A design method for flexible vibration monitoring points of rail transit vehicle body

Abstract

The present invention discloses a design method for flexible vibration monitoring points of rail transit vehicles. The method first determines the preliminary flexible vibration monitoring points of the vehicle body by analyzing the vibration mode of the flexible vibration mode that is easily excited by the vehicle body; extracts the vibration characteristics of the preliminary flexible vibration monitoring points of the vehicle body under the typical flexible vibration that is easily excited by the vehicle body, and constructs a data set for convolutional neural network; constructs a one-dimensional convolutional neural network for fitting the mapping relationship between the vibration characteristics of the measuring points and the flexible vibration mode of the vehicle body; according to the preliminary flexible vibration monitoring points of the vehicle body, further optimizes the number of measuring points, designs the arrangement scheme of monitoring measuring points with different numbers of vehicle bodies, uses the trained neural network to select the arrangement scheme of vehicle body measuring points with the highest recognition accuracy, installs the monitoring device, and monitors the flexible vibration characteristics of the vehicle body to realize early warning. The flexible vibration of the vehicle body of rail transit vehicles can be monitored in real time, the vehicle body structure mode that accounts for a particularly significant proportion of abnormal vibration can be determined, and the improved design of the vehicle body structure can be scientifically guided.

Description

Technical Field

[0001] The invention belongs to the technical field of rail transit, and in particular relates to a design method for flexible vibration monitoring points of a rail transit vehicle body. Background Art

[0002] With the continuous development of rail transit technology, green energy conservation has become an important development direction of rail transit equipment, and the key approach is to achieve lightweight design of vehicle structure. On the one hand, lightweight design of vehicle structure reduces the weight of the vehicle and reduces the pressure of the vehicle on the track, thereby slowing down the wear of wheels and rails, and further reducing the maintenance and repair costs of vehicles and lines. On the other hand, it reduces the traction and braking force of the vehicle, achieving energy saving and consumption reduction. However, with the continuous lightweight design of the body structure, the flexible vibration of the body is prominent, which will cause abnormal vibration under certain circumstances, greatly affecting the riding comfort of passengers. Therefore, real-time monitoring of the flexible vibration of the rail transit vehicle body and determining the body structure mode that accounts for a particularly significant proportion of abnormal vibration are very important for scientifically guiding the improvement of the body structure design and improving the comfort and safety of train operation.

[0003] According to research, existing monitoring of vehicle body vibration mainly focuses on monitoring stability indicators and abnormal vibration. Both are achieved by placing acceleration sensors on the floor of the vehicle body, collecting floor vibration acceleration, and analyzing different vibration frequency components of the vehicle body to obtain the stability indicators and vibration conditions of the vehicle body. This monitoring method has certain advantages, namely, a small number of sensors are required and the operation is simple. However, this method only collects the vibration conditions of the vehicle body floor. For flexible vibrations of the vehicle body structure that are not obviously deformed at the floor, this method cannot effectively identify the flexible vibration conditions of the vehicle body. Most other tests on flexible vibrations of the vehicle body are carried out by placing a large number of sensors on the vehicle body. Although this can effectively reflect the flexible vibration conditions of the vehicle body structure, the number of sensors is large, and the operation and maintenance are complicated, which is not suitable for long-term flexible vibration monitoring of the vehicle body structure.

[0004] In summary, how to provide a reasonable rail transit vehicle body flexible vibration monitoring method that can accurately and efficiently reflect the flexible vibration of the vehicle body structure and facilitate operation and maintenance has become a key issue that needs to be solved urgently. Summary of the invention

[0005] In order to overcome the above-mentioned shortcomings, the inventor of the present invention has continuously reformed and innovated through long-term exploration and attempts, multiple experiments and efforts, and proposed a design method for flexible vibration monitoring points of rail transit vehicle bodies. This method can not only accurately and efficiently monitor the flexible vibration of body parts, but also facilitate operation and maintenance. This method can monitor the flexible vibration of rail transit vehicle bodies in real time, determine the body structure mode with significant vibration, and provide strong support for further scientific guidance of body structure improvement design.

