Method and device for diagnosing faults in a multi-task rotary mechanical device

Abstract

The application relates to the technical field of fault diagnosis, in particular to a multi-task rotating mechanical equipment fault diagnosis method and device, wherein the method comprises the following steps: collecting vibration signals of a rotating mechanical equipment for data preprocessing to obtain a frequency spectrum, a continuous wavelet transform graph, an envelope spectrum, a wavelet packet decomposition graph and an axis trajectory graph, generating multi-class vibration feature data, taking the multi-class vibration feature data as model input, taking a fault label corresponding to a rotating fault type to be diagnosed as output, and training a rotating mechanical multi-task fault diagnosis model based on a convolutional neural network. According to the embodiment of the application, the neural network can be trained according to the rotating mechanical fault type, the multi-dimensional features of the preprocessed signal data are fused, the multi-class fault diagnosis results can be output at the same time, the pertinence and comprehensiveness of the fault diagnosis process are improved, and the diagnosis result is more accurate and reliable.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of fault diagnosis, in particular to a multi-task rotating mechanical equipment fault diagnosis method and device. BACKGROUND

[0002] With the continuous development of industry, fault diagnosis of key equipment for industrial production has gradually attracted attention. Rotating mechanical equipment is an important part of industrial production, and the safe operation of rotating mechanical equipment needs to be ensured.

[0003] In related technologies, with the popularization of online monitoring equipment, a large amount of online monitoring data has been accumulated in the industrial field. Rotating mechanical fault diagnosis can be performed based on deep learning, knowledge graph and automated machine learning algorithm.

[0004] However, the related art fails to fully consider the fault features and signal analysis when data is input into the model, the different vibration analysis methods corresponding to different components of the rotating machinery, and the relevance between the causes or characteristics of different component faults, resulting in low precision and poor generality of the obtained fault diagnosis results, and unable to achieve efficient, accurate and targeted rotating mechanical equipment fault diagnosis, which needs to be solved urgently. SUMMARY

[0005] The present application provides a multi-task rotating mechanical equipment fault diagnosis method and device to solve the problem that the related art fails to fully consider the fault features and signal analysis when data is input into the model, the different vibration analysis methods corresponding to different components of the rotating machinery, and the relevance between the causes or characteristics of different component faults, resulting in low precision and poor generality of the obtained fault diagnosis results, and unable to achieve efficient, accurate and targeted rotating mechanical equipment fault diagnosis.

[0006] The first aspect of the present application provides a multi-task rotating mechanical equipment fault diagnosis method, applied to a model training stage, wherein the method comprises the following steps: collecting at least one vibration signal of at least one rotating mechanical equipment during movement; performing data preprocessing on the at least one vibration signal to obtain a frequency spectrum, a continuous wavelet transform graph, an envelope spectrum, a wavelet packet decomposition graph and an axis trajectory graph of each vibration signal, and generating multi-class vibration feature data; inputting the multi-class vibration feature data as a model, and inputting the fault label corresponding to the rotating fault type to be diagnosed as an output, and training a rotating mechanical multi-task fault diagnosis model based on a convolutional neural network.

[0007] Optionally, in an embodiment of the present application, the training of the convolutional neural network-based rotating machinery multi-task fault diagnosis model comprises: for each vibration feature of the multi-class vibration feature data, constructing a separate residual convolutional neural network to perform feature extraction on each vibration feature of the multi-class vibration feature data to obtain a high-order representation feature map of each vibration feature.

[0008] Optionally, in an embodiment of the present application, the training of the convolutional neural network-based rotating machinery multi-task fault diagnosis model further comprises: after obtaining the high-order representation feature map of each vibration feature, introducing a residual network block with an attention mechanism to pay attention to different feature maps according to signals and fault features; based on the fault categories to be diagnosed, constructing a multi-task fault classification module with a corresponding number of residual network blocks to output the occurrence of each fault category.

[0009] The second aspect embodiment of the present application provides a multi-task rotating machinery equipment fault diagnosis method, applied to a fault diagnosis stage, wherein the method comprises the following steps: collecting at least one vibration signal of a target rotating machinery equipment; inputting the at least one vibration signal of the target rotating machinery equipment into a trained rotating machinery multi-task fault diagnosis model to output a fault diagnosis result of the target rotating machinery equipment, wherein the rotating machinery multi-task fault diagnosis model takes multi-class vibration feature data as model input and obtains a fault label corresponding to a rotating fault category to be diagnosed as output through training.

[0010] Optionally, in an embodiment of the present application, the fault category corresponding to the fault label comprises shaft system faults, bearing faults, and gear faults.

