Fault detection method, device, equipment, medium and product

Abstract

The invention discloses a fault detection method, device and equipment, a medium and a product. The method comprises the following steps: acquiring an acceleration signal of a rotating tool; determining a target image according to the acceleration signal; and inputting the target image into a fault detection model to obtain a fault detection result of the bearing in the rotating tool. Through the technical scheme of the invention, real-time detection can be carried out on the bearing fault of the near-drilling emission rotation test tool through the acceleration signal, so that timely shutdown can be carried out without human guarding when the bearing has a fault, the problem that manual inspection is needed when an existing near-drilling emission module carries out rotation test is solved, and the detection efficiency is improved. Meanwhile, the problems of man-made misjudgment, detection hysteresis and the like existing in existing fault detection are solved, the bearing fault of the near-drilling rotation test tool is effectively monitored, meanwhile, the rotation test of the near-drilling emission module can be conducted for 24 hours, the monitoring robustness is improved, and finally the intelligent level and reliability of the near-drilling emission rotation test are improved.

Description

Technical Field

[0001] The embodiments of the present invention relate to the technical field of bearing fault detection, and in particular to a fault detection method, device, equipment, medium and product. Background Art

[0002] The near-bit rotary test tooling is a key device for the ex-factory inspection of the near-bit launch module. The bearing is a component that drives its normal operation. Once faults such as jamming and fracture occur, it will directly affect the test of the near-bit launch module. If the machine cannot be stopped in time, it will damage the test tooling and the near-bit launch module, which will cause huge economic losses to the enterprise. The environment in the oil instrument test workshop is harsh and there are many interferences. At present, there is little research on the fault diagnosis of the near-bit rotary test tooling. The existing methods mainly rely on manual inspection for fault diagnosis, with low efficiency. Based on the existing bearing fault diagnosis methods for the near-bit rotary test tooling, faults cannot be diagnosed well, and the real-time performance is poor. There is an urgent need to design an intelligent bearing fault diagnosis method for the near-bit rotary test tooling in combination with the complex industrial field environment. Summary of the Invention

[0003] The embodiments of the present invention provide a fault detection method, device, equipment, medium and product, so as to realize real-time detection of the bearing faults of the near-bit launch rotary test tooling, and improve the intelligent level and reliability of the near-bit launch rotary test.

[0004] According to one aspect of the present invention, there is provided a fault detection method, including:

[0005] Obtain the acceleration signal of the rotary tooling;

[0006] Determine a target image according to the acceleration signal;

[0007] Input the target image into a fault detection model to obtain the fault detection result of the bearing in the rotary tooling.

[0008] According to another aspect of the present invention, there is provided a fault detection device, which includes:

[0009] An acquisition module, configured to acquire the acceleration signal of the rotary tooling;

[0010] A determination module, configured to determine a target image according to the acceleration signal;

[0011] An input module, configured to input the target image into a fault detection model to obtain the fault detection result of the bearing in the rotary tooling.

[0012] According to another aspect of the present invention, there is provided an electronic device, which includes:

[0013] At least one processor; and

[0014] A memory communicatively connected to the at least one processor; wherein,

[0015] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to execute the fault detection method according to any embodiment of the present invention.

[0016] According to another aspect of the present invention, there is provided a computer-readable storage medium storing computer instructions for causing a processor to implement the fault detection method according to any embodiment of the present invention when executed.

[0017] According to another aspect of the present invention, an embodiment of the present invention further provides a computer program product, which includes a computer program that implements the fault detection method according to any embodiment of the present invention when executed by a processor.

[0018] In the embodiment of the present invention, an acceleration signal of a rotating tooling is acquired, a target image is determined according to the acceleration signal, and the target image is input into a fault detection model to obtain a fault detection result of a bearing in the rotating tooling. Through the technical solution of the present invention, it is possible to perform real-time detection of the bearing fault of the near-drill emission rotating test tooling through the acceleration signal, so as to stop the machine in time when the bearing fails without human supervision, solving the problem that manual inspection is required during the rotation test of the existing near-drill emission module. At the same time, the problems of human misjudgment and detection lag in the existing fault detection are solved. While effectively monitoring the bearing fault of the near-drill rotating test tooling, the rotation test of the near-drill emission module can be carried out for 24 hours, improving the robustness of the monitoring, and ultimately improving the intelligent level and reliability of the near-drill emission rotation test.

[0019] It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the present invention, nor is it used to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. Description of the Drawings

[0020] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required in the embodiments. It should be understood that the following drawings only show some embodiments of the present invention, and therefore should not be regarded as limiting the scope. For those of ordinary skill in the art, without creative efforts, other related drawings can be obtained based on these drawings.