[0006] To achieve the above-mentioned purpose, the technical solution adopted by the present invention is to provide a design method for flexible vibration monitoring points of a rail transit vehicle body. The steps include:

[0007] S1, establish a finite element model of the rail transit vehicle body, and use finite element software to perform modal analysis to obtain the body modal information, including natural frequency, vibration mode, and damping ratio;

[0008] S2, a rigid-flexible coupling dynamic model of the vehicle system considering the flexibility of the vehicle body is established by using the hybrid coordinate method, and the vibration response of the vehicle body under typical wheel-rail excitation conditions is obtained by simulation analysis;

[0009] S3, perform spectrum analysis on the vibration response results of the vehicle body under typical wheel-rail excitation conditions, find out the flexible vibration modes with the largest energy proportion under different wheel-rail excitation conditions, and determine the flexible vibration modes of the rail transit vehicle body that are easily excited under typical wheel-rail excitation conditions;

[0010] S4, by analyzing the vibration modes of the flexible vibration modes that are easily excited by the vehicle body, and according to the locations of the typical vehicle body modal test measurement points, the locations of the four measurement points with the largest deformation in the vibration modes of the flexible vibration modes of the vehicle body that are easily excited by each order are determined, and all the determined measurement point locations are used as the preliminary selection of vehicle body flexible vibration monitoring measurement points;

[0011] S5, extracting vibration characteristics of the preliminary selected vehicle body flexible vibration monitoring measuring point positions under typical flexible vibrations that are easily excited by the vehicle body, and constructing a data set for convolutional neural network, including a training set and a test set;

[0012] S6, construct a one-dimensional convolutional neural network for fitting the mapping relationship between the vibration characteristics of the measuring points and the flexible vibration modes of the vehicle body. The network consists of three convolutional modules, one average pooling layer and one fully connected layer;

[0013] S7, based on the initially selected vehicle body flexible vibration monitoring measurement points, further optimize the number of measurement points, design different numbers of vehicle body monitoring measurement point layout schemes, and use the constructed training set and test set to train and test the neural network;

[0014] S8, compare the recognition accuracy of flexible vibration modes under different monitoring point layout schemes of the vehicle body, find a design scheme with a recognition accuracy of more than 95% and the least number of measurement points, and determine the final vehicle body measurement point layout scheme;

[0015] S9, install the monitoring device according to the optimal measurement point layout plan, monitor the flexible vibration characteristics of the vehicle body, and use the trained convolutional neural network to identify the main excited vehicle body flexible modes in real time and display early warnings.

[0016] According to the method for designing flexible vibration monitoring points of a rail transit vehicle body described in the present invention, a further preferred technical solution is: in step S1, a finite element model of a rail transit vehicle body is established, firstly a three-dimensional entity is established according to the geometric dimensions of the body, secondly the body is divided into solid grids using hexahedral grids, and finally material properties are assigned according to the type of body material, including density, elastic modulus, etc.

[0017] According to the method for designing flexible vibration monitoring points of a rail transit vehicle body described in the present invention, a further preferred technical solution is: in step S2, typical wheel-rail disturbance includes wheel disturbance and track disturbance, wherein wheel disturbance includes tread concave wear, wheel polygonal wear, wheel abrasion, etc., and track disturbance includes track geometric unevenness, turnout impact, abnormal rail wear, rail corrugation, etc.

[0018] According to the method for designing flexible vibration monitoring points of a rail transit vehicle body described in the present invention, a further preferred technical solution is: in step S4, the typical modal test points are selected from 5 test sections of a conventional vehicle body, and 8 measuring points are set on each section, which are respectively located at the four corners of the section and the middle point of each side line.

[0019] According to a method for designing flexible vibration monitoring points of a rail transit vehicle body described in the present invention, a further preferred technical solution is: in step S6, the convolutional neural network architecture is input layer-convolution module-average pooling layer-fully connected layer-output layer. The input layer can set different data depths according to the number of measuring points. There are three convolution modules, which are divided into convolution module a, convolution module b, and convolution module c. Each convolution module includes a convolution layer and a maximum pooling layer. In convolution module a, the convolution layer kernel size is 16, the number of kernels is 16, the maximum pooling layer kernel size is 8, and the step size is 2. In convolution module b, the convolution layer kernel size is 8, the number of kernels is 32, the maximum pooling layer kernel size is 4, and the step size is 2. In convolution module c, the convolution layer kernel size is 4, the number of kernels is 64, the maximum pooling layer kernel size is 2, and the step size is 2. The activation function of the convolution layer in each convolution module is the ReLU activation function.

[0020] According to a design method for monitoring points of flexible vibration of rail transit vehicle body described in the present invention, its further preferred technical scheme is: in step S7, different numbers of monitoring points are arranged, specifically, the positions of the four measuring points with the largest deformation in each order of the flexible vibration mode of the vehicle body that is easily excited are arranged according to the deformation amount. For example, the four measuring points of the first order mode are defined as 1-1, 1-2, 1-3 and 1-4 according to the deformation amount, wherein the deformation amount is 1-1>1-2>1-3>1-4, and the four measuring points of the nth order mode are n-1, n-2, n-3, and n-4 by analogy. First, one measuring point is taken for each order as the first monitoring point arrangement scheme, and scheme one is specifically 1-1, 2-1, 3-1...n-1. Scheme two is to take 2 measuring points for each order, scheme three is to take 3 measuring points for each order, and scheme four is to take 4 measuring points for each order, thereby completing the design of different numbers of monitoring points arrangement schemes.

[0021] Compared with the prior art, the technical solution of the present invention has the following advantages / benefits:

[0022] 1. The rail transit vehicle body flexible vibration monitoring method proposed in the present invention can determine the body structure mode with a significant proportion of vibration energy during operation through real-time flexible vibration of the body structure, which can provide strong support for guiding the improvement design of the body structure.

[0023] 2. The present invention designs the monitoring point arrangement scheme through simulation analysis and deep learning algorithm, thereby finding a measurement point arrangement scheme with high accuracy in identifying flexible deformation of the vehicle body and requiring fewer measurement points. Compared with the traditional scheme of arranging sensors in a few positions or a large number of positions, this measurement point arrangement scheme can more accurately and efficiently reflect the flexible vibration of the vehicle body, and is simple to operate and maintain. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0025] Figure 1 It is a flow chart of a method for monitoring flexible vibration of a rail transit vehicle body according to the present invention.

[0026] Figure 2 It is a schematic diagram of the finite element model of the vehicle body.

[0027] Figure 3 It is the modal information of the vehicle body.

[0028] Figure 4It is the time domain result of the vertical vibration acceleration of the middle floor of the car body under the excitation of vertical track unevenness.

[0029] Figure 5 It is the result of the vertical acceleration spectrum analysis of the middle floor of the car body under the excitation of vertical track unevenness.

[0030] Figure 6 It is a schematic diagram of the four positions with the maximum vertical bending modal deformation of the vehicle body.

[0031] Figure 7 It is the cross-sectional distribution diagram of the vehicle body structure test in conventional vehicle body modal test.

[0032] Figure 8 This is the layout diagram of the acceleration sensor in test section A of a conventional vehicle body modal test.

[0033] Fig. 9 It is the basic framework diagram of convolutional neural network.

[0034] Fig.10 It is a schematic diagram of a vehicle body flexible vibration monitoring device. DETAILED DESCRIPTION

[0035] In order to make the purpose, technical solutions and advantages of the present invention clearer, the technical solutions in the embodiments of the present invention are clearly and completely described below. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present invention. Therefore, the detailed description of the embodiments of the present invention provided below is not intended to limit the scope of the claimed invention, but merely represents the selected embodiments of the present invention.

[0036] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and thus, once an item is defined in one drawing, it may not be further defined or explained in the subsequent drawings.

