Wheel type carrying platform transmission device fault monitoring method based on multi-source data fusion under complex road conditions

Abstract

The invention relates to the technical field of wheel type carrying platforms, in particular to a wheel type carrying platform transmission device fault monitoring method based on multi-source data fusion under complex road conditions. Comprising the steps of determining a bumping grade of a wheel type carrying platform driving road surface; determining a deep learning diagnosis model according to the road bumping level; a fault signal of the transmission device is collected through a sensor; the collected fault signals are preprocessed; and fault monitoring of the wheel type carrying platform transmission device is completed. According to the method, the running state change of the transmission device can be more comprehensively captured, potential faults can be found in time, the downtime caused by the faults is shortened, and the use efficiency of equipment is improved. And the problem that the vibration frequency is abnormal due to the complex road condition environment of the wheel type carrying platform, especially muddy and hollow road sections and the like is solved. According to the method, the road condition state is judged through detection and comparison, so that the corresponding model is switched to carry out fault prediction, and the accuracy of fault prediction is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of wheeled transport platforms, and in particular to a method for monitoring transmission device faults of wheeled transport platforms based on multi-source data fusion under complex road conditions. Background Art

[0002] Wheeled transport platforms are the main combat equipment of modern armies. They integrate protection, mobility, and firepower, and have powerful off-road mobility, protection capabilities, and firepower systems. Mobility is the most basic performance of wheeled transport platforms, and the power transmission system is the core of the mobility of wheeled transport platforms. The working environment of wheeled transport platforms is usually very harsh, and the load carried by their power transmission systems varies greatly. For example, the resistance in complex areas is about 10 to 15 times that of flat areas.

[0003] The transmission system of a wheeled carrier platform is composed of multiple complex components, including elastic couplings, transmission boxes, main clutches, gearboxes, fan transmission devices, planetary steering gears, etc. These characteristics require the power transmission device of a special wheeled carrier platform to withstand greater load impact than a general power transmission system. From the analysis of fault data, the most frequent failures of wheeled carrier platforms are in the power unit, followed by electrical control, fire control, operating system and transmission system. Failures in the power and transmission devices account for more than half of the vehicle failures, accounting for about 51.21%. With the advent of the information age, the scale of equipment is increasing, the performance is increasing, and the structure is becoming more complex. The transmission system of a wheeled carrier platform is no exception, and its complexity makes fault diagnosis more difficult.

[0004] At present, the research on wheeled transport platform diagnosis in my country is mainly focused on single components under stable working conditions, while there is less research on the overall transmission system under complex working conditions, and the research on the failure mechanism of the transmission system is also rarely involved, and the accuracy and reliability are not ideal. Since the traditional wheeled transport platform transmission system fault monitoring method often relies on regular manual inspection or analysis of single sensor data, this method has problems such as long monitoring cycle, delayed response, low fault recognition rate and difficult monitoring, which is difficult to meet the requirements of modern warfare for rapid response and high reliability of equipment. Summary of the invention

[0005] The present invention discloses a method for monitoring transmission failure of a wheeled transport platform based on multi-source data fusion under complex road conditions. The specific method is as follows:

[0006] Determine the level of bumpiness on the road surface on which the wheeled platform is traveling;

[0007] Determine the deep learning diagnostic model based on the level of road bumps;

[0008] Collecting fault signals of the transmission device through sensors;

[0009] Preprocess the collected fault signals;

[0010] The preprocessed fault signal is input into the determined deep learning diagnostic model to complete the fault monitoring of the wheeled carrier platform transmission device.

[0011] Furthermore, the bumpiness level of the road surface on which the wheeled transport platform is traveling is determined in the following specific methods:

[0012] The distance between the road surface and the bottom of the wheeled transport platform is measured in real time by a laser distance measuring sensor to generate a distance data sequence;

[0013] Set the sliding window, calculate the variance of the distance data sequence of each sliding window, and construct the variance value sequence;

[0014] Smoothing the variance sequence;

[0015] According to the variance value sequence, the bumpiness level of the driving road is determined.

[0016] Furthermore, the bumpiness level of the driving road surface is divided into five levels, namely, a flat level, a level one bumpiness, a level two bumpiness, a level three bumpiness and an abnormal bumpiness.

[0017] Furthermore, the fault signal includes: a transmission vibration signal, a transmission operating temperature signal, a transmission operating noise signal and transmission external wear image data.

[0018] Furthermore, the preprocessing method of the transmission device vibration signal, the transmission device operating temperature signal and the transmission device operating noise signal is specifically as follows:

[0019] Eliminate abnormal values, repeated values ​​and missing values ​​of statistical characteristics of transmission device vibration signals, transmission device operating temperature signals and transmission device operating noise signals;

[0020] Smoothing the transmission device vibration signal, transmission device operating temperature signal and transmission device operating noise signal;

[0021] Extracting statistical features of the transmission device vibration signal, the transmission device operating temperature signal and the transmission device operating noise signal, the statistical features including mean features, variance features and peak features;

[0022] The extracted statistical features are time synchronized.