[0021] Figure 1 is a flowchart of a fault detection method in an embodiment of the present invention;

[0022] Figure 2 It is the process structure diagram of a convolutional neural network model in an embodiment of the present invention;

[0023] Figure 3 It is the flowchart of an exemplary fault detection method in an embodiment of the present invention;

[0024] Figure 4 It is the structural schematic diagram of a fault detection device in an embodiment of the present invention;

[0025] Figure 5 It is the structural schematic diagram of an electronic device for implementing the fault detection method of the embodiment of the present invention. Detailed implementation manners

[0026] In order to enable those skilled in the art to better understand the solution of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a 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 those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0027] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily need to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so that the embodiments of the present invention described here can be implemented in an order other than those illustrated or described here. In addition, the terms "comprising" and "having" and their derivatives are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device including a series of steps or units does not necessarily need to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0028] It can be understood that before using the technical solutions disclosed in the embodiments of the present disclosure, the types, usage scopes, usage scenarios, etc. of the personal information involved in the present disclosure should be informed to users and the authorization of users should be obtained in an appropriate manner in accordance with relevant laws and regulations.

[0029] Embodiment 1

[0030] Figure 1 It is the flowchart of a fault detection method in an embodiment of the present invention. This embodiment is applicable to the situation of bearing fault detection of a near-drill rotary test tooling. This method can be executed by the fault detection device in the embodiment of the present invention. The device can be implemented in a software and / or hardware manner, such asFigure 1 As shown in the figure, the method specifically includes the following steps:

[0031] S101. Obtain the acceleration signal of the rotating tooling.

[0032] In this embodiment, the rotating tooling is the near-drill rotating test tooling. It can be known that the near-drill rotating test tooling is a key device for the ex-factory inspection of the near-drill emission module, and the bearing is the component that drives its normal operation. This embodiment mainly conducts fault detection on the operating state of the bearing.

[0033] Among them, the acceleration signal can refer to the MEMS accelerometer acceleration signal collected by the near-drill emission module. Among them, MEMS (Micro-Electro-Mechanical Systems) is a micro-system that integrates micro-sensors, micro-actuators, and signal processing and control circuits. It integrates mechanical structures and electronic circuits on the same chip, realizing the deep integration of mechanical and electronic functions. MEMS technology is widely used in fields such as consumer electronics, automotive electronics, aerospace, biomedicine, and industrial automation.

[0034] In the actual operation process, the step of obtaining the acceleration signal of the rotating tooling can be described as the following two steps:

[0035] Step A. Install the near-drill emission module and the battery onto the near-drill emission rotating test tooling, then cover the cover plate and tighten the bolts.

[0036] Specifically, first write the MEMS accelerometer sensor control program, RS-232 communication program, and data acquisition program for the near-drill emission module, download the programs into the near-drill emission module, connect the near-drill emission module to the appropriate port of the upper computer through RS-232; then install the near-drill emission module into the near-drill emission rotating test tooling and install the battery into the near-drill emission rotating test tooling; finally, connect the RS-232 (Recommended Standard 232, a serial communication protocol widely used for data transmission between computers, industrial equipment, and embedded systems) communication line to the industrial control computer through the slip ring and the connector through the serial communication box, and collect the acceleration signal during the operation of the near-drill emission module in real time through the RS232 serial port.

[0037] Step B. Connect the TX (Transmit, for sending data), RX (Receive, for receiving data), and GND (Ground, for providing a grounding reference) pins of the RS-232 of the near-drill emission module to the industrial control computer through the slip ring, so that the MEMS accelerometer acceleration signal collected by the near-drill emission module can be transmitted to the industrial control computer, and save the acceleration signal of the near-drill emission module in the database.

[0038] Specifically, write the RS-232 serial communication protocol for the near-drill emission module, including the frame header, address, function code, checksum, and frame tail, so that it can be connected to the industrial control computer through the serial port, so as to transmit the acceleration signal of the near-drill emission module to the industrial control computer through RS-232; then connect the RS-232 serial cable to the serial communication box connected to the industrial control computer through the slip ring and connector, and transmit the acceleration signal of the near-drill emission module during operation to the industrial control computer through the serial port for fault diagnosis; finally, obtain the acceleration signal data of the near-drill emission module during operation collected on the industrial control computer, and the industrial control computer saves the received acceleration signal data of the near-drill emission module to the database.

[0039] During the actual operation process, the operating state of the near-drill emission rotation test tooling can be judged by processing the acceleration signal data of the near-drill emission module every minute.

[0040] S102. Determine the target image according to the acceleration signal.

[0041] In this embodiment, when detecting the fault of the bearing in the rotating tooling, it is necessary to convert the acceleration signal data into a two-dimensional image, and then determine the fault detection result of the bearing in the rotating tooling according to the two-dimensional image. Among them, the target image can be a two-dimensional image obtained after converting the acceleration signal.