[0037] Example:

[0038] Figure 1 The present invention is a flowchart of a method for designing flexible vibration monitoring points of a rail transit vehicle body. Figure 2 This is the distribution diagram of the test sections of the vehicle body structure for conventional vehicle body modal testing. Generally, 5 sections A~E are selected. Figure 3 This is the layout diagram of the acceleration sensor of the test section of the conventional vehicle body modal test. Generally, 8 measuring points A1~A8 are set for one section, which are located at the four corners of the section and the middle point of each side line. The specific steps of this method are as follows:

[0039] First, a finite element model of the rail transit vehicle body is established. A three-dimensional entity is established according to the geometric dimensions of the vehicle body. Then, the vehicle body is divided into solid grids using hexahedral grids with a grid size of 20 mm. Finally, material properties are assigned according to the type of vehicle body material, including material density and elastic modulus. The established vehicle body finite element model is as follows: Figure 2 The modal information of the vehicle body is obtained by modal analysis using finite element software, including natural frequency, vibration mode, damping ratio, etc. Some modal information is shown in Figure 3 shown.

[0040] The rigid-flexible coupling dynamic model of the vehicle system considering the flexibility of the vehicle body is established by the hybrid coordinate method. The vibration response of the vehicle body under typical wheel-rail excitation conditions is obtained by simulation analysis. Typical wheel-rail excitation includes wheel excitation and track excitation. Wheel excitation includes tread concave wear, wheel polygonal wear, wheel scratches, etc. Track excitation includes track geometric unevenness, turnout impact, rail abnormal wear, rail corrugation, etc. The time domain results of the vertical vibration acceleration of the middle floor of the vehicle body under the excitation of vertical track unevenness are shown in the following figure. Figure 4 shown.

[0041] The spectrum analysis of the vibration response results of the vehicle body under typical wheel-rail excitation conditions is carried out to find out the flexible vibration modes with the largest energy share under different wheel-rail excitation conditions, and determine the flexible vibration modes of the rail transit vehicle body that are easily excited under typical wheel-rail excitation conditions. The spectrum analysis results of the vertical acceleration of the middle floor of the vehicle body under the excitation of track vertical unevenness are shown in Figure 2. Figure 5 As shown in the figure, the flexible vibration mode with the largest energy share under this excitation is the vertical bending mode of the vehicle body; the simulation analysis results show that the modes that are easily excited under typical wheel-rail excitation conditions include 1 diamond mode, 2 vertical bending mode, 3 torsional mode, and 4 lateral bending mode.

[0042] By analyzing the vibration modes of the flexible vibration modes that are easily excited by the vehicle body, and according to the locations of the typical vehicle body modal test measuring points, the four measuring point locations with the largest deformation in the vibration modes of the flexible vibration modes of the vehicle body that are easily excited by each order are determined respectively, and all the determined measuring point locations are used as the preliminary vehicle body flexible vibration monitoring measuring points. The four locations with the largest deformation of the vertical bending mode of the vehicle body are as follows: Figure 6 As shown;

[0043] The vibration characteristics of the preliminary selected vehicle body flexible vibration monitoring measuring point positions under typical flexible vibrations that are easily excited by the vehicle body are extracted, and the extracted vibration responses of different measuring points are preprocessed to construct a data set for convolutional neural network, including a training set and a test set.

[0044] Construct a convolutional neural network for fitting the mapping relationship between vibration characteristics of different measuring points and flexible vibration of components. The basic framework is shown in the figure below. Fig. 9As shown in the figure, it includes an input layer, a convolution module, an average pooling layer, a fully connected layer, and an output layer. The input layer can set different data depths according to the number of measurement points. There are three convolution modules, namely convolution module a, convolution module b, and convolution module c. Each convolution module includes a convolution layer and a maximum pooling layer. In convolution module a, the convolution layer kernel size is 16, the number of kernels is 16, the maximum pooling layer kernel size is 8, and the step size is 2. In convolution module b, the convolution layer kernel size is 8, the number of kernels is 32, the maximum pooling layer kernel size is 4, and the step size is 2. In convolution module c, the convolution layer kernel size is 4, the number of kernels is 64, the maximum pooling layer kernel size is 2, and the step size is 2. The activation function of the convolution layer in each convolution module is the ReLU activation function.

[0045] According to the preliminarily selected vehicle body flexible vibration monitoring points, the number of monitoring points is further optimized, and different number of monitoring point layout schemes are designed for the vehicle body. Specifically, the four measurement points with the largest deformation in each order of the vehicle body flexible vibration mode that is easily excited are arranged from large to small according to the deformation amount to form a measurement point set of each order. The measurement point set includes the first-order mode to the n-th order mode. For example, the four measurement points of the first-order mode are defined as 1-1, 1-2, 1-3 and 1-4 according to the deformation amount, wherein the deformation amount is 1-1>1-2>1-3>1-4, and the four measurement points of the n-th order mode are n-1, n-2, n-3, and n-4 by analogy. Take one measurement point for each order as the first monitoring point layout scheme, take two measurement points for each order for the second scheme, take three measurement points for each order for the third scheme, and take four measurement points for each order for the fourth scheme, thereby completing the design of different number of monitoring point layout schemes. The measurement point layout scheme of this embodiment is as follows:

[0046] Plan 1: 1-1, 2-1, 3-1, 4-1.

[0047] Plan 2: 1-1, 1-2, 2-1, 2-2, 3-1, 3-2, 4-1, 4-2.

[0048] Plan three: 1-1, 1-2, 1-3, 2-1, 2-2, 2-3, 3-1, 3-2, 3-3, 4-1, 4-2, 4-3.

[0049] Plan 4: 1-1, 1-2, 1-3, 1-4, 2-1, 2-2, 2-3, 2-4, 3-1, 3-2, 3-3, 3-4, 4-1, 4-2, 4-3, 4-4.

[0050] According to the different schemes 1 to 4 designed, the input and output are set respectively, and the constructed data set is used to repeatedly test the neural network to complete parameter tuning. The recognition accuracy of the flexible vibration mode of the vehicle body under different schemes 1 to 4 is compared, and the optimal monitoring point layout scheme is selected. The recognition accuracy of the flexible vibration mode of the vehicle body under different schemes 1 to 4 is 70%, 85%, 95%, and 98%, respectively. The recognition accuracy of schemes 3 and 4 is above 95%, but scheme 3 requires fewer measurement points, so the final scheme selected in this embodiment is scheme 3.

[0051] Install the monitoring device according to the optimal monitoring point layout plan, as shown in the schematic diagram Fig.10 As shown, it includes an acceleration sensor, a signal acquisition system, a signal storage system, a signal analysis system, and a signal display system. The acceleration sensor is mainly used to test the vibration response of the measuring point position, the signal acquisition system collects the data tested by the acceleration sensor, the signal storage system stores the collected test data, the signal analysis system uses the collected data and the trained convolutional neural network to identify the flexible vibration mode of the vehicle body, and the signal display system mainly displays the current flexible vibration of the vehicle body and the main excited vehicle body mode in real time, and timely warns when encountering abnormal conditions.

[0052] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present invention.

[0053] In the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0054] In the present invention, unless otherwise clearly specified and limited, a first feature being "above" or "below" a second feature may include that the first and second features are in direct contact, or may include that the first and second features are not in direct contact but are in contact through another feature between them. Moreover, a first feature being "above", "above" and "above" a second feature includes that the first feature is directly above and obliquely above the second feature, or simply indicates that the first feature is higher in level than the second feature. A first feature being "below", "below" and "below" a second feature includes that the first feature is directly below and obliquely below the second feature, or simply indicates that the first feature is lower in level than the second feature.

[0055] The above are only preferred embodiments of the present invention. It should be noted that the above preferred embodiments should not be regarded as limiting the present invention, and the protection scope of the present invention should be based on the scope defined by the claims. For ordinary technicians in this technical field, several improvements and modifications can be made without departing from the spirit and scope of the present invention, and these improvements and modifications should also be regarded as the protection scope of the present invention.

Claims

1. A design method for flexible vibration monitoring points of a rail transit vehicle body, characterized in that: The steps include: S1, establishing a finite element model of a rail transit vehicle body, and using finite element software to perform modal analysis to obtain body modal information; S2, a rigid-flexible coupling dynamic model of the vehicle system considering the flexibility of the vehicle body is established by using the hybrid coordinate method, and the vibration response of the vehicle body under typical wheel-rail excitation conditions is obtained by simulation analysis. Typical wheel-rail excitation includes wheel excitation and track excitation. Wheel excitation includes tread concave wear, wheel polygonal wear, and wheel abrasion. Track excitation includes track geometry irregularity, turnout impact, abnormal rail wear, and rail corrugation. S3, perform spectrum analysis on the vibration response results of the vehicle body under typical wheel-rail excitation conditions, find out the flexible vibration modes with the largest energy proportion under different wheel-rail excitation conditions, and determine the flexible vibration modes of the rail transit vehicle body that are easily excited under typical wheel-rail excitation conditions; S4, by analyzing the vibration modes of the flexible vibration modes that are easily excited by the vehicle body, and according to the locations of the typical vehicle body modal test measurement points, the four measurement points with the largest deformation in the vibration modes of the flexible vibration modes of the vehicle body that are easily excited by each order are determined, and all the determined measurement point locations are used as the preliminary selection of vehicle body flexible vibration monitoring measurement points. The typical vehicle body modal test measurement points are selected from 5 test sections of the conventional vehicle body, and 8 measurement points are set for each section, which are located at the four corners of the section and the middle point of each side line; S5, extracting vibration characteristics of the preliminary selected vehicle body flexible vibration monitoring measuring point positions under typical flexible vibrations that are easily excited by the vehicle body, and constructing a data set for convolutional neural network, including a training set and a test set; S6, constructing a one-dimensional convolutional neural network for fitting the mapping relationship between the vibration characteristics of the measuring points and the flexible vibration modes of the vehicle body, which consists of 3 convolution modules, 1 average pooling layer and 1 fully connected layer; S7, based on the preliminarily selected vehicle body flexible vibration monitoring points, optimize the number of monitoring points, design different numbers of vehicle body monitoring point layout schemes, use the constructed training set and test set to train and test the neural network, and arrange the different numbers of monitoring points, specifically, arrange the four measurement points with the largest deformation in each order of the vehicle body flexible vibration mode that is easily excited according to the deformation from large to small to form each order of measurement point sets, the measurement point set includes the first order mode to the nth order mode, and take 1 measurement point for each order as the monitoring point layout. Scheme 1 and Scheme 2 take 2 measurement points for each order, Scheme 3 takes 3 measurement points for each order, and Scheme 4 takes 4 measurement points for each order, and completes the design of different numbers of monitoring point layout schemes; S8, compare the flexible vibration mode recognition accuracy under different vehicle body monitoring point layout schemes, find a design scheme with a recognition accuracy of more than 95% and the least number of measurement points, and determine the final vehicle body measurement point layout scheme; S9, install the monitoring device according to the optimal measurement point layout plan, monitor the flexible vibration characteristics of the vehicle body, and use the trained convolutional neural network to identify the main excited vehicle body flexible modes in real time and display early warnings.

2. A design method for flexible vibration monitoring points of a rail transit vehicle body according to claim 1, characterized in that: In step S1, a finite element model of a rail transit vehicle body is established. First, a three-dimensional entity is established according to the geometric dimensions of the body. Then, the body is solid meshed using a hexahedral mesh. Finally, material properties are assigned according to the body material type, including density and elastic modulus.

3. A design method for flexible vibration monitoring points of a rail transit vehicle body according to claim 1, characterized in that: The vehicle body modal information includes natural frequency, vibration mode, and damping ratio.

4. A design method for flexible vibration monitoring points of a rail transit vehicle body according to claim 1, characterized in that: In step S6, the convolutional neural network architecture is input layer-convolution module-average pooling layer-fully connected layer-output layer, and the input layer sets different data depths according to the number of measurement points.

5. A design method for flexible vibration monitoring points of a rail transit vehicle body according to claim 4, characterized in that: There are three convolution modules, namely convolution module a, convolution module b, and convolution module c. Each convolution module includes a convolution layer and a maximum pooling layer. In convolution module a, the convolution layer kernel size is 16, the number of kernels is 16, the maximum pooling layer kernel size is 8, and the step size is 2. In convolution module b, the convolution layer kernel size is 8, the number of kernels is 32, the maximum pooling layer kernel size is 4, and the step size is 2. In convolution module c, the convolution layer kernel size is 4, the number of kernels is 64, the maximum pooling layer kernel size is 2, and the step size is 2. The convolution layer activation function in each convolution module adopts the ReLU activation function.

Images