Gear state diagnosis method and device

Abstract

The application provides a gear state diagnosis method and device, and relates to the technical field of mechanical anomaly diagnosis.The method comprises the following steps: based on the vibration signals in the vibration monitoring data of the gear, source domain data sets and target domain data sets corresponding to the gear are prepared; the source domain data sets comprise a vibration signal set and a noise-added vibration signal set with gear state labels; the target domain data sets comprise a to-be-predicted vibration signal set; based on the noise-added vibration signal set with the gear state labels, the de-noising parameters of a U-shaped network are optimized to obtain an optimized residual U-shaped network; based on the residual U-shaped network, the to-be-predicted vibration signal set is de-noised to determine the gear state according to the obtained de-noised vibration signal, so that the de-noised vibration signal obtained by de-noising the to-be-predicted vibration signal set based on the residual U-shaped network can accurately determine the gear state, the accurate diagnosis of the vibration signals in the early abnormal period vibration monitoring data of the gear is realized, and the safety of the gear operation is ensured.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of mechanical anomaly diagnosis, and in particular to a gear state diagnosis method and device, an electronic device, and a storage medium. BACKGROUND

[0002] Vibration monitoring is one of the most effective tools for gear anomaly diagnosis. When a gear has local damage, different abnormal characteristic frequencies can be generated according to the surface affected by the anomaly. People diagnose the anomaly by analyzing the vibration signal to extract the abnormal characteristic frequency. However, during the early anomaly period of the gear, especially in weak anomaly diagnosis, the local defects and damage of the gear are very small, and the impact vibration caused is very weak. In addition, the interference (collectively referred to as noise) of the surrounding environment makes the vibration signal very complex. In the related art, the gear state diagnosis is a diagnosis method based on deep learning. A large amount of labeled fault data is used for training. However, it takes more manpower and material resources to obtain the labeled data than to simply record them. Moreover, the early anomaly period of the gear is very weak and complex, and the vibration signal cannot be diagnosed, which makes it difficult to achieve ideal results in the early weak anomaly diagnosis. Therefore, there is an urgent need for a more reliable gear state diagnosis method. SUMMARY

[0003] The present application aims to at least partially solve one of the technical problems in the related art.

[0004] To this end, a first object of the present application is to provide a gear state diagnosis method. The method is based on a residual U-shaped network to denoise the vibration signal set to be predicted to obtain a denoised vibration signal, accurately determine the gear state, and realize accurate diagnosis of the vibration signal in the vibration monitoring data during the early anomaly period of the gear, thereby ensuring the safety of the gear operation.

[0005] A second object of the present application is to provide a gear state diagnosis device.

[0006] A third object of the present application is to provide an electronic device.

[0007] A fourth object of the present application is to provide a non-transitory computer readable storage medium storing computer instructions.

[0008] To achieve the above objects, a gear state diagnosis method according to an embodiment of the first aspect of the present application is provided. The method comprises:

[0009] Based on the vibration signal in the vibration monitoring data of the gear, the source domain data set and the target domain data set corresponding to the gear are made. The source domain data set includes a vibration signal set with a gear state label and a noisy vibration signal set. The target domain data set includes a vibration signal set to be predicted.

[0010] The denoising parameters of the U-shaped network are optimized based on the noise-added vibration signal set of the vibration signal set with the gear state label, to obtain an optimized residual U-shaped network.

[0011] Based on the residual U-shaped network, the vibration signal set to be predicted is denoised to obtain a denoised vibration signal, and the gear state is determined based on the denoised vibration signal.

[0012] To achieve the above purpose, the second aspect of the present application provides a gear state diagnosis device, the device comprises:

[0013] The manufacturing module is configured to manufacture source domain data set and target domain data set corresponding to the gear based on the vibration signal in the vibration monitoring data of the gear, wherein the source domain data set comprises a vibration signal set with a gear state label and a noise-added vibration signal set, and the target domain data set comprises a vibration signal set to be predicted.

[0014] The generating module is configured to optimize the denoising parameters of the U-shaped network based on the noise-added vibration signal set of the vibration signal set with the gear state label, to obtain an optimized residual U-shaped network.

[0015] The determining module is configured to denoise the vibration signal set to be predicted based on the residual U-shaped network to obtain a denoised vibration signal, and determine the gear state based on the denoised vibration signal.

