A method for real-time estimation of train speed

Abstract

The application discloses a kind of train speed real-time estimation method, comprising the following steps: S1, when train operation, the acceleration signal of each axle box is collected;S2, with moving window intercept acceleration signal and band-pass filter is carried out to the acceleration signal of interception;S3, by generalized cross-correlation method, the correlation analysis of filtered acceleration signal of interception is carried out to obtain the time delay between acceleration signal of interception and correlation coefficient;S4, the constant longitudinal distance L between sensor and the time delay t sft The measured speed v of train operation is obtained i ;S5, using correlation coefficient, the multiple speed measurement values obtained are weighted average to obtain the estimated train speed v est The train running speed can be accurately and real-time estimated by the application, and the recognition accuracy and reliability are improved by using redundant design, and manpower and material resources are saved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of rail transit, in particular to a train speed real-time estimation method. BACKGROUND

[0002] In recent years, the rail transit industry in China has developed rapidly towards high speed and heavy load, which has brought great challenges to train operation safety and service operation and maintenance. As one of the core indicators of train operation, obtaining accurate train speed has practical application significance for train operation organization, anti-skid control, etc. Accurate acquisition of train running speed is the basis for train moving block, overspeed protection and other operation safety protection technologies. In addition, train absolute speed is also introduced in wheel slip detection, and if the train absolute speed is inaccurate, the wheel will slip and slide, which will seriously affect the service life of the wheel and the running quality of the train.

[0003] In the existing train speed estimation method, radar, speed sensor and ground equipment are mainly used for speed measurement. The inventors have found in practical use that the prior art at least has the following technical problems:

[0004] 1. Laying the balise on the ground for speed measurement requires a large amount of material and cost, and cannot realize continuous dynamic speed measurement.

[0005] 2. When the train is running, the traction and braking force acts on the rail through the creep force between the wheel and the rail. In the case of large traction and braking force (especially for heavy load locomotives), the wheel creep speed is higher than the actual train running speed, and if the wheel slips, its speed cannot be used as a basis for estimating the train running speed.

[0006] 3. When using radar to estimate the train running speed, it is easily affected by factors such as positioning accuracy, tunnels and weather conditions. SUMMARY

[0007] The purpose of the present application is to overcome the shortcomings of the prior art and provide a train speed real-time estimation method, which comprises the following steps:

[0008] S1, collecting the acceleration signals of each axle box when the train is running;

[0009] S2, intercepting the acceleration signals with a moving window and band-pass filtering the intercepted acceleration signals;

[0010] S3, performing correlation analysis on the filtered intercepted acceleration signals by a generalized cross-correlation method to obtain the time delay and correlation coefficient between the intercepted acceleration signals;

[0011] S4, combining the constant longitudinal distance L between the sensors and the time delay t sftObtaining the measured speed v of the train operation i ;

[0012] S5, using the correlation coefficient to obtain the estimated train speed v of the weighted average of the obtained multiple speed measurement values est .

[0013] Further, the acceleration signal is intercepted with a moving window and the intercepted acceleration signal is band-pass filtered, including the following processes:

[0014] The axle box vertical acceleration signal of the window selection is output during the train operation, wherein the length L of the window w , L m is the length of each window movement, and as the train moves forward, the window also follows the forward update output signal; the window needs to contain two sensors when the reference speed has not been calculated, and the window length is not less than L; after the reference speed is obtained, a single window can be changed into two independent windows containing only a single sensor; the signal is analyzed for coherence to obtain the coherence degree of the signal at different frequencies, and a Butterworth filter is used to band-pass filter the acceleration signal, and the filter band is selected from the part with a coherence exceeding a set value:

[0015]

[0016] In the formula, P 12 is the cross-correlation power spectrum of the input signal, P 11 and P 22 are the autocorrelation power spectrum of the input signal.

[0017] Further, the time delay and the correlation coefficient between the intercepted acceleration signals are obtained by performing correlation analysis on the filtered intercepted acceleration signals through the generalized cross-correlation method, using the following formula:

[0018]

[0019]

[0020]

[0021] In the formula, a i is the input acceleration signal; G 12 (w) is the convolution of the input signal; A(w) is the path weight of the generalized cross-correlation method; R 12 is the generalized cross-correlation function of the input signal.

