CN113240000B - Machine state monitoring method, readable storage medium and electronic device - Google Patents
Machine state monitoring method, readable storage medium and electronic device Download PDFInfo
- Publication number
- CN113240000B CN113240000B CN202110503055.4A CN202110503055A CN113240000B CN 113240000 B CN113240000 B CN 113240000B CN 202110503055 A CN202110503055 A CN 202110503055A CN 113240000 B CN113240000 B CN 113240000B
- Authority
- CN
- China
- Prior art keywords
- frequency domain
- time
- layer
- monitoring method
- machine
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/80—Management or planning
Abstract
The embodiment of the invention provides a machine state monitoring method, a readable storage medium and electronic equipment, wherein the machine state monitoring method comprises the following steps: extracting time-frequency domain characteristics of a sensor signal for monitoring the machine state; inputting the single sensor signal and the single time-frequency domain characteristic into a neural network to classify target variables to obtain the classification precision of the single sensor signal and the single time-frequency domain characteristic; respectively sequencing the contribution degrees of the single sensor signals and the contribution degrees of the single time-frequency domain features; selecting the first m1 sensor signals with large contribution degrees and the first m2 time-frequency domain features to form a reduced input data set, wherein m1 and m2 are positive integers; and inputting the reduced input data set into a deep residual error network for training. The embodiment of the invention can solve the problems of low state monitoring precision and high monitoring cost caused by inconsistent sampling rate of the sensor, huge data and diversified types in machine state monitoring.
Description
Technical Field
The invention relates to the technical field of machine state monitoring, in particular to a machine state monitoring method of relative contribution degree feature selection and a depth residual error network, a readable storage medium and electronic equipment.
Background
Machine condition monitoring is becoming more and more important, and it is possible to predict the health of a machine, improving the usability of the machine and the reliability of the process. The reliability of the operation process of the machine equipment is high under the intelligent manufacturing requirement. In order to improve the availability of machines and to prevent problems such as production equipment shutdown and system shutdown when a fault occurs, state-based maintenance strategies are increasingly emphasized. It is less costly than preventative maintenance. Therefore, monitoring the state of the machine equipment is an important item. The state monitoring utilizes a sensor and a measuring means to detect the running state of the equipment, judges whether the equipment is in a normal working state or not, and is the basic work for carrying out fault diagnosis on the equipment.
There are several disadvantages to current machine condition monitoring:
1. with the complexity and the intellectualization of machine equipment, more and more data are acquired by state monitoring. Even a professional technician may not be able to respond to a machine fault accurately in time.
2. With the development of sensor technology, state monitoring faces the increasing diversification of data types, and more than one-dimensional time sequence data, and more other types of data such as images, and meanwhile, inconsistent sampling rates of different sensors are also one of the challenges in practical machine state monitoring applications.
3. Under the development of industry 4.0, the traditional industry is impacted by intellectualization, the interaction between an equipment monitoring model and a real environment is more and more emphasized, and the design requirements on a production system and a machine equipment monitoring model are higher and higher.
Disclosure of Invention
In order to solve at least one of the above technical problems, embodiments of the present invention provide a machine state monitoring method, a readable storage medium, and an electronic device, which solve the problems of low state monitoring accuracy and high monitoring cost caused by inconsistent sampling rates of sensors, huge data, and diversified types in machine state monitoring.
In one aspect, an embodiment of the present invention provides a machine state monitoring method, including:
extracting time-frequency domain characteristics of a sensor signal for monitoring the machine state;
inputting the single sensor signal and the single time-frequency domain characteristic into a neural network to classify target variables to obtain the classification precision of the single sensor signal and the single time-frequency domain characteristic;
respectively sequencing the contribution degrees of the single sensor signals and the contribution degrees of the single time-frequency domain features;
selecting the first m1 sensor signals with large contribution degrees and the first m2 time-frequency domain features to form a reduced input data set, wherein m1 and m2 are positive integers;
and inputting the reduced input data set into a deep residual error network for training.
