CN116304954B - Mileage alignment method and system for high-frequency sampling data of high-speed railway dynamic inspection vehicle - Google Patents
Mileage alignment method and system for high-frequency sampling data of high-speed railway dynamic inspection vehicle Download PDFInfo
- Publication number
- CN116304954B CN116304954B CN202310505256.7A CN202310505256A CN116304954B CN 116304954 B CN116304954 B CN 116304954B CN 202310505256 A CN202310505256 A CN 202310505256A CN 116304954 B CN116304954 B CN 116304954B
- Authority
- CN
- China
- Prior art keywords
- data
- mileage
- window
- wheel
- speed
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000005070 sampling Methods 0.000 title claims abstract description 59
- 238000000034 method Methods 0.000 title claims abstract description 48
- 238000007689 inspection Methods 0.000 title claims description 6
- 230000001133 acceleration Effects 0.000 claims abstract description 115
- 238000012937 correction Methods 0.000 claims abstract description 68
- 239000011159 matrix material Substances 0.000 claims description 30
- 238000004364 calculation method Methods 0.000 claims description 15
- 230000008859 change Effects 0.000 claims description 13
- 238000011156 evaluation Methods 0.000 claims description 12
- 238000012545 processing Methods 0.000 claims description 11
- 230000011218 segmentation Effects 0.000 claims description 9
- 238000013075 data extraction Methods 0.000 claims description 2
- 238000004458 analytical method Methods 0.000 abstract description 4
- 238000010586 diagram Methods 0.000 description 11
- 238000012986 modification Methods 0.000 description 10
- 230000004048 modification Effects 0.000 description 10
- 230000008569 process Effects 0.000 description 9
- 230000006870 function Effects 0.000 description 7
- 238000001514 detection method Methods 0.000 description 6
- 238000004422 calculation algorithm Methods 0.000 description 4
- 230000006872 improvement Effects 0.000 description 3
- 238000005299 abrasion Methods 0.000 description 2
- 238000004590 computer program Methods 0.000 description 2
- 230000008602 contraction Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- NAWXUBYGYWOOIX-SFHVURJKSA-N (2s)-2-[[4-[2-(2,4-diaminoquinazolin-6-yl)ethyl]benzoyl]amino]-4-methylidenepentanedioic acid Chemical compound C1=CC2=NC(N)=NC(N)=C2C=C1CCC1=CC=C(C(=O)N[C@@H](CC(=C)C(O)=O)C(O)=O)C=C1 NAWXUBYGYWOOIX-SFHVURJKSA-N 0.000 description 1
- 241000579895 Chlorostilbon Species 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 230000004931 aggregating effect Effects 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 230000002457 bidirectional effect Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000010976 emerald Substances 0.000 description 1
- 229910052876 emerald Inorganic materials 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 239000010977 jade Substances 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- ZLIBICFPKPWGIZ-UHFFFAOYSA-N pyrimethanil Chemical compound CC1=CC(C)=NC(NC=2C=CC=CC=2)=N1 ZLIBICFPKPWGIZ-UHFFFAOYSA-N 0.000 description 1
- 239000010979 ruby Substances 0.000 description 1
- 229910001750 ruby Inorganic materials 0.000 description 1
- 238000003466 welding Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
Abstract
The invention provides a mileage alignment method and system for high-frequency sampling data of a high-speed railway motor car, which belongs to the technical field of mileage alignment of rail traffic data, wherein the method can be realized through the system and comprises the following steps: s1, extracting speed data in axle box acceleration data and speed data in wheel track force data; s2, carrying out long unit mileage error correction on the axle box acceleration data and the wheel track force data; s3, short unit mileage error correction is carried out on the axle box acceleration data and the wheel track force data. The invention can greatly improve the accuracy of analysis of the high-frequency sampling data of the motor car.
Description
Technical Field
The invention relates to the technical field of mileage alignment of rail transit data, in particular to a mileage alignment method and system of high-frequency sampling data of a high-speed railway motor car.
Background
In railway transportation, the geometry of the track is subject to common defects during railway operation. The deviation between the theoretical design value and the practical measurement value of the track geometry is generally used as an index (called track irregularity), and the operation condition of the railway track is evaluated based on the evaluation method of the track geometry, such as a peak value deduction method and a TQI method, so that the driving safety is ensured. At present, two methods based on track geometry cannot reflect the short wave irregularity type (such as micron-sized wave abrasion) with extremely small amplitude, but the two types of data of acceleration of the axle box part of the train and acting force between wheel tracks can well reflect the short wave irregularity of the lower part of the track due to the characteristic of high-frequency sampling, and the defect of the existing evaluation method in short wave irregularity detection is overcome. To comprehensively evaluate the track state by using multi-source track detection data (axle box acceleration, wheel track force, track irregularity, etc.), the method needs to be established on the basis of mileage information with accurate alignment between the data.
The method mainly comprises the steps that relative mileage errors commonly exist among multi-source orbit detection data, namely, the positions of corresponding measuring points among detection data are different, and the mileage positions among detection data are not corresponding to each other, so that the acquired data at present mainly depend on a global satellite navigation system (Global Navigation Satellite System, GNSS) to calibrate the positioning information of the odometer at a radio frequency tag at a specific position, and in the driving process, due to various influencing factors such as relative sliding, wheel abrasion, wheel axle grating encoder faults, acquisition system position differences, manual debugging faults and the like among wheel tracks, the mileage errors caused by the fact that the acquired data of the two systems are integrally misplaced and distributed unevenly locally are finally caused. The existence of the mileage error greatly increases the difficulty of exploring the mapping relation between two high-frequency sampling data, and also forms an obstacle for utilizing the multi-source data to comprehensively evaluate the track state, thereby obviously reducing the accurate evaluation of the track quality state and increasing the labor intensity of workers and the maintenance cost.
Because the mileage errors are difficult to be eliminated directly from the acquisition equipment level through multiparty cooperation, the mileage errors among the track detection data are usually solved by using a mileage correction algorithm at present. However, most of the methods are only suitable for low-frequency sampling data of the same type with high waveform repeatability, the partial section-by-section mileage correction algorithm needs to combine curve characteristic information carried by the data with a line ledger to correct the data, and operation efficiency and accuracy are difficult to maintain under huge data volume, so that mileage correction between high-frequency sampling data cannot be dealt with.
Disclosure of Invention
The invention provides a mileage alignment method and a mileage alignment system for high-frequency sampling data of a high-speed railway motor car, which utilize window length change as a break for reducing calculation amount and improving algorithm accuracy, utilize speed information to replace curve characteristics to establish connection between axle box acceleration and wheel track force data in a first stage, primarily correct waveform dislocation between the two data, greatly reduce calculation amount and greatly improve calculation efficiency and accuracy of correction in a second stage; in the second stage, the global data is subjected to short unit segmentation processing, so that the problem of waveform distortion caused by processing a large amount of data is avoided, the authenticity of the data is maintained, and meanwhile, the internal mileage residual error between the two data is eliminated.
The first aspect of the embodiments of the present specification discloses a mileage alignment method for high-frequency sampling data of a high-speed railway motor car, which includes:
s1, extracting speed data V in axle box acceleration data a And velocity data V in wheel-rail force data f ；
S2, correcting long unit mileage errors of the axle box acceleration data and the wheel track force data;
s21, based on the speed data V f Extracting a speed change interval, and taking the speed change interval as a reference window;
s22, based on the window length of the reference window, performing data V on the speed a Performing rectangular window segmentation processing to obtain the speed data V a Is a rectangular window of (2);
s23, for the speed data V a Is subjected to window data downsampling to obtain the speed data V a Is a collection of windows of (1);
s24, based on the speed data V a Performing search domain waveform similarity calculation to obtain a waveform similarity set;
s25, determining the center mileage of the optimal matching window based on the waveform similarity set;
s26, calculating mileage errors according to the center mileage of the reference window and the optimal matching window, and recalibrating the mileage of the axle box acceleration data based on the mileage errors, wherein the mileage of the wheel track force data is kept unchanged.
In some embodiments, the mileage alignment method of the high-frequency sampling data of the high-speed railway motor car further comprises:
s3, short unit mileage error correction is carried out on the axle box acceleration data and the wheel track force data;
s31, downsampling the axle box acceleration data to obtain acceleration sampling points;
s32, performing rectangular window segmentation processing on the axle box acceleration data and the wheel rail force data to respectively obtain a set W of rectangular windows of the axle box acceleration data a And a set W of rectangular windows of the wheel-rail force data f ；
S33, based on the set W f Performing search domain judgment, and determining the search domain;
s34, calculating a traversal similarity evaluation matrix based on the acceleration sampling points and the search domain to obtain a pearson correlation coefficient matrix;
s35, determining an optimal waveform matching matrix based on the Pearson correlation coefficient matrix;
s36, based on the optimal waveform matching matrix, matching the set W a Interpolation and expansion are carried out, and axle box acceleration data after mileage error correction is obtained.
In some embodiments, in S2, the speed variation interval is extracted using a quartile method.
In some embodiments, in S23, the speed data V is determined based on the length of the reference window a Is downsampled for the window data.
In some embodiments, in S24, a mileage section with the center mileage of the reference window as the center and the extreme point difference between the peaks or valleys 3 times before and after the mileage point is extracted from the axle box acceleration data is used as a search field, and the speed data V in the search field is obtained a And calculating the similarity of the window relative to the speed waveform in the reference window to obtain a waveform similarity set.
In some embodiments, in S25, the window where the pearson correlation coefficient maximum value position is located in the waveform similarity set is regarded as the optimal matching window, and the center mileage of the optimal matching window is determined.
In some embodiments, in S32, the natural multiple of this frequency is a unit length, and rectangular window segmentation processing is performed on the axle box acceleration data and the wheel rail force data.
In some embodiments, in S34, the pearson correlation coefficient is used to calculate the similarity of the rectangular window of the axle box acceleration data relative to the velocity waveform in the rectangular window of the wheel track force data in the search domain, determine the similarity sequence of all the rectangular windows of the axle box acceleration data and the rectangular window of the wheel track force data in the search domain, and repeat the above until the wheel track force data is traversed, so as to obtain the pearson correlation coefficient matrix of the wheel track force data.
In some embodiments, in S35, each row in the pearson correlation coefficient matrix represents all calculation results in the search domain determined by a rectangular window of the wheel-rail force data, a position of a rectangular window of the rectangular window of each wheel-rail force data, which is optimally matched with the axle-box acceleration data, in the search domain is determined by using a maximum value index function, a residual mileage error matrix under optimal matching is calculated by a center mileage of the rectangular window of the wheel-rail force data and a center mileage of the rectangular window of the axle-box acceleration data, and an optimal waveform matching matrix is determined by tracing back an optimal matching window path matrix of the axle-box acceleration data.
The second aspect of the embodiments of the present specification discloses a mileage alignment system for high-frequency sampling data of a high-speed railway motor train unit, which is used for executing the mileage alignment method for high-frequency sampling data of the high-speed railway motor train unit;
the mileage alignment system of the high-frequency sampling data of the high-speed railway dynamic inspection vehicle comprises:
the data extraction module is used for extracting the speed data V in the axle box acceleration data a And velocity data V in wheel-rail force data f ；
The long unit mileage error correction module is used for correcting the long unit mileage error of the axle box acceleration data and the wheel track force data;
and the short unit mileage error correction module is used for carrying out short unit mileage error correction on the axle box acceleration data and the wheel track force data.
In summary, the invention has at least the following advantages:
the invention utilizes the speed data to replace curve characteristic information to establish the connection between high-frequency sampling data, eliminates long mileage offset caused by different acquisition systems by identifying the speed change interval, effectively improves the calculation efficiency, avoids waveform distortion in the correction process, and simultaneously maintains the authenticity of correction data. Compared with the existing correction method, the method has higher accuracy and practicability, reduces the required correction time, and lays a foundation for exploring the association relation between data and comprehensively evaluating the track state by utilizing multi-source data.
In the first stage, the speed data in the axle box acceleration and wheel track force two sets of acquisition systems are used as mileage position references, a long unit section for extracting speed change is defined and extracted through a quartile threshold, and the integral mileage error between two high-frequency sampling data is eliminated based on a long unit speed window. In the second stage, the wheel rail force data is used as a reference, the axle box acceleration data is segmented in short units, the axle box acceleration data in each short unit is corrected according to a cross-correlation matching method, and finally the expansion and splicing processing is carried out on each section in the axle box acceleration waveform through linear interpolation, so that the residual mileage errors are eliminated section by section. The invention realizes the mileage alignment of the high-frequency sampling data of the motor car under the conditions of noise, drift and internal waveform extension, has high mileage correction efficiency and high accuracy, maintains the authenticity of the original data, and greatly improves the accuracy of the analysis of the high-frequency sampling data of the motor car.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a schematic diagram of steps of a mileage alignment method of high-frequency sampling data of a high-speed railway motor car according to the present invention.
Fig. 2 is a schematic flow chart of the long cell error correction and the short cell error correction according to the present invention.
Fig. 3 is a schematic diagram of linear interpolation correction of mileage errors according to the present invention.
Figure 4a is a schematic diagram of the acceleration system according to the present invention collecting velocity data.
Fig. 4b is a schematic diagram of the wheel and rail force system acquisition speed data according to the present invention.
Fig. 5a is a schematic diagram of a velocity image before correction of a long unit according to the present invention.
Fig. 5b is a schematic diagram of a velocity image after correction of a long cell according to the present invention.
Fig. 6a is a schematic diagram of the long unit corrected front axle box acceleration, wheel rail force image according to the present invention.
Fig. 6b is a schematic diagram of the axle box acceleration and wheel rail force images after correction of the long unit according to the present invention.
Fig. 7a is a schematic diagram of an image of the axle box acceleration and wheel rail force before the short cell correction according to the present invention.
Fig. 7b is a schematic diagram of the axle box acceleration and wheel rail force image after the short cell correction according to the present invention.
Fig. 8 is a schematic diagram of energy trend evaluation comparison of high-frequency sampled data before and after correction according to the present invention.
Detailed Description
Hereinafter, only certain exemplary embodiments are briefly described. As will be recognized by those of skill in the pertinent art, the described embodiments may be modified in numerous different ways without departing from the spirit or scope of the embodiments of the present invention. Accordingly, the drawings and description are to be regarded as illustrative in nature and not as restrictive.
The following disclosure provides many different implementations, or examples, for implementing different configurations of embodiments of the invention. In order to simplify the disclosure of embodiments of the present invention, components and arrangements of specific examples are described below. Of course, they are merely examples and are not intended to limit embodiments of the present invention. Furthermore, embodiments of the present invention may repeat reference numerals and/or letters in the various examples, which are for the purpose of brevity and clarity, and which do not themselves indicate the relationship between the various embodiments and/or arrangements discussed.
Embodiments of the present invention will be described in detail below with reference to the accompanying drawings.
As shown in fig. 1, the embodiment provides a mileage alignment method for high-frequency sampling data of a high-speed railway motor car, which specifically includes the following steps:
step 1, extracting speed data in axle box acceleration data (hereinafter referred to as acceleration data) and wheel-rail force data
And extracting a speed change interval by using a quartile method, and correcting long unit mileage errors on the axle box acceleration data and the wheel track force data. The quartile threshold definition method defines an inner limit and an outer limit by using upper and lower quartiles and a quartile difference, and considers outliers with small probability outside the inner limit as outliers. The upper and lower quartiles (data at 75% and 25% positions after the overall data is arranged from small to large) are respectively taken as Q3 and Q1; IQR is a four-bit difference, which is the difference between Q3 and Q1; the inner limit range is (Q1-1.5 IQR, Q3+1.5 IQR) and the probability of the data point being outside the inner limit is 0.7%. The speed image of axle box acceleration data and wheel track force data at the mileage change section and the corresponding quartile range limit are shown in fig. 2 (flow chart of long unit error correction and short unit error correction). In FIG. 2, l f 、l a Velocity data V, respectively wheel-rail force data f Speed data V of axle box acceleration data a The resulting closed window win in the quartile method f 、win a Number of sampling points in, l crest Is the mileage difference between the peaks or the troughs at the same position of two groups of speed data, L search Is based on the wheel-rail force window win f Center mileage Mile f L crest Defined search field, mile a Matching windows win for acceleration ai Center distance, delta a,f Long unit mileage errors for speed data.
Step 2, long unit mileage error correction
2.1 reference Window selection
Will V f Data change intervals, i.e. speed data V of wheel-rail force data f Cutting interval win with wheel rail force quartering limit value f As a reference window, the window length is l f 。
2.2 speed data Window segmentation of axlebox acceleration data
By window length l a For V a And carrying out rectangular window segmentation processing on the data. To ensure better repeatability between waveforms and accuracy of similarity evaluation, the window-to-window spacing should not be too great, with a step size of about one thousandth of the window length being recommended.
2.3V based on reference Window Length a Window data downsampling
Due to the difference of sampling frequency a Far greater than l f Therefore, before similarity calculation is performed on the two waveforms, the acceleration window obtained by cutting is divided according to the length l of the reference window f Downsampling to obtain V a Window set W of data a In the formula (1), n is V a Number of data windows, L search To window win according to wheel-rail force f Determined search area length, s a For the step size between adjacent windows.
2.4 search Domain waveform similarity calculation
Center mileage Mile with reference window f Centering on the velocity data V of acceleration a Extraction of the Chinese medicine Mileage point Mile f Extreme point difference l between front and back 3 times wave crest or wave trough crest Is taken as a searching field L search Acceleration window win for whole search domain ai Similarity s of the set of (c) relative to the velocity waveform within the wheel-rail force reference window a,f Calculating to obtain a waveform similarity set S a,f The expression (2), the expression (3), the expression (4) and the expression (5) are shown in the specification, wherein r (x, y) is a pearson correlation coefficient function.
L search ＝(Mile f -3l crest ,Mile f +3l crest ) (2)
s ai,f ＝r(win ai ,win f ) (4)
S a,f ＝{s ai,f |i＝1,2,...,n} (5)
2.5 determining the center mileage of the optimal matching Window
Aggregating waveform similarity S a,f The window where the maximum value position of the mesopearson correlation coefficient is located is regarded as the optimal matching window, and the mileage Mile of the central position of the matching window is obtained a As in formula (6).
Mile a ＝Mile(S a,fmax ) (6)
2.6 wave form dislocation mileage recalibration
Calculating mileage error delta according to center position mileage of reference window and matching window a,f Accordingly, recalibrating the acceleration data mileage of the axle box according to the formula (7) and the formula (8), wherein the wheel track force data mileage is kept unchanged, and Mile ai For each mileage data corresponding to axle box acceleration data,and (5) correcting mileage data for the long unit.
Δ a,f ＝Mile f -Mile a (7)
Step 3, short unit mileage error correction method
After the mileage calibration correction of the long unit, the whole dislocation of the original data waveform is eliminated, but the long unit correction does not solve the problem of the local uneven distribution of the data waveform. Therefore, the short unit mileage error in the high-frequency sampled axlebox acceleration, wheel-rail force data is corrected as follows.
3.1 acceleration data downsampling and Window division based on the wheel Rail force reference data sampling frequency
When similarity calculation is performed on two waveforms, the number of sampling points of two data in the window length is required to be ensured to be consistent, so that the sampling frequency f of the reference data is used s Downsampling the acceleration data to obtain an acceleration sampling point N a The sampling point of the reference data is N f Is unchanged.
3.2 acceleration, wheel Rail force data Window segmentation
To preserve the time domain characteristics as far as possible, dividing rectangular windows for axle box acceleration and wheel-rail force data by taking natural multiple n of the frequency as unit length, and corresponding to the length l=l a ＝l f =n×fs, wherein the step size s between the reference data window and the window f Acceleration data as matching data inter-window step size =lRectangular window win for obtaining axle box acceleration and wheel rail force ai 、win fj Set W of (2) a 、W f As shown in the formula (9) and the formula (10), m is V f Number of data windows.
3.3 search Domain determination
Search field L in short unit mileage error evaluation search Is longer than the text residual mileage error matrix delta A,fj Medium maximum value delta a,fjmax Is 2 times as large as the above. With each reference window win fj Center position mileage Mile of (2) fj Centered, search field L is set according to equation (11) search And judging. In N a,start ，N a,end Respectively an axle box acceleration data start point and an axle box acceleration data end point, N a,Milef For the corresponding axlebox acceleration data point at the current reference window (wheel rail force) center mileage location,is search field L search Half of the area corresponds to the mileage.
3.4 calculating a traversal similarity evaluation matrix based on the wheel-track force data window
From win f1 Initially, the set of acceleration windows W is applied to the entire search domain using the Pearson's correlation coefficient a′ Window win against wheel rail force f1 Similarity s of internal velocity waveforms ai′,f1 Calculating to obtain win f1 All acceleration rectangular windows and wheel-rail force reference windows win in determined search domain f1 Repeating the above steps until win fj Traversing the wheel-rail force data to finally obtain a pearson correlation coefficient matrix S of the whole wheel-rail force data ai′,fj . Formula (12), formula (13) and formula (14), wherein r (x, y) is represented by formula (3) and is a pearson correlation coefficient function, win ai′ Is win fj An i' th acceleration rectangular window within the determined search field.
s ai,fj ＝r(win ai ,win fj ) (13)
3.5 determining an optimal waveform matching matrix
Matrix of correlation coefficients S ai′,fj Each row of the wheel track force window represents all calculation results in a search domain determined by a wheel track force rectangular window, the position of an optimal matching acceleration window in the search domain of each wheel track force window can be determined by utilizing a maximum value index function, and a residual mileage error matrix delta under optimal matching is calculated through the center mileage of the wheel track force window and the center mileage of the optimal matching acceleration window A,fj Optimal matching window path matrix for simultaneously backtracking axle box acceleration dataAs shown in the formula (15) and the formula (16), wherein +.>The start and end path lines in the overall acceleration data are respectively acceleration windows that best match the jth wheel-rail force window.
Short unit waveform expansion correction method of 3.6 times of interpolation
The purpose of the mileage correction is to obtain real and continuous acceleration data, but from the correction path matrix N a It can be found that the acceleration window path that best matches the wheel-rail force data is not continuous if N aj，Ed ≠N aj+1，Bg It is explained that there is an overlapping or missing portion of the waveforms between the adjacent acceleration windows as shown in table 1.
TABLE 1 waveform overlap and miss Table
And (3) interpolating and stretching the acceleration data window, wherein a linear interpolation function is shown in a formula (17). In which x is ori Is the abscissa of the original data curve, y ori As the function value of the original data curve, x new Interpolation point abscissa, y of data curve new I.e. x new Lower data curve function value.
Will j-th reference window F fj Inner data path N fj ＝{N fh |(j-1)×l f ≤h<j×l f Setting the path as a reference path according to the optimal matching window path matrix N a Information extraction optimal axle box acceleration data matching window path N of jth row of medium fj ＝{N fh |(j-1)×l f ≤h<j×N aj,Ed Extracting the next best matching window path N by combining the j+1st line information aj+1 ＝{N ah |N aj+1,Bg ≤h<N aj+1,Ed According to N } aj And N aj+1 The waveform expansion and contraction state is determined, and the determination and linear interpolation correction process is shown in fig. 3.
The method for judging the adjacent windows of the acceleration data comprises the following steps:
st1, there is no missing or overlapping data segment between the matching windows. E.g. matching window A f1 And A is a f2 Satisfies the requirements therebetweenAt this time, window A is matched f2 Can be directly used as a correction window A' f2 With the start window A' f1 And (5) splicing as shown in a formula (18).
A′ f2 ＝A f2 (18)
St2, there is a data segment missing case between the matching windows. E.g. matching window A f2 And A is a f3 The relation between is thatAt this time, acceleration sampling point +.>The data of the segment is missing, the missing part should be included in the matching window A f3 In the process, the reference window length l is again calculated f For->(matching Window A) f3 Adding a missing segment)Is subjected to linear interpolation, and the whole data is compressed into a length l f Is a correction window A 'of (2)' f3 And correction window A' f2 And (3) splicing, as shown in a formula (19).
St3, there is a case where data sections overlap between the matching windows. E.g. matching window A f3 And A is a f4 The relation between is thatAt this time, acceleration sampling point +.>The data of the sections are overlapped, and the overlapped part is firstly selected from a matching window A f4 Delete, and then according to the length l of the reference window f For->(matching Window A) f4 Removing overlapping sections), and stretching the rest data to length l f Is a correction window A 'of (2)' f4 And correction window A' f3 And (5) splicing as shown in a formula (20).
Repeating the above process until all the matching windows are converted into correction windows, and then splicing to form axle box acceleration data after mileage error correction.
In one embodiment:
the axle box acceleration and wheel rail force data are shown in tables 2 and 3. It can be found that each acceleration and track force data acquired by the two sets of acquisition systems has corresponding speed data.
TABLE 2 left axle box vertical acceleration raw data
| Sampling point | Mileage (km) | Speed (km/h) | Acceleration (m/ss) |
| 1 | 80.000 | 238.000 | -3.120 |
| 2 | ··· | ··· | ··· |
| ··· | ··· | ··· | ··· |
| n-1 | ··· | ··· | ··· |
| n | 140.000 | 230.000 | -1.140 |
Table 3 left rail drop force raw data
| Sampling point | Mileage (km) | Speed (km/h) | Wheel rail force (kN) |
| 1 | 80.000 | 245.727 | 59.710 |
| 2 | ··· | ··· | ··· |
| ··· | ··· | ··· | ··· |
| n-1 | ··· | ··· | ··· |
| n | 140.000 | 215.755 | 59.828 |
In the first stage, the speed data is used to establish the relationship between axle box acceleration and wheel track force data. To separate the speed fine fluctuation from the speed change section, the original data is divided with Q1-2IQR as a boundary, a section outside the Q1-2IQR value range is regarded as the speed change section, and the closed area of the speed change is extracted as shown in fig. 4a and 4 b.
After extracting the closed interval velocity data, long unit mileage error correction is performed according to equations (1) - (8), and the step length between adjacent windows is set to l a According to the peak spacing l crest And determining the range of the search domain as a 3km mileage section before and after the reference window. The resulting velocity data waveform before long cell correction is shown in fig. 5a, where the overall misalignment between the red and black velocity waveforms indicates the overall data drift between the two types of high frequency sampled data. The long unit mileage error average value obtained in the speed closing section 1 and the speed closing section 2 is delta respectively a,f1 ＝-0.922km,Δ a,f2 = -0.992km, using delta a,f1 The acceleration data mileage is recalibrated, the speed data before and after the long unit correction is carried out are shown in fig. 5b, and the axle box acceleration and wheel track force K90+500-K93+500 section long unit correction data before and after the long unit correction are shown in fig. 6a and 6 b.
By delta a,f1 And starting short unit mileage correction of the calibrated axle box acceleration data and wheel track force data. The original axle box acceleration and wheel rail force data are sections of a certain bidirectional eastern freight special line K80+000-K140+000, and after correction by a long unit, the axle box acceleration data mileage is recalibrated to be K79+078-K139+078. And extracting overlapping interval sections K80+000-K139+078 of the corrected acceleration and wheel track force data as basic data of accurate mileage correction.
The two data used for mutual pairing in the proposed model are left rail vertical force data with 2000Hz as sampling frequency and left axle box vertical acceleration data with 5000Hz as sampling frequency, and in order to ensure the authenticity of corrected data, left rail vertical force with lower sampling frequency is used as reference data. The method comprises the steps of firstly carrying out band-pass filtering on two data by using an FIR filter, removing low-frequency information such as vehicle axle weight, wheel track force data fluctuation trend caused by a curve and high-frequency part above 1000Hz, and ensuring the consistency of the frequency characteristics of the two data.
In order to ensure the consistency of the two data time domains, 0.4 times of equidistant interpolation downsampling processing is carried out on the acceleration data of the axle box and the corresponding mileage, and the two preprocessed data can be regarded as sampling at 2000 Hz. Then dividing axle box acceleration and wheel track force rectangular windows by taking 5 sampling periods, namely 10000 sampling points in 5s as unit length, taking the wheel track force data as reference data, and the step length s between adjacent windows f =10000; acceleration data is used as data to be matched, and step length s between adjacent windows a ＝10。
Rectangular window set W of axle box acceleration and wheel rail force obtained a 、W f The short unit mileage error evaluation and correction is carried out by bringing in (11) - (20) because of the mileage error maximum value delta a,fjmax About 42m, thus defining the unilateral length of the search field as 100m.
The interpolation correction window length is consistent with the mileage error evaluation window length, and 10000 sampling points in a 5s section are taken for piecewise linear interpolation correction. Optimal matching window path matrix N formed by statistics results of mileage errors of all sections a As the interpolation basis. In order to ensure the calculation efficiency, the window length is reasonably selected, 2-5 welding seam impact signals can be considered as auxiliary positioning points in one window length, the recommended window length at the sampling frequency of 2000Hz at the speed of 240km/h is 4000-20000 sampling points, the calculation time required by the window length of 4000 is about 6 times of that required by the window length of 20000, and the waveform precision can be improved by about 5% compared with the waveform precision of 20000. The interpolation process follows the greedy algorithm idea, and in the correction process, the current matching window is always considered to be the correct correction result, so that the overall optimal correction data can be obtained after the expansion and contraction of all the local optimal windows.
The corrected axle box acceleration data and the wheel track force data have the same sampling frequency and mileage section, and can share mileage data. The detailed effects of the section K91+500-K92+500 acceleration and the track force data before and after correction are shown in FIGS. 7a and 7 b.
After the long unit correction in the first stage, the relative mileage error between any two data sections is reduced from 1km to within 42 m; and then after accurate correction by a short unit, the mean value of the mileage errors of each section is 0.32m, and the error of the high-frequency sampling data of any section of the line can be controlled within the range of [ -0.92m and 1.55m ] under the confidence of 99.7%.
And (3) carrying out data standardization on the axle box acceleration and the wheel rail force according to a formula (21), and eliminating the influence of data dimension, wherein x is data, mu is data mean value and sigma is data standard deviation. Then 200 sampling points within 0.1S are used as an analysis unit, the length is l, and root mean square value S in all data analysis units is calculated step by step according to the formula (22) i In which x' j Is a normalized acceleration or wheel-rail force data sampling point. After the root mean square calculation was completed, the root mean square energy trend of 50 analysis units per five seconds was extracted for the two data, and the similarity evaluation was performed, and the result is shown in fig. 8. The linear correlation coefficient of the overall energy trend of the two data is improved from-0.079 to 0.65, and the accurate correction data represented by the red curve is better than the original data almost in any global section, which shows that the correction model improves the overall correlation between the two data.
The linear correlation coefficient of the overall energy trend of the two data is improved from-0.079 to 0.65, and the accurate correction data represented by the red curve is better than the original data almost in any global section, which shows that the correction model improves the overall correlation between the two data.
The above embodiments are provided to illustrate the present invention and not to limit the present invention, so that the modification of the exemplary values or the replacement of equivalent elements should still fall within the scope of the present invention.
From the foregoing detailed description, it will be apparent to those skilled in the art that the present invention can be practiced without these specific details, and that the present invention meets the requirements of the patent statutes.
While preferred embodiments of the present invention have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. It is therefore intended that the following claims be interpreted as including the preferred embodiments and all such alterations and modifications as fall within the scope of the invention. The foregoing description of the preferred embodiment of the invention is not intended to be limiting, but rather to cover all modifications, equivalents, and alternatives falling within the spirit and principles of the invention.
It should be noted that the above description of the flow is only for the purpose of illustration and description, and does not limit the application scope of the present specification. Various modifications and changes to the flow may be made by those skilled in the art under the guidance of this specification. However, such modifications and variations are still within the scope of the present description.
While the basic concepts have been described above, it will be apparent to those of ordinary skill in the art after reading this application that the above disclosure is by way of example only and is not limiting of the present application. Although not explicitly described herein, various modifications, improvements, and adaptations of the present application are possible for those of ordinary skill in the art. Such modifications, improvements, and modifications are intended to be suggested within this application, and are therefore within the spirit and scope of the exemplary embodiments of this application.
Meanwhile, the present application uses specific words to describe embodiments of the present application. For example, "one embodiment," "an embodiment," and/or "some embodiments" means a particular feature, structure, or characteristic in connection with at least one embodiment of the present application. Thus, it should be emphasized and should be appreciated that two or more references to "an embodiment" or "one embodiment" or "an alternative embodiment" in various positions in this specification are not necessarily referring to the same embodiment. Furthermore, certain features, structures, or characteristics of one or more embodiments of the present application may be combined as suitable.
Furthermore, those of ordinary skill in the art will appreciate that aspects of the invention may be illustrated and described in terms of several patentable categories or circumstances, including any novel and useful processes, machines, products, or materials, or any novel and useful improvements thereof. Thus, aspects of the present application may be implemented entirely in hardware, entirely in software (including firmware, resident software, micro-code, etc.) or a combination of hardware and software. The above hardware or software may be referred to as a "unit," module, "or" system. Furthermore, aspects of the present application may take the form of a computer program product embodied in one or more computer-readable media, wherein the computer-readable program code is embodied therein.
Computer program code required for operation of portions of the present application may be written in any one or more programming languages, including an object oriented programming language such as Java, scala, smalltalk, eiffel, JADE, emerald, C ++, c#, vb.net, python, etc., a conventional programming language such as C programming language, visualBasic, fortran2103, perl, COBOL2102, PHP, ABAP, a dynamic programming language such as Python, ruby, and Groovy, or other programming languages, etc. The program code may execute entirely on the user's computer, or as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the latter scenario, the remote computer may be connected to the user's computer through any form of network, such as a Local Area Network (LAN) or a Wide Area Network (WAN), or the connection may be made to an external computer (for example, through the Internet), or the use of services such as software as a service (SaaS) in a cloud computing environment.
Furthermore, the order in which the elements and sequences are presented, the use of numerical letters, or other designations are used in the application and are not intended to limit the order in which the processes and methods of the application are performed unless explicitly recited in the claims. While certain presently useful inventive embodiments have been discussed in the foregoing disclosure, by way of various examples, it is to be understood that such details are merely illustrative and that the appended claims are not limited to the disclosed embodiments, but, on the contrary, are intended to cover all modifications and equivalent arrangements included within the spirit and scope of the embodiments of the present application. For example, while the implementation of the various components described above may be embodied in a hardware device, it may also be implemented as a purely software solution, e.g., an installation on an existing server or mobile device.