[0016] To achieve the above purpose, the third aspect of the present application provides an electronic device, comprising: at least one processor; and a memory connected in communication with the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the method of the first aspect.

[0017] In order to achieve the above purpose, the fourth aspect of the present application provides a non-transitory computer readable storage medium storing computer instructions, the computer instructions are used to make the computer execute the method of the first aspect.

[0018] The gear state diagnosis method, device, electronic equipment and storage medium provided by the embodiment of the present application are based on the vibration signals in the vibration monitoring data of the gear, the source domain data set and the target domain data set corresponding to the gear are made, the source domain data set includes the vibration signal set and the noise-added vibration signal set with the gear state label, and the target domain data set includes the vibration signal set to be predicted; the noise-added vibration signal set with the vibration signal set with the gear state label is used to optimize the denoising parameter of the U-shaped network to obtain the optimized residual U-shaped network; the vibration signal set to be predicted is denoised based on the residual U-shaped network, so that the gear state is determined according to the obtained denoised vibration signal, thereby the denoised vibration signal obtained by denoising the vibration signal set to be predicted based on the residual U-shaped network accurately determines the gear state, realizes the accurate diagnosis of the vibration signal in the early abnormal period vibration monitoring data of the gear, and guarantees the safety of the gear operation.

[0019] Additional aspects and advantages of the present application will be set forth in part in the description that follows, and in part will become apparent to those having ordinary skill in the art upon examination of the following or can be learned from practice of the present application. BRIEF DESCRIPTION OF DRAWINGS

[0020] The above and / or additional aspects and advantages of the present application will become apparent and be readily appreciated from the following description, taken in conjunction with the accompanying drawings, in which:

[0021] Figure 1 A flowchart of a gear state diagnosis method provided by an embodiment of the present application is shown in the figure;

[0022] Figure 2 A flowchart of another gear state diagnosis method provided by an embodiment of the present application is shown in the figure;

[0023] Figure 3 A flowchart of another gear state diagnosis method provided by an embodiment of the present application is shown in the figure;

[0024] Figure 4 A structural diagram of a gear state diagnosis device provided by an embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0025] The embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, in which the same or similar notations represent the same or similar elements or elements having the same or similar functions throughout. The embodiments described below by referring to the accompanying drawings are exemplary and are intended to explain the present application, and cannot be understood as a limitation of the present application.

[0026] It should be noted that the acquisition, storage, use, processing and the like of data in the technical solutions of the present application comply with the relevant provisions of national laws and regulations.

[0027] A gear state diagnosis method, device, electronic equipment and storage medium are described below with reference to the accompanying drawings.

[0028] Figure 1 A flowchart of a gear state diagnosis method provided by an embodiment of the application.

[0029] As Figure 1 shown, the method comprises the following steps:

[0030] In step 101, based on the vibration signals in the vibration monitoring data of the gear, a source domain data set and a target domain data set corresponding to the gear are made, wherein the source domain data set comprises a vibration signal set and a noisy vibration signal set with a gear state label, and the target domain data set comprises a to-be-predicted vibration signal set.

[0031] Optionally, the gear can be a gear in a planetary gear box, but is not limited thereto, and this embodiment does not make a specific limitation thereto.

[0032] Optionally, the vibration signals in the vibration monitoring data of the gear can be various planetary gear box state signals collected in a signal collection device, wherein the state signals comprise various types of abnormal state signals and normal signals.

[0033] The signal collection device can be a signal sensor, but is not limited thereto.

[0034] Optionally, the analysis bandwidth of the vibration signals should be able to cover the highest abnormal signal frequency, to meet this condition, the sampling frequency is not less than 2.56 times the maximum gear meshing frequency, the signal collection device supports the constant current source conditioning of the piezoelectric sensor, and the quantization accuracy of the vibration signals is 24 bits.

[0035] In some embodiments, one implementation of making a source domain data set and a target domain data set corresponding to the gear based on the vibration signals in the vibration monitoring data of the gear can be to obtain the vibration signals in the vibration monitoring data of the gear, and to add noise to the vibration signals to obtain noisy vibration signals, based on a preset sampling rule, to sample the vibration signals and the noisy vibration signals to obtain sampled signals and sampled noisy signals, wherein the sampling rule is determined based on the sampling sliding window length of the signals, to convert the sampled signals and the sampled noisy signals into a grayscale matrix, and to construct the source domain data set and the target domain data set corresponding to the gear based on the grayscale matrix, thereby realizing accurate making of the source domain data set and the target domain data set corresponding to the gear, and improving the reliability of the source domain data set and the target domain data set corresponding to the gear.