[0042] Specifically, after obtaining the acceleration signal, convert the acceleration signal to obtain the target image.

[0043] In a specific embodiment, the process of determining the target image according to the acceleration signal can be described as follows: preprocess the acceleration signal to obtain the target component of the acceleration signal; perform normalization processing on the target component to obtain the target sequence; perform sequential filling on the target sequence and convert it into the target image.

[0044] Further, the process of preprocessing the acceleration signal to obtain the target component of the acceleration signal, performing normalization processing on the target component to obtain the target sequence, and performing sequential filling on the target sequence and converting it into the target image can be described as follows: use the ensemble empirical mode decomposition method to preprocess the acceleration signal to obtain the IMF (Intrinsic Mode Functions) components of the acceleration signal; perform normalization processing on the IMF components obtained by the ensemble empirical mode decomposition method; perform sequential filling on the preprocessed IMF components and convert them into two-dimensional images.

[0045] S103. Input the target image into the fault detection model to obtain the fault detection result of the bearing in the rotating tooling.

[0046] Among them, the fault detection model can be a model for fault detection based on the target image. Preferably, in this embodiment, the fault detection model can be, for example, a trained convolutional neural network model.

[0047] In this embodiment, the fault detection results of the bearings in the rotation tooling can include, for example: the bearings of the rotation tooling are stuck, the bearings of the rotation tooling are broken, and the cover plate is loose.

[0048] Specifically, the obtained two-dimensional image data is input into the trained convolutional neural network model for judgment, and the state of the near-drill emission rotation test tooling is judged through the output result of the convolutional neural network model, and the fault detection results of the bearings in the rotation tooling are determined.

[0049] In the actual operation process, after judging the running state of the near-drill emission rotation test tooling through the convolutional neural network model, the real-time running state of the near-drill emission rotation test tooling can also be displayed on the industrial control computer interface. Exemplarily, various data including the fault diagnosis information of the near-drill emission rotation test tooling can be displayed on the software interface of the fault diagnosis system of the near-drill emission rotation test tooling on the industrial control computer, and the current state and historical fault information can be displayed in real time.

[0050] In the actual operation process, when it is detected that a near-drill emission rotation test tooling fails, the motor will be shut down to stop rotating, and an alarm message will be sent to prompt the staff that the near-drill emission rotation test tooling has failed and corresponding countermeasures should be taken.

[0051] In the embodiment of the present invention, the acceleration signal of the rotation tooling is obtained, the target image is determined according to the acceleration signal, and the target image is input into the fault detection model to obtain the fault detection results of the bearings in the rotation tooling. Through the technical solution of the present invention, the bearing faults of the near-drill emission rotation test tooling can be detected in real time through the acceleration signal, so that the machine can be stopped in time without human supervision when the bearing fails, solving the problem of manual inspection required during the rotation test of the existing near-drill emission module. At the same time, the problems of human misjudgment and detection lag in the existing fault detection are solved. While effectively monitoring the bearing faults of the near-drill rotation test tooling, the rotation test of the near-drill emission module can be carried out for 24 hours, improving the robustness of the monitoring, and finally improving the intelligent level and reliability of the near-drill emission rotation test.

[0052] Optionally, determining the target image according to the acceleration signal includes:

[0053] Preprocess the acceleration signal to obtain the target component of the acceleration signal.

[0054] In this embodiment, the target component can be the IMF component of the acceleration signal.

[0055] It is known that the IMF component is a core concept in EMD (Empirical Mode Decomposition). EMD is an adaptive time-frequency analysis method suitable for processing non-linear and non-stationary signals. The IMF component is the basic component extracted from the original signal (in this embodiment, the original signal is the acceleration signal) through EMD, and each IMF component represents the characteristics of the signal at different frequencies and time scales.

[0056] In the actual operation process, the acceleration signal data of the near-drill emission module collected on the industrial control computer, with a frequency of 250 hz, is obtained. Through the analysis of the time-frequency domain of the data, it is known that the high-frequency signal part of the acceleration changes significantly during a fault.

[0057] Specifically, the ensemble empirical mode decomposition is used to preprocess the acceleration signal data of each group of idler runs to obtain the IMF components of the acceleration signal.

[0058] The target component is normalized to obtain the target sequence.

[0059] In this embodiment, the target component can be in the form of a sequence. Among them, the target sequence can be the sequence obtained after normalizing the IMF components obtained by the ensemble empirical mode decomposition method.

[0060] Specifically, the IMF components obtained by the ensemble empirical mode decomposition method are normalized to obtain the target sequence, preparing for the subsequent generation of images.

[0061] The target sequence is sequentially filled to be transformed into the target image.

