RU2699942C1 - Method for prediction of development of anomalies in rail head - Google Patents

Abstract

FIELD: defectoscopy.

SUBSTANCE: use: for evaluation of rails state. Summary of invention consists in the fact that defectoscopic means are moved along rail track, rails head is probed by them, received signals are evaluated, anomalies are detected and recorded with reference to track coordinates, additionally, according to the results of probing, integral parameter of each anomaly is formed, during subsequent movements of the flaw detection devices along the rail track, probing is repeated, the integral parameters of the current and previously found anomalies are compared, the dynamics of the change of the integral parameter of each anomaly is estimated, the prospects of its development and the repair measures are planned.

EFFECT: possibility of reliable estimation of anomalies in a rail head, as well as possibility of predetermining development of anomalies before critical situations occur.

1 cl, 3 dwg

Description

A method for predicting the development of anomalies in the rail head relates to methods and means of non-destructive testing of materials and can be used to assess the condition of rails in order to take timely measures to eliminate problems.

It is known that the largest (up to 70%) number of dangerous defects in the rail track occurs for the rail head [1], which determines the relevance of the invention.

Defects can have various: shape, size, orientation, [1, 2] which complicates their detection and evaluation.

Defectoscopy is carried out by moving search tools along the rail track by various transport devices (wagons, railcars, carts) [3, 4].

Flaw detection is carried out by various technical means:

1. Magnetic-dynamic (MD), including multi-channel.

2. Ultrasonic (ultrasound), as a rule, multi-channel.

3. Eddy current (VT), used not often due to the small depth of detection of defects.

Each of the means has its own capabilities, application features, advantages and disadvantages, and many random factors affect the results of sounding.

Improvement of flaw detection tools has led to the possibility of their use on high-speed vehicles moving at speeds of up to 80 km / h and higher, which does not require long occupancy of rail tracks and can increase the frequency of travel and the probability of detecting defects, which means traffic safety. As a result, rail tracks are flawed several times a month (from 20 to 60 times a year [3]).

Sensing results are s signals from flaw detectors, indicating anomalies in the rail head. Hereinafter, anomalies are understood as violations of the state of the rail head (potential defects), which according to [3] are classified as “suspicious” sections of rails, on the defectograms of which there are signals from defects that do not allow us to unambiguously assess the degree of their danger (see page 10 [ 3]).

These signals are presented to the operator and stored in the form of a scan A (on the coordinate plane: “amplitude - propagation time of ultrasonic vibrations) and / or scan B (“ propagation time of ultrasonic vibrations - transducer travel time ”) in a certain observation window corresponding to a small portion of the path where signals from the desired defect are received [4, 5].

For evaluation, s signals in accordance with the regulatory documents of the Russian Railways [6], as a rule, are compared with two levels:

- noise S _W , allowing to exclude from further consideration interference signals with small amplitude;

- defect S _op , in which the anomaly should be considered a dangerous defect with the adoption of appropriate measures.

Such an analysis can be carried out automatically and promptly displayed on the operator’s screen during the movement of flaw detection tools.

All signals s obtained in the process of sounding are stored with reference to the longitudinal coordinates of the rail track for subsequent analysis in the diagnostic laboratory. To solve the binding task, you must:

- determine the current coordinate of flaw detection tools on the rail using: global navigation systems (GPS, GLONAS, etc.), speed sensor and odometer, signals from structural elements of the rail (joints, rail linings, etc.) and others methods for achieving positioning accuracy to the level of discreteness of sounding of the rail head, for example, 2 mm;

- combine the results of sounding s with all channels of spatially distributed means of defectoscopy to the rail cross sections (scan B for all channels in the “Reduction into a single section” mode [4]).

Laboratory analysis of sounding results is associated with significant labor costs.

Known mobile flaw detection complex [7] (patent RU 2438903), which suggests the presence of various means of flaw detection of railway infrastructure. This decision contains fundamental views on a mobile diagnostic complex with many diagnosed parameters.

The disadvantages of the method [7] are:

1) the lack of technical solutions to many specific problems of flaw detection due to the declarative-advertising nature of the patent;

2) the diagnostic complex is designed for high-speed flaw detection, and for analysis uses only the results of current soundings, without taking into account the previous ones.

There is a method of evaluating anomalies in the rail head [8] (Patent RU 2446971), which consists in probing it by moving flaw detection tools along the rail track and evaluating the received signals, detecting anomalies and registering them with reference to the coordinates of the rail track. The obtained sounding results are analyzed promptly and in laboratory conditions, and these actions are performed by the operator based on experience.