In another aspect, the present invention further provides a readable storage medium, which has executable instructions thereon, and when the executable instructions are executed, the computer is caused to execute the steps in the machine state monitoring method described in any one of the foregoing embodiments.
In another aspect, an embodiment of the present invention further provides an electronic device, where the device includes a processor and a memory, where the memory stores computer program instructions adapted to be executed by the processor, and the computer program instructions, when executed by the processor, perform the steps in the machine state monitoring method according to any one of the above descriptions.
The machine state monitoring method provided by the embodiment of the invention solves the problem of inconsistent data length caused by inconsistent sampling rate of heterogeneous sensors of machine equipment in actual industry by extracting the time-frequency domain characteristics of the sensors, and can be applied to the state monitoring data processing and analysis of the machine equipment in different industrial scenes. Through the relative contribution degree feature selection step, the sensor and the feature quantity are selected, redundant sensors and time-frequency domain features are reduced, the calculation cost is reduced, and energy and resources are saved. The selected sensors and characteristics are trained through the depth residual error network model, and the state monitoring of the industrial machine equipment can be realized with high precision. Meanwhile, the method has expandability, and in the actual industry, different numbers of sensors and time-frequency domain characteristics can be used for monitoring the machine state by adopting the method.
Drawings
The accompanying drawings, which are included to provide a further understanding of embodiments of the invention and are incorporated in and constitute a part of this specification, illustrate exemplary embodiments of the invention and together with the description serve to explain the principles of the embodiments of the invention.
FIG. 1 is a schematic flow chart illustrating a method for monitoring machine condition according to an embodiment of the present invention;
FIG. 2 is a schematic processing flow diagram of a single hidden layer feedforward neural network according to an embodiment of the present invention;
fig. 3 is a schematic diagram of an exemplary structure of the electronic device of the present invention.
Detailed Description
The embodiments of the present invention will be described in further detail with reference to the drawings and the following description. It should be understood that the detailed description and specific examples, while indicating the embodiments of the invention, are given by way of illustration only. It should be further noted that, for the convenience of description, only the portions related to the embodiments of the present invention are shown in the drawings.
It should be noted that, in the embodiments of the present invention, features in the embodiments may be combined with each other without conflict. Embodiments of the present invention will be described in detail below with reference to the accompanying drawings in conjunction with embodiments.
It should be noted that, the step numbers in the text are only for convenience of explanation of the specific embodiments, and do not serve to limit the execution sequence of the steps.
The method provided by the embodiment of the present invention can be executed by a relevant processor, and the following description takes the processor as an execution subject as an example. The execution subject can be adjusted according to the specific case, such as a server, an electronic device, a computer, and the like.
The embodiment of the invention provides a machine state monitoring method, which is used for respectively analyzing single-factor contribution degrees of sensors and characteristics, selecting better sensor signals and time-frequency domain characteristics, and carrying out state monitoring on an industrial machine by using a relative contribution degree characteristic selection and depth residual error network method.
In one aspect, referring to fig. 1, a method for monitoring a machine state is provided, including:
and S11, extracting the time-frequency domain characteristics of the sensor signals for monitoring the machine state. Optionally, the time-frequency domain features include one or more of a mean value, a variance, a root mean square, a peak value, a crest factor, a kurtosis factor, a form factor, a pulse factor, and a margin factor; the frequency domain features are selected from one or more of center of gravity frequency, frequency variance, and mean square frequency. By performing the above time-frequency domain processing on the sensor signals to extract the health characteristics of the machine equipment, the sampling rates of heterogeneous sensors can be kept consistent.
And S12, inputting the single sensor signal and the single time-frequency domain feature into the neural network for target variable classification to obtain the classification accuracy of the single sensor signal and the single time-frequency domain feature. By calculating the classification accuracy of the single sensor signal and the single time-frequency domain feature, which sensor signals and which time-frequency domain feature have a larger relative contribution to the classification accuracy can be obtained, and the specific process is described in detail in an embodiment later.
And S13, respectively sequencing the contribution degrees of the single sensor signals and the contribution degrees of the single time-frequency domain features.