[0036] One implementation method for converting the sampled signal and the sampled noisy signal into a grayscale matrix, and constructing the source domain dataset and target domain dataset corresponding to the gear based on the grayscale matrix, can be as follows: convert the sampled signal and the sampled noisy signal into a grayscale matrix, determine the noisy vibration signal set of the vibration signal set based on the grayscale matrix, divide the vibration signal set into a first vibration signal set and a second vibration signal set based on a preset ratio threshold, use the second vibration signal set as the target domain dataset, label the first vibration signal set with gear states to obtain a vibration signal set with gear state labels, and use the vibration signal set with gear state labels and the noisy vibration signal set as the source domain dataset.

[0037] In this context, it's understandable that, to better understand the source and target domain datasets corresponding to gears, we can take the gear vibration signal X from the gear vibration monitoring data as an example and add additive Gaussian white noise to the vibration signal X to obtain the noisy vibration signal X. n Then, the above vibration signal X and the noise-added vibration signal X are analyzed. n Sliding window sampling is used to convert the obtained sampled signal and the sampled noisy signal into grayscale matrices, which are then used to construct vibration signal sets with gear status labels. Noisy vibration signal set Predicted vibration signal set Thus, the aforementioned source domain dataset is obtained. Target domain dataset

[0038] Wherein, the sliding window length for sliding window sampling is a, and the sampling length is n. 2 D x ={x1,x2,...,x i},

[0039] Step 102: Based on the noisy vibration signal set with gear state label, optimize the denoising parameters of the U-shaped network to obtain the optimized residual U-shaped network.

[0040] In some embodiments, the denoising parameters of the U-shaped network are optimized based on the noisy vibration signal set with gear state labels to obtain the optimized residual U-shaped network. One implementation method is to train the U-shaped network using the noisy vibration signal set with gear state labels in the source domain dataset, optimize the denoising parameters of the U-shaped network until the state signals of each noisy vibration signal in the noisy vibration signal set are extracted from the U-shaped network, so as to achieve accurate optimization of the denoising parameters of the U-shaped network and obtain the optimized residual U-shaped network.

[0041] In addition, after obtaining the optimized residual U-shaped network, the historical vibration signals corresponding to the vibration monitoring data of the gear can also be obtained, and based on the historical vibration signals, the verification source domain data set and the test source domain data set corresponding to the gear are made, so as to improve the reliability of the residual U-shaped network through the verification and test of the residual U-shaped network based on the verification source domain data set and the test source domain data set.

[0042] In step 103, based on the residual U-shaped network, the noise reduction is performed on the vibration signal set to be predicted to obtain a noise-reduced vibration signal, and based on the noise-reduced vibration signal, the gear state is determined.

[0043] In some embodiments, after the noise reduction is performed on the vibration signal set to be predicted based on the residual U-shaped network to obtain a noise-reduced vibration signal, the state signal extraction is performed on the noise-reduced vibration signal to obtain a feature state signal, and based on the feature state signal, the gear state is determined, thereby realizing the accurate diagnosis of the gear state.

[0044] In this embodiment, the state signal extraction can be performed on the noise-reduced vibration signal through Hilbert transform, but is not limited thereto.

[0045] The gear state diagnosis method provided in the embodiment of the present application comprises the following steps: based on the vibration signals in the vibration monitoring data of the gear, a source domain data set and a target domain data set corresponding to the gear are made, the source domain data set comprises a vibration signal set with a gear state label and a noise-added vibration signal set, and the target domain data set comprises a vibration signal set to be predicted; the de-noising parameters of a U-shaped network are optimized based on the noise-added vibration signal set of the vibration signal set with the gear state label, to obtain an optimized residual U-shaped network; and based on the residual U-shaped network, the noise reduction is performed on the vibration signal set to be predicted, so as to accurately determine the gear state based on the noise-reduced vibration signal obtained through the noise reduction of the vibration signal set to be predicted based on the residual U-shaped network, realize the accurate diagnosis of the vibration signal in the early abnormal period of the gear based on the vibration monitoring data, and ensure the safety of the gear operation.