[0062] In this embodiment, the sequential filling of the target sequence into the target image is essentially the conversion of a one-dimensional array (the target sequence) into a two-dimensional matrix (the target image is a two-dimensional image).

[0063] Specifically, the target sequence is sequentially filled to be transformed into a two-dimensional image.

[0064] Optionally, the preprocessing of the acceleration signal to obtain the target component of the acceleration signal includes:

[0065] The acceleration signal is decomposed by the ensemble empirical mode decomposition method to obtain the target component of the acceleration signal.

[0066] It is known that Ensemble Empirical Mode Decomposition (EEMD) is an improved Empirical Mode Decomposition (EMD) method. EEMD solves the problem of mode mixing that may occur in EMD by adding white noise to the original signal, thereby improving the stability and accuracy of decomposition. EEMD is suitable for processing non-stationary signals and can adaptively decompose the signal into multiple Intrinsic Mode Functions (IMFs).

[0067] In this embodiment, the Ensemble Empirical Mode Decomposition method is used to preprocess the acceleration data of the near-drill emission module to obtain the IMF components of the acceleration signal.

[0068] Optionally, the acceleration signal is decomposed by the Ensemble Empirical Mode Decomposition method to obtain the target components of the acceleration signal, including:

[0069] Obtain a set of white noise, and add each white noise in the set of white noise to the acceleration signal respectively to obtain a set of acceleration signals.

[0070] In this embodiment, the set of white noise can be a set composed of several white noises obtained by setting white noises obeying Gaussian distribution to different amplitudes respectively.

[0071] Among them, the set of acceleration signals can be a set composed of acceleration signals obtained by adding the white noises in the set of white noise to the acceleration signal respectively.

[0072] In the actual operation process, the original acceleration signal can be set as x(t), the number of times N of adding white noise to x(t) and the amplitude A of the white noise are set, and the white noise Wn(t) obeying Gaussian distribution is added to x(t) to generate N new acceleration signals, obtaining a set of acceleration signals. That is, the set of white noise includes N white noises with different amplitudes A, and the set of acceleration signals includes N updated acceleration signals.

[0073] Decompose each updated acceleration signal in the set of acceleration signals respectively to obtain the initial components corresponding to each updated acceleration signal.

[0074] It should be noted that the updated acceleration signal can be an acceleration signal formed by adding white noise to the original acceleration signal.

[0075] Among them, the initial component can be the IMF component corresponding to the updated acceleration signal obtained by performing EMD decomposition on the updated acceleration signal.

[0076] In the actual operation process, for the N updated acceleration signals obtained, EMD decomposition is performed on each of them in turn to obtain the IMF components corresponding to each of the N updated acceleration signals.

[0077] The average value of the initial components corresponding to each updated acceleration signal is calculated to obtain the target component.

[0078] Specifically, the average value of the IMF components of the corresponding mode is calculated to obtain a series of IMF components of EEMD.

[0079] Optionally, the target component is normalized to obtain a target sequence, including:

[0080] Obtain the maximum value and the minimum value in the target component.

[0081] In this embodiment, the target component can be in the form of a sequence. Suppose the target component obtained by ensemble empirical mode decomposition is a sequence I(t) of length t, and find its maximum value and minimum value I max and I min .

[0082] Subtract the maximum value from each data point in the target component to obtain an initial sequence.

[0083] It should be noted that the initial sequence can be a new sequence obtained by subtracting the maximum value in the target component from each data point in the target component sequence.

[0084] Specifically, the calculation method of the initial sequence I new (t) can be expressed as: I new (t) = I max - I(t).

[0085] The initial sequence is uniformly mapped to the image gray value interval to obtain the target sequence.

[0086] It can be known that the image gray value interval is [0, 255].

[0087] Specifically, since the image gray value interval is [0, 255], the values of the obtained initial sequence are uniformly mapped between [0, 255] to obtain the target sequence.

[0088] Optionally, the initial sequence is uniformly mapped to the image gray value interval to obtain the target sequence, including:

[0089] Subtract the minimum value from the maximum value to obtain the target difference.

[0090] Among them, the target difference can be the maximum value in the target component minus the minimum value I max and Imin The difference obtained by taking the difference.

[0091] The product of the quotient of each data point in the initial sequence and the target difference and the target value forms the target sequence.

[0092] Among them, the target value is the maximum value of the image gray value interval. Preferably, the target value is 255.

[0093] In this embodiment, I now (t) can be used to represent the target sequence. Specifically, the calculation method of the target sequence I now (t) can be expressed as:

[0094] Optionally, performing sequential filling on the target sequence to convert it into a target image includes:

[0095] Decompose the target sequence into a set of subsequences.

[0096] Among them, the set of subsequences can be a set composed of several subsequences obtained by decomposing the target sequence according to the size of the target image.