The disadvantage of this method is the subjective nature of the assessment of anomalies, depending on the experience of the operator, his care, etc.

Closest to the claimed method for evaluating anomalies in the rail head is the method [9] (Patent RU 2521095), which consists in probing it by moving flaw detection tools along the rail track and evaluating the received signals, detecting anomalies and registering them with reference to the coordinates of the rail the way. The obtained sounding results are analyzed promptly and in laboratory conditions. The method involves the use of multichannel ultrasound and MD flaw detection tools, indicates the appropriateness of their joint analysis. In addition, the need for storing the results of soundings in the “diagnostic map” with subsequent analysis by the operator is indicated.

The disadvantage of this method is the subjective nature of the assessment of anomalies, depending on the experience of the operator, his care, etc. The multi-channel means of flaw detection involves taking into account the signals of each channel, measuring its parameters (up to 5 parameters of each signal) and analyzing them to assess the degree of danger of the detected anomalies. This significantly complicates the processing of signals and their evaluation when making decisions.

Thus, the common disadvantages of the methods known to the authors for evaluating anomalies in the rail head are:

1) the lack of objective criteria for evaluating signals from anomalies, except for the instantaneous amplitudes of the received signals in different channels, which leads to the need for visual evaluation of the results of flaw detection by experienced operators and prevents the automation of the processing of sounding results;

2) the absence of specific methods for jointly taking into account the results of successive journeys of sections of the rail track by flaw detection means.

To solve these problems in a method for predicting the development of anomalies in the rail head, namely, they are probed by moving flaw detectors along the rail track and evaluating the received signals, anomalies are detected and recorded with reference to the rail track coordinates, additionally, based on the sounding results, an integral the parameter of each anomaly, with subsequent soundings, the dynamics of changes in the integral parameter of the anomaly is estimated, the prospects for the development of the anomaly and iruyut repair activities.

Significant differences of the proposed method for predicting the development of anomalies in the head of the rails are as follows.

Based on the results of sounding, an S - integral parameter of each anomaly is formed. The integral parameter is, for example, a number that combines abnormal signals into a single scalar quantity, allowing one to evaluate the degree of their danger and other characteristics, including in automatic mode.

In the prototype, the integral estimate of S is not used.

During subsequent movements of flaw detection tools along a given rail track (after a certain time interval determined by the schedule of periodic monitoring), the soundings are repeated to obtain an integral estimate S of the anomaly.

The prototype also provides for consecutive rail travel with the goal of defectoscopy, but with only the amplitudes of the signals of the anomalies (without an integral estimate of the S anomaly).

The integral parameters S of current and previously found anomalies are compared, which allows us to determine the dynamics of their changes.

The prototype also provides for the possibility of comparing “new” sounding signals with “old” - previous ones, according to the results of comparisons they make some assessment and replace the “old” signals with “new” ones. Thus, the evaluation of the results is carried out on a limited time interval.

The dynamics of changes in the integral parameter S of each anomaly is estimated and the prospects for the development of each anomaly are predicted, which allows one to determine in advance the time it will reach critical values.

In the prototype forecasting is not provided.

Repair activities are planned, which allows for appropriate proactive measures. Scheduled work (installation of reinforcing defective section of overlays, removal of a rail section, etc.) to restore the integrity of the track until a critical defect size is achieved is much cheaper than immediate restoration of an acute defective rail.

The prototype provides only a reaction to detected defects.

The technical result of using the proposed method for predicting anomalies in the rail head is the ability to predetermine the development of anomalies before critical situations arise that require immediate intervention, including in automatic mode.

The inventive method for predicting the development of anomalies in the rail head is illustrated by the following graphic materials:

FIG. 1. The algorithm (simplified) of the proposed method in predicting the development of anomalies.

FIG. 2. An example of an integrated assessment of multichannel ultrasonic flaw detection.

FIG. 3. Schedule for predicting the development of anomalies.

Consider the possibility of implementing the proposed method.

Multichannel flaw detection tools of one or several types (MD, UZ, etc.) are installed on a wagon, motor-track, car, or any mobile unit on a rail track, providing ease of attachment, electromagnetic compatibility, and other circumstances. Flaw detectors are moved along the rail, providing diagnostics of the head over the entire distance of the rail.