[0046] In order to clearly illustrate the above embodiment, Figure 2 The flowchart of another gear state diagnosis method provided in the embodiment of the present application is shown.

[0047] In step 201, based on the vibration signals in the vibration monitoring data of the gear, a source domain data set and a target domain data set corresponding to the gear are made, wherein the source domain data set comprises a vibration signal set with a gear state label and a noise-added vibration signal set, and the target domain data set comprises a vibration signal set to be predicted.

[0048] It should be noted that the specific implementation of step 201 can be referred to the related description in the above embodiments.

[0049] Step 202, input the set of noise-added vibration signals into the U-shaped network for noise reduction processing of different signal intensity thresholds to obtain the state signal of each noise-added vibration signal in the set of noise-added vibration signals, and judge the gear state based on the state signal.

[0050] In some embodiments, different state signals represent different gear states, for example, when the state signal matches the signal corresponding to the abnormal gear, the type of abnormality corresponding to the state signal is judged and recorded, and when the state signal matches the signal corresponding to the normal operation of the gear, the gear operates normally.

[0051] Step 203, according to the difference between the candidate gear state and the gear state label of the vibration signal set, the denoising parameters of the U-shaped network are optimized to obtain the optimized residual U-shaped network.

[0052] In some embodiments, the residual U-shaped network mainly consists of an encoder, a decoder, and a skip connection, and the encoder and the decoder can be respectively composed of 4 down-sampling modules and 4 up-sampling modules in cascade, but not limited to this.

[0053] Optionally, the down-sampling module and the up-sampling module of the residual U-shaped network: each down-sampling module corresponds to four residual blocks (Res-block) of an 18-layer residual network (Resnet18) combined with an adaptive threshold acquisition layer (TL), and the up-sampling module is consistent with the U-shaped network.

[0054] Optionally, the working steps of the down-sampling module can be: the down-sampling feature map (signal map of the vibration signal on the displayable platform) first passes through a pre-trained residual block to double the channel of the down-sampling feature map and reduce the size by half; then pass through j channel spatial threshold acquisition layers to reduce the interference of noise in the down-sampling feature map.

[0055] Optionally, the working steps of the up-sampling module can be: the up-sampling feature map passes through the up-sampling (Bilinear Interpolation, BI) layer to double its size, and considering that the detail information of the shallow network is well preserved, in order to improve the accuracy of data restoration, the corresponding down-sampling feature map of the same resolution is fused by using the skip connection; then pass through a double convolution layer to reduce the channel by half. Each time an up-sampling module is passed through, the size of the up-sampling feature map is doubled, and the number of channels is halved.

[0056] The threshold obtaining layer is used for adaptively learning a set of thresholds when the residual U-shaped network down-samples data, reduces the interference of noise in the feature map, and improves the performance of feature extraction (denoising). The steps of obtaining the adaptive threshold by the threshold obtaining layer can be: inputting the down-sampled feature map into the threshold obtaining layer, on the channel threshold obtaining branch, obtaining a 1-D vector containing channel information by compressing the spatial information of the down-sampled feature map through the global average pooling (GAP) acting on the absolute value of the down-sampled feature map; then, the 1-D vector passes through two layers of fully connected (FC) network to obtain the channel scale parameter sigma (c), and the channel scale parameter sigma (c) is scaled to (0, 1) through the sigmoid function; finally, the channel scale parameter sigma (c) is multiplied by the 1-D vector to obtain the channel threshold containing the channel information of the down-sampled feature map, on the spatial threshold obtaining branch, a set of 2-D vectors containing spatial information are extracted. The 2-D vector is transmitted into two layers of convolution network to obtain the spatial scale parameter sigma (s), and the value of the spatial scale parameter sigma (s) is also scaled to (0, 1) by using the sigmoid function; finally, the spatial scale parameter sigma (s) is multiplied by the 2-D vector to obtain the threshold containing the spatial information of the down-sampled feature map, the channel threshold and the spatial threshold are fused to obtain the threshold of the down-sampled feature map, and the soft threshold function is used for denoising to obtain the denoised down-sampled feature map.

