A detection interval identification method for subway train onboard sensors

Abstract

The present invention belongs to the field of train interval detection technology, and specifically relates to a detection interval identification method for subway train on-board sensors, and the specific steps are as follows: using sensors to collect vibration acceleration data of subway trains; using multiple moving average filtering methods to filter the longitudinal acceleration and vertical acceleration of the train collected by the sensor; adding the MMA filtering results of the longitudinal acceleration and the vertical acceleration to determine the cutting threshold; the signal corresponding to the signal less than the threshold can be judged as the stop time of the train, and the time corresponding to the signal greater than the threshold can be judged as the time when the train is running in the interval; the average value of each stop time is taken as the segmentation point of the train interval; extracting the curve information therein and comparing it with the curve ledger information of each interval; driving interval identification; adjusting the parameters of the filtering algorithm. The present invention can greatly improve the analysis efficiency of subway train on-board data and ensure the accuracy of the average filtering processing algorithm.

Description

Technical Field

[0001] The invention belongs to the technical field of train section detection, and in particular relates to a detection section identification method for a subway train onboard sensor. Background Art

[0002] Traditional subway line operation and maintenance inspection mainly relies on manual inspection, which is inefficient. At present, many studies have proved that installing sensors in carriages can reflect the service status of subway tracks. At the same time, installing sensors on subway trains can quickly detect the entire subway line within a few hours, greatly improving the operational efficiency of subway management departments.

[0003] However, it is difficult to receive GPS signals in subway environments, which makes it difficult to map the collected train vibration data to specific driving sections. Figure 1 As shown in the figure, due to the short stop time of the train, the collected acceleration data is difficult to accurately locate the specific subway driving section. At present, in most cases, we can only rely on manual section cutting and manual identification of the specific driving section corresponding to the cut section. Therefore, there is an urgent need for an identification method to improve the analysis efficiency of subway train on-board data. Summary of the invention

[0004] The purpose of the present invention is to provide a detection interval identification method for subway train on-board sensors, which can greatly improve the analysis efficiency of subway train on-board data and ensure the accuracy of the average filtering processing algorithm.

[0005] The technical solution adopted by the present invention is as follows:

[0006] A detection interval identification method for a subway train onboard sensor, the specific steps are as follows:

[0007] Step 1: Use sensors to collect vibration acceleration data of subway trains, which are used as train-mounted vibration data;

[0008] Step 2: Use the multiple moving average filtering method to filter the longitudinal acceleration and vertical acceleration of the train collected by the sensor;

[0009] Step 3: Add the MMA filtering results of the longitudinal acceleration and the vertical acceleration to determine the cutting threshold;

[0010] Step 4: less than the threshold The signal corresponding to can be judged as the train stop time, greater than The time corresponding to the signal is judged as the time when the train is running in the section;

[0011] Step 5: Take the average value of each stop time as the segmentation point of the train section;

[0012] Step 6: When the train passes through the section, it will generate a large angular velocity of the shaking head. Therefore, the curve information is further extracted from the angular velocity of the shaking head collected by the sensor and compared with the curve information of each section, so as to determine the specific section where the data segmented in Step 1-Step 5 is located;

[0013] Step 7: Driving range identification;

[0014] Step 8: Adjust the parameters of the filtering algorithm by analyzing the distance and speed changes between trains.

[0015] In step 1, the train-borne vibration data consists of 6-axis data, mainly including: lateral, longitudinal, vertical acceleration, nodding, shaking, and rolling angular velocity.

[0016] The sensor is mainly installed in the B09 box of the train.

[0017] The multiple moving average filtering method is described as follows:

[0018] The first moving average filter:

[0019]

[0020] No. Heavy moving average filtering:

[0021]

[0022] In the formula, For the The re-filtering result is at time The output when For the The re-filtering result is at time The output when is the filtering window length, is the input acceleration signal.

[0023] In step 3, the MMA filtering results of the longitudinal acceleration and the vertical acceleration are added, and the added signal is recorded as , determine the cutting threshold:

[0024] In the formula, is the maximum peak-to-peak value of signal A, ; It is a percentage value, and the value range here is ≤5%.