[0011] The third aspect embodiment of the present application provides a multi-task rotating machinery equipment fault diagnosis device, applied to a model training stage, wherein the device comprises: a collection module configured to collect at least one vibration signal of at least one rotating machinery equipment during movement; a preprocessing module configured to perform data preprocessing on the at least one vibration signal to obtain a frequency spectrum, a continuous wavelet transform graph, an envelope spectrum, a wavelet packet decomposition graph, and an axis trajectory graph of each vibration signal, and generate multi-class vibration feature data; and a training module configured to take the multi-class vibration feature data as model input and take a fault label corresponding to a rotating fault category to be diagnosed as output, and train a convolutional neural network-based rotating machinery multi-task fault diagnosis model.

[0012] Optionally, in an embodiment of the present application, the training module is further configured to, for each vibration feature of the multi-class vibration feature data, construct a separate residual convolutional neural network to perform feature extraction on each vibration feature of the multi-class vibration feature data to obtain a high-order representation feature map of each vibration feature.

[0013] Optionally, in an embodiment of the present application, the training module further comprises: a classification unit, configured to introduce a residual network block of attention mechanism after obtaining the high-order representation feature map of each vibration feature, so as to pay attention to different feature maps according to signals and fault features; and an output unit, configured to construct a corresponding number of residual network blocks based on the fault category to be diagnosed, so as to output the occurrence of each fault category.

[0014] The fourth aspect embodiment of the present application provides a fault diagnosis device for a multi-task rotating machinery, applied to a fault diagnosis stage, wherein the device comprises: an acquisition module, configured to collect at least one vibration signal of a target rotating machinery; and a diagnosis module, configured to input the at least one vibration signal of the target rotating machinery into a trained rotating machinery multi-task fault diagnosis model, and output a fault diagnosis result of the target rotating machinery, wherein the rotating machinery multi-task fault diagnosis model is trained by taking multi-class vibration feature data as model input and taking a fault label corresponding to a rotating fault category to be diagnosed as output.

[0015] Optionally, in an embodiment of the present application, the fault category corresponding to the fault label comprises a shaft system fault, a bearing fault and a gear fault.

[0016] The fifth aspect embodiment of the present application provides an electronic device, comprising: a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the fault diagnosis method for a multi-task rotating machinery as described in the above embodiments.

[0017] The sixth aspect embodiment of the present application provides a computer readable storage medium, which stores a computer program executable by a processor to implement the fault diagnosis method for a multi-task rotating machinery as described above.

[0018] The embodiments of the present application can train a neural network according to the rotating machinery fault type, fuse the multi-dimensional features of the preprocessed signal data, and simultaneously output the multi-class fault diagnosis result, thereby improving the pertinence and comprehensiveness of the fault diagnosis process and making the diagnosis result more accurate and reliable. Thus, the problems in the related art that the fault features and signal analysis when the data are input into the model are not fully considered, different vibration analysis methods corresponding to different components of the rotating machinery, and the relevance between different component fault causes or features are not considered, resulting in low accuracy and poor universality of the obtained fault diagnosis result, and the rotating machinery fault diagnosis cannot be efficiently and accurately and pertinently performed are solved.

[0019] Additional aspects and advantages of the present application will be in part apparent and in part pointed out hereinafter. Attached Figure Description

[0020] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:

[0021] Figure 1 This is a flowchart of a fault diagnosis method for a multi-tasking rotating machine according to an embodiment of this application;

[0022] Figure 2 This is a diagram of the original vibration waveform of a rotating mechanical device according to an embodiment of this application.

[0023] Figure 3 This is a spectrum diagram of the vibration signal of a rotating mechanical device according to an embodiment of this application;

[0024] Figure 4 This is the envelope spectrum of a vibration signal from a rotating mechanical device according to an embodiment of this application;

[0025] Figure 5 This is a continuous wavelet transform diagram of the vibration signal of a rotating mechanical device according to an embodiment of this application;

[0026] Figure 6 This is a wavelet packet decomposition diagram of a vibration signal from a rotating mechanical device according to an embodiment of this application;

[0027] Figure 7 This is a diagram of the shaft center trajectory of vibration signals from a rotating mechanical device according to an embodiment of this application;

[0028] Figure 8 This is a schematic diagram of the attention mechanism network for vibration signals of rotating machinery according to an embodiment of this application;

[0029] Figure 9 This is a flowchart of another fault diagnosis method for multi-tasking rotating machinery provided according to an embodiment of this application;

[0030] Figure 10 This is a schematic diagram of a multi-task fault diagnosis model for vibration signals of rotating machinery according to an embodiment of this application;

[0031] Figure 11 This is a schematic diagram of a fault diagnosis device for a multi-task rotating machine according to an embodiment of this application;

[0032] Figure 12 This is a schematic diagram of the structure of another fault diagnosis device for multi-tasking rotating machinery provided according to an embodiment of this application;

[0033] Figure 13 This is a schematic diagram of the structure of an electronic device according to an embodiment of this application. DETAILED DESCRIPTION

[0034] Embodiments of the present application are described below in detail, examples of which are shown in the drawings, wherein 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 drawings are exemplary and are intended to explain the present application, and cannot be understood as a limitation of the present application.