[0057] The number of the four residual blocks combined with the threshold obtaining layer is 1, 2, 2 and 1 respectively.

[0058] In step 204, based on the residual U-shaped network, the denoising is performed on the to-be-predicted vibration signal set to obtain a denoised vibration signal, and the gear state is determined based on the denoised vibration signal.

[0059] The gear state diagnosis method of the embodiment of the application is based on the vibration signal in the vibration monitoring data of the gear, and the source domain data set and the target domain data set corresponding to the gear are made. The source domain data set includes a vibration signal set and a noisy vibration signal set with gear state labels, and the target domain data set includes a to-be-predicted vibration signal set. The noisy vibration signal set is input into the U-shaped network for denoising processing of different signal intensity thresholds to obtain state signals of each noisy vibration signal in the noisy vibration signal set. The gear state is judged based on the state signals. The denoising parameters of the U-shaped network are optimized according to the difference between the candidate gear state and the gear state label of the vibration signal set to obtain an optimized residual U-shaped network. The to-be-predicted vibration signal set is denoised based on the residual U-shaped network to determine the gear state according to the obtained denoised vibration signal. Thus, the residual U-shaped network based on the optimized U-shaped network of the source domain data set denoises the to-be-predicted vibration signal set, accurately determines the denoised vibration signal, accurately establishes the residual U-shaped network, and accurately diagnoses the gear state.

[0060] In summary, the application further provides a gear state diagnosis method, as shown in the following formula (1) : Figure 3 The gear state diagnosis method includes data set making, signal denoising, and fault diagnosis. The data set making includes obtaining a vibration signal in vibration monitoring data of a gear, adding a Gaussian white noise to the vibration signal to obtain a noise-added vibration signal, sampling the vibration signal and the noise-added vibration signal based on a preset sampling sliding window length to obtain a sampling signal and a sampling noise-added signal, converting the sampling signal and the sampling noise-added signal into a gray matrix, determining a noise-added vibration signal set of the vibration signal set based on the gray matrix, and further constructing a source domain data set and a target domain data set corresponding to the gear. The signal denoising includes optimizing a denoising parameter of a U-shaped network based on the noise-added vibration signal set of the vibration signal set with a gear state label, until the U-shaped network extracts a state label of the noise-added vibration signal set, to obtain an optimized residual U-shaped network, and denoising a vibration signal set to be predicted based on the residual U-shaped network to obtain a denoised vibration signal. The fault diagnosis includes performing Hilbert transform on the denoised vibration signal, diagnosing a gear state, judging whether there is a fault, and realizing accurate diagnosis of a vibration signal in early abnormal period vibration monitoring data of the gear.

[0061] To realize the above-mentioned embodiments, the application further provides a gear state diagnosis device.

[0062] Figure 4 A structural schematic diagram of a gear state diagnosis device provided by an embodiment of the application is shown in the following figure.

[0063] As shown in the following formula (1) : Figure 4 The gear state diagnosis device 40 includes a making module 41, a generating module 42, and a determining module 43.

[0064] The making module 41 is configured to make a source domain data set and a target domain data set corresponding to a gear based on a vibration signal in vibration monitoring data of the gear. The source domain data set includes a vibration signal set with a gear state label and a noise-added vibration signal set, and the target domain data set includes a vibration signal set to be predicted.

[0065] The generating module 42 is configured to optimize a denoising parameter of a U-shaped network based on the noise-added vibration signal set of the vibration signal set with the gear state label, to obtain an optimized residual U-shaped network.

[0066] The determining module 43 is configured to denoise the vibration signal set to be predicted based on the residual U-shaped network, to obtain a denoised vibration signal, and determine a gear state based on the denoised vibration signal.

[0067] Further, in a possible implementation manner of the embodiment of the application, the making module 41 includes:

[0068] The acquisition unit is configured to acquire a vibration signal in vibration monitoring data of a gear and add noise to the vibration signal to obtain a noise-added vibration signal.

[0069] The sampling unit is configured to sample the vibration signal and the noise-added vibration signal based on a preset sampling rule to obtain a sampled signal and a sampled noise-added signal, wherein the sampling rule is determined based on a sampling sliding window length of the signal.

[0070] The construction unit is configured to convert the sampled signal and the sampled noise-added signal into a grayscale matrix and construct a source domain data set and a target domain data set corresponding to the gear based on the grayscale matrix.

