Ballast bed hardening detection method, device, equipment, storage medium and computer program
By decomposing and adjusting the train vibration signal and detecting the track bed slab disease, the problems of low detection efficiency and high cost in the existing technology are solved, and efficient and accurate track bed slab disease detection is achieved.
Patent Information
- Application Number
- CN202411512782.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-28
- Publication Date
- 2025-03-07
AI Technical Summary
The prior art has high cost and low detection efficiency when detecting diseases of the tray bed plate, making it difficult to achieve efficient and accurate detection.
By obtaining the vibration signal and vibration signal energy of the train, decompose the vibration signal, determine the modal component and its pulse indicators, dynamically adjust the pulse and energy thresholds, and detect whether the pulse index and vibration signal energy exceed the threshold to determine the abnormal section of the track bed.
Without affecting the normal operation of the train, the inspection cycle is shortened, the detection efficiency and accuracy are improved, and the track defects are detected in a timely manner and maintenance are carried out.
Smart Images
Figure CN119577559A_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates to the field of track detection technology, and in particular to a method, device, equipment, storage medium and computer program for detecting track bed compaction. Background Art
[0002] In recent years, track defect detection methods based on vehicle vibration response have attracted widespread attention. With the continuous development of railway transportation, on-board vibration detection technology has gradually become a key means to improve railway transportation safety and maintenance efficiency.
[0003] In some technologies, the detection of ballast compaction is usually carried out through manual inspection, inspection vehicle inspection or by installing sensors at specific positions of the track. However, the above technologies have the problems of high cost and low detection efficiency, and are not convenient for detecting ballast compaction. Summary of the invention
[0004] The present invention provides a method, device, equipment, storage medium and computer program for detecting ballast compaction to monitor and diagnose ballast compaction defects, thereby shortening the detection cycle and effectively improving the detection efficiency and detection accuracy without affecting the normal operation of trains.
[0005] In a first aspect, the present disclosure provides a method for detecting ballast hardening, comprising:
[0006] Obtaining the vibration signal and vibration signal energy of the train;
[0007] Decomposing the vibration signal, determining modal components, and determining a pulse index of each modal component;
[0008] Adjusting the pulse threshold to be adjusted based on the pulse index to obtain a candidate pulse threshold;
[0009] Adjusting the energy threshold to be adjusted based on the vibration signal energy to obtain a candidate energy threshold;
[0010] Detecting whether the pulse index is greater than a candidate pulse threshold and whether the vibration signal energy is greater than a candidate energy threshold;
[0011] When the pulse index is greater than the candidate pulse threshold and the vibration signal energy is greater than the candidate energy threshold, the section in the track bed where the pulse index is greater than the candidate pulse threshold and the vibration signal energy is greater than the candidate energy threshold is determined to be an abnormal track bed section.
[0012] In some embodiments, the step of obtaining the vibration signal and vibration signal energy of the train includes:
[0013] Obtain the vibration acceleration data of the train;
[0014] Performing sliding window analysis on the vibration acceleration data to determine the vibration signal of each time window;
[0015] Based on the vibration signal in each time window, the vibration signal energy in each time window is determined.
[0016] In some embodiments, the vibration acceleration data is vibration acceleration data of an axle box of the train.
[0017] In some embodiments, it also includes:
[0018] Get historical vibration acceleration data;
[0019] The candidate threshold corresponding to the most recently obtained vibration acceleration data is used as the threshold to be adjusted, and the threshold to be adjusted is adaptively adjusted based on the historical vibration acceleration data to obtain the candidate threshold, wherein the threshold to be adjusted includes a pulse threshold to be adjusted and an energy threshold to be adjusted, and the candidate threshold includes a candidate pulse threshold and a candidate energy threshold.
[0020] In some embodiments, it also includes:
[0021] Perform multi-dimensional feature extraction based on each modal component to obtain multiple multi-dimensional component features;
[0022] Based on the multi-dimensional component features, a preset support vector machine model is trained to obtain a disease recognition model;
[0023] The vibration acceleration data corresponding to the abnormal section of the ballast bed is input into the disease identification model for analysis to obtain a ballast bed compaction identification result.
[0024] In some embodiments, after the step of extracting multidimensional features based on each modal component to obtain multiple multidimensional component features, the method further includes:
[0025] Constructing a high-dimensional feature matrix data set based on a plurality of the multi-dimensional component features;
[0026] Based on the high-dimensional feature matrix data set, determining a feature training set;
[0027] The step of training a preset support vector machine model based on the multi-dimensional component features to obtain a disease recognition model comprises:
[0028] The preset support vector machine model is trained based on the feature training set to obtain a disease recognition model.