Likewise, it should be noted that in order to simplify the presentation disclosed herein and thereby aid in understanding one or more inventive embodiments, various features are sometimes grouped together in a single embodiment, figure, or description thereof for the purpose of streamlining the disclosure. This method of disclosure, however, is not to be interpreted as reflecting an intention that the claimed subject matter requires more features than are expressly recited in each claim. Rather, the inventive subject matter should be provided with fewer features than the single embodiments described above.
Claims (8)
1. The mileage alignment method of the high-frequency sampling data of the high-speed railway dynamic inspection vehicle is characterized by comprising the following steps of:
s1, extracting speed data V in axle box acceleration data a And velocity data V in wheel-rail force data f ；
S2, correcting long unit mileage errors of the axle box acceleration data and the wheel track force data;
s21, based on the speed data V f Extracting a speed change section, namely speed data V of wheel track force data f Cutting interval win with wheel rail force quartering limit value f As a reference window;
s22, based on the window length of the reference window, performing data V on the speed a Performing rectangular window segmentation processing to obtain the speed data V a Is a rectangular window of (2);
s23, for the speed dataV a Is subjected to window data downsampling to obtain the speed data V a Is a collection of windows of (1);
s24, based on the speed data V a Performing search domain waveform similarity calculation to obtain a waveform similarity set;
s24, taking the center mileage of the reference window as the center, extracting mileage sections of which the distance points are 3 times of the peak or trough extreme point difference between the front and rear of the mileage points from the axle box acceleration data as a search domain, and carrying out data V on the speed in the search domain a Calculating the similarity of the window relative to the speed waveform in the reference window to obtain a waveform similarity set;
s25, determining the center mileage of the optimal matching window based on the waveform similarity set;
s25, identifying a window in which the maximum value position of the pearson correlation coefficient in the waveform similarity set is located as an optimal matching window, and determining the center mileage of the optimal matching window;
s26, calculating mileage errors according to the center mileage of the reference window and the optimal matching window, and recalibrating the mileage of the axle box acceleration data based on the mileage errors, wherein the mileage of the wheel track force data is kept unchanged.
2. The mileage alignment method of high-frequency sampling data of a high-speed railway motor car according to claim 1, further comprising:
s3, short unit mileage error correction is carried out on the axle box acceleration data and the wheel track force data;
s31, downsampling the axle box acceleration data to obtain acceleration sampling points;
s32, performing rectangular window segmentation processing on the axle box acceleration data and the wheel rail force data to respectively obtain a set W of rectangular windows of the axle box acceleration data a And a set W of rectangular windows of the wheel-rail force data f ；
S33, based on the set W f Performing search domain judgment, and determining the search domain;
s34, calculating a traversal similarity evaluation matrix based on the acceleration sampling points and the search domain to obtain a pearson correlation coefficient matrix;
s35, determining an optimal waveform matching matrix based on the Pearson correlation coefficient matrix;
s36, based on the optimal waveform matching matrix, matching the set W a Interpolation and expansion are carried out, and axle box acceleration data after mileage error correction is obtained.
3. The mileage alignment method of high-frequency sampling data of high-speed railway motor car according to claim 1, wherein in S2, the speed change section is extracted by adopting a quartile method.
4. The mileage alignment method for high-speed railway motor car high-frequency sampling data according to claim 1, wherein in S23, the speed data V is compared based on the length of the reference window a Is downsampled for the window data.
5. The mileage alignment method of high-speed railway motor car high-frequency sampling data according to claim 2, wherein in S32, the natural multiple of the frequency is the unit length, and rectangular window division processing is performed on the axle box acceleration data and the wheel track force data.
6. The mileage alignment method of high-frequency sampling data of high-speed railway motor train unit inspection according to claim 2, wherein in S34, pearson correlation coefficients are used to calculate the similarity of rectangular windows of axle box acceleration data relative to velocity waveforms in rectangular windows of wheel-rail force data in the search domain, and the similarity sequence of rectangular windows of all axle box acceleration data and rectangular windows of wheel-rail force data in the search domain is determined until the wheel-rail force data is traversed, so as to obtain pearson correlation coefficient matrix of the wheel-rail force data.
7. The mileage alignment method of high-frequency sampling data of high-speed railway motor train unit according to claim 2, wherein in S35, each row in the pearson correlation coefficient matrix represents all calculation results in the search domain determined by a rectangular window of the wheel-rail force data, a position of a rectangular window of the best matching axle-box acceleration data in the search domain of the rectangular window of each wheel-rail force data is determined by using a maximum value index function, a residual mileage error matrix under the best matching is calculated by a center mileage of the rectangular window of the wheel-rail force data and a center mileage of the rectangular window of the best matching axle-box acceleration data, and a best matching window path matrix of the axle-box acceleration data is traced back at the same time, so as to determine a best waveform matching matrix.