[0035] A fault diagnosis method and device of a multi-task rotating machinery equipment are described below with reference to the drawings. In view of the fact that the related art fails to fully consider the fault features and signal analysis when the data input model is not fully considered, different vibration analysis methods corresponding to different components of the rotating machinery, and the correlation between the causes or features of different component faults, resulting in low accuracy and poor universality of the obtained fault diagnosis results, and the problem that efficient, accurate and targeted fault diagnosis of the rotating machinery equipment cannot be achieved, the present application provides a fault diagnosis method of a multi-task rotating machinery equipment. In the method, the neural network can be trained according to the rotating machinery fault type, the multi-dimensional features of the preprocessed signal data are fused, and the simultaneous output of the multi-class fault diagnosis results is realized, thereby improving the pertinence and comprehensiveness of the fault diagnosis process, and making the diagnosis results more accurate and reliable. Thus, the related art fails to fully consider the fault features and signal analysis when the data input model is not fully considered, different vibration analysis methods corresponding to different components of the rotating machinery, and the correlation between the causes or features of different component faults, resulting in low accuracy and poor universality of the obtained fault diagnosis results, and the problem that efficient, accurate and targeted fault diagnosis of the rotating machinery equipment cannot be achieved is solved.

[0036] Specifically, Figure 1 A flowchart of a fault diagnosis method of a multi-task rotating machinery equipment provided by an embodiment of the present application is shown in the figure.

[0037] As Figure 1 shown, the fault diagnosis method of the multi-task rotating machinery equipment is applied to a model training stage, wherein the method comprises the following steps:

[0038] In step S101, at least one vibration signal of at least one rotating machinery equipment during movement is collected.

[0039] It can be understood that the at least one vibration signal of the rotating machinery equipment during movement in the embodiment of the present application can be displacement, speed or acceleration vibration data of a part of the rotating machinery equipment, and the form of the collected vibration signal can be determined according to the frequency of the collected part vibration.

[0040] The embodiment of the application can collect at least one vibration signal of the rotating mechanical equipment during movement, thereby providing required original data for model training in the following steps, and establishing a data basis for rotating mechanical equipment fault diagnosis.

[0041] In step S102, data preprocessing is performed on the at least one vibration signal to obtain a spectrum diagram, a continuous wavelet transform diagram, an envelope spectrum, a wavelet packet decomposition diagram and an axis trajectory diagram of each vibration signal, and to generate multi-class vibration feature data.

[0042] It can be understood that the spectrum diagram, the continuous wavelet transform diagram, the envelope spectrum, the wavelet packet decomposition diagram and the axis trajectory diagram of each vibration signal in the embodiment of the application can be drawn by data processing on the vibration signal.

[0043] In actual execution, as shown in Figure 2 , the original vibration waveform diagram of the rotating mechanical equipment vibration signal of one embodiment of the application can be data preprocessed by using the vibration signal data provided by the diagram. As shown in Figure 3 , the spectrum diagram of the rotating mechanical equipment vibration signal of one embodiment of the application can be obtained by using a signal analysis method to extract vibration time domain indicators and frequency domain indicators, and then performing Fourier transform on the vibration signal.

[0044] Figure 4 The envelope spectrum of the rotating mechanical equipment vibration signal of one embodiment of the application can be obtained by using envelope analysis.

[0045] Figures 5-6 The continuous wavelet transform diagram of the rotating mechanical equipment vibration signal of one embodiment of the application and the wavelet packet decomposition diagram of the rotating mechanical equipment vibration signal of one embodiment of the application can be obtained by using wavelet packet decomposition.

[0046] Figure 7 The axis trajectory diagram of the rotating mechanical equipment vibration signal of one embodiment of the application can be obtained by synthesizing vibration signals of two sensors to draw the axis trajectory diagram of the rotor.

[0047] The embodiment of the application can perform data preprocessing on the at least one vibration signal to obtain a spectrum diagram, a continuous wavelet transform diagram, an envelope spectrum, a wavelet packet decomposition diagram and an axis trajectory diagram of each vibration signal, and to generate multi-class vibration feature data, thereby providing diversified information selection for training of the neural network in the following steps, enriching the use form of the collected vibration signal data, and enhancing the perfection of the data basis of the neural network framework.

[0048] In step S103, the multi-class vibration feature data is taken as the model input, and the fault label corresponding to the rotating fault type to be diagnosed is taken as the output, and a rotating machinery multi-task fault diagnosis model based on a convolutional neural network is trained.