[0023] Furthermore, the transmission device vibration signal, the transmission device operating temperature signal and the transmission device operating noise signal are smoothed, and the specific method is as follows:

[0024] Determine the range of the smoothing coefficient, which is between 0 and 1. The larger the smoothing coefficient, the closer the smoothed data is to the original data, and the smaller the smoothing coefficient, the smoother the smoothed data is.

[0025] Within the range of smoothing coefficients, set the candidate set;

[0026] For each smoothing coefficient in the candidate set, use the training set data to predict the smoothing index, and calculate the mean square error and mean absolute error of the prediction results on the test set;

[0027] Select the optimal smoothing coefficient based on the mean square error and mean absolute error.

[0028] Furthermore, the transmission device external wear image data is preprocessed as follows:

[0029] The transmission device external wear image data is adjusted to a fixed size by using an interpolation method;

[0030] Image feature extraction is performed through convolutional neural networks.

[0031] Further, there are three deep learning diagnostic models;

[0032] The first deep learning model is used for diagnosis of flat grade driving road surfaces;

[0033] The second deep learning model is used for the diagnosis of driving road surfaces with level 1 bumps, level 2 bumps, and level 3 bumps;

[0034] The third deep learning model is used for abnormally bumpy driving road surface diagnosis.

[0035] Further, the deep learning diagnostic model includes an input layer, a hidden layer, and an output layer;

[0036] The input layers of the first deep learning model and the second deep learning model take the transmission vibration signal, the transmission operating temperature signal and the transmission operating noise signal as input; the third deep learning model takes the transmission operating temperature signal, the transmission operating noise signal and the transmission external wear image data as input; the outputs of the deep learning diagnosis models are all probabilistic predictions of the fault type.

[0037] Due to the adoption of the above technical solution, the present invention has the following beneficial effects:

[0038] 1. The present invention can more comprehensively capture the changes in the operating status of the transmission device, timely discover potential faults, reduce downtime caused by faults, and improve the use efficiency of equipment.

[0039] 2. The complex road conditions of wheeled transport platforms, especially muddy and potholed roads, can cause abnormal vibration frequencies. This patent determines the road conditions by detection and comparison, and then switches to the corresponding model for fault prediction, thereby improving the accuracy of fault prediction.

[0040] 3. By using advanced prediction model algorithms, the present invention can not only monitor current faults, but also predict possible future faults based on historical data and real-time data, thereby achieving predictive maintenance, avoiding unplanned downtime, and reducing maintenance costs.

[0041] Other advantages, objectives and features of the present invention will be described in the following description to some extent, and to some extent, will be obvious to those skilled in the art based on the following examination and study, or can be taught from the practice of the present invention. The objectives and other advantages of the present invention can be achieved and obtained through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The accompanying drawings of the present invention are as follows.

[0043] Figure 1 It is a schematic diagram of the overall process of the present invention.

[0044] Figure 2 Schematic diagram of the deep learning diagnostic model process. DETAILED DESCRIPTION

[0045] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0046] A method for monitoring transmission failure of wheeled transport platform based on multi-source data fusion under complex road conditions. Figure 1 As shown, the specific steps are as follows:

[0047] S1. Determine the bumpiness level of the road surface on which the wheeled transport platform is traveling.

[0048] Specifically, the bump level of the road surface on which the wheeled transport platform is traveling is determined as follows:

[0049] S11, measuring the distance between the road surface and the bottom of the wheeled transport platform in real time through a laser distance measuring sensor to generate a distance data sequence;

[0050] S12, setting a sliding window, calculating the variance of the distance data sequence of each sliding window, and constructing a variance value sequence;

[0051] S13, smoothing the variance sequence;

[0052] S14. Determine the bumpiness level of the driving road according to the variance value sequence.

[0053] S2. Determine the deep learning diagnostic model based on the level of road bumps.

[0054] In this embodiment, the bumpiness level of the driving road is divided into five levels, namely, gentle level, level 1 bumpiness, level 2 bumpiness, level 3 bumpiness and abnormal bumpiness. There are three deep learning diagnosis models, the first deep learning model is used for the diagnosis of gentle level driving road surface; the second deep learning model is used for the diagnosis of level 1 bumpiness, level 2 bumpiness and level 3 bumpiness driving road surface; the third deep learning model is used for the diagnosis of abnormal bumpiness driving road surface.

[0055] S3. Collect the fault signal of the transmission device through the sensor.