[0097] Specifically, decompose the target sequence according to the size of the target image to obtain a set of subsequences.

[0098] Construct the target image according to the set of subsequences.

[0099] Specifically, construct a two-dimensional image according to each data point in the set of subsequences.

[0100] Optionally, decomposing the target sequence into a set of subsequences includes:

[0101] Obtain the size information of the target image.

[0102] Among them, the size information includes a length value and a width value.

[0103] In the specific implementation process, the target image is a two-dimensional image. It can be assumed that the size information of the target image is n×m, that is, the length value of the target image is n and the width value is m.

[0104] Decompose the target sequence into a first number of subsequences with a second number of lengths to obtain a set of subsequences.

[0105] Among them, the first number is the length value and the second number is the width value.

[0106] Specifically, decompose the target sequence I now (t) into n subsequences with a length of m, that is Among them, r does not participate in the construction of the target image and is the data point that has not been extracted.

[0107] Finally, the obtained target image can be expressed as:

[0108] Optionally, the fault detection model is obtained by iteratively training a convolutional neural network model with a target sample set.

[0109] Wherein, the target sample set includes: an image sample set and the corresponding target fault types of the image sample set.

[0110] In this embodiment, the image sample set can be a two-dimensional image generated after processing the previously collected historical acceleration signals. The processing process is the same as that of the method in this embodiment and will not be elaborated here one by one. Specifically, the image sample set can be divided into a training set, a validation set, and a test set to support the training of the model, the adjustment of hyperparameters, and the evaluation of the generalization ability. Specifically, the training set is used to train the model to learn the feature representation of the images; the validation set is used to adjust hyperparameters (such as learning rate, regularization parameter, etc.) and select the optimal model structure; the test set is used to evaluate the generalization ability of the model, that is, the performance of the model on unseen data.

[0111] It should be noted that the target fault types can be the bearing fault types corresponding to each image sample in the image sample set. In the actual operation process, the target fault types can be pre-annotated by the user according to experience. Exemplarily, the target fault type corresponding to image sample A is bearing seizure, and the target fault type corresponding to image sample B is bearing fracture. The fault types of the image samples can be pre-annotated.

[0112] Iteratively training the convolutional neural network model with the target sample set includes:

[0113] Establish a convolutional neural network model.

[0114] In this embodiment, an efficient convolutional neural network model is constructed, and image features are extracted through convolutional layers, pooling layers, and residual blocks.

[0115] Input the image samples in the target sample set into the convolutional neural network model to obtain predicted fault types.

[0116] Wherein, the predicted fault types can be the bearing fault types output by the convolutional neural network model after predicting the fault types based on the input image samples.

[0117] Figure 2 is a flow structure diagram of a convolutional neural network model in an embodiment of the present invention. As Figure 2As shown, in the specific implementation process, the network structure that the input image sample goes through includes: First, it passes through a convolutional layer 1 with a convolutional kernel size of 5×5, 32 convolutional kernels, and a stride of 2 (denoted as / 2 in the figure), and the output size is halved to extract preliminary features. Then, the convolutional result passes through a max-pooling layer with a pooling kernel size of 2×2 and a stride of 2 to further reduce the size of the feature map and reduce the computational amount. After that, it passes through two ResBlock (residual block) structures (residual block 1 and residual block 2). Each ResBlock contains two Basic Block (basic module), and each Basic Block includes two 3×3 convolutional layers, and the number of convolutional kernels is the same as the upper-layer input. Next, the ReLU activation function and the BN layer (Batch Normalization layer) are used. Finally, a residual connection is made: the input x is added to the output O(x) to enhance the network's non-linear ability and training stability.

[0118] Among them, the basic structure of the ResBlock is that the input x passes through two Basic Blocks and the ReLU function to obtain the output O(x). The basic structure of the Basic Block is that the input x first passes through a 3×3 convolutional kernel to obtain feature information, then the BN layer is used to standardize the parameters, and then the ReLU activation function introduces non-linearity. Then it passes through another 3×3 convolutional kernel to extract deeper features. Finally, the output O(x) is added to the input x to form a residual connection, avoiding the problem of gradient disappearance.

[0119] Among them, two 3×3 convolutional kernels are continuously used in the Basic block to achieve the convolutional effect of a 5×5 convolutional kernel. The advantage of this is that it not only increases the receptive field range and improves the network's operation efficiency, but also reduces the number of parameters, which is much less than that of a 5×5 convolutional kernel.