S14, selecting the first m1 sensor signals with large contribution degrees and the first m2 time-frequency domain features to form a reduced input data set, wherein m1 and m2 are positive integers. Through the relative contribution degree feature selection step, the sensor signals and the feature quantity are selected, redundant sensor signals and time-frequency domain features are reduced, the calculation cost is reduced, and energy and resources are saved.
And S15, inputting the reduced input data set into a deep residual error network for training. The selected sensor signals and the time-frequency domain characteristics are trained through the depth residual error network model, and the state monitoring of the industrial machine equipment can be realized with higher precision.
In an implementation manner of the embodiment of the present invention, in step S12, in the step of inputting the single sensor signal and the single time-frequency domain feature into the neural network to perform target variable classification, and obtaining the classification accuracy of the single sensor signal and the single time-frequency domain feature, the neural network used is a single hidden layer feedforward neural network. Different neural network architectures can be adopted to calculate the classification accuracy of the sensor signals and the time-frequency domain features.
Referring to fig. 2, the selection of the sensor signals and the time-frequency domain features includes two parts, one is the selection of the number of the sensor signals, and the other is the selection of the number of the extracted time-frequency domain features. The specific method is described below by taking the selection of the number of sensor signals as an example, and the method for selecting the number of extracted time-frequency domain features is the same as the step for selecting the number of sensor signals, except that the time-frequency domain features are used for replacing the sensor signals therein as processing objects.
And (3) inputting the single sensor signal and the single time-frequency domain characteristic into a neural network to classify the target variable, and observing the classification precision of each sensor signal. Specifically, for a given training set:
D＝{(x 1 ,y 1 ),(x 2 ,y 2 )...(x m ,y m )}
wherein x is i Is a d-dimensional vector, y i Vectors in l dimensions, i.e. arguments consisting of d dimensions, with output values in l dimensionsAnd correspondingly, constructing a single hidden layer feedforward neural network consisting of d input layer neurons, q hidden layer neurons and l output layer neurons. Wherein the threshold value of the jth neuron of the output layer is theta j The threshold of the h-th neuron of the hidden layer is represented by gamma h The connection weight between the ith neuron of the input layer and the h neuron of the hidden layer is represented as v ih The connection weight between the h-th neuron of the hidden layer and the h-th neuron of the output layer is w hj Note that the input received by the h-th neuron of the hidden layer is:
let b be h If the output of the h-th neuron in the hidden layer is positive, the input received by the j-th neuron in the output layer is:
hidden and output layers each functional neuron uses a Sigmoid-type function:
for any (x) in the training set k ,y k ) Assume that the output of the neural network is:
namely, it is
Then the network is at (x) k ,y k ) The mean square error above is:
assume that the connection weight from the hidden layer to the output layer is w hj Given a learning rate η, there are
Then
All final results w hj The update formula of (2):
Δw hj ＝ηg j b h
similarly, the update formula of other parameters:
Δθ j ＝-ηg j
Δv ih ＝ηe h x i
Δγ h ＝-ηe h
wherein e h Comprises the following steps:
optionally, an early-stop and regularization technique is used in the training process to prevent the occurrence of an overfitting condition. Assuming that a fault X has A, B, C degradation states in total, a single sensor signal is classified in each degradation state with the classification accuracy:
wherein n is a positive integer; in this example of sensors n denotes the number of sensors, for example: if 17 sensors collect state signals, n is 17, three types of faults ABC exist in total,
there should be a classification accuracy for each sensor after a single sensor analysis,
A＝{a1.a2…a17}
B＝{b1.b2…b17}
C＝{c1.c2…c17}
according to the sorting of the sorting precision from high to low, according to different actual equipment conditions, some equipment may only need to select the first three complementary sets with the highest sorting to achieve 90% of sorting precision, and in this case, the first three sorting precision with the highest sorting is selected as a reduced input data set, that is, m is 3.
The above is only an example for a sensor, and the time-frequency domain features and the sensor are processed separately and in parallel, in the same way, n represents the number of time-frequency domain features in the example of the time-frequency domain features.
