Research on the measurement of the rail straightness based on the outer boundary support point model

Abstract

Rail straightness is an important index of the quality of large components of steel rails, and is of great significance for product quality testing in rail plants. At present, the measurement of straightness in rail factories still depends on the manual use of straightness rulers for measurement, which leads to low measurement speeds and precision. Therefore, we propose a non-contact measurement method based on the outer boundary support point model. Using laser profilers to measure the original curve of the rail straightness, the proposed straightness curve processing algorithm in this paper is used to obtain the median filter curve and calculate the maximum difference values. Results show excellent performance and indicate that the proposed method presents a better straightness measurement approach than traditional manual measurements using physical straightness rulers.

Introduction

In recent years, high-speed railways have undergone rapid development, resulting not only in economic growth, but also in higher requirements for high-quality railway construction. Straightness is a geometrical feature to evaluate rail quality [1], and refers to the roughness perpendicular to the rail reference plane, which directly affects the running speed, and therefore the safety of train [2].

Many straightness measurement methods for rails have been proposed [3]. In the recent years, systems tend to use noncontact techniques, mainly based on automated visual inspection using structured light [[4], [5], [6], [7], [8]]. For example, machine vision methods that use different light sources to scan the surface of the rails, are also used to detect straightness. Other methods are based on a laser triangular ranging principle using distance sensors [9].

The most widely used method in rail welding plants is to use an electronic straightness ruler, such as the SEC-RC electronic ruler, which is placed on the rail ridge lines, and the measurement results are sent wirelessly via Bluetooth to an embedded handled computer, which records the straightness results.

In this paper, an accurate and convenient approach is proposed to measure rail straightness. This method avoids human

errors during the rail quality evaluation, and significantly promotes the development of fully automatic non-contact rail straightness detection. The rest of the paper is organized as follows: Section summarizes the related standards, Section describes the proposed approach, Section shows the experimental results and Section presents conclusions.