[0120] In the specific implementation process, subsequently, to further process the features and reduce the dimension in preparation for the classification task, it passes through a convolutional layer 2 with a convolutional kernel size of 3×3, 128 convolutional kernels, and a stride of 2 to increase the feature dimension and further reduce the size of the feature map. Subsequently, through GAP (Global Average Pooling), global average pooling is performed on each channel of the feature map to obtain a fixed-length feature vector. After dimensionality reduction, two FC (Fully Connected Layer) (fully connected layer 1 and fully connected layer 2, where fully connected layer 1 contains 256 neurons and fully connected layer 2 contains 2 neurons) are used to reduce the number of network parameters. Subsequently, the outputs of the two fully connected layers are concatenated together and classified through softmax to obtain the output.

[0121] Train the parameters of the convolutional neural network model with an objective function formed based on the predicted fault type and the target fault type corresponding to the image sample set.

[0122] In this embodiment, the objective function can be a loss function, and the specific form of the loss function is not limited in this embodiment. Specifically, train the parameters of the convolutional neural network model with a loss function formed based on the predicted fault type and the target fault type corresponding to the image sample set.

[0123] Return and perform the operation of inputting the image samples in the target sample set into the convolutional neural network model to obtain the predicted fault type until a fault detection model is obtained.

[0124] Specifically, loop through the above steps to train the convolutional neural network model until the iteration termination condition is met, such as convergence or the number of iterations reaches a threshold (the specific threshold can be preset by the user according to the actual situation or empirical values, which is not limited in this embodiment), and obtain a trained fault detection model.

[0125] As an exemplary description of an embodiment of the present invention, Figure 3 is a flowchart of an exemplary fault detection method in an embodiment of the present invention. As Figure 3 shown, the process of the fault detection method can be described as follows:

[0126] Step a: Collect acceleration signals;

[0127] Step b: Perform ensemble empirical mode decomposition on the acceleration signals to obtain target components;

[0128] Step c: Normalize the target components to obtain a target sequence;

[0129] Step d: Sequentially pad the target sequence and convert it into a target image;

[0130] Step e: Input the trained convolutional neural network model;

[0131] Step f: Obtain the bearing fault type.

[0132] The technical solution of the embodiment of the present invention collects acceleration signals through MEMS plus-meter measurement of the near-drill emission module, decomposes the received acceleration signals using the ensemble empirical mode decomposition method, converts the IMF components obtained by the ensemble empirical mode decomposition into images, and inputs the images into a pre-trained fault detection model for detecting bearing faults of the near-drill emission rotation test tooling to obtain the bearing fault detection result. Through the technical solution of the present invention, the bearing faults of the near-drill emission rotation test tooling can be detected in real time through acceleration signals, so as to stop the machine in time without human supervision when the bearing fails, solving the problem of manual inspection required during the rotation test of the existing near-drill emission module. At the same time, problems such as human misjudgment and detection lag in the existing fault detection are solved. While effectively monitoring the bearing faults of the near-drill rotation test tooling, the rotation test of the near-drill emission module can be carried out for 24 hours, improving the robustness of the monitoring, and ultimately improving the intelligent level and reliability of the near-drill emission rotation test.

[0133] Embodiment 2

[0134] Figure 4 It is a schematic structural diagram of a fault detection device in an embodiment of the present invention. This embodiment is applicable to the situation of bearing fault detection of the near-drill rotation test tooling. The device can be implemented in software and / or hardware, and can be integrated in any device providing fault detection functions, such as Figure 4 As shown, the fault detection device specifically includes: an acquisition module 201, a determination module 202, and an input module 203.

[0135] Among them, the acquisition module 201 is used to acquire the acceleration signal of the rotation tooling;

[0136] The determination module 202 is used to determine the target image according to the acceleration signal;

[0137] The input module 203 is used to input the target image into the fault detection model to obtain the fault detection result of the bearing in the rotation tooling.

[0138] Optionally, the determination module 202 includes:

[0139] A preprocessing unit for preprocessing the acceleration signal to obtain the target component of the acceleration signal;

[0140] A normalization unit for normalizing the target component to obtain a target sequence;

[0141] An order filling unit for filling the target sequence in order and converting it into a target image.

[0142] Optionally, the preprocessing unit includes:

[0143] The first decomposition subunit is configured to decompose the acceleration signal by using the ensemble empirical mode decomposition method to obtain the target component of the acceleration signal.

[0144] Optionally, the first decomposition subunit is specifically configured to:

[0145] Obtain a white noise ensemble, and respectively add each white noise in the white noise ensemble to the acceleration signal to obtain an acceleration signal ensemble;

[0146] Respectively decompose each updated acceleration signal in the acceleration signal ensemble to obtain the initial component corresponding to each updated acceleration signal;

[0147] Average the initial components corresponding to each updated acceleration signal to obtain the target component.