Optionally, the single-factor precision of each type of degradation state is normalized:
respectively taking the m-before contribution degree set of each type of degraded state single sensor signals after being sorted as A * ,B * ,C * Wherein m is less than or equal to n,
the complement of the three types of degradation states is the number of the selected sensor signals, and similarly, the processing of the single time-frequency domain features is also the same, so that the selected reduced input data set is finally obtained.
In one implementation of the embodiment of the present invention, the deep residual network includes a residual block, a global average pooling layer, and a Softmax layer. The Softmax function is a function commonly used in deep learning, and is generally arranged at the last layer of a neural network to serve as a classifier. Adding the output and the input of the residual block and transmitting the sum to the next layer; the global average pooling layer is used for receiving a result of adding the output and the input of the residual error block and transmitting a processing result to the Softmax layer; the Softmax layer is used to get the results of the tag prediction.
Optionally, there are at least two residual blocks, and one of the global average pooling layer and one of the Softmax layers are provided.
Further, the residual block is composed of three basic blocks including a convolution layer, a batch normalization layer, and a ReLU activation layer. A recirculation (Rectified Linear Unit), also called a modified Linear Unit, is an activation function (activation function) commonly used in an artificial neural network, and generally refers to a nonlinear function represented by a ramp function and a variation thereof.
Specifically, the deep residual network links the output of the residual block to the input through a quick connection between successive convolutional layers, so that the gradient directly flows through the bottom layer, the gradient disappearance effect is reduced, and the network can be expanded to a very deep structure. The residual network follows two rules in the design process, first, layers have the same number of filters for the same output feature size; second, if the signature graph size is halved, the number of filters is doubled to preserve the temporal complexity of each layer.
The residual structure can be expressed in the form of
x l+1 ＝x l +F(x l ,W l )
X herein l And x l+1 Is the input and output vector of the residual structure under consideration, F (x) l ,W l ) Representing the residual mapping to be learned. By recursion, the expression of the characteristics of any deep unit L can be obtained:
i.e. feature x for arbitrarily deep cell L L Feature x, which can be expressed as shallow cell l l Is added with a shape as
Indicating the arbitrary units L andl all have residual properties between them. Similarly, for an arbitrarily deep cell L, its characteristic is denoted x L ＝x 0 +
I.e. the sum of the outputs of all previous residual functions plus x 0 Characteristic x of L Is the product of a series of matrix vectors, and the computation of summation is much less than the computation of product. For back propagation, assuming a loss function of E, the following formula can be obtained according to the chain rule of back propagation:
the method not only ensures that the signal can be directly transmitted back to any shallow layer, but also ensures that the phenomenon of gradient disappearance can not occur.
Specifically, the input layer of the depth residual error network is a matrix composed of sensor signals and time-frequency domain characteristics, each row represents the time-frequency domain characteristics selected by the sensor signals in one loading period, and each column represents the selected sensor signals. The deep residual network may be composed of three residual blocks, one global averaging pooling layer and one Softmax layer, each residual block being composed of three basic blocks, performing convolution and batch normalization operations, respectively, and feeding the results to the ReLU activation function. The output of the residual block is added to the input and passed to the next layer, with filters of 64, 128, respectively. The output of the residual block is sent to the global average pooling layer and finally the Softmax layer (classifier) is fully connected to the output of the global average pooling layer to derive the result of the label prediction.
In summary, the machine state monitoring method according to the embodiment of the present invention performs state monitoring on the industrial machine by using a relative contribution feature selection and a deep residual error network method, and by selecting appropriate sensor signals and time-frequency domain features, can implement state monitoring of the machine equipment with higher accuracy for the problems of various actual industrial machine sensors, inconsistent sampling rates, huge data volumes, various data types, and the like, and mainly has the following functions and advantages:
1. by extracting the time-frequency domain characteristics of the sensor, the problem of inconsistent data length caused by inconsistent sampling rate of heterogeneous sensors of the machine equipment in the actual industry is solved, and the method can be applied to the state monitoring data processing and analysis of the machine equipment in different industrial scenes.