[0071] Further, in a possible implementation manner of the embodiment of the present application, the construction unit is specifically configured to:

[0072] convert the sampled signal and the sampled noise-added signal into a grayscale matrix and determine a noise-added vibration signal set of a vibration signal set based on the grayscale matrix;

[0073] divide the vibration signal set based on a preset proportion threshold to obtain a first vibration signal set and a second vibration signal set, and take the second vibration signal set as the target domain data set;

[0074] label the first vibration signal set based on a gear state to obtain a vibration signal set with a gear state label, and take the vibration signal set with the gear state label and the noise-added vibration signal set as the source domain data set.

[0075] Further, in a possible implementation manner of the embodiment of the present application, the generation module 42 is specifically configured to:

[0076] input the noise-added vibration signal set into the U-shaped network for noise reduction processing based on different signal intensity thresholds to obtain a state signal of each noise-added vibration signal in the noise-added vibration signal set, and determine the gear state based on the state signal;

[0077] optimize a denoising parameter of the U-shaped network according to a difference between the candidate gear state and a gear state label of the vibration signal set to obtain an optimized residual U-shaped network.

[0078] Further, in a possible implementation manner of the embodiment of the present application, the determination module 43 is specifically configured to:

[0079] after denoising a vibration signal set of the target domain data set based on the residual U-shaped network to obtain a denoised vibration signal, extract a state signal from the denoised vibration signal to obtain a feature state signal;

[0080] A gear state is determined according to the feature state signal.

[0081] It should be noted that the foregoing description of the method embodiments also applies to the device embodiments, and thus will not be repeated here.

[0082] The gear state diagnosis device of the embodiment of the application is based on the vibration signal in the vibration monitoring data of the gear, and a source domain data set and a target domain data set corresponding to the gear are made, the source domain data set includes a vibration signal set with a gear state label and a noisy vibration signal set, and the target domain data set includes a to-be-predicted vibration signal set; the noisy vibration signal set based on the vibration signal set with the gear state label is used to optimize the denoising parameter of the U-shaped network to obtain an optimized residual U-shaped network; the to-be-predicted vibration signal set is denoised based on the residual U-shaped network to determine the gear state according to the obtained denoised vibration signal, thereby the denoised vibration signal obtained by denoising the to-be-predicted vibration signal set based on the residual U-shaped network accurately determines the gear state, realizes accurate diagnosis of the vibration signal in the early abnormal period vibration monitoring data of the gear, and guarantees the safety of the gear operation.

[0083] To achieve the above-mentioned embodiments, the application further provides an electronic device, comprising:

[0084] at least one processor; and

[0085] a memory in communication with the at least one processor; wherein

[0086] The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the foregoing method.

[0087] To achieve the above-mentioned embodiments, the application further provides a non-transitory computer readable storage medium storing computer instructions, the computer instructions being used to make the computer execute the foregoing method.

[0088] In the description of the present specification, the description of the terms "one embodiment", "some embodiments", "an example", "a specific example", or "some examples" and the like means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In the present specification, the illustrative description of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any appropriate manner in any one or more embodiments or examples. In addition, the person skilled in the art can combine and combine the different embodiments or examples described in the present specification and the features of the different embodiments or examples without contradiction.

[0089] Moreover, the terms "first", "second", "third", etc. are used herein only to describe different steps or categories of steps in a claim for patent purposes, and are not to be construed as implying or implying relative importance or a number of indicated technical features. Thus, features defined with "first", "second" can explicitly or implicitly include at least one of the features. In the description of the present application, the meaning of "a plurality" is at least two, for example, two, three, etc., unless otherwise explicitly specified.

[0090] Any process or method descriptions or descriptions of the flow diagrams described herein or otherwise described herein can be understood as representing modules, segments, or portions of code that include one or more executable instructions for implementing specific logical functions or steps in the process, and the preferred embodiments of the present application include additional implementations that can not be described in detail in the description of the flow diagrams or otherwise described herein, and that can include the implementation of the functions according to the involved functions in a substantially simultaneous manner or in reverse order, which should be understood by those skilled in the art to which the embodiments of the present application belong.

[0091] The logic and / or steps represented in the flow diagrams or otherwise described herein, for example, can be considered as a list of executable instructions for implementing logical functions, which can be specifically implemented in any computer readable medium for use by or in conjunction with an instruction execution system, device or apparatus, such as a computer-based system, a system including a processor, or other system that can take instructions from an instruction execution system, device or apparatus and execute them. For the purposes of this specification, a "computer readable medium" can be any device that can contain, store, communicate, propagate or transport a program for use by or in conjunction with an instruction execution system, device or apparatus. More specific examples (non-exhaustive list) of computer readable medium include the following: electrical connections having one or more wires (electronic devices), portable computer diskette (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), fiber optic devices, and portable compact disc read-only memory (CD ROM). In addition, a computer readable medium can even be paper or other suitable medium on which the program can be printed, as the program can be electronically obtained, for example, by optical scanning of the paper or other medium, followed by electronic conversion, interpretation or processing, if necessary, in other suitable manner, and then stored in a computer memory.

[0092] It should be understood that each part of the present application can be realized by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of steps or methods can be realized by software or firmware stored in a memory and executed by a suitable instruction execution system. As in another embodiment, if realized by hardware, any one or a combination of the following technologies known in the art can be used: discrete logic circuit with logic gate circuit for implementing logic function on data signal, application specific integrated circuit with suitable combination logic gate circuit, programmable gate array (PGA), field programmable gate array (FPGA) and the like.

[0093] Those skilled in the art of the present technology can understand that all or part of the steps carried out by the above-mentioned embodiment method can be completed by a program instructing the relevant hardware, and the program can be stored in a computer readable storage medium. When the program is executed, it includes one of the steps of the method embodiment or a combination thereof.

[0094] In addition, each functional unit in each embodiment of the present application can be integrated in one processing module, or each unit can exist physically alone, or two or more units can be integrated in one module. The above-mentioned integrated module can be realized in the form of hardware or in the form of software functional module. When the integrated module is realized in the form of software functional module and sold or used as an independent product, it can also be stored in a computer readable storage medium.

[0095] The above-mentioned storage medium can be a read-only memory, a magnetic disk or an optical disk, etc. Although the embodiments of the present application have been shown and described above, it should be understood that the above-mentioned embodiments are exemplary and cannot be understood as limiting the present application, and those skilled in the art can make changes, modifications, replacements and variations to the above-mentioned embodiments within the scope of the present application.

Claims

1. A method for diagnosing the condition of gears, characterized in that, The method includes: Based on the vibration signals in the vibration monitoring data of gears, a source domain dataset and a target domain dataset corresponding to the gears are created. The source domain dataset includes a set of vibration signals with gear state labels and a set of vibration signals with added noise. The target domain dataset includes a set of vibration signals to be predicted. Based on the noisy vibration signal set of the gear-state-labeled vibration signal set, the denoising parameters of the U-shaped network are optimized to obtain the optimized residual U-shaped network. Based on the residual U-shaped network, the set of vibration signals to be predicted is denoised to obtain a denoised vibration signal, and the gear state is determined based on the denoised vibration signal. The optimization of the denoising parameters of the U-shaped network based on the noisy vibration signal set with gear state labels to obtain the optimized residual U-shaped network includes: The set of noisy vibration signals is input into the U-shaped network for noise reduction processing at different signal strength thresholds to obtain the state signal of each noisy vibration signal in the set of noisy vibration signals. Based on the state signal, the state of the gear is determined. Based on the difference between the candidate gear state and the gear state label of the vibration signal set, the denoising parameters of the U-shaped network are optimized to obtain the optimized residual U-shaped network. The step of denoising the set of vibration signals to be predicted based on the residual U-shaped network to obtain a denoised vibration signal, and determining the gear state based on the denoised vibration signal, includes: After denoising the vibration signal set of the target domain dataset based on the residual U-shaped network to obtain the denoised vibration signal, state signal extraction is performed on the denoised vibration signal to obtain the feature state signal. The gear state is determined based on the characteristic state signal.