[0025] In step Step 6, the curve information extracted from the shaking angular velocity includes the number of curves and the direction of the curves.

[0026] In step 6, multiple moving average filtering is further performed on the shaking angular velocity, and the peaks and troughs of the filtered angular velocity are identified by using the multiple moving average filtering method;

[0027] The peak value must meet the following conditions at the same time:

[0028] The trough value must meet the following conditions at the same time:

[0029] In the formula, is the location of the peak / trough, and They are exist The first and second order derivatives at .

[0030] In step 7, the extracted shaking angle peak and trough sequence is recorded as: ,in Indicates The peak / trough information of the element, if the elements are peaks, then On the contrary, if elements are troughs, then ;

[0031] Similarly, the interval The curve recorded in the curve ledger information is converted into a sequence , if elements are left-opening / right-opening curves, then On the contrary, if elements are right-opening / left-opening curves, then ;

[0032] Finally, the sequence with sequence Match if and Exactly the same, that is , then the driving interval of the cut data is interval.

[0033] In step 8, the parameters of the filtering algorithm are adjusted by analyzing the distance and speed changes between trains. The specific steps are as follows:

[0034] S1: Obtain real-time data: It is necessary to obtain the speed and position information of the train in real time;

[0035] S2: Calculate relative distance and speed: Calculate the relative distance and speed difference between the front and rear trains based on the acquired data;

[0036] S3: Analyze dynamic changes: Analyze the impact of these dynamic changes on the filtering algorithm;

[0037] S4: Adjust filtering parameters: Based on the above analysis, dynamically adjust the parameters of the filtering algorithm;

[0038] S5: Implement adaptive control: Combined with the PID control algorithm, the adjusted filter parameters are applied to the train speed control system.

[0039] The PID control algorithm compares the actual speed with the set target speed and calculates the speed tracking error , use proportional regulation to amplify the deviation at the current moment, use integral regulation to integrate the deviations of all past times, and then use differential regulation to adjust the rate of change of the deviation. Finally, the outputs of the three links of proportional, integral and differential are weighted summed to obtain the final control amount. :

[0040]

[0041] in: is the output of the controller; is the input of the controller, i.e., the deviation between the set value and the actual value; is the proportionality coefficient; is the integration coefficient; is the differential coefficient; is the integral of the deviation; is the differential of the deviation.

[0042] The technical effects achieved by the present invention are:

[0043] The present invention firstly performs a moving average filter on the train-borne vibration data (6 axes: lateral, longitudinal, vertical acceleration, nod, shake, and roll angular velocity) collected by the sensor to achieve data smoothing; then the smoothed longitudinal acceleration and vertical acceleration data are added to calculate the maximum peak-to-peak value; finally, the segmentation threshold is determined according to the percentage of the peak value to achieve rapid cutting of the sensor data; finally, the shake angular velocity generated when the train passes through the curve is matched with the line curve ledger to identify the correct driving section. Through this automated sensor data section cutting and identification method, the analysis efficiency of subway train-borne data can be greatly improved.

[0044] The present invention can improve the safety and efficiency of train operation by acquiring train data in real time and adjusting the filtering algorithm parameters accordingly. The optimized filtering algorithm can more accurately reflect the actual state of the train and can accurately reflect complex vibration modes, especially in the case of multi-train interaction, thereby ensuring the accuracy of the average filtering processing algorithm. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 It is a linear graph of the data of vertical vibration acceleration of the train collected by the sensor;

[0046] Figure 2 is a flow chart of automatic cutting of sensor data provided by an embodiment of the present invention;

[0047] Figure 3 is a linear graph of data before vertical acceleration movement filtering provided by an embodiment of the present invention;

[0048] Figure 4 is a linear graph of vertical acceleration movement filtered data provided by an embodiment of the present invention;

[0049] Figure 5 It is a schematic diagram of determining the train stop time provided by an embodiment of the present invention;

[0050] Figure 6 is a schematic diagram of a train section cutting result provided by an embodiment of the present invention;

[0051] Figure 7 is a schematic diagram of the angular velocity generated when a train passes through a curve provided by an embodiment of the present invention;

[0052] Figure 8 It is a graph of extracting interval curve information using the shaking angle velocity of a sensor provided by an embodiment of the present invention;

[0053] Fig. 9 It is a flow chart of adjusting filtering algorithm parameters provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0054] In order to make the purpose and advantages of the present invention more clearly understood, the present invention is specifically described below in conjunction with embodiments. It should be understood that the following text is only used to describe one or several specific embodiments of the present invention, and does not strictly limit the scope of protection of the specific claims of the present invention.

[0055] like Figures 1 to 9 As shown, a detection interval identification method for a subway train onboard sensor, the specific steps are as follows:

[0056] Step 1: Use sensors to collect vibration acceleration data of subway trains, which are used as train-mounted vibration data;

[0057] In step 1, the train vibration data consists of 6 axial data, mainly including: lateral, longitudinal, vertical acceleration, nodding, shaking, and rolling angular velocity;

[0058] The sensor is mainly installed in the B09 box of the train. B09 is the number of the air brake cut-off valve, which is similar to B05, B07, etc., and is the mark of specific components in the train braking system.

[0059] Step 2: Use the multiple moving average filtering method to filter the longitudinal acceleration and vertical acceleration of the train collected by the sensor;

[0060] The multiple moving average filtering method is introduced as follows:

[0061] The first moving average filter:

[0062]

[0063] No. Heavy moving average filtering:

[0064]

[0065] In the formula, For the The re-filtering result is at time The output when For the The re-filtering result is at time The output when is the filtering window length, is the input acceleration signal; refer to the attached Figure 3 and attached Figure 4 ,The two figures show the comparison of vertical acceleration before and after MMA filtering. It can be found that after MMA filtering, the segmentation points of the interval are clearer.

[0066] Step 3: Add the MMA filtering results of the longitudinal acceleration and the vertical acceleration to determine the cutting threshold;

[0067] In step 3, the MMA filtering results of the longitudinal acceleration and the vertical acceleration are added, and the added signal is recorded as , determine the cutting threshold:

[0068] In the formula, is the maximum peak-to-peak value of signal A, ; It is a percentage value, and the value range here is ≤5%.

[0069] Step 4: less than the threshold The signal corresponding to can be judged as the train stop time, greater than The time corresponding to the signal is judged as the time when the train is running in the section. The results are shown in the attached figure. Figure 5 shown.

[0070] Step 5: Take the average value of each stop time as the segmentation point of the train section. The result is as shown in the attached figure. Figure 6 As shown, the segmentation points can well divide the train driving range.

[0071] Step 6: When the train passes through the section, it will generate a large angular velocity (see attached Figure 7 ), Figure 7 (a) is the angular velocity of the train when it passes through a certain section, Figure 7 (b) is the curvature recorded in the interval curve record. Therefore, the curve information is further extracted from the angular velocity of the head shaking collected by the sensor. The curve information extracted from the angular velocity of the head shaking includes the number of curves and the direction of the curves, and is compared with the information of the curve record of each interval to determine the specific interval where the data segmented in Step 1-Step 5 is located.

[0072] In step 6, multiple moving average filtering is further performed on the shaking angular velocity, and the multiple moving average filtering method (Formula With ), as attached Figure 8 As shown, the peaks and troughs of the filtered angular velocity are identified;

[0073] The peak value must meet the following conditions at the same time:

[0074] The trough value must meet the following conditions at the same time:

[0075] In the formula, is the location of the peak / trough, and They are exist The first and second order derivatives at .

[0076] Step 7: Identify the driving section, and record the extracted head shaking angle peak and trough sequence as: ,in Indicates The peak / trough information of the element, if the elements are peaks, then On the contrary, if elements are troughs, then ;

[0077] Similarly, the interval The curve recorded in the curve ledger information is converted into a sequence , if elements are left-opening / right-opening curves, then On the contrary, if elements are right-opening / left-opening curves, then ;

[0078] Finally, the sequence with sequence Match if and Exactly the same, that is , then the driving interval of the cut data is interval.

[0079] According to the above steps Step 1-Step 7: the present invention first performs a moving average filter on the train onboard vibration data (6 axes: lateral, longitudinal, vertical acceleration, nod, shake, and roll angular velocity) collected by the sensor to achieve data smoothing; then the smoothed longitudinal acceleration and vertical acceleration data are added to calculate the maximum peak-to-peak value; finally, the segmentation threshold is determined according to the percentage of the peak value to achieve rapid cutting of the sensor data; finally, the shake angular velocity generated when the train passes through the curve is matched with the line curve ledger to identify the correct driving section. Through this automated sensor data section cutting and identification method, the analysis efficiency of subway train onboard data can be greatly improved.

[0080] The working principle of the present invention is: using the MMA filtering method to filter the vertical acceleration and the longitudinal acceleration; using the peak percentage of the sum of the vertical acceleration and the longitudinal acceleration filtering results as the cutting threshold; using the curve information extracted from the shaking angular velocity measured by the sensor to perform train section identification.

[0081] Step 8: Adjust the parameters of the filtering algorithm by analyzing the distance and speed changes between trains;

[0082] like Fig. 9 As shown, in step 8, the parameters of the filtering algorithm are adjusted by analyzing the distance and speed changes between trains. The specific steps are as follows:

[0083] S1: Obtaining real-time data: It is necessary to obtain the speed and position information of the train in real time, which is usually achieved through on-board sensors and communication systems to ensure the timeliness and accuracy of the data.

[0084] S2: Calculate relative distance and speed: Based on the acquired data, calculate the relative distance and speed difference between the front and rear trains, which includes considering the current speed, acceleration of the trains and the rate of change of the distance between them.

[0085] S3: Analyze dynamic changes: Analyze the impact of these dynamic changes on the filtering algorithm. For example, when the train spacing decreases or the speed difference increases, the filtering strength may need to be increased to reduce noise interference; conversely, the filtering strength may need to be reduced to retain more useful signals.

[0086] S4: Adjust filtering parameters: Based on the above analysis, dynamically adjust the parameters of the filtering algorithm, which may involve adjusting the size of the filtering window, weight allocation, or the type of filter.

[0087] S5: Implement adaptive control: Combined with the PID control algorithm, the adjusted filter parameters are applied to the train speed control system. The PID controller can correct the control amount according to the speed tracking error and error deviation, thereby achieving more accurate speed control;

[0088] The PID control algorithm compares the actual speed with the set target speed and calculates the speed tracking error , use proportional regulation to amplify the deviation at the current moment, use integral regulation to integrate the deviations of all past times, and then use differential regulation to adjust the rate of change of the deviation. Finally, the outputs of the three links of proportional, integral and differential are weighted summed to obtain the final control amount. :

[0089]

[0090] in: is the output of the controller; is the input of the controller, i.e., the deviation between the set value and the actual value; is the proportionality coefficient; is the integration coefficient; is the differential coefficient; is the integral of the deviation; is the differential of the deviation.

[0091] The working principle of the present invention is: by acquiring train data in real time and adjusting the filtering algorithm parameters accordingly, the safety and efficiency of train operation can be improved. The optimized filtering algorithm can more accurately reflect the actual state of the train and can accurately reflect complex vibration modes, especially in the case of multi-train interaction, thereby ensuring the accuracy of the average filtering processing algorithm.

[0092] The above is only a preferred embodiment of the present invention. It should be noted that, for those skilled in the art, several improvements and modifications can be made without departing from the principles of the present invention, and these improvements and modifications should also be considered as the protection scope of the present invention. The structures, devices and operating methods not specifically described and explained in the present invention shall be implemented according to the conventional means in the art unless otherwise specified and limited.

