data collected messy, wi...

AI Technical Summary

Problems solved by technology

[0003] However, in practice, the massive raw data collected by the subway track index collection system is affected by the diversity, uncertainty and complexity of the environment, making the actual data collected messy, with missing and abnormal phenomena, many The case does not meet the specification requirements for the modeling of the smoothness evaluation model of the subway track

Images