[0029] In a second aspect, the present disclosure provides a ballast compaction detection device, comprising:
[0030] An acquisition module, used for acquiring the vibration signal and vibration signal energy of the train;
[0031] A signal decomposition module, used to decompose the vibration signal, determine the modal components, and determine the pulse index of each modal component;
[0032] A pulse threshold adjustment module adjusts the pulse threshold to be adjusted based on the pulse index to obtain a candidate pulse threshold;
[0033] An energy threshold adjustment module, which adjusts the energy threshold to be adjusted based on the energy of the vibration signal to obtain a candidate energy threshold;
[0034] A detection module, used to detect whether the pulse index is greater than a candidate pulse threshold and whether the vibration signal energy is greater than a candidate energy threshold;
[0035] The abnormal section determination module is used to determine that a section in the track bed where the pulse index is greater than the candidate pulse threshold and the vibration signal energy is greater than the candidate energy threshold is a track bed abnormal section when the pulse index is greater than the candidate pulse threshold and the vibration signal energy is greater than the candidate energy threshold.
[0036] In a third aspect, the present disclosure provides a computer device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the method described in the above aspects.
[0037] In a fourth aspect, the present disclosure provides a computer-readable storage medium having a computer program stored thereon, which implements the steps of the method described in the above aspects when executed by a processor.
[0038] In a fifth aspect, the present disclosure provides a computer program product, including a computer program / instructions, which implements the steps of the method described in the above aspects when the computer program is executed by a processor.
[0039] The present disclosure provides a method, device, equipment, storage medium and computer program for detecting ballast compaction, which obtains the vibration signal and vibration signal energy of the train, decomposes the vibration signal, determines the modal components, and determines the pulse index of each modal component, then adjusts the pulse threshold to be adjusted based on the pulse index to obtain the candidate pulse threshold, and adjusts the energy threshold to be adjusted based on the vibration signal energy to obtain the candidate energy threshold, and then detects whether the pulse index is greater than the candidate pulse threshold and whether the vibration signal energy is greater than the candidate energy threshold. When the pulse index is greater than the candidate pulse threshold and the vibration signal energy is greater than the candidate energy threshold, the section in the track ballast where the pulse index is greater than the candidate pulse threshold and the vibration signal energy is greater than the candidate energy threshold is determined to be an abnormal ballast section. In this way, the track state detection is performed based on the dynamic response of the operating vehicle, and the ballast compaction disease is monitored and diagnosed. Without affecting the normal operation of the train, the detection cycle is shortened, the detection efficiency and detection accuracy are effectively improved, and the track defects are detected in time and maintenance is carried out. BRIEF DESCRIPTION OF THE DRAWINGS
[0040] The present disclosure will be described in more detail below based on embodiments and with reference to the accompanying drawings:
[0041] Figure 1 A schematic diagram of a process flow of a method for detecting ballast compaction provided in an embodiment of the present disclosure;
[0042] Figure 2 is a structural block diagram of a ballast bed compaction detection device in one embodiment;
[0043] Figure 3 is an internal structure diagram of an electronic device in an embodiment;
[0044] Figure 4 is a schematic flow chart of a method for detecting ballast compaction in another embodiment;
[0045] Figure 5 It is a schematic diagram of the flow of a ballast bed hardening disease identification algorithm in one embodiment;
[0046] Figure 6 A schematic diagram of a disease recognition model training result in one embodiment;
[0047] Figure 7 Schematic diagram of a vehicle-track coupling system in one embodiment. DETAILED DESCRIPTION
[0048] In order to enable those skilled in the art to better understand the technical solution of the present disclosure, and to fully understand and implement how the present disclosure applies technical means to solve technical problems and achieve the corresponding technical effects, the technical solution in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are only embodiments of a part of the present disclosure, not all of the embodiments. The embodiments of the present disclosure and the various features in the embodiments can be combined with each other without conflict, and the technical solutions formed are all within the scope of protection of the present disclosure. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in this field without making creative work should fall within the scope of protection of the present disclosure.
[0049] It should be noted that the terms "first", "second", etc. in the specification and claims of the present disclosure and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present disclosure described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products, or devices.
[0050] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.
[0051] Embodiment 1
[0052] Figure 1 The following is a flow chart of a method for detecting ballast compaction provided by an embodiment of the present disclosure. Figure 1 As shown, a method for detecting ballast compaction includes:
[0053] Step 110, obtaining the vibration signal and vibration signal energy of the train.
[0054] In this embodiment, by installing vibration sensors in operating vehicles, the vibration signals and vibration signal energy of operating vehicles can be obtained in real time, without the need for additional expensive equipment, thereby reducing costs and improving economic benefits. It is understandable that the information of the vibration signal includes various dynamic characteristics of the interaction between the vehicle and the track, such as the interaction between the wheel and the rail, the health of the track structure, etc., in order to judge the disease condition of the track and warn of potential problems in advance, thereby effectively improving the safety of railway transportation and the convenience of maintenance. In some embodiments, the track bed is an important component of the track. As the basis of the track frame, when obtaining the vibration signal and vibration signal energy of the train, the running path of the train on the track is divided into multiple sections, that is, the track bed is divided into multiple sections, and then the vibration signal and vibration signal energy of the train in each section during operation are obtained.
[0055] Step 120, decomposing the vibration signal, determining modal components, and determining a pulse index of each modal component.