[0022] Further, the time delay value is the time corresponding to the maximum value of the signal correlation function, using the following formula:

[0023] t sft= arg max R 12 (τ)

[0024] Further, the train running speed is obtained through the relationship between the time delay between the two sensor signals and their longitudinal distance, and the following formula is used:

[0025]

[0026] In the formula, v ref is the reference speed estimated at the last time, and when the reference speed has not been estimated at the initial time, the calculation formula is as follows:

[0027]

[0028] Further, the estimated train speed v est is obtained by using the correlation coefficient to weight average the obtained multiple speed measurement values, and the following formula is used:

[0029]

[0030] In the formula, t sft is the time delay value; f i is the speed estimation weighting value, v i is the speed value estimated by a single group of data, and n is the number of estimated speeds.

[0031] Further, the speed estimation weighting coefficient is the maximum value of the coherence function, and the following formula is used:

[0032] f = max (C ij )

[0033] In the formula, C ij is the coherence function of the two input signals, and f is the weighting value used for speed estimation using this group of signals.

[0034] The beneficial effects of the present application are: the present application is not affected by weather conditions, is not limited by line conditions, does not need to install ground equipment, can accurately and timely estimate the train running speed, and uses redundant design to improve the identification accuracy and reliability, and saves manpower and resources. BRIEF DESCRIPTION OF DRAWINGS

[0035] Figure 1 It is a flowchart of a train speed real-time estimation method;

[0036] Figure 2 It is an algorithm flowchart;

[0037] Figure 3 It is a signal window selection diagram;

[0038] Figure 4is an acceleration signal coherence analysis schematic diagram;

[0039] Figure 5 is a schematic diagram of a plurality of vehicle speed weighted average estimation processes;

[0040] Figure 6 is a recognition contrast curve schematic diagram of the train running speed under the uniform speed working condition in the embodiment of the application;

[0041] Figure 7 is a recognition error curve schematic diagram of the train running speed under the uniform speed working condition in the embodiment of the application;

[0042] Figure 8 is a recognition contrast curve schematic diagram of the train running speed under the variable speed working condition in the embodiment of the application;

[0043] Figure 9 is a recognition error curve schematic diagram of the train running speed under the variable speed working condition in the embodiment of the application. DETAILED DESCRIPTION

[0044] The technical solutions of the application will be further described in detail below with reference to the accompanying drawings, but the protection scope of the application is not limited to the following description.

[0045] In order to make the purpose, technical solutions and advantages of the application clearer, the application will be further described in detail in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the application, and are not used to limit the application, that is, the described embodiments are only a part of the embodiments of the application, but not all the embodiments. The components of the embodiments of the application described and shown in the drawings can be arranged and designed in various different configurations.

[0046] Therefore, the detailed description of the embodiments of the application provided in the drawings below is not intended to limit the scope of the claimed application, but only represents selected embodiments of the application. Based on the embodiments of the application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the application. It should be noted that the relationship terms such as "first" and "second" and the like are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between the entities or operations.

[0047] Moreover, the term "comprising" or "including" or any other variant thereof is intended to cover a non-exclusive inclusion, such that processes, methods, articles, or apparatuses that comprise a list of elements are not required to comprise only those elements in the list, but can include other elements not expressly listed, or inherent to such processes, methods, articles, or apparatuses. In other words, any process, method, article, or apparatus that comprises a list of elements is deemed to include not only those elements in the list, but also other elements not expressly listed, or inherent to such process, method, article, or apparatus.

[0048] The features and characteristics of the present application will be further understood from the following detailed description of the embodiments.