[0049] It can be understood that the fault label content in the embodiments of the present application can come from the maintenance and repair records of the rotating machinery equipment from which the vibration signals are collected, the fault label types can be defined and divided by the rotating fault types to be diagnosed, and each multi-class vibration feature data can have a corresponding fault label. If there is a new fault type, the existing multi-task fault diagnosis model can be quickly updated and trained by adding a residual network block.

[0050] In actual execution, historical monitoring data of the rotating machinery equipment can be collected, vibration signals and corresponding fault contents thereof are obtained, and multi-class vibration feature data obtained by data processing of the vibration signals is taken as input data for model training, and the labeled results obtained by classifying the fault contents are taken as output data for model training.

[0051] The embodiments of the present application can take multi-class vibration feature data as model input, and take the fault label corresponding to the rotating fault type to be diagnosed as output, and train a rotating machinery multi-task fault diagnosis model based on a convolutional neural network, thereby using the collected rotating equipment data to realize the construction and training of the model, improve the data processing capability of the model for the vibration signals of the rotating machinery equipment, and enhance the accuracy of the model in actual fault judgment.

[0052] Optionally, in an embodiment of the present application, the rotating machinery multi-task fault diagnosis model based on a convolutional neural network is trained, including: for each vibration feature of the multi-class vibration feature data, constructing a separate residual convolutional neural network to extract features of each vibration feature of the multi-class vibration feature data, to obtain a high-order representation feature map of each vibration feature.

[0053] It can be understood that each vibration feature of the multi-class vibration feature data in the embodiments of the present application can correspond to the original waveform, the frequency spectrum graph, the continuous wavelet transform graph, the envelope spectrum, the wavelet packet decomposition graph and the axis trajectory graph of each vibration signal obtained in the above steps as a form of expression, and a separate residual convolutional neural network is constructed corresponding thereto, and then the corresponding vibration features are extracted by the separate residual convolutional neural network to obtain the corresponding high-order representation feature map.

[0054] The embodiments of the present application can construct a separate residual convolutional neural network for each vibration feature of the multi-class vibration feature data to extract features of each vibration feature of the multi-class vibration feature data, to obtain a high-order representation feature map of each vibration feature, thereby further exploring the data information of the vibration signals and improving the data processing level of fault diagnosis.

[0055] Optionally, in an embodiment of the present application, the training of the rotating machinery multi-task fault diagnosis model based on the convolutional neural network further comprises: after obtaining the high-order representation feature map of each vibration feature, introducing a residual network block of an attention mechanism to focus on different feature maps according to the signal and the fault feature; and based on the fault types to be diagnosed, constructing a corresponding number of residual network blocks to output the occurrence of each fault type.

[0056] It can be understood that the residual network block of the attention mechanism in the embodiment of the present application can calculate the weight of each input feature map according to the signal and the fault feature, so as to focus on the feature map with a larger weight, and input the obtained weighted feature map into an equal residual network block for fault determination, so as to output the fault occurrence corresponding to the vibration signal.

[0057] In actual execution, the attention mechanism network is as shown in Figure 8 Fig. 1, which is a schematic diagram of the attention mechanism network of the rotating machinery equipment vibration signal in an embodiment of the present application. After inputting the Feature Map feature map F, the maximum pooling MaxPooling and the average pooling Avg Pooling are performed respectively, and then the full connection layer FC layer is inputted to output the channel-based attention ChannelAttention C.

[0058] The embodiment of the present application can introduce a residual network block of an attention mechanism after obtaining the high-order representation feature map of each vibration feature, so as to focus on different feature maps according to the signal and the fault feature, and construct a corresponding number of residual network blocks based on the fault types to be diagnosed to output the occurrence of each fault type, so as to realize the targeted attention to different features of the obtained vibration signal data, thereby improving the accuracy of the rotating machinery fault diagnosis result, and more efficiently realizing the mechanical fault diagnosis.

[0059] According to the multi-task rotating machinery equipment fault diagnosis method proposed in the embodiment of the present application, the neural network can be trained according to the rotating machinery fault type, the multi-dimensional features of the preprocessed signal data are fused, and the multi-class fault diagnosis result is outputted at the same time, so as to improve the pertinence and comprehensiveness of the fault diagnosis process, and make the diagnosis result more accurate and reliable. Therefore, the problems in the related art that the fault characteristics and signal analysis when the data is inputted into the model are not fully considered, the different vibration analysis methods corresponding to different components of the rotating machinery, and the relevance between the causes or characteristics of different component faults are not considered, resulting in low precision and poor universality of the obtained fault diagnosis result, and the rotating machinery equipment fault diagnosis cannot be efficiently and accurately and targetedly realized are solved.

[0060] The above embodiment is described as a model training stage, and the following describes an embodiment of a fault diagnosis stage.

[0061] Figure 9 Another flowchart of a fault diagnosis method of a multi-task rotating machinery equipment provided in the embodiment of the application.

[0062] As shown in the figure, the fault diagnosis method of the multi-task rotating machinery equipment is applied to a fault diagnosis stage, and the method comprises the following steps: Figure 9

[0063] In step S901, at least one vibration signal of a target rotating machinery equipment is collected.

[0064] In actual execution, the vibration signal can be acquired by installing a vibration sensor on the object on which the rotating machinery equipment fault detection is performed, and the vibration sensor can be a vibration speed sensor, a vibration acceleration sensor, or a vibration displacement sensor.

[0065] The embodiment of the application can collect at least one vibration signal of at least one rotating machinery equipment during movement, thereby providing the required original data for model training in the following steps and establishing a data basis for rotating machinery equipment fault diagnosis.

[0066] In step S902, the at least one vibration signal of the target rotating machinery equipment is input into the trained rotating machinery multi-task fault diagnosis model, and a fault diagnosis result of the target rotating machinery equipment is output, wherein the rotating machinery multi-task fault diagnosis model is trained by taking multi-class vibration feature data as model input and taking a fault label corresponding to a rotating fault type to be diagnosed as output.

[0067] Specifically, the vibration signal data of the target object collected in the above step can be input into the rotating machinery multi-task fault diagnosis model that has completed model training for fault diagnosis. The vibration signal after data preprocessing is input in the form of original waveform, frequency spectrum, continuous wavelet transform graph, envelope spectrum, wavelet packet decomposition graph, and shaft center trajectory graph as multi-class vibration feature data, the final diagnosis result is obtained, the classification label of the fault is taken as output content, and finally the fault diagnosis result is obtained.

[0068] The embodiment of the application can input the at least one vibration signal of the target rotating machinery equipment into the trained rotating machinery multi-task fault diagnosis model, and output the fault diagnosis result of the target rotating machinery equipment, thereby realizing diagnosis output with high efficiency and high coupling, making the diagnosis of the rotating machinery equipment more practical, and further ensuring the safe operation of the rotating machinery equipment.

[0069] ​Optionally, in an embodiment of the present application, the fault type corresponding to the fault label includes shaft system faults, bearing faults and gear faults.

[0070] It can be understood that in the embodiments of the present application, the faults of the fault type can come from the shaft system, bearings and gears of the rotating machinery equipment, for example, the shaft system faults can include imbalance, misalignment, rubbing and the like, the bearing faults can include inner ring faults, outer ring faults, rolling element faults and the like, and the gear faults can include broken teeth, missing teeth, tooth surface wear and the like.

[0071] In particular, if the fault diagnosis result does not belong to any of the shaft system faults, bearing faults and gear faults, the fault label is classified into other types.

[0072] The fault type corresponding to the fault label in the embodiments of the present application includes shaft system faults, bearing faults and gear faults, and by specifically dividing the fault type, the obtained result can reflect the specific fault position and fault content of the rotating machinery equipment, thereby improving the pertinence of the rotating machinery equipment fault diagnosis result, and further making the fault processing have an efficient information basis.

[0073] According to the multi-task rotating machinery equipment fault diagnosis method proposed in the embodiments of the present application, the neural network can be trained according to the rotating machinery fault type, the multi-dimensional features of the preprocessed signal data are fused, and the multi-class fault diagnosis result is output at the same time, thereby improving the pertinence and comprehensiveness of the fault diagnosis process, and making the diagnosis result more accurate and reliable. Thus, the problems in the related art that the fault features and signal analysis when the data are input into the model are not fully considered, different vibration analysis methods corresponding to different components of the rotating machinery, and the relevance between the causes or features of different components are not fully considered, resulting in low precision and poor universality of the obtained fault diagnosis result, and the rotating machinery equipment fault diagnosis cannot be efficiently and accurately and pertinently implemented are solved.

[0074] The working content of the embodiments of the present application will be described in detail below in combination with the model training stage and the fault diagnosis stage of the embodiments of the present application.

[0075] First, the collected rotating machinery equipment vibration signal is processed to obtain multi-class vibration feature data in the form of original vibration signal, frequency spectrum, continuous wavelet transform graph, envelope spectrum, wavelet packet decomposition graph and shaft center trajectory graph, which are used as input data, and the fault content corresponding to the vibration signal is classified into the fault type for fault labeling, which is used as output data, thereby training the constructed multi-task fault diagnosis model.

[0076] Secondly, the trained multi-task fault diagnosis model is deployed on the rotating machinery equipment that needs fault diagnosis, and the vibration data of the equipment is input into the model for real-time fault diagnosis.