2. Through the selection of the relative contribution degree characteristics, the quantity of the sensor signals and the time-frequency domain characteristics is selected, redundant sensor signals and time-frequency domain characteristics are reduced, the calculation cost is reduced, and energy and resources are saved.
3. The selected sensor signals and the time-frequency domain characteristics are trained through a deep residual error network model, the model comprises three residual error blocks, each residual error block is composed of three basic blocks including a convolutional layer, a batch normalization layer and a ReLU activation layer, and the output of the residual error block is added into the input and transmitted to the next layer. The final prediction result is output through a global average pool and a softmax classifier, and the condition monitoring of the industrial machine equipment can be realized with higher precision.
4. The method has expandability, and in the actual industry, different quantities and sensor signals and time-frequency domain characteristics can be used for monitoring the machine state by the method.
In yet another aspect of the embodiments of the present invention, there is also provided a readable storage medium having executable instructions thereon, when executed, cause a computer to perform the steps of the machine condition monitoring method described in any one of the preceding claims.
In another aspect of the embodiment of the present invention, an electronic device is further provided, and an exemplary structural diagram of the electronic device shown in fig. 3 includes a communication interface 1000, a memory 2000 and a processor 3000. The communication interface 1000 is used for communicating with an external device to perform data interactive transmission. The memory 2000 has stored therein a computer program that is executable on the processor 3000. The number of the memory 2000 and the processor 3000 may be one or more.
If the communication interface 1000, the memory 2000, and the processor 3000 are implemented independently, the communication interface 1000, the memory 2000, and the processor 3000 may be connected to each other through a bus and perform communication with each other. The bus may be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, an Extended ISA (Extended Industry Standard Component) bus, or the like. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, only one thick line is shown, but this does not represent only one bus or one type of bus.
Optionally, in a specific implementation, if the communication interface 1000, the memory 2000, and the processor 3000 are integrated on a chip, the communication interface 1000, the memory 2000, and the processor 3000 may complete communication with each other through an internal interface.
The processor is configured to perform one or more steps of the machine condition monitoring method according to any of the embodiments. The processor may be a Central Processing Unit (CPU), or other general-purpose processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component, or the like. Wherein a general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
The memory stores computer program instructions adapted to be executed by the processor, and the computer program instructions, when executed by the processor, perform one or more steps of the machine condition monitoring method according to any of the above embodiments.
The Memory may be, but is not limited to, a Read-Only Memory (ROM) or other type of static storage device that can store static information and instructions, a Random Access Memory (RAM) or other type of dynamic storage device that can store information and instructions, an Electrically Erasable Programmable Read-Only Memory (EEPROM), a Compact Disc Read-Only Memory (CD-ROM) or other optical Disc storage, optical Disc storage (including Compact Disc, laser Disc, optical Disc, digital versatile Disc, blu-ray Disc, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer. The memory may be self-contained and coupled to the processor via a communication bus. The memory may also be integral to the processor.
In the description herein, reference to the description of the terms "one embodiment/mode," "some embodiments/modes," "example," "specific example," or "some examples," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment/mode or example is included in at least one embodiment/mode or example of the application. In this specification, the schematic representations of the terms used above are not necessarily intended to be the same embodiment/mode or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments/modes or examples. Furthermore, the various embodiments/modes or examples and features of the various embodiments/modes or examples described in this specification can be combined and combined by one skilled in the art without being mutually inconsistent.
Furthermore, the terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of the embodiments of the present invention, "a plurality" means at least two, e.g., two, three, etc., unless specifically limited otherwise. Meanwhile, in the description of the embodiments of the present invention, unless explicitly specified or limited otherwise, the terms "connected" and "connected" should be interpreted broadly, for example, as being fixedly connected, detachably connected, or integrally connected; the connection can be mechanical connection or electrical connection; may be directly connected or indirectly connected through an intermediate. Specific meanings of the above terms in the embodiments of the present invention can be understood by those of ordinary skill in the art according to specific situations.