[0148] Optionally, the normalization unit includes:

[0149] An acquisition subunit, configured to acquire the maximum value and the minimum value in the target component;

[0150] A difference-making subunit, configured to subtract each data point in the target component from the maximum value to obtain an initial sequence;

[0151] A mapping subunit, configured to uniformly map the initial sequence to the image gray value interval to obtain a target sequence.

[0152] Optionally, the mapping subunit is specifically configured to:

[0153] Subtract the minimum value from the maximum value to obtain a target difference;

[0154] Multiply the quotient of each data point in the initial sequence and the target difference by a target value to form a target sequence; the target value is the maximum value of the image gray value interval.

[0155] Optionally, the sequential filling unit includes:

[0156] A second decomposition subunit, configured to decompose the target sequence into a subsequence ensemble;

[0157] A construction subunit, configured to construct a target image according to the subsequence ensemble.

[0158] Optionally, the second decomposition subunit is specifically configured to:

[0159] Obtain the size information of the target image, where the size information includes a length value and a width value;

[0160] Decompose the target sequence into a first numerical number of subsequences each with a length of a second numerical value, obtaining a set of subsequences; the first numerical value is the length value, and the second numerical value is the width value.

[0161] Optionally, the fault detection model is obtained by iteratively training a convolutional neural network model with a target sample set, where the target sample set includes: an image sample set and the corresponding target fault types of the image sample set;

[0162] The device is used for:

[0163] Establish a convolutional neural network model;

[0164] Input the image samples in the target sample set into the convolutional neural network model to obtain predicted fault types;

[0165] Train the parameters of the convolutional neural network model according to an objective function formed by the predicted fault types and the target fault types corresponding to the image sample set;

[0166] Return to perform the operation of inputting the image samples in the target sample set into the convolutional neural network model to obtain predicted fault types until a fault detection model is obtained.

[0167] The above product can execute the fault detection method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0168] Embodiment III

[0169] Figure 5 FIG. shows a schematic structural diagram of an electronic device 30 that can be used to implement the embodiments of the present invention. The electronic device is intended to represent various forms of digital computers, such as, laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as, personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or claimed herein.

[0170] As Figure 5As shown, the electronic device 30 includes at least one processor 31 and a memory communicatively connected to the at least one processor 31, such as a read-only memory (ROM) 32, a random access memory (RAM) 33, etc. The memory stores a computer program executable by the at least one processor. The processor 31 can execute various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 32 or the computer program loaded from the storage unit 38 into the random access memory (RAM) 33. In the RAM 33, various programs and data required for the operation of the electronic device 30 can also be stored. The processor 31, the ROM 32, and the RAM 33 are connected to each other via a bus 34. An input / output (I / O) interface 35 is also connected to the bus 34.

[0171] Multiple components in the electronic device 30 are connected to the I / O interface 35, including: an input unit 36, such as a keyboard, a mouse, etc.; an output unit 37, such as various types of displays, speakers, etc.; a storage unit 38, such as a disk, an optical disc, etc.; and a communication unit 39, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 39 allows the electronic device 30 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0172] The processor 31 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the processor 31 include but are not limited to a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 31 executes the various methods and processes described above, such as a fault detection method:

[0173] Obtain the acceleration signal of the rotary tooling;

[0174] Determine a target image according to the acceleration signal;

[0175] Input the target image into a fault detection model to obtain the fault detection result of the bearing in the rotary tooling.

[0176] In some embodiments, the fault detection method may be implemented as a computer program tangibly embodied in a computer-readable storage medium, such as storage unit 38. In some embodiments, part or all of the computer program may be loaded and / or installed onto the electronic device 30 via the ROM 32 and / or the communication unit 39. When the computer program is loaded into the RAM 33 and executed by the processor 31, one or more steps of the fault detection method described above may be performed. Alternatively, in other embodiments, the processor 31 may be configured to perform the fault detection method by any other suitable means (e.g., by means of firmware).

[0177] The various embodiments of the systems and techniques described above in this document can be implemented in digital electronic circuitry, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-chip (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special-purpose or general-purpose programmable processor that receives data and instructions from a storage system, at least one input device, and at least one output device, and transmits the data and instructions to the storage system, the at least one input device, and the at least one output device.

[0178] The computer programs for implementing the methods of the present invention can be written in any combination of one or more programming languages. These computer programs can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when the computer programs are executed by the processor, the functions / operations specified in the flowcharts and / or block diagrams are implemented. The computer programs can be executed entirely on the machine, partially on the machine, as a stand-alone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0179] In the context of the present invention, a computer-readable storage medium can be a tangible medium that can contain or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. The computer-readable storage medium can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. Alternatively, the computer-readable storage medium can be a machine-readable signal medium. More specific examples of the machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0180] To provide for interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the electronic device. Other kinds of devices can also be used to provide for interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, speech input, or tactile input).