[0077] Finally, according to the model output result, the fault content is obtained as the final fault diagnosis result. As shown in Figure 10 the figure is a schematic diagram of a multi-task fault diagnosis model of a rotating mechanical equipment vibration signal according to an embodiment of the present application. The collected original vibration signal, including the data collected by the vibration speed sensor, the vibration acceleration sensor and the vibration displacement sensor, is subjected to a data preprocessing process through signal analysis, to obtain a spectrum, a continuous wavelet transform graph, an envelope spectrum, a wavelet packet decomposition graph and an axis trajectory graph and an original waveform, which are respectively input into a corresponding residual convolutional neural network for feature extraction, to obtain a high-order representation feature map of each feature, and are introduced into a residual network block of an attention mechanism to calculate the weight of each input feature map and focus on the feature map with a larger weight. Finally, the input is input into a residual block to obtain different fault type judgment results and obtain the final output content of the model.

[0078] Next, a multi-task fault diagnosis device of a rotating mechanical equipment according to an embodiment of the present application is described with reference to the accompanying drawings.

[0079] Figure 11 is a structural schematic diagram of a multi-task fault diagnosis device of a rotating mechanical equipment provided by an embodiment of the present application.

[0080] As shown in Figure 11 , the multi-task fault diagnosis device 10 of the rotating mechanical equipment is applied to the model training stage and includes an acquisition module 110, a preprocessing module 120 and a training module 130.

[0081] The acquisition module 110 is configured to acquire at least one vibration signal of at least one rotating mechanical equipment during movement.

[0082] The preprocessing module 120 is configured to perform data preprocessing on the at least one vibration signal to obtain a spectrum, a continuous wavelet transform graph, an envelope spectrum, a wavelet packet decomposition graph and an axis trajectory graph of each vibration signal, and generate multi-class vibration feature data.

[0083] The training module 130 is configured to input the multi-class vibration feature data as a model and output a fault label corresponding to a rotating fault type to be diagnosed, and train a rotating mechanical multi-task fault diagnosis model based on a convolutional neural network.

[0084] Optionally, in an embodiment of the present application, the training module 130 is further configured to construct a separate residual convolutional neural network for each vibration feature of the multi-class vibration feature data to extract features of each vibration feature of the multi-class vibration feature data, to obtain a high-order representation feature map of each vibration feature.

[0085] Optionally, in an embodiment of the present application, the training module 130 further comprises a classification unit and an output unit.

[0086] The classification unit is configured to introduce a residual network block with an attention mechanism after obtaining the high-order representation feature map of each vibration feature, so as to pay attention to different feature maps according to the signal and the fault feature.

[0087] The output unit is configured to construct a corresponding number of residual network blocks based on the fault type to be diagnosed, so as to output the occurrence of each fault type.

[0088] It should be noted that the foregoing description of the embodiment of the method for diagnosing faults of a multi-task rotating machinery device also applies to the embodiment of the device for diagnosing faults of a multi-task rotating machinery device, which will not be described here again.

[0089] The device for diagnosing faults of a multi-task rotating machinery device according to the embodiment of the present application can train a neural network according to the rotating machinery fault type, fuse the multi-dimensional features of the preprocessed signal data, and simultaneously output the multi-class fault diagnosis results, thereby improving the pertinence and comprehensiveness of the fault diagnosis process and making the diagnosis results more accurate and reliable. Thus, the problems in the related art that the fault features and signal analysis when data is input into a model are not fully considered, different vibration analysis methods corresponding to different components of a rotating machinery, and the relevance between different component fault causes or features are not fully considered, resulting in low accuracy and poor generality of the obtained fault diagnosis results, and the rotating machinery equipment fault diagnosis cannot be efficiently and accurately and pertinently performed are solved.

[0090] The foregoing embodiment describes the model training stage, and the following embodiment describes the fault diagnosis stage.

[0091] Figure 12 FIG. 2 is a structural schematic diagram of another device for diagnosing faults of a multi-task rotating machinery device provided by the embodiment of the present application.

[0092] As shown in FIG. 2, the device for diagnosing faults of a multi-task rotating machinery device 20 is applied to the fault diagnosis stage and comprises an acquisition module 210 and a diagnosis module 220. Figure 12

[0093] The acquisition module 210 is configured to collect at least one vibration signal of a target rotating machinery device.

[0094] The diagnosis module 220 is configured to input the at least one vibration signal of the target rotating machinery device into the trained rotating machinery multi-task fault diagnosis model, and output a fault diagnosis result of the target rotating machinery device, wherein the rotating machinery multi-task fault diagnosis model takes multi-class vibration feature data as model input and takes a fault label corresponding to a rotating fault type to be diagnosed as output training.​

[0095] Optionally, in an embodiment of the present application, the fault type corresponding to the fault label includes shaft system fault, bearing fault and gear fault.

[0096] It should be noted that the foregoing explanation and description of the embodiment of the fault diagnosis method of the multi-task rotating machinery equipment also applies to the fault diagnosis device of the multi-task rotating machinery equipment of the embodiment, which will not be described here.

[0097] The fault diagnosis device of the multi-task rotating machinery equipment according to the embodiment of the present application can train the neural network according to the rotating machinery fault type, fuse the multi-dimensional features of the preprocessed signal data, and simultaneously output the multi-class fault diagnosis result, thereby improving the pertinence and comprehensiveness of the fault diagnosis process and making the diagnosis result more accurate and reliable. Thus, the problems in the related art that the fault characteristics and signal analysis when data is input into the model are not fully considered, different vibration analysis methods corresponding to different components of the rotating machinery, and the relevance between different component fault causes or characteristics are not considered, resulting in low accuracy and poor generality of the obtained fault diagnosis result, and the rotating machinery equipment fault diagnosis cannot be efficiently and accurately and pertinently implemented are solved.

[0098] Figure 13 The structure schematic diagram of the electronic device provided by the embodiment of the present application is shown. The electronic device can include:

[0099] The memory 1301, the processor 1302 and the computer program stored in the memory 1301 and executable on the processor 1302.

[0100] The processor 1302 implements the fault diagnosis method of the multi-task rotating machinery equipment provided in the above embodiment when executing the program.

[0101] Further, the electronic device further includes:

[0102] The communication interface 1303 is used for communication between the memory 1301 and the processor 1302.

[0103] The memory 1301 is used to store the computer program executable on the processor 1302.

[0104] The memory 1301 can include a high-speed RAM memory, and can also include a non-volatile memory, for example, at least one disk memory.

[0105] If the memory 1301, the processor 1302 and the communication interface 1303 are implemented independently, the communication interface 1303, the memory 1301 and the processor 1302 can be connected with each other through a bus and complete communication between each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For convenience of representation, Figure 13 Only one thick line is used to represent the bus in the figure, but it does not mean that there is only one bus or only one type of bus.

[0106] Optionally, in a specific implementation, if the memory 1301, the processor 1302 and the communication interface 1303 are integrated on a chip, the memory 1301, the processor 1302 and the communication interface 1303 can complete communication between each other through an internal interface.

[0107] The processor 1302 can be a Central Processing Unit (CPU), or an Application Specific Integrated Circuit (ASIC), or one or more integrated circuits configured to implement one or more embodiments of the present application.

[0108] The embodiment also provides a computer readable storage medium, which stores a computer program. The program is executed by a processor to implement the fault diagnosis method of the multi-task rotating machine equipment as described above.

[0109] In the description of the specification, the description of the terms "one embodiment", "some embodiments", "an example", "a specific example" or "some examples" means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In the 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 one or N embodiments or examples in a suitable manner. In addition, the person skilled in the art can combine and combine the different embodiments or examples described in the specification and the features of the different embodiments or examples without contradiction.

[0110] Moreover, the terms "first", "second", "third", etc. are used herein only to describe different steps or categories of steps. Thus, the use of the term "first" does not imply that different steps must be in a time sequence. Nor is it implied that a "first" step must precede a "second" step, that a "second" step, etc. must follow a "first" step, etc. Furthermore, when a process or method is described herein with several steps or several categories of steps, it should be understood that these are merely illustrative of the steps that can be employed in the process or method. Not all of the steps can be required, and in some cases, additional steps can be employed. The order of the steps can be varied, and some of the steps can be performed simultaneously. The steps can be performed in an order different than that described herein. The steps can be performed in any order, unless otherwise specified.

[0111] Any process or method described in flowcharts or otherwise described herein can be understood as representing a module, segment, or portion of code that includes one or N executable instructions for implementing the specified logical function or process. It will be understood that the scope of the preferred embodiments of the present application encompasses additional implementations that can not be expressly shown or described herein, including implementations that can be performed in an order different than that shown or discussed, including substantially simultaneously, or in reverse order, as appropriate, depending on the functionality involved.

[0112] The logic and / or steps represented in flowcharts or otherwise described herein, for example, can be considered as a list of executable instructions for implementing the logic function, and can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, processor- containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions. For purposes of this specification, a "computer-readable medium" can be any apparatus that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device. The computer-readable medium can be a product of nature, a manufacture, or combinations of both. More specific examples (a non-exhaustive list) of the computer-readable medium include the following: an electronic connection having one or N wires (electronic devices), a portable computer diskette (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber (optical devices), and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium can be paper or another suitable medium upon which the program is printed, as the program can be electronically captured, via the optically scanning of the paper or other suitable medium, then compiled, interpreted, or otherwise processed in a suitable manner, if necessary, and stored in a computer memory to execute the computer program.