The operation algorithm of the proposed method is shown in FIG. 1. In accordance with the algorithm, a rail (its head) is probed, for which exciting signals s are emitted and receive corresponding signals s reflected from possible defects. At the same time, the various coordinates of the navigation tools discussed above determine the current coordinates of the flaw detection tools on the rail, taking into account the speed of movement, relative position, etc. The frequency (periodicity) of sounding of the rail head during the scanning process is chosen to achieve the required resolution along the length of the rail, for example, every 2 mm At the first stage, the signals s are filtered by comparing them with the noise threshold S _w . Signals s received on various channels of a multichannel flaw detector and exceeding the noise threshold in amplitude indicate the possible presence of anomalies in the rail head and are displayed as a type A sweep, which is difficult to observe in the on-line mode.

At the second stage, the results of measurements of s are brought to the cross section of the rail with the formation of a reamer B, which is displayed to the operator promptly during the movement, and is stored for subsequent analysis.

In the presence of known signs of signals from the defect (inclined bursts of signals), an integral assessment of the defective cross section is formed. In the absence of such signs (for example, the simultaneous appearance of signals in the entire time zone from electromagnetic interference), the signals are classified as "random" and excluded from further consideration.

Thus, in the analysis, the signals s must be divided into (Fig. 1):

1. “Random”, which should be excluded from consideration, because by indirect signs, they do not apply to signals from defects (for example, signals from short-term electromagnetic interference - occur in the entire time interval at the same time, in contrast to signals from reflectors, which shift along the time scale as the EAT moves).

2. “Dangerous or ODR according to [2]” - for which the integral assessment is higher than the threshold value, indicating the need for an immediate reaction.

3. “Suspicious” according to [2] - which are not signals from dangerous defects, but require closer attention during subsequent diagnostic trips.

In accordance with these obvious rules, the indicated signal differentiation operations are carried out promptly in the process of moving flaw detection tools.

At the third stage, an integral S _int estimate of the signals s received by various means and their sounding channels is formed and stored for subsequent analysis. This problem can be solved in various ways:

- by all means of sounding;

- for each channel of a specific sensing means;

- by groups of sensing equipment.

In the case of a single-channel MD defectoscopy, the amplitude ΔS _m - the amplitude of the anomaly signal can act as S _int . If there are several - n MD channels with signals ΔS _mi, you can use the average amount:

The signals of each ultrasound channel that implements reflection methods (echo method and mirror control methods), as noted in [5], can be described by at least four parameters: the conditional height ΔS of the _ultrasound and the conditional duration ΔL of the signal, the amplitude U (detection coefficient k _d ) and depth h _d reflector. For example, for a system consisting of n = 6 ultrasound channels that sound the rail head, 24 measured parameters are formed in total. It is natural to operate with so many parameters when analyzing signals, especially with multiple inspections (drives), it is very problematic.

In this case, the type B scan has the form shown in FIG. 2b. The channels of ultrasonic flaw detection differ in directivity. They can be grouped according to the principle of “driving over” - “leaving”, but according to the general sounding plane and according to other principles.

As an example in FIG. Figure 2a shows the sounding pattern of the rail head 1 with a transverse defect 2, voiced as it is scanned by six (n = 6) electro-acoustic transducers (EAP): “driving in” No. 3, 4 and 5, and “driving off” No. 6, 7 and 8 with different angles of entry of ultrasonic vibrations into the metal of the rail. In accordance with the well-known technical solutions [4], the registration of signals from each EAA is carried out in a separate zone (“track”) of registration. Moreover, in order to save the area of defectograms 9, signals from the so-called "eponymous" EAP (at the same angles, but against the movement, for example, EAP 3 and 6) are recorded on one common track. Due to the different directions of sounding of defect 2 on defectogram 9, groups are formed in the form of a packet of signals in the form of oblique lines of different orientations (see, for example, the packet of signals 3p and 6p in Fig. 2b) and can be visually (and programmatically) distinguishable.

In the considered example, for the formation of the integral parameter, the most informative parameter of the echo signal pack from a possible defect is _selected - the conditional height ΔS _{ultrasound of the} signal pack (Fig. 2 b), which is measured in microseconds (or in arbitrary units). In contrast to the conventional length ΔL (in mm), taken as the main parameter in [6], this parameter does not depend on the error (slippage) of the odometer. The choice of ΔS _{ultrasound is} also due to the fact that in the presence of a group of signals (the so-called crushed pack) from the analyzed section, it is possible to add up the values of their parameters and obtain a single signal parameter from a specific channel.

могут быть автоматизированы.Physically, the images of oblique packets of signals with parameters ΔS _ultrasound scan type B represent the number of echo signals with amplitudes above the threshold level obtained from the reflector within the radiation pattern of the EAP during scanning by the search system (EAP No. 3 - 8 in Fig. 2). The sum of all signals received from the defect from all possible sides by a plurality of converters reflects to a certain extent the degree of development and potential danger of the defect. This total value can be used as an integral estimate of the S _int defect in multichannel scanning. Measurement operations of parameters ΔS _ultrasound , their summation and obtaining averaged parameters

can be automated.

Thus, each of the groups of received reflected signals can be estimated as its conditional height - ΔS _ultrasound . Then, as an integral estimate of S _int , the averaged sum of the combined signal groups can act

Sufficiently frequent passages (in loaded areas up to 5 times a month) by means of flaw detectors along the track distances allow us to evaluate the dynamics of the anomaly in time T, which consists in approximating the integral estimates S _int obtained during each inspection passage, for example, a linear polynomial (Fig. 3 ), using the least squares method, which gives an unbiased estimate of random components with minimal dispersion. In FIG. 3 along the horizontal axis shows the serial numbers of inspection passages of the flaw detector along a portion of the path containing anomalous sections D and G in the rail head 1 when analyzing control signals according to the above algorithm (Fig. 1).

The obtained polynomial allows one to carry out a development forecast (Fig. 1) of the integral estimate of the anomaly S _int , i.e. time to reach the defective (dangerous for further use) level S _op . For example, to anomalies D development perspective S _D sufficiently distant (favorable for the further operation of the route section), and for anomalies G dangerous level S _op integral evaluation defect S _G can be achieved quite quickly, and after the 25th inspection directions (FIG. 3 ) it is necessary to plan anticipatory (repair) work. The composition and scope of the repair work depends on many external factors: the load and missing tonnage, such as the track (link or jointless) of the section with the detected anomalies, and is not included in the subject of the application.

The above sequence of formation of the integral parameter in multichannel ultrasonic testing is given as an example proving the feasibility of the proposed method. In the general case, when generating a generalizing parameter, other signal parameters can be used. For example, in addition to the relative height, the amplitudes of each echo can also be taken into account. When summing the parameters, one can additionally introduce channel significance coefficients (for example, signals of mirror channels are more likely to indicate developed defects than signals from echo channels).

In a similar way, a generalizing integral parameter can be obtained for several control methods: for example, magnetic and ultrasonic methods. In any case, the parameters of individual methods should be normalized, using the corresponding coefficients, the significance of each method is determined, and a numerical value is obtained that combines the anomalous signals into a single scalar quantity. The magnitude of the danger of the anomalous cross section is estimated from this value and the dynamics of the development of the anomaly with multiple (periodic) monitoring is monitored.

The analysis of many defectograms and the subsequent development of events with rails shows that the rate of development of transverse (most dangerous) cracks in the heads of rails is much higher than longitudinal. As a result, estimation of the defect development rate in the rail head allows, among other things, determining the defect orientation.

Thus, the claimed method can be implemented and allows you to evaluate the anomalies in the head of the rails and determine the prospects for their development. Planned repairs are much cheaper than the emergency removal of defective rails from the track.

Information sources:

1. Shur EA Rail damage. - M: Intekt, 2012 .-- 192 p.

2. Instruction “Rail defects. Classification, catalog and parameters of defective and acute defective rails. " Approved Russian Railways by order No. 2499 r of 10.23.2014. - 140 p.

3. Regulation on the system of non-destructive testing of rails and the operation of means of rail defectoscopy in the track facilities of railways of JSC Russian Railways (approved by order of JSC Russian Railways dated July 26, 2017 No. 1471 / r).

4. Markov A.A., Kuznetsova E.A. Defectoscopy of rails. The formation and analysis of signals. Book 2. Explanation of defectograms. St. Petersburg: Ultra Print. 2014.

5. Markov A.A., Kuznetsova E.A. Defectoscopy of rails. The formation and analysis of signals. Book 1. Basics. - SPb .: KultInform-Press, 2010.290 s.

6. Regulation on the interpretation of the results of NK rails (dec. JSC Russian Railways dated 09.01.2018 No. TsDI-1 / p). Changes to the Regulation (dated May 29, 2018 No. TsDI-558 / p).

7. RU 2438903.

8. RU 2446971.

9. RU 2521095.

Claims (1)

A method for predicting the development of anomalies in the rail head, consisting in the fact that defectoscopic tools are moved along the rail, probing the rail head with them, evaluating the received signals, detecting anomalies and registering them with reference to the coordinates of the rail track, characterized in that they form an integral the parameter of each anomaly, during subsequent movements of flaw detectors along the rail track, repeat soundings, compare the integral parameters of current and earlier found anomalies, evaluate the dynamics of changes in the integral parameter of each anomaly, predict the prospects for its development and plan repair measures.

Images