[0049] As shown in Figure 1 , a method for real-time estimation of train speed comprises the following steps:

[0050] S1, collecting acceleration signals of each axle box when the train is running;

[0051] S2, intercepting the acceleration signals with a moving window and band-pass filtering the intercepted acceleration signals;

[0052] S3, performing correlation analysis on the filtered intercepted acceleration signals by a generalized cross-correlation method to obtain time delay and correlation coefficient between the intercepted acceleration signals;

[0053] S4, combining the constant longitudinal distance L between sensors and the time delay t delay to obtain the measured speed v i of the train running;

[0054] S5, performing weighted average on the obtained multiple speed measurements by using the correlation coefficient to obtain the estimated train speed v est .

[0055] The algorithm flow chart is shown in Figure 2 .

[0056] The acceleration signals are intercepted with a moving window and the intercepted acceleration signals are band-pass filtered, including the following processes:

[0057] The axle box vertical acceleration signals selected by the window are outputted when the train is running, wherein the length of the window is L w , L mThe length of the window is the length of each movement, and the window is updated forwardly with the train moving forwardly. The window needs to contain two sensors when the reference speed is not calculated, and the length of the window is not less than L; after the reference speed is obtained, the single window can be changed into two independent windows containing single sensors. The coherence of the signal at different frequencies is obtained by performing coherence analysis on the signal, and the acceleration signal is band-pass filtered by using a Butterworth filter, and the filter band is selected from the part with the coherence exceeding a set value.

[0058]

[0059] In the formula, P 12 is the cross-correlation power spectrum of the input signal, P 11 is the autocorrelation power spectrum of the input signal. 22

[0060] The time delay and the correlation coefficient between the intercepted acceleration signals are obtained by performing correlation analysis on the filtered intercepted acceleration signals by using the generalized cross-correlation method, and the following formula is used:

[0061]

[0062]

[0063]

[0064] In the formula, a i is the input acceleration signal; G 12 (w) is the convolution of the input signal; A(w) is the path weight of the generalized cross-correlation method; R 12 is the generalized cross-correlation function of the input signal.

[0065] The correlation coefficient is used to perform weighted average on the obtained multiple speed measurement values to obtain the estimated train speed v est , and the following formula is used:

[0066] t sft = arg max R 12 (τ)

[0067]

[0068] In the formula, t sft is the time delay value; f i is the speed estimation weighted value, v i is the speed value estimated by a single group of data, and n is the number of estimated speeds.

[0069] Specifically, the present application comprises the following steps:

[0070] ​S1, lay acceleration sensors on the train running part; measure the acceleration signal of each axle box when the train is running;

[0071] S2, intercept the signal in the form of moving window and band-pass filter the signal data;

[0072] S3, obtain the time delay and correlation coefficient between signals by correlation analysis of the signal through the generalized cross-correlation method;

[0073] S4, combine the constant longitudinal distance L between sensors and the time delay t sft Backwardly obtain the speed v of the train running i ;

[0074] S5, use the correlation coefficient to obtain the weighted average of the multiple speed measurement values to obtain the estimated train speed v est .

[0075] The axle box vertical acceleration signal of the window selected when the train is running is output in real time, the window form is as shown in Figure 3 , wherein L w is the length of the window, and L m is the length of the window each time. As the train moves forward, the window also follows the output data update. The output acceleration time domain signal is as shown in Figure 4 (a), the coherence analysis of the signal is obtained, the coherence degree of the signal at different frequencies is as shown in Figure 4 (b), the band-pass filter is used to filter the acceleration signal, and the filter band is selected to be the part with the coherence exceeding 0.5. The generalized cross-correlation calculation is performed on the filtered signal, and the calculation method is as shown below.

[0076]

[0077]

[0078]

[0079] In the formula, a i is the input acceleration signal; G 12 (w) is the convolution of the input signal; A(w) is the path weight of the generalized cross-correlation method; R 12 is the generalized cross-correlation function of the input signal.

[0080] The correlation calculation result is as shown in Figure 5 . In the figure, the maximum of the correlation peak value is the time delay value, and the corresponding peak value (correlation coefficient) is used as the weight to obtain the weighted average of the multiple speed values.

[0081] t sft = arg max R 12(τ) (4)

[0082]

[0083] wherein t sft is the time delay value; f i is the speed estimation weighting value, v i is the estimated speed value of a single set of data, and n is the number of estimated speeds.