[0181] The systems and techniques described herein can be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer having a graphical user interface or a web browser through which the user can interact with an implementation of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0182] A computing system may include a client and a server. The client and the server are generally far from each other and usually interact via a communication network. The client-server relationship is generated by computer programs running on respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or a cloud host, which is a host product in the cloud computing service system, and solves the defects of difficult management and weak business scalability existing in traditional physical hosts and VPS services.

[0183] In one embodiment, the embodiment of the present invention further includes a computer program product, which includes a computer program. When the computer program is executed by a processor, it implements the fault detection method of any embodiment of the present invention.

[0184] In the process of implementation, the computer program product can be written in one or more programming languages or combinations thereof to write computer program code for performing the operations of the present invention. The programming languages include object-oriented programming languages such as Java, Smalltalk, C++, and also include conventional procedural programming languages such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, executed as an independent software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network - including a local area network (LAN) or a wide area network (WAN) - or can be connected to an external computer (for example, by using an Internet service provider to connect through the Internet).

[0185] It should be understood that various forms of the processes shown above can be used, with steps reordered, added, or deleted. For example, the steps recited in the present invention can be executed in parallel, sequentially, or in a different order, as long as the desired results of the technical solution of the present invention can be achieved, and no limitation is imposed herein.

[0186] The above specific embodiments do not constitute a limitation to the protection scope of the present invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A fault detection method, characterized in that: include: Obtaining acceleration signals of rotating tooling; determining a target image according to the acceleration signal; The target image is input into a fault detection model to obtain a fault detection result of the bearing in the rotating tooling.

2. The method according to claim 1, characterized in that Determining a target image according to the acceleration signal includes: Preprocessing the acceleration signal to obtain a target component of the acceleration signal; Normalizing the target components to obtain a target sequence; The target sequence is sequentially filled and converted into a target image.

3. The method according to claim 2, characterized in that Preprocessing the acceleration signal to obtain a target component of the acceleration signal includes: The acceleration signal is decomposed by an ensemble empirical mode decomposition method to obtain a target component of the acceleration signal.

4. The method according to claim 3, characterized in that Decomposing the acceleration signal by using an ensemble empirical mode decomposition method to obtain a target component of the acceleration signal includes: Acquire a white noise set, and add each white noise in the white noise set to the acceleration signal to obtain an acceleration signal set; Decomposing each updated acceleration signal in the acceleration signal set respectively to obtain an initial component corresponding to each updated acceleration signal; The initial components corresponding to each updated acceleration signal are averaged to obtain the target component.

5. The method according to claim 2, characterized in that: The target components are normalized to obtain a target sequence, including: Obtaining the maximum and minimum values ​​of the target component; Subtract the maximum value from each data point in the target component to obtain an initial sequence; The initial sequence is uniformly mapped to the image gray value interval to obtain a target sequence.

6. The method according to claim 5, characterized in that The initial sequence is uniformly mapped to the image gray value interval to obtain a target sequence, including: Subtract the maximum value from the minimum value to obtain a target difference value; The target sequence is formed by multiplying the quotient of each data point in the initial sequence and the target difference value by the target value; the target value is the maximum value of the image grayscale value interval.

7. The method according to claim 2, characterized in that The target sequence is sequentially filled and converted into a target image, including: Decomposing the target sequence into a set of subsequences; A target image is constructed according to the set of subsequences.

8. The method according to claim 7, characterized in that Decomposing the target sequence into a set of subsequences, including: Obtaining size information of a target image, wherein the size information includes a length value and a width value; The target sequence is decomposed into a first value number of subsequences whose length is a second value to obtain a subsequence set; the first value is the length value, and the second value is the width value.

9. The method according to claim 1, characterized in that: The fault detection model is obtained by iteratively training a convolutional neural network model with a target sample set, wherein the target sample set includes: an image sample set and a target fault type corresponding to the image sample set; Iteratively training the convolutional neural network model through the target sample set includes: Build a convolutional neural network model; Inputting the image samples in the target sample set into the convolutional neural network model to obtain a predicted fault type; Training the parameters of the convolutional neural network model according to an objective function formed by the predicted fault type and the target fault type corresponding to the image sample set; Return to executing the operation of inputting the image samples in the target sample set into the convolutional neural network model to obtain the predicted fault type until a fault detection model is obtained.

10. A fault detection device, characterized in that: include: An acquisition module, used for acquiring an acceleration signal of a rotating tooling; A determination module, used to determine a target image according to the acceleration signal; The input module is used to input the target image into the fault detection model to obtain the fault detection result of the bearing in the rotating tooling.

11. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can perform the fault detection method according to any one of claims 1 to 9.

12. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the fault detection method according to any one of claims 1 to 9 when executed.

13. A computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements the fault detection method according to any one of claims 1 to 9.

Images