2. The method according to claim 1, characterized in that, The vibration signals from the gear-based vibration monitoring data are used to create a source domain dataset and a target domain dataset corresponding to the gear. The source domain dataset includes a set of vibration signals with gear state labels and a set of vibration signals with added noise. The target domain dataset includes a set of vibration signals to be predicted, including: The vibration signal in the vibration monitoring data of the gear is acquired, and the vibration signal is denoised to obtain a denoised vibration signal; Based on a preset sampling rule, the vibration signal and the noisy vibration signal are sampled to obtain a sampled signal and a sampled noisy signal, wherein the sampling rule is determined based on the sampling window length of the signal; The sampled signal and the sampled noisy signal are converted into grayscale matrices, and the source domain dataset and target domain dataset corresponding to the gear are constructed based on the grayscale matrices.

3. The method according to claim 2, characterized in that, The step of converting the sampled signal and the sampled noisy signal into a grayscale matrix, and constructing the source domain dataset and target domain dataset corresponding to the gear based on the grayscale matrix, includes: The sampled signal and the sampled noise-added signal are converted into a grayscale matrix, and the noise-added vibration signal set of the vibration signal set is determined based on the grayscale matrix; The vibration signal set is divided into two sets based on a preset ratio threshold to obtain a first vibration signal set and a second vibration signal set, and the second vibration signal set is used as the target domain dataset. Gear state labeling is performed on the first vibration signal set to obtain a vibration signal set with gear state labels, and the vibration signal set with gear state labels and the noisy vibration signal set are used as the source domain dataset.

4. A gear condition diagnostic device, characterized in that, The device includes: The module is used to generate vibration signals from vibration monitoring data based on gears, and to generate source domain datasets and target domain datasets corresponding to the gears. The source domain dataset includes a set of vibration signals with gear state labels and a set of vibration signals with added noise. The target domain dataset includes a set of vibration signals to be predicted. The generation module is used to optimize the denoising parameters of the U-shaped network based on the noisy vibration signal set of the gear state label vibration signal set, so as to obtain the optimized residual U-shaped network. The determination module is used to denoise the set of vibration signals to be predicted based on the residual U-shaped network to obtain a denoised vibration signal, and to determine the gear state based on the denoised vibration signal; The optimization of the denoising parameters of the U-shaped network based on the noisy vibration signal set with gear state labels to obtain the optimized residual U-shaped network includes: The set of noisy vibration signals is input into the U-shaped network for noise reduction processing at different signal strength thresholds to obtain the state signal of each noisy vibration signal in the set of noisy vibration signals. Based on the state signal, the state of the gear is determined. Based on the difference between the candidate gear state and the gear state label of the vibration signal set, the denoising parameters of the U-shaped network are optimized to obtain the optimized residual U-shaped network. The step of denoising the set of vibration signals to be predicted based on the residual U-shaped network to obtain a denoised vibration signal, and determining the gear state based on the denoised vibration signal, includes: After denoising the vibration signal set of the target domain dataset based on the residual U-shaped network to obtain the denoised vibration signal, state signal extraction is performed on the denoised vibration signal to obtain the feature state signal. The gear state is determined based on the characteristic state signal.

5. The apparatus according to claim 4, characterized in that, The production module includes: The acquisition unit is used to acquire the vibration signal from the vibration monitoring data of the gear, and add noise to the vibration signal to obtain a noisy vibration signal; A sampling unit is used to sample the vibration signal and the noisy vibration signal based on a preset sampling rule to obtain a sampled signal and a sampled noisy signal, wherein the sampling rule is determined based on the sampling sliding window length of the signal; The construction unit is used to convert the sampled signal and the sampled noisy signal into a grayscale matrix, and to construct the source domain dataset and target domain dataset corresponding to the gear based on the grayscale matrix.

6. The apparatus according to claim 5, characterized in that, The building unit is specifically used for: The sampled signal and the sampled noise-added signal are converted into a grayscale matrix, and the noise-added vibration signal set of the vibration signal set is determined based on the grayscale matrix; The vibration signal set is divided into two sets based on a preset ratio threshold to obtain a first vibration signal set and a second vibration signal set, and the second vibration signal set is used as the target domain dataset. Gear state labeling is performed on the first vibration signal set to obtain a vibration signal set with gear state labels, and the vibration signal set with gear state labels and the noisy vibration signal set are used as the source domain dataset.

7. An electronic device, characterized in that, include: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-3.

8. A non-transitory computer-readable storage medium storing computer instructions, characterized in that, The computer instructions are used to cause the computer to perform the method according to any one of claims 1-3.

Images