CN112836346A - Motor fault diagnosis method based on CN and PCA, electronic equipment and medium - Google Patents
Motor fault diagnosis method based on CN and PCA, electronic equipment and medium Download PDFInfo
- Publication number
- CN112836346A CN112836346A CN202110020431.4A CN202110020431A CN112836346A CN 112836346 A CN112836346 A CN 112836346A CN 202110020431 A CN202110020431 A CN 202110020431A CN 112836346 A CN112836346 A CN 112836346A
- Authority
- CN
- China
- Prior art keywords
- data
- channel
- original data
- binary data
- bit binary
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 70
- 238000003745 diagnosis Methods 0.000 title claims abstract description 55
- 239000011159 matrix material Substances 0.000 claims abstract description 46
- 238000013507 mapping Methods 0.000 claims abstract description 33
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 19
- 230000009467 reduction Effects 0.000 claims abstract description 7
- 238000004590 computer program Methods 0.000 claims description 13
- 239000000470 constituent Substances 0.000 claims description 12
- 238000003860 storage Methods 0.000 claims description 11
- 238000012549 training Methods 0.000 claims description 10
- 238000010606 normalization Methods 0.000 claims description 9
- 230000001133 acceleration Effects 0.000 claims description 5
- 238000006243 chemical reaction Methods 0.000 claims description 5
- 238000005070 sampling Methods 0.000 claims description 5
- 230000009977 dual effect Effects 0.000 claims description 2
- 238000001514 detection method Methods 0.000 abstract description 7
- 238000000513 principal component analysis Methods 0.000 description 34
- 230000008569 process Effects 0.000 description 15
- 230000000875 corresponding effect Effects 0.000 description 14
- 238000012545 processing Methods 0.000 description 11
- 238000005516 engineering process Methods 0.000 description 6
- 238000000605 extraction Methods 0.000 description 6
- 238000012360 testing method Methods 0.000 description 5
- 230000009466 transformation Effects 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 238000004519 manufacturing process Methods 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 238000013527 convolutional neural network Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000012544 monitoring process Methods 0.000 description 3
- 238000002759 z-score normalization Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 230000009347 mechanical transmission Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000009877 rendering Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000005096 rolling process Methods 0.000 description 1
- 238000005549 size reduction Methods 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 239000000758 substrate Substances 0.000 description 1
- 239000013598 vector Substances 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
- 238000004804 winding Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/34—Testing dynamo-electric machines
- G01R31/343—Testing dynamo-electric machines in operation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/21—Design, administration or maintenance of databases
- G06F16/215—Improving data quality; Data cleansing, e.g. de-duplication, removing invalid entries or correcting typographical errors
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- Databases & Information Systems (AREA)
- Mathematical Optimization (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Analysis (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Pure & Applied Mathematics (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Health & Medical Sciences (AREA)
- Computational Mathematics (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Computer Hardware Design (AREA)
- Algebra (AREA)
- Geometry (AREA)
- Quality & Reliability (AREA)
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
- Control Of Electric Motors In General (AREA)
Abstract
The embodiment of the invention relates to the technical field of fault detection, and discloses a motor fault diagnosis method based on CN and PCA, electronic equipment and a medium. The method comprises the following steps: receiving raw data of a motor collected by a sensor; normalizing the original data and converting the normalized original data into binary data with 15 bits; performing space mapping on the binary data by using a CN algorithm to obtain a 10-dimensional space mapping matrix; performing dimensionality reduction on the space mapping matrix by using a PCA algorithm to obtain a 3-dimensional feature matrix; and inputting the characteristic matrix into a pre-trained network model, and outputting a diagnosis result. By implementing the embodiment of the invention, the spatial separability of the fault types can be improved, and the detection efficiency of the fault diagnosis algorithm can be improved.
Description
Technical Field
The invention relates to the technical field of fault detection, in particular to a motor fault diagnosis method based on CN and PCA, electronic equipment and a medium.
Background
Advanced manufacturing is the primary engine and prime mover for innovation-driven development and high-quality development of the economic society. In the field of high-end equipment and intelligent manufacturing, an electric motor can directly convert electric energy into mechanical energy of linear motion, and has the advantages of large thrust, high force density, long stroke, low inertia, fast dynamic response, simple mechanical structure and the like.
The motor directly drives the motion equipment, a mechanical transmission mechanism is omitted, the physical limit limits of the speed and the acceleration of a mechanical transmission element are completely eliminated, and the motor is widely applied to a reciprocating servo system, an industrial robot and a high-precision positioning direct drive system.
The fault diagnosis technology comprises the following contents that whether the equipment works normally or not is judged by various monitoring means under the running state or working state of the equipment; if the fault is abnormal, the fault is indicated through analysis and judgment, and the maintenance is convenient for managers; or before the fault does not occur, a forecast of possible faults is provided, so that management personnel can take measures as early as possible to avoid the faults or avoid major faults, thereby causing shutdown and production halt and bringing major economic losses to the engineering. This is the task of the fault diagnosis technology and the purpose of developing the fault diagnosis technology of the equipment.
Under the background of the generation of big data, the method is accompanied by great computational complexity and modeling complexity, and the data-driven intelligent fault diagnosis method is more applicable to the directness and the effectiveness of statistical analysis and information extraction on massive, multi-source and high-dimensional data. The technology takes the collected monitoring data of different sources and different types as a substrate, and various data mining technologies are utilized to obtain implicit useful information, so as to represent a normal mode and a fault mode of system operation, and further achieve the purposes of detection and diagnosis.