[0113] It should be understood that portions of the application can be realized with a combination of hardware, software, firmware, or their combination. In the above-described embodiments, the N steps or methods can be realized with software or firmware stored in a memory and executed by a suitable instruction execution system. As in another embodiment, if realized with hardware, any one or their combination of the following technologies known in the art can be used: discrete logic circuit with logic gate circuit for implementing logic functions on data signals, application specific integrated circuit with suitable combination logic gate circuit, programmable gate array (PGA), field programmable gate array (FPGA), etc.

[0114] 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 methods 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.

[0115] In addition, each functional unit in each embodiment of the present application can be integrated into one processing module, or each unit can exist physically alone, or two or more units can be integrated into one module. The above-mentioned integrated module can be realized in the form of hardware or in the form of a software functional module. When the integrated module is realized in the form of a software functional module and sold or used as an independent product, it can also be stored in a computer readable storage medium.

[0116] 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 fault diagnosis method for multi-tasking rotating machinery, characterized in that, Applied to the model training phase, the method includes the following steps: Collect at least one vibration signal from at least one rotating mechanical device during its motion; Data preprocessing is performed on the at least one vibration signal to obtain the spectrum, continuous wavelet transform, envelope spectrum, wavelet packet decomposition diagram, and axisymmetric trajectory diagram for each vibration signal, generating multiple types of vibration feature data; and The multi-type vibration feature data is used as the model input, and the fault labels corresponding to the types of rotating faults to be diagnosed are used as the output to train a multi-task fault diagnosis model for rotating machinery based on a convolutional neural network. The training of the multi-task fault diagnosis model for rotating machinery based on convolutional neural networks includes: For each vibration feature of the multi-class vibration feature data, a separate residual convolutional neural network is constructed to extract features for each vibration feature of the multi-class vibration feature data, so as to obtain a high-order representation feature map of each vibration feature.

2. The method according to claim 1, characterized in that, The training of the multi-task fault diagnosis model for rotating machinery based on convolutional neural networks also includes: After obtaining the high-order representation feature map of each vibration feature, a residual network block with an attention mechanism is introduced to focus on different feature maps according to the signal and fault features. Based on the types of faults that need to be diagnosed, a corresponding number of residual network blocks are constructed to output the occurrence status of each type of fault.

3. A fault diagnosis method for multi-tasking rotating machinery, characterized in that, Applied to the fault diagnosis stage, the method includes the following steps: Collect at least one vibration signal from the target rotating mechanical equipment; At least one vibration signal of the target rotating machinery is input into the trained rotating machinery multi-task fault diagnosis model as described in claim 1, and the fault diagnosis result of the target rotating machinery is output. The rotating machinery multi-task fault diagnosis model is trained by using multiple types of vibration feature data as model input and fault labels corresponding to the types of rotating faults to be diagnosed as output.

4. The method according to claim 3, characterized in that, The fault types corresponding to the fault labels include shaft system faults, bearing faults, and gear faults.

5. A fault diagnosis device for multi-tasking rotating machinery, characterized in that, Applied to the model training phase, wherein the apparatus includes: The acquisition module is used to acquire at least one vibration signal of at least one rotating mechanical device during its motion. The preprocessing module is used to preprocess the at least one vibration signal to obtain the spectrum, continuous wavelet transform, envelope spectrum, wavelet packet decomposition, and axisymmetric trajectory of each vibration signal, generating multiple types of vibration feature data; and The training module is used to take the multi-type vibration feature data as model input and the fault labels corresponding to the types of rotating faults to be diagnosed as output, and train a multi-task fault diagnosis model for rotating machinery based on a convolutional neural network. The training module is further used to construct a separate residual convolutional neural network for each vibration feature of the multi-class vibration feature data to extract features for each vibration feature of the multi-class vibration feature data, so as to obtain a high-order representation feature map of each vibration feature.

6. A fault diagnosis device for multi-tasking rotating machinery, characterized in that, Applied to the fault diagnosis stage, the device includes: Acquisition module, used to acquire at least one vibration signal of the target rotating mechanical equipment; The diagnostic module is used to input at least one vibration signal of the target rotating machinery into the trained rotating machinery multi-task fault diagnosis model as described in claim 1, and output the fault diagnosis result of the target rotating machinery. The rotating machinery multi-task fault diagnosis model is trained by using multiple types of vibration feature data as model input and fault labels corresponding to the types of rotating faults to be diagnosed as output.

7. An electronic device, characterized in that, include: A memory, a processor, and a computer program stored in the memory and executable on the processor, the processor executing the program to implement the fault diagnosis method for a multi-tasking rotating machine as described in any one of claims 1-2 or 3-4.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by the processor to implement the fault diagnosis method for multi-tasking rotating machinery as described in any one of claims 1-2 or 3-4.

Images