[0084] Figures 6-7 is the comparison and error between the identified speed and the actual speed of the train when the train is running at an initial speed of 40 km / h. It can be seen that when running at the initial speed, the running speed of the vehicle decreases slightly, the identified speed can reflect the slight change in the vehicle speed, and the maximum identification error is less than 0.3%; Figures 8-9 is the comparison and error between the identified speed and the actual speed of the train when the train is running at an initial speed of 40 km / h. It can be seen that when running at the initial speed, the running speed of the vehicle decreases slightly, the identified speed can reflect the slight change in the vehicle speed, and the maximum identification error is less than 0.3%;

[0085] The above description is only the preferred embodiment of the present application, and it should be understood that the present application is not limited to the form disclosed herein, and should not be considered as excluding other embodiments, but can be used in various other combinations, modifications and environments, and can be modified within the scope of the concept described herein, by the above teaching or related art or knowledge. Any modification and change made by those skilled in the art without departing from the spirit and scope of the present application shall be within the protection scope of the appended claims of the present application.

Claims

1. A method of real-time estimation of train speed, characterized in that, The method comprises the following steps: S1, collecting acceleration signals of each axle box during train operation; S2, intercepting the acceleration signals with a moving window and performing band-pass filtering on the intercepted acceleration signals; S3, performing correlation analysis on the filtered intercepted acceleration signals by a generalized cross-correlation method to obtain time delay and correlation coefficient between the intercepted acceleration signals; S4, the longitudinal distance L between the sensors and the time delay t sft obtaining a measured speed v of the train operation i ; S5, using a correlation coefficient to weight average the multiple speed measurement values to obtain an estimated train speed v est ; The step of intercepting the acceleration signals with a moving window and performing band-pass filtering on the intercepted acceleration signals comprises the following process: The axle box vertical acceleration signal of the train operation output window option, wherein the length of the window L w , L m is the length of each window movement, as the train moves forward, the window also follows the forward update output signal, the window needs to contain two sensors when the reference speed has not been calculated, the window length is not less than L; after obtaining the reference speed, the single window changes to two independent windows containing only a single sensor; the signal is analyzed for coherence to obtain the coherence degree of the signal at different frequencies, and a Butterworth filter is used to band-pass filter the acceleration signal, and the filter band is selected as the part with coherence exceeding a set value: where P 12 is the cross-correlation power spectrum of the input signals, P 11 and P 22 are the autocorrelation power spectra of the input signals.

2. The method for real-time estimation of train speed according to claim 1, wherein, The step of performing correlation analysis on the filtered intercepted acceleration signals by the generalized cross-correlation method to obtain time delay and correlation coefficient between the intercepted acceleration signals adopts the following formula: where a i is the input acceleration signal; G 12 (w) is the convolution of the input signals; A(w) is the path weight of the generalized cross-correlation method; R 12 is the generalized cross-correlation function of the input signals.

3. The method for real-time estimation of train speed according to claim 1, wherein, The time delay is the time corresponding to the maximum value of the signal correlation function. t sft = argmax R 12 (τ).

4. The method for real-time estimation of train speed according to claim 3, wherein, combining the longitudinal distance L between the sensors and the time delay t sft obtaining a measured speed v of the train operation i using the following equation: In the formula, v ref is the reference speed estimated at the previous time, and when the reference speed has not been estimated at the initial time, the following formula is used:

5. The method for real-time estimation of train speed according to claim 1, wherein, The estimated train speed v is obtained by weighted average of the plurality of speed measurement values using the correlation coefficient est using the following formula: In the formula, t sft is the time delay; f i is the velocity estimation weighting value, v i is the estimated velocity value of a single group of data, and n is the number of estimated velocities.

6. The method for real-time estimation of train speed according to claim 5, wherein, The speed estimation weighting value f i The maximum value obtained by the coherence analysis is obtained by using the following formula: f = max(C ij ) In the formula, C ij is the coherence function of the two input signals, and f is the weighting value used for velocity estimation.

Images