[0056] In this embodiment, the vibration signals of each section of the ballast are modally decomposed by VMD (Variational Mode Decomposition), and the vibration signals are decomposed into multiple modal components, so as to extract information of different frequencies and vibration modes, thereby better analyzing the internal structure of the vibration signal. In addition, after determining the multiple modal components corresponding to each vibration signal, the pulse index of each modal component is calculated, wherein the pulse index is used to measure whether the vibration signal is abrupt or abnormal. By calculating the pulse index of the modal component, the influence of different frequency components on the overall vibration signal can be deeply understood, providing more detailed signal characteristics for subsequent detection.
[0057] Step 130: Adjust the pulse threshold to be adjusted based on the pulse indicator to obtain a candidate pulse threshold.
[0058] In this embodiment, the pulse threshold to be adjusted can be set according to the actual situation, or can be obtained by analyzing the collected train travel data. After calculating the pulse index of each modal component, the pulse threshold to be adjusted is adjusted based on the pulse index to obtain a candidate pulse threshold for subsequent track disease detection.
[0059] Step 140: Adjust the energy threshold to be adjusted based on the vibration signal energy to obtain a candidate energy threshold.
[0060] In this embodiment, the energy threshold to be adjusted can be set according to the actual situation, or can be obtained by analyzing the collected train driving data. After obtaining the vibration signal energy of the train, the energy threshold to be adjusted is adjusted based on the vibration signal energy to obtain a candidate energy threshold for subsequent track disease detection.
[0061] Step 150 , detecting whether the pulse index is greater than a candidate pulse threshold and whether the vibration signal energy is greater than a candidate energy threshold.
[0062] It should be understood that the ballast is an important part of the track. As the basis of the track frame, when judging whether the track has ballast bed hardening disease, the pulse index is judged to determine whether the pulse index exceeds the normal range. In this embodiment, after adjusting the pulse threshold to be adjusted to obtain the candidate pulse threshold, the pulse index is detected to be greater than the candidate pulse threshold, that is, the pulse index is detected to be beyond the normal range, so as to make a preliminary judgment on whether the track has ballast bed hardening disease.
[0063] In this embodiment, the pulse index exceeds the normal range, that is, the pulse index is greater than the candidate pulse threshold. It can be considered that the section of the track bed where the pulse index is greater than the candidate pulse threshold may have a track bed hardening disease. Subsequently, it is necessary to detect whether the vibration signal energy is within the threshold range, that is, whether the vibration signal energy is greater than the candidate energy threshold, in order to further determine whether the track has a track bed hardening disease. When the pulse index is greater than the candidate pulse threshold and the vibration signal energy is greater than the candidate energy threshold, the section of the track bed where the pulse index is greater than the candidate pulse threshold and the vibration signal energy is greater than the candidate energy threshold is determined to be an abnormal track bed section.
[0064] Step 160, when the pulse index is greater than the candidate pulse threshold and the vibration signal energy is greater than the candidate energy threshold, determine that the section in the track bed where the pulse index is greater than the candidate pulse threshold and the vibration signal energy is greater than the candidate energy threshold is an abnormal track bed section.
[0065] In this embodiment, the pulse index exceeds the normal range and the vibration signal energy is within the threshold range, that is, the pulse index is greater than the candidate pulse threshold and the vibration signal energy is greater than the candidate energy threshold. It can be considered that the section of the track bed where the pulse index is greater than the candidate pulse threshold and the vibration signal energy is greater than the candidate energy threshold has a track bed hardening disease. The section of the track bed where the vibration signal energy is greater than the candidate energy threshold is an abnormal track bed section, thereby performing a track disease detection. In some embodiments, when the vibration signal energy exceeds the threshold range, it indicates that a short-wave unevenness disease, such as corrugation disease, may have occurred, which will mask the vibration response of the track bed hardening disease.
[0066] In some embodiments, it is possible to first detect whether the pulse index is greater than the candidate pulse threshold, and then detect whether the vibration signal energy is greater than the candidate energy threshold; specifically, detect whether the pulse index is greater than the candidate pulse threshold; when the pulse index is greater than the candidate pulse threshold, detect whether the vibration signal energy is greater than the candidate energy threshold; when the vibration signal energy is greater than the candidate energy threshold, determine that the section in the track bed where the vibration signal energy is greater than the candidate energy threshold is an abnormal track bed section; it is also possible to first detect whether the vibration signal energy is greater than the candidate energy threshold, and then detect whether the pulse index is greater than the candidate pulse threshold, specifically, detect whether the vibration signal energy is greater than the candidate energy threshold; when the vibration signal energy is greater than the candidate energy threshold, detect whether the pulse index is greater than the candidate pulse threshold; when the pulse index is greater than the candidate pulse threshold, determine that the section in the track bed where the pulse index is greater than the candidate pulse threshold is an abnormal track bed section.
[0067] In the above embodiment, the vibration signal and vibration signal energy of the train are obtained, the vibration signal is decomposed, the modal components are determined, and the pulse index of each modal component is determined. Then, the pulse threshold to be adjusted is adjusted based on the pulse index to obtain the candidate pulse threshold, and the energy threshold to be adjusted is adjusted based on the vibration signal energy to obtain the candidate energy threshold. Then, it is detected whether the pulse index is greater than the candidate pulse threshold and whether the vibration signal energy is greater than the candidate energy threshold. When the pulse index is greater than the candidate pulse threshold and the vibration signal energy is greater than the candidate energy threshold, the section in the track bed where the pulse index is greater than the candidate pulse threshold and the vibration signal energy is greater than the candidate energy threshold is determined to be an abnormal section of the track bed. In this way, the track state detection is performed based on the dynamic response of the operating vehicle, and the track bed compaction disease is monitored and diagnosed. Without affecting the normal operation of the train, the detection cycle is shortened, the detection efficiency and detection accuracy are effectively improved, and track defects are detected in time and maintenance is carried out.
[0068] In one embodiment, the step of obtaining the vibration signal and vibration signal energy of the train includes:
[0069] Obtain the vibration acceleration data of the train;
[0070] Performing sliding window analysis on the vibration acceleration data to determine the vibration signal of each time window;
[0071] Based on the vibration signal in each time window, the vibration signal energy in each time window is determined.
[0072] In this embodiment, a vibration sensor is installed in the train (operating vehicle) to obtain the vibration acceleration data of the train in real time during driving. After obtaining the vibration acceleration data of the train, the vibration acceleration data is subjected to sliding window analysis, that is, the vibration acceleration data is segmented according to a fixed length to obtain vibration signals of multiple time windows to capture the time domain changes of the vibration signal, and then the vibration signal energy corresponding to each time window and the vibration signal is calculated based on the vibration signal of each time window. The distribution of the vibration signal energy helps to determine the possible abnormal position in the track bed. In this way, without affecting the normal operation of the train, the real-time monitoring of the track status is realized through the on-board acceleration sensor data, which effectively improves the detection frequency and detection efficiency, detects track defects in time, grasps the service status of the line and performs maintenance and repair.
[0073] In one embodiment, the vibration acceleration data is vibration acceleration data of an axle box of the train.
[0074] It should be understood that in actual applications, there are certain difficulties in installing sensors on wheelsets. The installation space of wheelsets is limited. Due to the intense vibration, the loss of sensors at this location is also large. In addition, daily maintenance and other work require additional capital investment, which is costly. The axle box, as a component directly connected to the wheelset, can accurately reflect the vibration of the wheelset, and the vibration signal is relatively strong. By installing sensors on the axle box, the vibration data of the wheelset can be efficiently obtained, analyzed and processed, and accurate evaluation and early warning of the health status of the railway line can be provided. In this embodiment, by installing a vibration sensor on the axle box, the vibration acceleration data of the train during driving is obtained. In this way, the maintenance difficulty and cost are reduced, and a high detection accuracy can be maintained, thereby realizing the monitoring of the roadbed compaction disease.
[0075] In one embodiment, the roadbed hardening detection method further includes:
[0076] Get historical vibration acceleration data;
[0077] The candidate threshold corresponding to the most recently obtained vibration acceleration data is used as the threshold to be adjusted, and the threshold to be adjusted is adaptively adjusted based on the historical vibration acceleration data to obtain the candidate threshold, wherein the threshold to be adjusted includes a pulse threshold to be adjusted and an energy threshold to be adjusted, and the candidate threshold includes a candidate pulse threshold and a candidate energy threshold.
[0078] In this embodiment, the acquisition of historical vibration acceleration data can be obtained by collecting vibration acceleration data of the train during previous travel and integrating the collected vibration acceleration data to obtain historical vibration acceleration data. In some embodiments, the vibration acceleration data obtained last time is merged into the historical vibration acceleration data to increase the data volume of the historical vibration acceleration data, thereby improving the accuracy of the threshold adjustment.
[0079] In this embodiment, the candidate threshold includes a candidate pulse threshold and an energy threshold to be adjusted, and the candidate threshold includes a candidate pulse threshold and a candidate energy threshold. The candidate pulse threshold corresponding to the pulse index in the most recently obtained vibration acceleration data is used as the pulse threshold to be adjusted, and the pulse threshold to be adjusted is adaptively adjusted based on the pulse index in the historical vibration acceleration data, thereby obtaining the candidate pulse threshold corresponding to the pulse index in the most recently obtained vibration acceleration data. In addition, the candidate energy threshold corresponding to the vibration signal energy in the most recently obtained vibration acceleration data is used as the energy threshold to be adjusted, and the energy threshold to be adjusted is adaptively adjusted based on the vibration signal energy in the historical vibration acceleration data, thereby obtaining the candidate energy threshold corresponding to the vibration signal energy in the most recently obtained vibration acceleration data.
[0080] In one embodiment, the roadbed hardening detection method further includes:
[0081] Perform multi-dimensional feature extraction based on each modal component to obtain multiple multi-dimensional component features;
[0082] Based on the multi-dimensional component features, a preset support vector machine model is trained to obtain a disease recognition model;
[0083] The vibration acceleration data corresponding to the abnormal section of the ballast bed is input into the disease identification model for analysis to obtain a ballast bed compaction identification result.