It should be understood by those skilled in the art that the foregoing embodiments are merely for clearly illustrating the examples of the present invention, and are not intended to limit the scope of the examples of the present invention. Other variations or modifications will occur to those skilled in the art based on the foregoing disclosure and are within the scope of the embodiments of the invention.
Claims (10)
1. A method of monitoring a condition of a machine, comprising:
extracting time-frequency domain characteristics of a sensor signal for monitoring the machine state;
inputting the single sensor signal and the single time-frequency domain characteristic into a neural network to classify target variables to obtain the classification precision of the single sensor signal and the single time-frequency domain characteristic;
respectively sequencing the contribution degrees of the single sensor signals and the contribution degrees of the single time-frequency domain features;
selecting the first m1 sensor signals with large contribution degrees and the first m2 time-frequency domain features to form a reduced input data set, wherein m1 and m2 are positive integers;
and inputting the reduced input data set into a deep residual error network for training.
2. A machine condition monitoring method as claimed in claim 1, wherein the time-frequency domain features comprise time-domain features and/or frequency-domain features; the time domain features include at least one of a mean, a variance, a root mean square, a peak, a crest factor, a kurtosis factor, a form factor, a pulse factor, and a margin factor; the frequency domain features include at least one of a center of gravity frequency, a frequency variance, and a mean square frequency.
3. The machine state monitoring method according to claim 1, wherein in the step of inputting the single sensor signal and the single time-frequency domain feature into a neural network for target variable classification to obtain the classification accuracy of the single sensor signal and the single time-frequency domain feature, the neural network used is a single hidden layer feedforward neural network.
4. The machine condition monitoring method according to claim 3, wherein after the single sensor signal and the single time-frequency domain feature are input into the neural network for target variable classification to obtain the classification accuracy of the single sensor signal and the single time-frequency domain feature, the method further comprises:
and carrying out normalization processing on the classification precision.
5. The machine state monitoring method of any of claims 1-4, wherein the deep residual network comprises a residual block, a global average pooling layer, and a Softmax layer;
adding the output and the input of the residual block and transmitting the sum to the next layer; the global average pooling layer is used for receiving a result of adding the output and the input of the residual error block and transmitting a processing result to the Softmax layer; the Softmax layer is used to obtain the result of the tag prediction.
6. The machine-state monitoring method according to claim 5, wherein the residual block is provided in at least two, and one each of the global averaging pooling layer and the Softmax layer is provided.
7. The machine state monitoring method of claim 5, wherein the residual block is comprised of three basic blocks, the basic blocks including a convolutional layer, a batch normalization layer, and a ReLU activation layer.
8. The machine-state monitoring method of claim 7, wherein an input layer of the deep residual network is a matrix of sensor signals and time-frequency domain features.
9. A readable storage medium having executable instructions thereon which, when executed, cause a computer to perform the steps in the machine condition monitoring method of any one of claims 1-8.
10. An electronic device, characterized in that the device comprises a processor and a memory, in which computer program instructions adapted to be executed by the processor are stored, which computer program instructions, when executed by the processor, perform the steps in the machine condition monitoring method according to any of claims 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110503055.4A CN113240000B (en) | 2021-05-10 | 2021-05-10 | Machine state monitoring method, readable storage medium and electronic device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110503055.4A CN113240000B (en) | 2021-05-10 | 2021-05-10 | Machine state monitoring method, readable storage medium and electronic device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113240000A CN113240000A (en) | 2021-08-10 |
| CN113240000B true CN113240000B (en) | 2022-08-23 |
Family
ID=77132981