8. A mileage alignment system for high-frequency sampling data of a high-speed railway motor car, characterized by being used for executing the mileage alignment method for high-frequency sampling data of a high-speed railway motor car according to any one of claims 2 to 7;
the mileage alignment system of the high-frequency sampling data of the high-speed railway dynamic inspection vehicle comprises:
the data extraction module is used for extracting the speed data V in the axle box acceleration data a And velocity data V in wheel-rail force data f ；
The long unit mileage error correction module is used for correcting the long unit mileage error of the axle box acceleration data and the wheel track force data;
and the short unit mileage error correction module is used for carrying out short unit mileage error correction on the axle box acceleration data and the wheel track force data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310505256.7A CN116304954B (en) | 2023-05-08 | 2023-05-08 | Mileage alignment method and system for high-frequency sampling data of high-speed railway dynamic inspection vehicle |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310505256.7A CN116304954B (en) | 2023-05-08 | 2023-05-08 | Mileage alignment method and system for high-frequency sampling data of high-speed railway dynamic inspection vehicle |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116304954A CN116304954A (en) | 2023-06-23 |
| CN116304954B true CN116304954B (en) | 2023-07-28 |
Family
ID=86788956
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310505256.7A Active CN116304954B (en) | 2023-05-08 | 2023-05-08 | Mileage alignment method and system for high-frequency sampling data of high-speed railway dynamic inspection vehicle |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116304954B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104711912A (en) * | 2013-12-11 | 2015-06-17 | 中航西安飞行自动控制技术有限公司 | High speed railway mileage benchmarking method |
| CN108595374A (en) * | 2018-03-26 | 2018-09-28 | 中国铁道科学研究院 | High speed railway track geometry minor change recognition methods and device |
| CN109188385A (en) * | 2018-08-31 | 2019-01-11 | 西安电子科技大学 | High speed weak object detection method under clutter background |
| WO2022036284A1 (en) * | 2020-08-13 | 2022-02-17 | Invensense, Inc. | Method and system for positioning using optical sensor and motion sensors |
| CN114987564A (en) * | 2022-06-16 | 2022-09-02 | 西南交通大学 | Portable high-speed turnout detection trolley based on binocular recognition and detection method |
| CN115390062A (en) * | 2022-08-25 | 2022-11-25 | 中国人民解放军国防科技大学 | High-speed target migration compensation method, device and equipment under air resistance |
-
2023
- 2023-05-08 CN CN202310505256.7A patent/CN116304954B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104711912A (en) * | 2013-12-11 | 2015-06-17 | 中航西安飞行自动控制技术有限公司 | High speed railway mileage benchmarking method |
| CN108595374A (en) * | 2018-03-26 | 2018-09-28 | 中国铁道科学研究院 | High speed railway track geometry minor change recognition methods and device |
| CN109188385A (en) * | 2018-08-31 | 2019-01-11 | 西安电子科技大学 | High speed weak object detection method under clutter background |
| WO2022036284A1 (en) * | 2020-08-13 | 2022-02-17 | Invensense, Inc. | Method and system for positioning using optical sensor and motion sensors |
| CN114987564A (en) * | 2022-06-16 | 2022-09-02 | 西南交通大学 | Portable high-speed turnout detection trolley based on binocular recognition and detection method |
| CN115390062A (en) * | 2022-08-25 | 2022-11-25 | 中国人民解放军国防科技大学 | High-speed target migration compensation method, device and equipment under air resistance |
Non-Patent Citations (5)
| Title |
|---|
| On-board detection of longitudinal track irregularity via axle box acceleration in HSR;Sun X等;《 IEEE Access》;第9卷;14025-14037 * |
| Xiao J L等.Analysis on the track quality evolution law of polyurethane-reinforced ballasted track in high-speed railway.《Proceedings of the Institution of Mechanical Engineers, Part F: Journal of Rail and Rapid Transit,》.2020,第235卷(第8期), 993-1005. * |
| 基于动静态检测数据的轨道弹性状态评估及平顺性调整方法;杨飞等;《铁道学报》;1-9 * |
| 基于多次波形匹配的高速铁路动检数据里程误差评估与修正;汪鑫等;《铁道学报》;第42卷(第2期);110-116 * |
| 高速铁路动检车检测数据里程误差评估与修正;汪鑫等;《铁道标准设计》;第62卷(第7期);46-51 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116304954A (en) | 2023-06-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wei et al. | Evaluating degradation at railway crossings using axle box acceleration measurements | |
| Mori et al. | Condition monitoring of railway track using in-service vehicle | |
| Tsai et al. | Railway track inspection based on the vibration response to a scheduled train and the Hilbert–Huang transform | |
| CN108595374B (en) | Method and device for identifying geometric tiny changes of high-speed railway track | |
| CN107314776B (en) | Method for mileage calibration based on railway track detection data | |
| Khosravi et al. | Reducing the positional errors of railway track geometry measurements using alignment methods: A comparative case study | |
| CN108592853A (en) | Track plates arch upward position identifying method, device, storage medium and equipment | |
| CN113705412B (en) | Track state detection method and device for multi-source data fusion based on deep learning | |
| Balouchi et al. | Development of railway track condition monitoring from multi-train in-service vehicles | |
| Tsunashima et al. | Japanese railway condition monitoring of tracks using in-service vehicle | |
| DE102012219111A1 (en) | Method for localization of rail vehicle in rail network, involves matching corresponding period of reference signal of route section of rail network with selected portion of actual time signal | |
| Tanaka et al. | Practical application of portable trolley for the continuous measurement of rail surface roughness for rail corrugation maintenance | |
| CN113281414B (en) | Method and device for identifying short-wave disease types of steel rails and electronic equipment | |
| CN116304954B (en) | Mileage alignment method and system for high-frequency sampling data of high-speed railway dynamic inspection vehicle | |
| WO2017055838A1 (en) | Method and system for predicting railway track quality | |
| Carnevale et al. | An algorithm for precise localization of measurements in rolling stock-based diagnostic systems | |
| CN117119588B (en) | Wi-Fi6 technology-based method for realizing positioning of vehicle in track parking lot | |
| Men et al. | Health assessment of high-speed train wheels based on group-profile data | |
| Wang et al. | Development of metro track geometry fault diagnosis convolutional neural network model based on car-body vibration data | |
| Xu et al. | A novel description method for track irregularity evolution | |
| Bragança et al. | Drive-by Methodologies Applied to Railway Infrastructure Subsystems: A Literature Review—Part II: Track and Vehicle | |
| Alten et al. | Detecting and classifying rail corrugation based on axle bearing vibration | |
| CN113343919A (en) | Method and device for detecting continuous equidistant rubbing damage of steel rail and computer equipment | |
| Bhardwaj et al. | Railroad track irregularities position accuracy assessments using low-cost sensors on a hi-rail vehicle | |
| Xu et al. | A novel short-range prediction model for railway track irregularity |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |