Mileage Positioning Method for High-Speed ​​Railway Track Inspection Data Based on Multi-Resolution Singular Value Decomposition

Abstract

The present invention provides a method for locating the mileage of high-speed railway track inspection data based on multi-resolution singular value decomposition, comprising: collecting height or track direction signals of a first preset length of track tread as a data sample to be detected; specifying the depth of multi-resolution singular value decomposition, and performing multi-resolution singular value decomposition on the data sample to be detected; obtaining the singular values ​​of the weld and extracting the weld characteristics; locating the weld mileage in combination with historical data and equipment inventory, locating the weld mileage using weld characteristics, and locating the track inspection data mileage in sections. The present invention uses the singularity of the track tread to determine the weld characteristics, and then determines the track inspection data mileage in sections. The mileage positioning has low cost and high accuracy. In particular, after track operation, it can still provide a stable mileage benchmark for the precise positioning of track inspection data, which is conducive to promoting the realization of comprehensive management of the dynamic and static performance of high-speed railway tracks.

Description

Technical Field

[0001] The present invention relates to the field of track detection technology, and in particular to a high-speed railway track inspection data mileage positioning method based on multi-resolution singular value decomposition. Background Art

[0002] Repetitive loads and uneven variations in road foundations can lead to track irregularities, compromising driving comfort and safety. Therefore, the "High-Speed ​​Railway Line Maintenance Rules" (TG / GW115-2023) mandate that high-speed railways conduct track inspections based on the principle of "primarily dynamic inspections, combining dynamic and static inspections." Through dynamic and static inspections, engineering departments collect a vast amount of track irregularity data. However, the precise location of track irregularities remains an unresolved issue, directly impacting the application of this data.

[0003] Conventional positioning methods based on encoders or GNSS can have positioning errors of up to 10 0 ~10 2 m, and cannot provide stable and accurate mileage data. To achieve accurate mileage positioning of track inspection data, relevant patents such as CN202311619288.6, CN201210262250.3, CN202210648194.0, CN202210648355.6, CN202310686345.6, CN202211037409.1, CN201810845096.X, CN202110043468.9, CN201910949473.9, and CN202010514649.0 have developed mileage positioning technology or equipment for track inspection data through tread geometry, axle box acceleration, or radio frequency identification. However, positioning methods based on tread geometry or axlebox acceleration all presuppose similarity between data, such as dynamic time warping or time-delay cross-correlation. Track operations can disrupt this similarity, directly impacting the accuracy of these mileage positioning technologies. Radio frequency identification-based positioning methods are less sensitive to track operations, but their accuracy is limited and require equipment installation.

[0004] High-speed railway track structures are typically seamless, spanning sections. Under normal conditions, a weld, also known as a welded joint, is present every 100 or 500 meters. Clearly, effectively detecting welds would provide a relatively stable and sufficiently dense mileage benchmark for track inspection data, even when waveform similarity is compromised. However, the signal from high-speed railway welds is weak (straightness ≤ 0.3 mm), and processing them using methods such as wavelet transform, spectral analysis, empirical mode decomposition, and variational mode decomposition can easily result in smoothing or nonlinear phase distortion, making weld features difficult to detect and mileage location difficult to determine from the detected signal. Related researchers have proposed a method for correcting the mileage error of track irregularity data based on rail weld extraction. This method extracts weld features through dynamic standard deviation and sliding average. However, the calculation of dynamic standard deviation has super-parameters and cannot maintain phase, making it difficult to accurately locate weld mileage. Related researchers have also used chord length combination and singular value decomposition (SVD) to obtain the sliding peak state of singular features and then locate rail tread damage. However, this decomposition result and the original signal belong to the same vector space, and its detailed features are not stable in a noisy environment. Therefore, it is necessary to find a method that can accurately detect weld features and achieve precise positioning of track inspection data when waveform similarity is destroyed. Summary of the Invention

[0005] In view of the above situation, it is necessary to provide a high-speed railway track inspection data mileage positioning method based on multi-resolution singular value decomposition when the waveform similarity is destroyed.

[0006] A high-speed railway track inspection data mileage positioning method based on multi-resolution singular value decomposition, comprising:

[0007] Collecting a height or track direction signal of a track tread of a first preset length as a data sample to be detected;

[0008] Specifying the depth of multi-resolution singular value decomposition, and performing multi-resolution singular value decomposition on the data sample to be detected;

[0009] Obtain the singular values ​​of the weld and extract the weld features;

[0010] Combine historical data and equipment records to locate weld mileage, use weld features to locate weld mileage, and locate track inspection data mileage in sections.

[0011] Furthermore, collecting the height or track direction signal of the track tread of the first preset length as a data sample to be detected specifically includes:

[0012] Collecting track inspection data of a first preset length N;

[0013] Select short chord height or track direction signal from multi-channel track inspection data as the data sample to be detected X=(x1,x2,…x i ,…,x N ), where x1, x2, x i 、x N They are the short chord height or track direction data at mileage indexes 1, 2, i, and N respectively.

[0014] Furthermore, the depth of the multi-resolution singular value decomposition is specified, and the multi-resolution singular value decomposition is performed on the data sample to be detected, specifically including:

[0015] Specify the multi-resolution singular value decomposition depth J;

[0016] Perform multi-resolution singular value decomposition on the data sample X to be detected to obtain J layers of detail signals and 1 layer of approximate signal A J ,Right now Among them, D j Represents the j-th layer detail signal.

[0017] Furthermore, the singular values ​​of the weld are obtained and the weld features are extracted, specifically including:

[0018] Select the j-th layer detail signal D j =(d j,1 ,d j,2 ,…d j,k ,…,d j,N ), where d j,1 d j,2 d j,k d j,N They are the j-th layer detail signal D j Detailed component data at mileage indices 1, 2, k, and N;

[0019] From the j-th layer detail signal D j Filter out m outliers P={p j,1 ,p j,2 ,…,p j,m}, where p j,1 、p j,2 、p j,m Represent the 1st, 2nd, and mth outliers respectively, m≤N;

[0020] The outlier P is considered as the weld feature.

[0021] Furthermore, the outlier value P is calculated and determined based on statistical features or density features.

[0022] Furthermore, the height or track direction signal of the track tread is collected by a level 0 track inspection instrument or a track inspection vehicle.

[0023] According to the high-speed railway track inspection data mileage positioning method based on multi-resolution singular value decomposition provided by the present invention, the singularity of the track tread is used to determine the weld characteristics, and then the track inspection data mileage is determined in sections. The mileage positioning has low cost and high accuracy. In particular, after track operation, it can still provide a stable mileage benchmark for the precise positioning of track inspection data, which is conducive to promoting the realization of comprehensive management of the dynamic and static performance of high-speed railway tracks. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 This is a flow chart of a method for high-speed railway track inspection data mileage positioning based on multi-resolution singular value decomposition in a first embodiment of the present invention;

[0025] Figure 2 Schematic diagram of the binary recursive SVD decomposition process of the signal in the first embodiment of the present invention;

[0026] Figure 3 Schematic diagram of the left 1.25m high and low chord measurement values ​​and their various order singular characteristic components of the track inspection instrument in the first embodiment of the present invention;

[0027] Figure 4 Schematic diagram of the singular characteristic components of each order of the left 1.25m high and low chords of the track inspection instrument in the second embodiment of the present invention. DETAILED DESCRIPTION

[0028] To facilitate understanding of the present invention, the present invention will be described more fully below with reference to various embodiments. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, the purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive.

[0029] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this invention pertains. The terms used in this specification of the present invention are for the purpose of describing specific embodiments only and are not intended to limit the present invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0030] See also Figure 1 The first embodiment of the present invention provides a method for mileage location of high-speed railway track inspection data based on multi-resolution singular value decomposition. This method, based on track weld identification, can provide a stable mileage benchmark for track inspection data even when data similarity is destroyed, thereby achieving accurate location of the track inspection data. This method for mileage location of high-speed railway track inspection data based on multi-resolution singular value decomposition includes steps S11 to S14.

[0031] Step S11 : collecting height or track direction signals of a track tread of a first preset length as data samples to be detected.

[0032] Track tread height or track direction signals can be collected using existing detection equipment, including but not limited to level 0 track inspection instruments, track measuring instruments, and track inspection vehicles. Track inspection is a long-term and tedious task. To ensure the accuracy and efficiency of the inspection, the track to be inspected can be inspected in sections in this embodiment. In this embodiment, a first preset length of track tread signals is collected as the data sample to be inspected. Specifically, considering the high-speed railway track structure, there will generally be one factory-welded joint every 100m and one on-site welded joint every 500m. To ensure that the track inspection data contains weld information, the first preset length should be ≥100m. Based on the sampling step of 0.25m for dynamic inspection and 0.125m for static inspection, the first preset length N of dynamic and static track inspection data should be ≥400 and 800 respectively. For example, if the first preset length is 500m, the first preset length N of the data sample is 4000.

[0033] Dynamic and static track inspection data include height and low or track direction signals of different chord lengths, such as 1.25m, 10m, 20m, 60m, and 70m chord lengths. Since the weld characteristics of height and low or track direction signals are more obvious at shorter chord lengths, the height and low or track direction signals of the 1.25m chord can be selected as the data sample to be tested X = (x1, x2, ... x i ,…,x N ), where x1, x2, x i 、x N They are the short chord height or track direction data at mileage indexes 1, 2, i, and N respectively.

[0034] Step S12: specifying the depth of multi-resolution singular value decomposition, and performing multi-resolution singular value decomposition on the data sample to be detected.

[0035] As a relatively stable structural feature of high-speed railway tracks, welds can provide a relatively stable and sufficiently dense mileage benchmark for track inspection data when waveform similarity is destroyed. However, the signal of high-speed railway welds is weak (straightness ≤ 0.3mm), so the key to accurately positioning track inspection data after track operation lies in the selection of characteristic vectors that characterize the welds. A singular signal is a signal in which the signal itself or one of its derivatives has a sudden change at a certain moment. If the signal f(t) has a sudden change at a certain point or a sudden change in one of its derivatives, the signal is said to be singular at that point. Obviously, the track inspection data near the weld has a sudden change, and singular value decomposition can be applied to extract weld features.

[0036] Singular value decomposition is a decomposition algorithm based on mathematical morphology. The basic idea is that any m×n matrix H can be transformed into the product of three matrices after singular value decomposition:

[0037] H=USV T

[0038] Among them: U and V are m×m and n×n unitary matrices respectively; S is an m×n diagonal matrix, whose values ​​are arranged from large to small on the diagonal, and are called singular values ​​of matrix H; and H is mostly constructed in the form of a Hankel matrix.

[0039] SVD has a wide range of applications in data compression, feature extraction, fault diagnosis, signal denoising and other fields. The key to applying SVD to signal processing is how to construct a suitable matrix using the signal. Different matrix construction methods will result in different signal processing effects. In order to show the overview and detailed features of different levels of the signal, in 2001, KAKARALA R et al. (KAKARALA R, OGUNBONA P. Signal analysis using a multiresolution form of the singular value decomposition [J]. IEEE Transactions on Image Processing, 2001, 10 (5): 724-735.) constructed multiresolution singular value decomposition (Multiresolution SVD); in 2010, Zhao Xuezhi et al. (Zhao Xuezhi, Ye Bangyan, Chen Tongjian. Multiresolution singular value decomposition theory and its application in signal processing and fault diagnosis [J]. Journal of Mechanical Engineering, 2010, 46 (20): 64-75.) constructed a bisection recursive multiresolution singular value decomposition. Research shows that the binary recursive multi-resolution singular value decomposition can accurately detect the location of singular points in the signal and has excellent noise reduction capabilities, and can be used for weak fault feature extraction.

[0040] For details, please refer to Figure 2 By specifying the multi-resolution singular value decomposition depth J, the data sample X to be detected is subjected to multi-resolution singular value decomposition to obtain J-layer detail signals and 1-layer approximate signal A J ,Right now Among them, D j Represents the j-th layer detail signal.

[0041] For specific implementation, please refer to Figure 3 , is a schematic diagram of the left 1.25m high and low chord measurement values ​​and their various order singular characteristic components of the track inspection instrument in the first embodiment of the present invention. Decomposition depth J = 3, then the data sample X to be detected is decomposed into {D1, D2, D3} ∪ A3. Figure 3 It can be seen that the detail signals D of each order j It has a significant 100m periodicity characteristic and no phase shift at the peak.

[0042] Step S13: obtaining the singular values ​​of the weld and extracting weld features.

[0043] Specifically, select the j-th layer detail signal D j =(d j,1 ,d j,2 ,…d j,k ,…,d j,N ), where d j,1 d j,2 d j,k d j,N They are the j-th layer detail signal D j Detailed component data at mileage indices 1, 2, k, and N;

[0044] From the j-th layer detail signal D j Filter out m outliers P={p j,1 ,p j,2 ,…,p j,m}, where p j,1 、p j,2 、p j,m Represent the 1st, 2nd, and mth outliers respectively, m≤N;

[0045] The outlier P is considered as the weld feature.

[0046] For specific implementation, please refer to Figure 3 , select the third layer detail signal D3 = (d 3,1 ,d 3,2 ,…,d 3,N ); According to the equipment inventory, there are welds near mileage 1475.615km, 1475.712km, and 1475.812km; find the peak from the detail signal D3 according to certain rules, where d 3,1023 =0.00376mm, d 3,1801 =0.00246mm, d 3,2600 =0.00341mm; the outlier set P is constructed from the above three points as the weld feature.

[0047] Step S14: locate the weld mileage in combination with historical data and equipment inventory, and locate the track inspection data mileage in sections.

[0048] For specific implementation, please refer to Figure 3 Combining historical data and equipment records, according to the index value of the outlier, it is determined that the mileage of each weld in the data sample X to be tested is 1475.61512km, 1475.71237km and 1475.81225km.

[0049] The embodiment of the present invention utilizes the singularity of the track tread to determine the weld characteristics, and then determines the track inspection data mileage in sections. The mileage positioning has low cost and high accuracy. In particular, after track operation, it can still provide a stable mileage benchmark for the precise positioning of track inspection data, which is conducive to promoting the realization of comprehensive management of the dynamic and static performance of high-speed railway tracks.

[0050] A second embodiment of the present invention provides a method for mileage positioning of high-speed railway track inspection data based on multi-resolution singular value decomposition, comprising steps S21 to S24.

[0051] Step S21 : collecting height or track direction signals of a track tread of a first preset length as data samples to be detected.

[0052] In this embodiment, the track tread signal of the first preset length is collected as the data sample to be detected. Specifically, the first preset length is 200m. According to the sampling step of 0.125m for static inspection, the first preset length of the track inspection data N=1600. And the 1.25m chord height or track direction signal is selected as the data sample to be detected X=(x1, x2,…, x N ).

[0053] Step S22 , specifying the depth of multi-resolution singular value decomposition, and performing multi-resolution singular value decomposition on the data sample to be detected.

[0054] For specific implementation, please refer to Figure 4 , is a schematic diagram of the singular characteristic components of each order of the left 1.25m high and low chord of the track inspection device in the second embodiment of the present invention. Decomposition depth J = 20, then the data sample X to be detected is decomposed into {D1, ..., D 18 ,D 19 ,D 20}∪A 20 .Depend on Figure 4 It can be seen that the detail signals D of each order j It has a significant 100m periodicity characteristic and no phase shift at the peak.

[0055] Step S23: obtaining the singular values ​​of the weld and extracting weld features.

[0056] Specifically, select the 20th layer detail signal D 20 =(d 20,1 ,d 20,2 ,…,d 20,N ); from the 20th layer detail signal D 20 Filter outliers P {(318.59996, 0.00331), (318.69958, 0.00330)} that are close to the mileage of the weld in the original equipment inventory; use the outlier P as the weld feature.

[0057] Step S24, locate the weld mileage in combination with historical data and equipment inventory, and locate the track inspection data mileage in sections.

[0058] For specific implementation, please refer to Figure 4 Combining historical data and equipment records, according to the index value of the outlier, it is determined that the mileage of each weld in the data sample X to be tested is 318.59996 km and 318.69958 km.

[0059] It can be understood that, in order to enhance weld features, in other embodiments of the present invention, outliers may be determined based on statistical features or density features.

[0060] In summary, the high-speed railway track inspection data mileage positioning method based on multi-resolution singular value decomposition provided by the present invention utilizes the singularity of the track tread to determine the weld characteristics, and then determines the track inspection data mileage in sections. The mileage positioning has low cost and high accuracy. In particular, after track operation, it can still provide a stable mileage benchmark for the precise positioning of track inspection data, which is conducive to promoting the realization of comprehensive management of the dynamic and static performance of high-speed railway tracks.

[0061] Throughout this specification, reference to terms such as "one embodiment," "some embodiments," "examples," "specific examples," or "some examples" means that a specific feature, structure, material, or characteristic described in conjunction with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0062] The above-described embodiments merely illustrate several implementations of the present invention, and while their descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art would be able to make numerous variations and improvements without departing from the spirit of the present invention, all of which fall within the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be determined by the appended claims.

Claims

1. A high-speed railway track inspection data mileage positioning method based on multi-resolution singular value decomposition, characterized by: include: Collecting a height or track direction signal of a track tread of a first preset length as a data sample to be detected; Specifying the depth of multi-resolution singular value decomposition, and performing multi-resolution singular value decomposition on the data sample to be detected; Obtain the singular values ​​of the weld and extract the weld features; Combine historical data and equipment records to locate weld mileage, use weld features to locate weld mileage, and locate track inspection data mileage in sections; Collecting the height or track direction signal of the track tread of the first preset length as a data sample to be detected, specifically including: Collecting track inspection data of a first preset length N; Select short chord height or track direction signal from multi-channel track inspection data as the data sample to be detected X=(x1,x2,…x i ,…,x N ), where x1, x2, x i 、x N They are the short chord height or track direction data at mileage indexes 1, 2, i, and N respectively; Specifying the depth of multi-resolution singular value decomposition and performing multi-resolution singular value decomposition on the data sample to be detected, specifically including: Specify the multi-resolution singular value decomposition depth J; Perform multi-resolution singular value decomposition on the data sample X to be detected to obtain J layers of detail signals and 1 layer of approximate signal A J ,Right now Among them, D j Represents the j-th layer detail signal; Obtain the singular values ​​of the weld and extract the weld features, including: Select the j-th layer detail signal D j =(d j,1 ,d j,2 ,…d j,k ,…,d j,N ), where d j,1 d j,2 d j,k d j,N They are the j-th layer detail signal D j Detailed component data at mileage indices 1, 2, k, and N; From the j-th layer detail signal D j Filter out m outliers P={p j,1 ,p j,2 ,…,p j,m }, where p j,1 、p j,2 、p j,m Represent the 1st, 2nd, and mth outliers respectively, m≤N; The outlier P is used as the weld feature; The outlier value P is calculated and determined based on statistical features or density features.

2. The high-speed railway track inspection data mileage positioning method based on multi-resolution singular value decomposition according to claim 1 is characterized in that: The height or track direction signal of the track tread is collected by a level 0 track inspection instrument or a track inspection vehicle.

Images