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110503055.4A Active CN113240000B (en) | 2021-05-10 | 2021-05-10 | Machine state monitoring method, readable storage medium and electronic device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113240000B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180284743A1 (en) * | 2016-05-09 | 2018-10-04 | StrongForce IoT Portfolio 2016, LLC | Methods and systems for industrial internet of things data collection for vibration sensitive equipment |
| CN108647707A (en) * | 2018-04-25 | 2018-10-12 | 北京旋极信息技术股份有限公司 | Probabilistic neural network creation method, method for diagnosing faults and device, storage medium |
| US20190294987A1 (en) * | 2018-03-21 | 2019-09-26 | Jiangnan University | Multilevel Pattern Monitoring Method for Industry Processes |
| CN110674892A (en) * | 2019-10-24 | 2020-01-10 | 北京航空航天大学 | Fault feature screening method based on weighted multi-feature fusion and SVM classification |
| CN111398820A (en) * | 2020-05-07 | 2020-07-10 | 南京凯奥思数据技术有限公司 | Motor health state online monitoring method |
-
2021
- 2021-05-10 CN CN202110503055.4A patent/CN113240000B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180284743A1 (en) * | 2016-05-09 | 2018-10-04 | StrongForce IoT Portfolio 2016, LLC | Methods and systems for industrial internet of things data collection for vibration sensitive equipment |
| US20190294987A1 (en) * | 2018-03-21 | 2019-09-26 | Jiangnan University | Multilevel Pattern Monitoring Method for Industry Processes |
| CN108647707A (en) * | 2018-04-25 | 2018-10-12 | 北京旋极信息技术股份有限公司 | Probabilistic neural network creation method, method for diagnosing faults and device, storage medium |
| CN110674892A (en) * | 2019-10-24 | 2020-01-10 | 北京航空航天大学 | Fault feature screening method based on weighted multi-feature fusion and SVM classification |
| CN111398820A (en) * | 2020-05-07 | 2020-07-10 | 南京凯奥思数据技术有限公司 | Motor health state online monitoring method |
Non-Patent Citations (1)
| Title |
|---|
| Machinery condition monitoring in the era of industry 4.0: A relative degree of contribution feature selection and deep residual network combined approach;Shuaijie Mei et al.;《Computers & Industrial Engineering》;20220326;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113240000A (en) | 2021-08-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Chen et al. | An information granulation based data mining approach for classifying imbalanced data | |
| Ko et al. | Fault classification in high-dimensional complex processes using semi-supervised deep convolutional generative models | |
| US11650968B2 (en) | Systems and methods for predictive early stopping in neural network training | |
| CN111352965B (en) | Training method of sequence mining model, and processing method and equipment of sequence data | |
| CN111665819A (en) | Deep learning multi-model fusion-based complex chemical process fault diagnosis method | |
| KR102531645B1 (en) | Computer program for performance testing of models | |
| CN112651418B (en) | Data classification method, classifier training method and system | |
| CN115687925A (en) | Fault type identification method and device for unbalanced sample | |
| CN114154557A (en) | Cancer tissue classification method, apparatus, electronic device, and storage medium | |
| CN114118138A (en) | Bearing composite fault diagnosis method based on multi-label field self-adaptive model | |
| CN110111365B (en) | Training method and device based on deep learning and target tracking method and device | |
| Al-Akkam et al. | Plants Leaf Diseases Detection Using Deep Learning | |
| Xiu et al. | Variational disentanglement for rare event modeling | |
| Zhong et al. | CAE‐WANN: A novel anomaly detection method for gas turbines via search space extension | |
| CN109858245A (en) | A kind of intrusion detection method based on improvement depth confidence network | |
| CN113240000B (en) | Machine state monitoring method, readable storage medium and electronic device | |
| Ebenezer et al. | A comprehensive analysis of handling imbalanced dataset | |
| CN116739100A (en) | Vulnerability detection method of quantum neural network and automatic driving vulnerability detection method | |
| CN113935413A (en) | Distribution network wave recording file waveform identification method based on convolutional neural network | |
| Ushio et al. | The application of deep learning to predict corporate growth | |
| Rout | A Novel Grid Ann for Prediction of Heart Disease | |
| Adhikarla | Automated bug classification.: Bug report routing | |
| Calma | Active Learning with Uncertain Annotators: Towards Dedicated Collaborative Interactive Learning | |
| Joshi et al. | A Principal Component Analysis, Sampling and Classifier strategies for dealing with concerns of class imbalance in datasets with a ratio greater than five | |
| Yin et al. | Fault diagnosis method based on CWGAN-GP-1DCNN |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |