Magnetic levitation vehicle, wheel polygon detection, active suppression method and related systems

Abstract

The application discloses a kind of magnetic levitation vehicle, wheel polygon detection, active inhibition method and related system, collect wheel out-of-roundness signal, vibration acceleration signal and noise signal data;The wheel out-of-roundness signal is introduced into collaborative simulation model, and the vibration acceleration signal and noise signal after simulation are output;Utilize the vibration acceleration signal and noise signal after simulation to build dataset S;Part of sample in dataset S is randomly selected as training sample set, the training sample set is used as the input of support vector machine, the support vector machine is trained, and the mapping model of polygon wavelength, amplitude relative vibration noise signal feature is established;The sample to be detected is input into mapping model, and polygon amplitude prediction result is obtained;If the polygon amplitude prediction result exceeds critical threshold, it is judged that there is wheel polygon problem.The detection method of the application can truly reflect the wheel polygon problem, simplify the detection process, and improve the detection accuracy.

Description

Technical Field

[0001] This invention relates to the field of rail transit technology, and in particular to a magnetic levitation vehicle, a wheel polygon detection and active suppression method and related system. Background Technology

[0002] With the increase in train speed and axle load, electric locomotives have repeatedly experienced wheel polygonal issues during operation, leading to abnormal vibrations and noise, severely impacting passenger comfort. Further development of wheel polygonal issues can cause cracks in vehicle components, posing significant safety and noise hazards. Currently, the improvement and control of wheel polygonal issues mainly relies on passive measures, including shortening the wheel turning cycle, improving wheel turning quality, variable speed operation, and adding grinding tools. However, these passive measures are merely remedial methods, and the wheel polygonal problem continues to plague locomotive operation. Therefore, adopting a forward design approach to prevent and control the occurrence of wheel polygonal issues is of great significance for improving vehicle operating quality.

[0003] In the prior art, there are methods for diagnosing wheel polygons by measuring the order and depth of the wheel polygons. However, these methods are complex to implement and have limited detection accuracy. The prior art does not provide a solution for preventing and controlling wheel deformation problems in advance. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a method and system for detecting and actively suppressing polygons of magnetic levitation vehicles and wheels, which simplifies the detection process of polygons of wheels, improves detection accuracy, and achieves active suppression of polygons of wheels, in order to address the shortcomings of the existing technology.

[0005] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: a wheel polygon detection method, comprising the following steps:

[0006] S1. Collect wheel out-of-roundness signal, vibration acceleration signal, and noise signal data; import the wheel out-of-roundness signal into the co-simulation model, and output the simulated vibration acceleration signal and noise signal; construct dataset S using the simulated vibration acceleration signal and noise signal.

[0007] S2. Randomly select a portion of samples from dataset S as a training sample set, use the training sample set as input to the support vector machine, train the support vector machine, and establish a mapping model of polygon wavelength and amplitude relative to vibration noise signal features.

[0008] S3. Input the sample to be tested into the mapping model to obtain the polygon amplitude prediction result; if the polygon amplitude prediction result exceeds the critical threshold, it is determined that there is a wheel polygon problem.

[0009] This invention utilizes wheel out-of-roundness signals to simulate vibration acceleration and noise signals, and uses the simulated signals to build a dataset. A subset of samples from this dataset is then used to train a support vector machine, resulting in a mapping model. This mapping model is then used to detect wheel polygons. Because this invention uses real out-of-roundness signals to indirectly generate the dataset, its detection method accurately reflects the wheel polygon problem, simplifies the detection process, and improves detection accuracy.

[0010] In this invention, to minimize the error between the simulation results and the experimental results, and to ensure that the simulation model can realistically simulate the line operation to the greatest extent possible, the following steps are included after step S2 and before step S3:

[0011] S21. Take the remaining samples in the dataset S as the samples to be tested, and take the samples to be tested as the input of the mapping model to obtain the prediction results corresponding to the samples to be tested.