[0056] In this embodiment, the fault signal includes: a transmission vibration signal, a transmission operating temperature signal, a transmission operating noise signal and transmission external wear image data.

[0057] The transmission vibration signal can reflect the wear, looseness, imbalance and other faults inside the transmission.

[0058] The operating temperature signal of the transmission device, especially the easily worn parts such as bearings and gears. Abnormal temperature rise may mean faults such as poor lubrication, overload or wear.

[0059] Transmission operation noise signals, especially cylinder combustion noise, gear meshing noise, etc. Noise data can reflect abnormal friction, impact and other faults inside the transmission.

[0060] S4. Preprocess the collected fault signal.

[0061] The preprocessing method for the transmission device vibration signal, transmission device operating temperature signal and transmission device operating noise signal is as follows:

[0062] S41. Eliminate abnormal values, repeated values, and missing values ​​of statistical characteristics of the transmission device vibration signal, the transmission device operating temperature signal, and the transmission device operating noise signal.

[0063] S42, smoothing the transmission device vibration signal, the transmission device operating temperature signal, and the transmission device operating noise signal.

[0064] In step S42, smoothing is performed, and the specific method is as follows:

[0065] S421, determining a smoothing coefficient range, where the smoothing coefficient ranges from 0 to 1; the larger the smoothing coefficient, the closer the smoothed data is to the original data, and the smaller the smoothing coefficient, the smoother the smoothed data is;

[0066] S422, setting a candidate set within the smoothing coefficient range;

[0067] S423. For each smoothing coefficient in the candidate set, use the training set data to perform smoothing index prediction, and calculate the mean square error and mean absolute error of the prediction result on the test set;

[0068] S424. Select an optimal smoothing coefficient according to the mean square error and the mean absolute error.

[0069] In this embodiment, the evaluation index of the optimal smoothing coefficient is specifically formulated as follows:

[0070] S t =α.X t +(1-α).S t-1

[0071] Among them, X t is the original data value at time t, S t-1 is the smoothing value at time t-1, and α is the smoothing coefficient.

[0072] S43, extracting statistical features of the transmission device vibration signal, the transmission device operating temperature signal, and the transmission device operating noise signal, where the statistical features include mean features, variance features, and peak features.

[0073] S44. Perform time synchronization on the extracted statistical features.

[0074] In this embodiment, the transmission device external wear image data is adjusted to a fixed size through interpolation method; then image features are extracted through convolutional neural network, and the transmission device external wear image data processing model includes: input layer, convolution layer, pooling layer, fully connected layer and output layer.

[0075] Input layer: receives preprocessed image data.

[0076] Convolutional layer 1: Use multiple convolution kernels to perform convolution operations to extract low-level features of the image.

[0077] Pooling layer 1: Pooling operation is performed on the output of convolutional layer 1 to reduce the size of the feature map.

[0078] Convolutional layer 2: Use more convolution kernels to perform convolution operations to extract mid-level features of the image.

[0079] Pooling layer 2: Pooling operation is performed on the output of convolutional layer 2.

[0080] Convolutional layer 3 and pooling layer 3: further extract high-level features of the image.

[0081] Fully connected layer: Maps the features extracted by the convolutional layer to the output space to obtain the feature vector of the image.

[0082] S5. Input the preprocessed fault signal into the determined deep learning diagnosis model to complete the fault monitoring of the transmission device of the wheeled carrier platform.

[0083] In this embodiment, the deep learning diagnostic model workflow is as follows: Figure 2 As shown, the deep learning model includes an input layer, a hidden layer, and an output layer;

[0084] The input layers of the first deep learning model and the second deep learning model take the transmission vibration signal, the transmission operating temperature signal and the transmission operating noise signal as input; the third deep learning model takes the transmission operating temperature signal, the transmission operating noise signal and the transmission external wear image data as input; the outputs of the deep learning diagnosis models are all probabilistic predictions of the fault type.

[0085] In this embodiment, the hidden layer is designed as follows:

[0086] The first hidden layer

[0087] Number of neurons: Determined by the input feature dimension and model complexity, usually set to several times the input feature dimension.

[0088] Activation function: ReLU function, increasing nonlinearity.

[0089] Regularization: L1 regularization can be introduced to prevent overfitting.

[0090] The second hidden layer

[0091] Number of neurons: less than the first layer to gradually extract high-level features.

[0092] Activation function: ReLU function.

[0093] Dropout layer: Set the dropout ratio to 0.5 to prevent overfitting.

[0094] The third hidden layer

[0095] Activation function: ReLU function

[0096] Output layer: Use the softmax function to output the probability distribution of fault types.

[0097] The output layer predicts the failure probability of the transmission.

[0098] The output layer is a regression layer.

[0099] The fused data is output in a unified format to facilitate subsequent data analysis and fault prediction. The trained model is deployed to the real-time monitoring system. Through the above steps, a deep learning-based fault prediction model for the transmission device of a wheeled carrier platform can be constructed. The model can monitor the failure probability of the transmission device in real time and accurately, providing strong support for the maintenance and maintenance of the wheeled carrier platform.

[0100] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for monitoring transmission device faults of a wheeled carrier platform based on multi-source data fusion under complex road conditions, characterized in that: The specific method is as follows: Determine the level of bumpiness on the road surface on which the wheeled platform is traveling; Determine the deep learning diagnostic model based on the level of road bumps; Collecting fault signals of the transmission device through sensors; Preprocess the collected fault signals; The preprocessed fault signal is input into the determined deep learning diagnostic model to complete the fault monitoring of the wheeled carrier platform transmission device.

2. The method for monitoring transmission failure of a wheeled carrier platform based on multi-source data fusion under complex road conditions according to claim 1, characterized in that: Determine the bumpiness level of the road surface on which the wheeled carrier platform is traveling. The specific method is as follows: The distance between the road surface and the bottom of the wheeled transport platform is measured in real time by a laser distance measuring sensor to generate a distance data sequence; Set the sliding window, calculate the variance of the distance data sequence of each sliding window, and construct the variance value sequence; Smoothing the variance sequence; According to the variance value sequence, the bumpiness level of the driving road is determined.

3. The method for monitoring transmission failure of a wheeled carrier platform based on multi-source data fusion under complex road conditions as claimed in claim 2, characterized in that: The bumpiness levels of the driving road surface are divided into five levels, namely, a flat level, a level one bumpiness, a level two bumpiness, a level three bumpiness and an abnormal bumpiness.

4. The method for monitoring transmission failure of a wheeled carrier platform based on multi-source data fusion under complex road conditions according to claim 1, characterized in that: The fault signal includes: a transmission vibration signal, a transmission operating temperature signal, a transmission operating noise signal and transmission external wear image data.

5. The method for monitoring transmission device failure of a wheeled carrier platform based on multi-source data fusion under complex road conditions as claimed in claim 4, characterized in that: The preprocessing method of the transmission device vibration signal, transmission device operating temperature signal and transmission device operating noise signal is as follows: Eliminate abnormal values, repeated values ​​and missing values ​​of statistical characteristics of transmission device vibration signals, transmission device operating temperature signals and transmission device operating noise signals; Smoothing the transmission device vibration signal, the transmission device operating temperature signal and the transmission device operating noise signal; Extracting statistical features of the transmission device vibration signal, the transmission device operating temperature signal and the transmission device operating noise signal, the statistical features including mean features, variance features and peak features; The extracted statistical features are time synchronized.

6. The method for monitoring transmission failure of a wheeled carrier platform based on multi-source data fusion under complex road conditions as claimed in claim 5, characterized in that: The transmission device vibration signal, transmission device operating temperature signal and transmission device operating noise signal are smoothed. The specific method is as follows: Determine the range of the smoothing coefficient, which is between 0 and 1. The larger the smoothing coefficient, the closer the smoothed data is to the original data, and the smaller the smoothing coefficient, the smoother the smoothed data is. Within the range of smoothing coefficients, set the candidate set; For each smoothing coefficient in the candidate set, use the training set data to predict the smoothing index, and calculate the mean square error and mean absolute error of the prediction results on the test set; Select the optimal smoothing coefficient based on the mean square error and mean absolute error.

7. The method for monitoring transmission failure of a wheeled carrier platform based on multi-source data fusion under complex road conditions as claimed in claim 4, characterized in that: The preprocessing method of transmission device external wear image data is as follows: The transmission device external wear image data is adjusted to a fixed size by using an interpolation method; Image feature extraction is performed through convolutional neural networks.

8. The method for monitoring transmission device failure of a wheeled carrier platform based on multi-source data fusion under complex road conditions as claimed in claim 3, characterized in that: There are three deep learning diagnostic models; The first deep learning model is used for diagnosis of flat grade driving road surfaces; The second deep learning model is used for the diagnosis of driving road surfaces with level 1 bumps, level 2 bumps, and level 3 bumps; The third deep learning model is used for abnormally bumpy driving road surface diagnosis.

9. The method for monitoring transmission device failure of a wheeled carrier platform based on multi-source data fusion under complex road conditions according to claim 7, characterized in that: The deep learning diagnostic model includes an input layer, a hidden layer, and an output layer; The input layers of the first deep learning model and the second deep learning model take the transmission vibration signal, the transmission operation temperature signal and the transmission operation noise signal as input; The third deep learning model takes the transmission operating temperature signal, transmission operating noise signal and transmission external wear image data as input; the output of the deep learning diagnosis model is the probabilistic prediction of the fault type.

Images