[0084] In this embodiment, multidimensional feature extraction is performed based on each modal component, and feature extraction is performed on each modal component using multidimensional indicators to obtain multiple multidimensional component features of each modal component, wherein the multidimensional indicators include mean, skewness factor, waveform factor, peak factor, pulse factor, standard deviation, root mean square, root amplitude, kurtosis, energy, spectrum amplitude mean, and spectrum amplitude standard deviation. The multidimensional component features are trained on a preset support vector machine model (SVM model) to obtain a disease recognition model. After the section in the track bed where the pulse index is greater than the candidate pulse threshold and the vibration signal energy is greater than the candidate energy threshold is defined as the abnormal section of the track bed, the vibration acceleration data corresponding to the abnormal section of the track bed is input into the disease recognition model for analysis to obtain the track bed hardening recognition result, and the specific abnormality and location point of the abnormal section of the track bed are determined to further identify the track bed hardening disease, thereby improving the accuracy of the identification of the track bed hardening disease.
[0085] In one embodiment, after the step of extracting multidimensional features based on each modal component to obtain multiple multidimensional component features, the method further includes:
[0086] Constructing a high-dimensional feature matrix data set based on a plurality of the multi-dimensional component features;
[0087] Based on the high-dimensional feature matrix data set, determining a feature training set;
[0088] The step of training a preset support vector machine model based on the multi-dimensional component features to obtain a disease recognition model comprises:
[0089] The preset support vector machine model is trained based on the feature training set to obtain a disease recognition model.
[0090] In this embodiment, a high-dimensional feature matrix data set is constructed based on multiple multidimensional component features of each modal component, and the high-dimensional feature matrix data set is segmented to generate a feature training set. The feature training set is used to train a preset support vector machine model to obtain a disease recognition model. In addition, the support vector machine model can be further optimized using a particle swarm optimization algorithm to improve the reliability and accuracy of the results. In some embodiments, the high-dimensional feature matrix data set is segmented, and a feature verification set and a feature test set are also generated. After the support vector machine model is trained multiple times, the feature verification set and the feature test set are input into the disease recognition model to adjust the model and evaluate the ability of the disease recognition model, thereby evaluating the reliability and accuracy of the disease recognition model.
[0091] It should be understood that although Figure 1 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 1 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.
[0092] Embodiment 2
[0093] Based on the above embodiments, this embodiment provides an application example.
[0094] A method for detecting ballast compaction, comprising:
[0095] The present invention firstly realizes on-board detection by installing vibration acceleration sensors on operating vehicles, thereby realizing on-board detection of roadbed compaction diseases, providing timely monitoring for disease identification, and solving the problems of long detection cycle and untimely detection of roadbed compaction diseases in traditional methods.
[0096] Among them, the source of vehicle vibration response information: in operating vehicles, the main vibration acceleration data is obtained from the vehicle body vibration acceleration sensor, frame vibration acceleration sensor and axle box vibration acceleration sensor. Figure 7 As shown, Figure 7 The schematic diagram of the vehicle-track coupling system is shown. Since the vehicle body adopts a dual-system spring damping system for vibration reduction, the acceleration of the vehicle body is basically unchanged in the time domain; the acceleration change of the frame is also small because it uses a single-system suspension vibration reduction system. In contrast to the two, the wheelset is in direct contact with the track, and is affected by the load and the uneven track, so its acceleration change is relatively obvious. In practical applications, there are certain difficulties and challenges in installing sensors directly on the wheelset. First, the installation space of the wheelset is limited. Due to the relatively severe vibration, the loss of the sensor there is also relatively large. In addition, daily maintenance and other work require additional capital investment, making the installation of sensors on the wheelset a complex and expensive task. The axle box, as a component directly connected to the wheelset, can accurately reflect the vibration of the wheelset, and the vibration signal is relatively strong. By installing sensors on the axle box, the vibration data of the wheelset can be efficiently obtained, analyzed and processed, and accurate evaluation and early warning of the health status of the railway line can be provided. Therefore, the present invention comprehensively considers the limited installation space of the wheelset, severe vibration, maintenance cost and other issues, and finally chooses to install sensors on the axle box. Doing so will not only greatly reduce the difficulty and cost of maintenance, but also maintain a high level of detection accuracy, thereby enabling the monitoring and evaluation of ballast bed compaction diseases.
[0097] like Figure 4 As shown in the figure, after obtaining the vibration acceleration information data, signal processing is required to mine the implicit rules from the data and extract relevant features for judgment, so as to realize the identification of the roadbed compaction disease. In order to more accurately identify the roadbed compaction disease, a variety of processing methods such as sliding window segmentation data and VMD decomposition are adopted, and the calculation of energy and pulse indicators and the setting of thresholds are combined to determine whether the roadbed compaction disease occurs in the section. Specifically, the algorithm flow is as follows:
[0098] (1) Sliding window segmentation of vibration data:
[0099] The vehicle axle box vibration acceleration data is divided into fixed-length sliding windows to obtain local vibration characteristics. Through sliding window analysis, the time domain changes of the vibration signal are captured and the vibration signal energy within each window is calculated. This energy distribution helps to determine where anomalies may exist. If the energy within the window exceeds the threshold, it may indicate the presence of short-wave irregularities, such as corrugation, which will mask the vibration response of the ballast bed compaction disease.