[0012] S22. Compare the deviation between the prediction results corresponding to the training sample set and the prediction results corresponding to the sample to be tested. If the deviation is less than the set error, proceed to step S3; otherwise, return to step S2.

[0013] In step S2 of the present invention, the vibration noise signal characteristics include waveform indicators, peak indicators, impulse indicators, margin indicators, kurtosis indicators of vibration acceleration signals, and time-domain average value, time-domain median, time-domain standard deviation, and dominant frequency of noise signals.

[0014] To facilitate the application of the above-mentioned detection method to vehicles and thus to facilitate the early prevention of wheel polygonal problems, the method of the present invention further includes:

[0015] S4. Import the mapping model into the active controller.

[0016] As an inventive concept, the present invention also provides a wheel polygon detection system, which includes a memory, a processor, and a computer program stored in the memory; the processor executes the computer program to implement the steps of the detection method described above.

[0017] As an inventive concept, the present invention also provides a method for actively suppressing wheel polygons, which includes:

[0018] A1. Obtain the amplitude of the wheel polygon. If the amplitude exceeds the critical threshold, proceed to step A2.

[0019] A2. At regular intervals Tw, the collected vibration acceleration signal and noise signal data are extracted, and the vibration acceleration signal and noise signal data are used to predict the time-domain signal waveform within the time range of Ty1+Ty2+Tw; where Ty1 is the acquisition and fault diagnosis delay time; Ty2 is the delay time for applying active control response and generating active excitation;

[0020] A3. Generate active control excitation with anti-phase, equal amplitude, and equal frequency within the time range of Ty1+Ty2+Tw;

[0021] A4. Apply active control excitation to the magnetic levitation wheel pair system until the polygon amplitude is less than the critical threshold.

[0022] The amplitude of the wheel polygon and the time-domain signal waveform are obtained according to the method described above in this invention.

[0023] This invention proposes a polygon active control method that applies active control excitation to counteract superimposed polygon vibration excitation, actively cutting off the inducing conditions of the polygon from the source of excitation, blocking the further development of the wheel polygon, realizing comprehensive control of the wheel polygon, reducing maintenance costs, reducing wheel-rail vibration and noise, and improving vehicle operation quality.

[0024] To facilitate signal acquisition, this invention utilizes a vibration acceleration sensor and a wireless acceleration sensor to acquire the vibration acceleration signal. The vibration acceleration sensor is located at both ends of the axle stator and close to the electromagnet, while the wireless acceleration sensor is mounted on the wheel rotor.

[0025] As an inventive concept, the present invention also provides a wheel polygon active suppression system, including 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 above-described active suppression method of the present invention.

[0026] As an inventive concept, the present invention also provides a rail transit vehicle that employs the wheel polygon detection system described in the present invention, and / or that employs the active suppression system described in the present invention.

[0027] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0028] 1) This invention uses real out-of-roundness signals to indirectly generate a dataset, which can truly reflect the polygon problem of wheels, simplify the detection process, and improve detection accuracy;

[0029] 2) This invention proposes a polygon active control method, which applies active control excitation to counteract superimposed polygon vibration excitation, actively cuts off the inducing conditions of polygon from the source of excitation, blocks the further development of wheel polygon, realizes comprehensive control of wheel polygon, reduces maintenance costs, reduces wheel-rail vibration and noise, and improves vehicle operation quality. Attached Figure Description

[0030] Figure 1 This is a flowchart of the wheel polygon detection and active suppression method according to an embodiment of the present invention;

[0031] Figure 2 This is a schematic diagram illustrating the principle of active suppression of polygonal shapes in wheels according to an embodiment of the present invention. Detailed Implementation

[0032] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0033] In this paper, the terms "contains," "includes," and similar words are intended to indicate logical relationships, not spatial relationships. For example, "A includes B" means that logically B belongs to A, not that spatially B is located inside A. Furthermore, the meanings of the terms "contains," "includes," and similar words should be considered open-ended, not closed-ended. For example, "A includes B" means that B belongs to A, but B does not necessarily constitute the entirety of A; A may also include other elements such as C, D, and E.

[0034] The method of this invention can be used on the wheelset system disclosed in CN114312129A and other magnetic levitation vehicle wheelset systems.

[0035] Example 1

[0036] like Figure 1 As shown, this embodiment provides a method for detecting wheel polygons, which includes the following steps:

[0037] Step S1: Collect a large amount of sample data on wheel out-of-roundness, vibration acceleration, and noise signals on actual operating lines in advance.

[0038] In this embodiment, the testing instrument is placed against the wheel, and then the wheel is rotated once to measure the circumferential non-roundness of the wheel. During the test, the vehicle is stationary, and a jack is needed to lift the axle box so that the wheel can rotate freely in the air.

[0039] Step S2: Establish a multidisciplinary collaborative simulation model based on multibody dynamics software and noise software.

[0040] In this embodiment, the multi-disciplinary collaborative simulation model can use the multibody dynamics software Simpack for dynamic simulation (outputting vibration acceleration data) and the noise software Vone for noise simulation (outputting noise data).

[0041] Step S3: Import the wheel out-of-roundness signal collected from the line into the model, simulate and output vibration acceleration signal and noise signal, compare and verify the noise signal and vibration acceleration signal output by simulation with the noise signal and vibration acceleration signal collected by test, and correct the simulation model to minimize the error between the simulation result and the test, so as to ensure that the simulation model can simulate the line operation as realistically as possible.

[0042] In this embodiment, when correcting the simulation model, the least squares method is used to continuously modify the simulation model so that the error between the simulation results and the experimental results is controlled within a reasonable range (Reference: Wang Hongshan, Zhang Xing, Yang Shuying, et al. Simulation of vector control of asynchronous motor based on online parameter identification of least squares method [J]. Journal of Hefei University of Technology: Natural Science Edition, 2009, 32(4):5.).

[0043] Step S5: Preprocess and extract features from the simulation output dataset S of vibration acceleration and noise signals. Extract waveform parameters, peak values, impulse parameters, margin parameters, and kurtosis parameters for the vertical vibration acceleration a; extract the time-domain average value, time-domain median, time-domain standard deviation, and dominant frequency for the noise signal y. Among these, the waveform parameters... Peak index K2 = a max / a rms Pulse index Margin index K4 = a max / a r kurtosis index In the above formula, the average value Root mean square value Maximum value a max =max(a i ), root mean square amplitude N is the length of the vibration acceleration data, a i The vertical vibration acceleration data corresponding to the i-th data point; the time-domain average value in the noise signal characteristic parameters. Time domain standard deviation Perform Fourier transform on the noise signal y y i (t) represents the noise signal corresponding to the i-th data point at time t. The six dominant frequencies in the Fourier transform signal F(w) correspond to K7 to K10 respectively.13 .

[0044] 80% of the samples in the dataset S are randomly selected as training samples. Based on the support vector machine (SVM) algorithm, polygon wavelength λ and amplitude A are used to establish the vibration noise signal features K1~K1~K2~K3~K4~K5~K6~K7 ... 13 The mapping model, λ=p(K1,…,K5,…,K 13 A = g(K1,…,K5,…,K) 13 (The mapping model establishment process is referenced in: [1] Lu Wencong, Chen Nianyi, Ye Chenzhou, et al. Introduction to support vector machine algorithm and software ChemSVM [J]. Computer and Applied Chemistry, 2002, 19(6):6.), the mapping model is a fitted curve function.

[0045] S6. Using the remaining 20% ​​of the dataset as samples to be tested, calculate the polygon wavelength and amplitude of the samples to be tested according to the mapping model, calculate the prediction results λ2, A2, and calculate the deviation between the polygon wavelength and amplitude λ1, A1 of the sample points to be tested. If the deviation is less than the set error, the mapping model is feasible; otherwise, the mapping model needs to be corrected.

[0046] Through the above process, a realistic and reliable self-learning diagnostic model of the wheel polygon was established; the intelligent diagnostic model of the wheel polygon was then input into the electromagnetic control module of the maglev vehicle and incorporated into the wheel polygon active controller.

[0047] It should be noted that step S6 is optional. That is, if the mapping model obtained in step S5 meets the requirements, step S6 can be skipped, and the wheel polygon can be directly identified using the mapping model obtained in step S5.

[0048] Example 2

[0049] This embodiment provides an active method for suppressing wheel polygons, including the following steps:

[0050] Step T1: Place vibration acceleration sensors and noise sensors / microphones near the electromagnets at both ends of the stator of the inner ring axle of the magnetic levitation vehicle. Install wireless acceleration sensors and wheel speed sensors on the outer ring wheel rotor. Output the collected speed signals, vibration acceleration signals, and noise signals to the signal acquisition module in real time.

[0051] Step T2: The signal acquisition module 12 transmits the real-time acquired speed signal, vibration acceleration and noise signal to the wheel polygon active controller in the electromagnetic control module 13;

[0052] Step T3: The wheel polygon active controller retrieves the wheel polygon self-learning diagnostic model to quickly identify the wavelength, amplitude, and frequency characteristics of the polygon;

[0053] Step T4: When the amplitude of the wheel polygon exceeds the critical threshold, the controller sends a polygon active control command;

[0054] Step T5: Statistically analyze the time delay Ty1 of the system's data acquisition and fault diagnosis process and the time delay Ty2 of the system's application of active control response and generation of active excitation, and set the control time window length Tw. In this embodiment, since there is a delay in the real-time data acquisition and fault diagnosis process, if the active control excitation is applied directly based on the characteristics of the acquired data, the control error will be too large due to the time delay. Therefore, relevant delay times and time windows are set.

[0055] Step T6: After receiving the polygon active control command, the system extracts the amplitude, frequency and phase characteristics of the collected polygon vibration and noise time domain signal every Tw time interval. Based on the polygon wavelength, amplitude and frequency characteristics diagnosed in step T3, the system predicts the time domain signal waveform of the signal within the subsequent Ty1+Ty2+Tw time range based on machine self-learning.

[0056] Step T7: The controller quickly generates active control excitation with anti-phase, equal amplitude, and equal frequency within the time range of Ty1+Ty2+Tw;

[0057] Step T8: Adjust the input current of the polygon controller to apply active control excitation to the magnetic levitation wheel system, achieving the cancellation and superposition of polygon vibrations (e.g., Figure 2 As shown), the conditions that induce polygons are actively cut off from the source, thus preventing the further development of wheel polygons.

[0058] Step T9: The polygon controller tracks the effect of the polygon active control in real time and adjusts the control logic adaptively according to the effect until the active control stops when the monitoring amplitude of the polygon is less than the critical threshold, thereby realizing precise active control of the wheel polygon.

[0059] Example 3

[0060] Embodiment 3 of the present invention provides a wheel polygon detection system corresponding to Embodiment 1 above. The detection system of this embodiment includes a memory, a processor and a computer program stored in the memory; the processor executes the computer program in the memory to implement the steps of the method of Embodiment 1 above.

[0061] In some implementations, the memory may be high-speed random access memory (RAM), and may also include non-volatile memory, such as at least one disk storage device.

[0062] In other implementations, the processor can be any type of general-purpose processor, such as a central processing unit (CPU) or a digital signal processor (DSP), and there is no limitation here.

[0063] Example 4

[0064] Embodiment 4 of the present invention provides a wheel polygon active suppression system corresponding to Embodiment 2 above. The active suppression system of this embodiment includes a memory, a processor and a computer program stored in the memory; the processor executes the computer program in the memory to implement the steps of the method of Embodiment 2 above.

[0065] In some implementations, the memory may be high-speed random access memory (RAM), and may also include non-volatile memory, such as at least one disk storage device.

[0066] In other implementations, the processor can be any type of general-purpose processor, such as a central processing unit (CPU) or a digital signal processor (DSP), and there is no limitation here.

[0067] Example 5

[0068] Embodiment 5 of the present invention provides a rail transit vehicle corresponding to Embodiment 3 or 4 above, which adopts the system of Embodiment 3 and / or Embodiment 4 above.

[0069] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0070] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for detecting polygonal shapes of wheels, characterized in that, Includes the following steps: S1. Collect wheel out-of-roundness signal, vibration acceleration signal, and noise signal data; import the wheel out-of-roundness signal into the co-simulation model, and output the simulated vibration acceleration signal and noise signal; construct dataset S using the simulated vibration acceleration signal and noise signal. S2. Randomly select a portion of samples from dataset S as a training sample set, use the training sample set as input to the support vector machine, train the support vector machine, and establish a mapping model of polygon wavelength and amplitude relative to vibration noise signal features. S3. Input the sample to be tested into the mapping model to obtain the polygon amplitude prediction result; if the polygon amplitude prediction result exceeds the critical threshold, it is determined that there is a wheel polygon problem; wherein, 80% of the samples in the dataset S are randomly selected as training samples, and the polygon wavelength is established based on the support vector machine algorithm. and amplitude Characteristics of vibration noise signals ~ Mapping model; waveform index Peak index Pulse index Margin indicators kurtosis index ;average value Root mean square value maximum value Root mean square amplitude N is the length of the vibration acceleration data. For the vertical vibration acceleration data corresponding to the i-th data point, the time-domain average value in the noise signal characteristic parameters is... Time domain standard deviation Perform Fourier transform on the noise signal y , The noise signal corresponding to the i-th data point at time t, and the Fourier transform signal. The 6th order main frequency corresponds to ~ .

2. The active suppression method for wheel polygons according to claim 1, characterized in that, After step S2 and before step S3, the following also applies: S21. Take the remaining samples in the dataset S as the samples to be tested, and take the samples to be tested as the input of the mapping model to obtain the prediction results corresponding to the samples to be tested. S22. Compare the deviation between the prediction results corresponding to the training sample set and the prediction results corresponding to the sample to be tested. If the deviation is less than the set error, proceed to step S3; otherwise, return to step S2.

3. The wheel polygon detection method according to claim 1, characterized in that, In step S2, the vibration noise signal characteristics include waveform indicators, peak indicators, impulse indicators, margin indicators, kurtosis indicators of the vibration acceleration signal, and time-domain average value, time-domain median, time-domain standard deviation, and dominant frequency of the noise signal.

4. The wheel polygon detection method according to any one of claims 1 to 3, characterized in that, Also includes: S4. Import the mapping model into the active controller.

5. A wheel polygon detection system, characterized in that, The method includes 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 4.

6. A method for actively suppressing polygonal shapes on wheels, characterized in that, include: A1. Obtain the amplitude of the wheel polygon. If the amplitude exceeds the critical threshold, proceed to step A2. A2. At regular intervals Tw, the collected vibration acceleration signal and noise signal data are extracted, and the vibration acceleration signal and noise signal data are used to predict the time-domain signal waveform within the time range of Ty1+Ty2+Tw; where Ty1 is the acquisition and fault diagnosis delay time; Ty2 is the delay time for applying active control response and generating active excitation; A3. Generate active control excitation with anti-phase, equal amplitude, and equal frequency within the time range of Ty1+Ty2+Tw; A4. Apply active control excitation to the magnetic levitation wheel pair system until the polygon amplitude is less than the critical threshold. The amplitude of the wheel polygon and the time-domain signal waveform are obtained by the method according to any one of claims 1 to 4.

7. The active suppression method for wheel polygons according to claim 6, characterized in that, The vibration acceleration signal is acquired using a vibration acceleration sensor and a wireless acceleration sensor. The vibration acceleration sensor is located at both ends of the axle stator and close to the electromagnet, and the wireless acceleration sensor is installed on the wheel rotor.

8. A wheel polygon active suppression system, characterized in that, It includes 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 of claim 6.

9. A rail transit vehicle, characterized in that, It employs the wheel polygon detection system of claim 5, and / or the active suppression system of claim 8.

Images