Claims

1. A detection interval identification method for a subway train onboard sensor, characterized in that: The specific steps are as follows: Step 1: Use sensors to collect vibration acceleration data of subway trains, which are used as train-mounted vibration data; Step 2: Use the multiple moving average filtering method to filter the longitudinal acceleration and vertical acceleration of the train collected by the sensor; Step 3: Add the MMA filtering results of the longitudinal acceleration and the vertical acceleration to determine the cutting threshold; Step 4: less than the threshold The signal corresponding to can be judged as the train stop time, greater than The time corresponding to the signal is judged as the time when the train is running in the section; Step 5: Take the average value of each stop time as the segmentation point of the train section; Step 6: When the train passes through the section, it will generate a large angular velocity of the shaking head. Therefore, the curve information is further extracted from the angular velocity of the shaking head collected by the sensor and compared with the curve information of each section, so as to determine the specific section where the data segmented in Step 1-Step 5 is located; Step 7: Driving range identification; Step 8: Adjust the parameters of the filtering algorithm by analyzing the distance and speed changes between trains; The multiple moving average filtering method is described as follows: The first moving average filter: No. Heavy moving average filtering: In the formula, For the The re-filtering result is at time The output when For the The re-filtering result is at time The output when is the filtering window length, is the input acceleration signal.

2. The detection interval identification method for a subway train onboard sensor according to claim 1 is characterized in that: In step 1, the train-borne vibration data consists of 6-axis data, including: lateral, longitudinal, vertical acceleration, nodding, shaking, and rolling angular velocity.

3. The detection interval identification method for a subway train onboard sensor according to claim 2 is characterized in that: The sensor is installed in the B09 box of the train.

4. The detection interval identification method for a subway train onboard sensor according to claim 3 is characterized in that: In step 3, the MMA filtering results of the longitudinal acceleration and the vertical acceleration are added, and the added signal is recorded as , determine the cutting threshold: In the formula, is the maximum peak-to-peak value of signal A, ; It is a percentage value, and the value range here is ≤5%.

5. The detection interval identification method for a subway train onboard sensor according to claim 4 is characterized in that: In step Step 6, the curve information extracted from the shaking angular velocity includes the number of curves and the direction of the curves.

6. The detection interval identification method for a subway train onboard sensor according to claim 5 is characterized in that: In step 6, multiple moving average filtering is further performed on the shaking angular velocity, and the peaks and troughs of the filtered angular velocity are identified by using the multiple moving average filtering method; The peak value must meet the following conditions at the same time: The trough value must meet the following conditions at the same time: In the formula, is the location of the peak / trough, and They are exist The first and second order derivatives at .

7. The detection interval identification method for a subway train onboard sensor according to claim 6 is characterized by: In step 7, the extracted shaking angle peak and trough sequence is recorded as: ,in Indicates The peak / trough information of the element, if the elements are peaks, then On the contrary, if elements are troughs, then ; Similarly, the interval The curve recorded in the curve ledger information is converted into a sequence , if elements are left-opening / right-opening curves, then On the contrary, if elements are right-opening / left-opening curves, then ; Finally, the sequence with sequence Match if and Exactly the same, that is , then the driving interval of the cut data is interval.

8. The detection interval identification method for a subway train onboard sensor according to claim 7, characterized in that: In step 8, the parameters of the filtering algorithm are adjusted by analyzing the distance and speed changes between trains. The specific steps are as follows: S1: Obtain real-time data: It is necessary to obtain the speed and position information of the train in real time; S2: Calculate relative distance and speed: Calculate the relative distance and speed difference between the front and rear trains based on the acquired data; S3: Analyze dynamic changes: Analyze the impact of these dynamic changes on the filtering algorithm; S4: Adjust filtering parameters: Based on the above analysis, dynamically adjust the parameters of the filtering algorithm; S5: Implement adaptive control: Combined with the PID control algorithm, the adjusted filter parameters are applied to the train speed control system.

9. The detection interval identification method for a subway train onboard sensor according to claim 8, characterized in that: The PID control algorithm compares the actual speed with the set target speed and calculates the speed tracking error , use proportional regulation to amplify the deviation at the current moment, use integral regulation to integrate the deviations of all past times, and then use differential regulation to adjust the rate of change of the deviation. Finally, the outputs of the three links of proportional, integral and differential are weighted summed to obtain the final control amount. : in: is the output of the controller; is the input of the controller, i.e., the deviation between the set value and the actual value; is the proportionality coefficient; is the integration coefficient; is the differential coefficient; is the integral of the deviation; is the differential of the deviation.

Images