[0100] (2) VMD variational mode decomposition:
[0101] VMD is used to decompose the vibration signal in each sliding window into multiple modal components. The use of VMD helps to extract information of different frequencies and vibration modes, so as to better understand the inherent structure of the vibration signal. Each component is analyzed to gain a deeper understanding of the contribution of different frequency components to the overall vibration, providing more detailed signal characteristics for subsequent steps.
[0102] (3) Pulse index calculation and threshold judgment:
[0103] Calculate the pulse index of each component to measure the sudden change or abnormality of the vibration signal.
[0104] (4) Dynamically adjust threshold and loop processing:
[0105] The adaptive threshold adjustment module is introduced to dynamically adjust the threshold by monitoring the energy changes and pulse indicators of real-time data. This helps to adapt to the characteristics of vibration signals under different conditions. The above steps are executed in a loop to ensure that all data are processed. Through iterative processing, the possible existence of ballast bed hardening diseases is continuously monitored to improve the robustness of the algorithm. By monitoring the changes in vibration data in real time, the threshold is continuously adjusted to flexibly adapt to the characteristics of vibration signals under different working conditions.
[0106] Specifically, the initial threshold of the pulse index is set at the same time to determine whether the pulse index exceeds the normal range. Combined with the threshold of the vibration signal energy, if the pulse index is abnormal and the energy is within the set threshold range, it is determined that the section has a ballast bed hardening disease. Combined with the pulse index and energy information of each component, a dynamic comparison is made with the current threshold to determine in real time whether the section has a ballast bed hardening disease. Continuously optimize the threshold setting to improve the accuracy and stability of the algorithm. The real-time and flexibility of this step helps to capture changes in ballast bed hardening diseases in a timely manner, ensuring that the algorithm can effectively distinguish in different situations.
[0107] In one embodiment, after the axle box vibration acceleration is decomposed by VMD, the mean, skewness factor, standard deviation, waveform factor, root mean square, peak factor, root square amplitude, pulse factor, energy, spectrum amplitude mean, spectrum amplitude standard deviation and other indicators are used to extract features of each signal component and construct a training set of a support vector machine. The specific feature calculation formula is shown in the following table.
[0108] Characteristic calculation formula table
[0109]
[0110]
[0111] The overall process of the roadbed compaction disease identification algorithm based on high-dimensional feature matrix is as follows Figure 5 As shown. First, VMD decomposition is performed on the axle box vibration acceleration under the action of ballast compaction, and the signal is decomposed into multiple IMF (Intrinsic Mode Function) components. Then, for each IMF component, energy feature extraction is performed to construct a high-dimensional feature matrix data set for subsequent model training and verification. In the training stage, the training set is input into the SVM model for learning and training, and the SVM model is further optimized using the particle swarm optimization algorithm to improve the reliability and accuracy of the results. After the model training is completed, the verification set is input into the model for model verification, and the model is verified using the test set to evaluate its recognition reliability and accuracy.
[0112] A simulation data set is constructed according to the table below, and the roadbed compaction disease is identified using the disease identification model.
[0113] Table of parameters related to the trackbed status binary classification dataset
[0114]
[0115] The sorted data set is input into the SVM model for training and learning. The final training results are as follows Figure 6 The confusion matrix is shown in Figure 2. By observing the confusion matrix, it can be found that after using VMD to decompose the axle box vibration signal and extracting the IMF energy features, combined with the recognition model of the support vector machine, significant results have been achieved in the recognition task of ballast bed compaction disease. Specifically, the recognition accuracy of ballast bed compaction disease reached more than 96%, of which the recognition accuracy of the second category reached 100%, and only 3.6% of the samples in the first category were misclassified as the second category. These results show that this method has high feasibility and accuracy.
[0116] The present invention detects the track status based on the dynamic response of the operating vehicle and monitors and diagnoses the roadbed compaction disease. The method can realize real-time monitoring of the track status through the vehicle-mounted acceleration sensor data without affecting the normal operation of the train. This method can improve the detection frequency and efficiency, detect track defects in time, grasp the service status of the line and perform maintenance and repair, which is of great significance for maintaining the smoothness of the line and ensuring the safety of train operation.
[0117] (1) This method first solves the problem of long detection cycles of manual inspections and track inspection vehicles. Traditional methods require regular manual inspections or the use of special track inspection vehicles for inspections, which have long detection cycles and low efficiency. However, the method of the present invention can utilize the dynamic response of operating vehicles for real-time monitoring, greatly shortening the detection cycle.
[0118] (2) Secondly, the method reduces the cost of detection equipment and improves economic benefits. Traditional detection methods require a large amount of investment to install and maintain specific detection equipment, and to collect and process data, which will cause a high cost burden. The present invention uses a certain number of sensors installed on operating vehicles, without the need to invest in additional expensive equipment, thereby reducing costs and improving economic benefits.
[0119] Finally, this method can improve the detection accuracy of roadbed compaction diseases and solve the problem of low detection efficiency of traditional detection methods. Traditional methods can often only be used for detection in a limited area, cannot fully cover the railway line, and are easily affected by the environment and operators, resulting in inaccurate detection results. The present invention detects through the dynamic response of operating vehicles, can monitor roadbed compaction diseases in real time and comprehensively, and improves detection accuracy and efficiency.
[0120] Embodiment 3
[0121] like Figure 2 As shown, based on the above embodiment, this embodiment provides a roadbed compaction detection device, including:
[0122] An acquisition module 210 is used to acquire a vibration signal and vibration signal energy of a train;
[0123] A signal decomposition module 220, used to decompose the vibration signal, determine modal components, and determine the pulse index of each modal component;
[0124] A pulse threshold adjustment module 230 adjusts the pulse threshold to be adjusted based on the pulse indicator to obtain a candidate pulse threshold;
[0125] An energy threshold adjustment module 240 adjusts the energy threshold to be adjusted based on the energy of the vibration signal to obtain a candidate energy threshold;
[0126] A detection module 250, configured to detect whether the pulse index is greater than a candidate pulse threshold and whether the vibration signal energy is greater than a candidate energy threshold;
[0127] The abnormal section determination module 270 is used to determine that the section in the track bed where the vibration signal energy is greater than the candidate energy threshold is a track bed abnormal section when the pulse index is greater than the candidate pulse threshold and the vibration signal energy is greater than the candidate energy threshold.
[0128] In one embodiment, the acquisition module 210 includes:
[0129] A data acquisition unit, used for acquiring vibration acceleration data of the train;
[0130] A signal determination unit, used to perform sliding window analysis on the vibration acceleration data to determine the vibration signal of each time window;
[0131] The energy determination unit is used to determine the vibration signal energy of each time window based on the vibration signal of each time window.
[0132] In one embodiment, the ballast bed compaction detection device further comprises:
[0133] A historical data acquisition module is used to acquire historical vibration acceleration data;
[0134] An adaptive adjustment module is used to use the candidate threshold corresponding to the most recently obtained vibration acceleration data as the threshold to be adjusted, and to perform adaptive threshold adjustment on the threshold to be adjusted based on the historical vibration acceleration data to obtain the candidate threshold, wherein the threshold to be adjusted includes a pulse threshold to be adjusted and an energy threshold to be adjusted, and the candidate threshold includes a candidate pulse threshold and a candidate energy threshold.
[0135] In one embodiment, the ballast bed compaction detection device further comprises:
[0136] An extraction module, used for performing multi-dimensional feature extraction based on each modal component to obtain multiple multi-dimensional component features;
[0137] A training module, used for training a preset support vector machine model based on the multi-dimensional component features to obtain a disease recognition model;
[0138] The analysis module is used to input the vibration acceleration data corresponding to the abnormal section of the ballast bed into the disease identification model for analysis to obtain the ballast bed compaction identification result.
[0139] In one embodiment, the ballast bed compaction detection device further comprises:
[0140] A data construction module, used for constructing a high-dimensional feature matrix data set based on a plurality of the multi-dimensional component features;
[0141] The data determination module is used to determine a feature training set based on the high-dimensional feature matrix data set.
[0142] In one embodiment, the training module is further used to train a preset support vector machine model based on the feature training set to obtain a disease recognition model.
[0143] Embodiment 4
[0144] like Figure 3 As shown, based on the above embodiments, this embodiment provides a computer device, including a memory, a processor, and a computer program stored in the memory, and the processor executes the computer program to implement the steps of the method described in the above embodiments.
[0145] In some implementations of this embodiment, a computer-readable storage medium is provided, on which a computer program is stored, characterized in that when the computer program is executed by a processor, the steps of the method described in the above embodiment are implemented.
[0146] In some implementations of this embodiment, a computer program product is provided, including a computer program / instructions, characterized in that when the computer program is executed by a processor, the steps of the method described in the above embodiment are implemented.
[0147] The processor may include, but is not limited to, one or more processors or microprocessors, etc. Each processor may be an application specific integrated circuit (ASIC), a digital signal processor (DSP), a digital signal processing device (DSPD), a programmable logic device (PLD), a field programmable gate array (FPGA), a controller, a microcontroller, a microprocessor or other electronic components to execute the methods in the above embodiments.
[0148] The computer-readable storage medium may be implemented by any type of volatile or non-volatile storage device or a combination thereof, and the computer-readable storage medium may include but is not limited to, for example, random access memory (RAM), read-only memory (ROM), flash memory, EPROM memory, EEPROM memory, registers, computer storage media (e.g., hard disk, floppy disk, solid-state drive, removable disk, CD-ROM, DVD-ROM, Blu-ray disc, etc.).
[0149] The computer-readable storage medium may also store at least one computer executable program / instruction, which may be, for example, a computer-readable instruction. The computer-readable storage medium includes, but is not limited to, for example, a volatile memory and / or a non-volatile memory. The volatile memory may include, for example, a random access memory (RAM) and / or a cache memory (cache), etc. The computer-readable storage medium may include, for example, a read-only memory (ROM), a hard disk, a flash memory, etc. For example, a non-transitory computer-readable storage medium may be connected to a computing device such as a computer, and then, when the computing device runs the computer-readable instructions stored on the computer-readable storage medium, the various methods described above may be performed.
[0150] In addition, the computer device may also include (but not limited to) a data bus, an input / output (I / O) bus, a display, and input / output devices (eg, keyboard, mouse, speaker, etc.), etc.
[0151] The processor may communicate with external devices via an I / O bus via a wired or wireless network.
[0152] In one embodiment, the at least one computer executable instruction may also be compiled into or constitute a software product / computer program product, wherein one or more computer executable instructions are executed by a processor to perform the various functions and / or method steps in the embodiments described in the present technology.
[0153] In the embodiments provided in the present disclosure, it should be understood that the disclosed devices and methods can also be implemented in other ways. The device embodiments described above are merely schematic. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architecture, functions and operations of the devices, methods and computer program products according to multiple embodiments of the present disclosure. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of a code, and the above-mentioned module, program segment or a part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of boxes in the block diagram and / or flowchart can be implemented with a dedicated hardware-based system that performs a specified function or action, or can be implemented with a combination of dedicated hardware and computer instructions.
[0154] It should be noted that in the present disclosure, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element limited by the sentence "includes a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element.
[0155] Although the embodiments disclosed in the present disclosure are as above, the above contents are only embodiments adopted for facilitating the understanding of the present disclosure and are not intended to limit the present disclosure. Any technician in the technical field to which the present disclosure belongs can make any modifications and changes in the form and details of the implementation without departing from the spirit and scope disclosed in the present disclosure, but the scope of patent protection of the present disclosure shall still be subject to the scope defined in the attached claims.
Claims
1. A method for detecting ballast compaction, characterized in that: include: Obtaining the vibration signal and vibration signal energy of the train; Decomposing the vibration signal, determining modal components, and determining a pulse index of each modal component; Adjusting the pulse threshold to be adjusted based on the pulse index to obtain a candidate pulse threshold; Adjusting the energy threshold to be adjusted based on the vibration signal energy to obtain a candidate energy threshold; Detecting whether the pulse index is greater than a candidate pulse threshold and whether the vibration signal energy is greater than a candidate energy threshold; When the pulse index is greater than the candidate pulse threshold and the vibration signal energy is greater than the candidate energy threshold, the section in the track bed where the pulse index is greater than the candidate pulse threshold and the vibration signal energy is greater than the candidate energy threshold is determined to be an abnormal track bed section.
2. The method for detecting ballast compaction according to claim 1, characterized in that: The step of obtaining the vibration signal and vibration signal energy of the train comprises: Obtain the vibration acceleration data of the train; Performing sliding window analysis on the vibration acceleration data to determine the vibration signal of each time window; Based on the vibration signal in each time window, the vibration signal energy in each time window is determined.
3. The method for detecting ballast compaction according to claim 2, characterized in that: The vibration acceleration data is vibration acceleration data of an axle box of the train.
4. The method for detecting ballast compaction according to claim 1, characterized in that: Also includes: Get historical vibration acceleration data; The candidate threshold corresponding to the most recently obtained vibration acceleration data is used as the threshold to be adjusted, and the threshold to be adjusted is adaptively adjusted based on the historical vibration acceleration data to obtain the candidate threshold, wherein the threshold to be adjusted includes a pulse threshold to be adjusted and an energy threshold to be adjusted, and the candidate threshold includes a candidate pulse threshold and a candidate energy threshold.
5. The method for detecting ballast compaction according to any one of claims 1 to 4, characterized in that: Also includes: Perform multi-dimensional feature extraction based on each modal component to obtain multiple multi-dimensional component features; Based on the multi-dimensional component features, a preset support vector machine model is trained to obtain a disease recognition model; The vibration acceleration data corresponding to the abnormal section of the ballast bed is input into the disease identification model for analysis to obtain a ballast bed compaction identification result.
6. The method for detecting ballast compaction according to claim 5, characterized in that: After the step of extracting multidimensional features based on each modal component to obtain multiple multidimensional component features, the method further includes: Constructing a high-dimensional feature matrix data set based on a plurality of the multi-dimensional component features; Based on the high-dimensional feature matrix data set, determining a feature training set; The step of training a preset support vector machine model based on the multi-dimensional component features to obtain a disease recognition model comprises: The preset support vector machine model is trained based on the feature training set to obtain a disease recognition model.
7. A device for detecting ballast compaction, characterized in that: include: An acquisition module, used for acquiring the vibration signal and vibration signal energy of the train; A signal decomposition module, used to decompose the vibration signal, determine the modal components, and determine the pulse index of each modal component; A pulse threshold adjustment module adjusts the pulse threshold to be adjusted based on the pulse index to obtain a candidate pulse threshold; An energy threshold adjustment module, which adjusts the energy threshold to be adjusted based on the energy of the vibration signal to obtain a candidate energy threshold; A detection module, used to detect whether the pulse index is greater than a candidate pulse threshold and whether the vibration signal energy is greater than a candidate energy threshold; The abnormal section determination module is used to determine that the section in the track bed where the vibration signal energy is greater than the candidate energy threshold is a track bed abnormal section when the pulse index is greater than the candidate pulse threshold and the vibration signal energy is greater than the candidate energy threshold.
8. A computer device comprising a memory, a processor and a computer program stored in the memory, characterized in that: The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 6.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.
10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.