The performance of the intelligent fault diagnosis method greatly depends on the quality of the extracted features, including real-time change, stage change, trend change, fault modes and the like of the features, namely, the representation learning of data is the core of the intelligent fault diagnosis technology. The conventional feature expression learning method has the following problems:
(1) the proper characteristic extraction method can be designed only by needing prior information, professional knowledge and deep mathematical basis of the field;
(2) most of the extracted features are shallow features, and the generalization capability of the extracted features is limited to a certain extent aiming at the problem of complex classification;
(3) subject to the physical characteristics of the mechanical system, component or fault condition variations may significantly alter the feature extraction method or its evaluation criteria;
(4) the feature extraction depends on the original features and evaluation standards, and has certain limitation on the mining of new features.
For massive state data and monitoring variables in the production process and equipment operation, information acquisition is usually in the form of multidimensional vectors, such as: vibration information of the base in the running process of the rotary equipment, vibration information of the driving end, current and voltage information in the running process of the equipment and the like. For multi-channel sensor information, how to fuse fault information acquired by sensors into a multi-channel matrix form is a key for realizing fault diagnosis based on a convolutional neural network. In the convolutional neural network fault diagnosis method based on matrix form input, a matrix form sample processing method realized by traditional fault data splicing has certain randomness, and a deep network for fault diagnosis is inevitably required to have stronger feature extraction robustness and stronger classification generalization capability. Stronger feature extraction robustness and stronger generalization capability inevitably require a relatively complex network structure to meet the requirements. Complex convolutional networks necessarily require enough samples for model pre-training and stronger hardware support in the real-time fault diagnosis process.
Disclosure of Invention
Aiming at the defects, the embodiment of the invention discloses a motor fault diagnosis method based on CN and PCA, electronic equipment and a medium, wherein a CN (color Names) multi-color space model method is applied to map original data to a high-dimensional space, the dimension of a high-dimensional space mapping result is reduced by a PCA method, and finally, a fault data high-dimensional space principal element is obtained, the spatial separability of a fault type is improved, and the detection efficiency of a fault diagnosis algorithm is improved.
The embodiment of the invention discloses a motor fault diagnosis method based on CN and PCA in the first aspect, which comprises the following steps:
receiving raw data of a motor collected by a sensor;
normalizing the original data and converting the normalized original data into binary data with 15 bits;
performing space mapping on the binary data by using a CN algorithm to obtain a 10-dimensional space mapping matrix;
performing dimensionality reduction on the space mapping matrix by using a PCA algorithm to obtain a 3-dimensional feature matrix;
and inputting the characteristic matrix into a pre-trained network model, and outputting a diagnosis result.
As an optional implementation manner, in the first aspect of the embodiment of the present invention, the receiving raw data of the motor collected by the sensor includes:
receiving single-channel or multi-channel raw data of the motor collected by one or more types of sensors;
the sensor is any one or more of a vibration sensor, a voltage transformer, a current transformer and an acceleration sensor.
As an optional implementation manner, in the first aspect of the embodiment of the present invention, the original data is single-channel original data;
the normalizing the original data and converting the normalized original data into binary data with 15 bits includes:
normalizing the single-channel original data to be 0-32, namely expressing each single-channel original data by using 5-bit binary data;
selecting adjacent 3 binary data with 5 bits to be connected in sequence to form binary data with 15 bits:
Si＝{Mi-1,Mi,Mi+1}
wherein S isiFor the constituent i-th 15-bit binary data, Mi5-bit binary data normalized for the ith original data of a single channel; i is more than or equal to 0 and less than or equal to L, and L is the length of single-channel original data; when i is 0, obtaining M by using a linear difference modei-1When i is equal to L, M is obtained by using a linear difference modei+1。
As an optional implementation manner, in the first aspect of the embodiment of the present invention, the original data is dual-channel original data;
the normalizing the original data and converting the normalized original data into binary data with 15 bits includes:
normalizing the two-channel original data to be 0-127 and 0-255 respectively, namely expressing the original data of one channel by adopting 7-bit binary data, and expressing the original data of the other channel by adopting 8-bit binary data;
selecting 7-bit binary data after one channel is specified to be connected with 8-bit binary data after the other channel is specified to form 15-bit binary data:
Si＝{Mi,Ni}
wherein S isiFor the constituent i-th 15-bit binary data, MiNormalized 7-bit binary data of the ith original data of one channel; n is a radical ofiNormalized 8-bit binary data of the ith original data of the other channel; i is more than or equal to 0 and less than or equal to L, and L is the length of the original data of each channel.
As an optional implementation manner, in the first aspect of the embodiment of the present invention, the original data is three-channel original data;
the normalizing the original data and converting the normalized original data into binary data with 15 bits includes:
normalizing the three-channel original data to be 0-32, namely expressing the original data of each channel in the three-channel original data by using 5-bit binary data;
selecting the corresponding original data of three channels to connect, and forming 15-bit binary data:
Si＝{Mi,Ni,Oi}
wherein S isiFor the constituent i-th 15-bit binary data, Mi、Ni、OiRespectively 5-bit binary data of normalized ith original data of the three channels; i is more than or equal to 0 and less than or equal to L, and L is the length of the original data of each channel.
As an optional implementation manner, in the first aspect of the embodiment of the present invention, performing spatial mapping on the binary data by using a CN algorithm to obtain a 10-dimensional spatial mapping matrix, includes:
and performing high-dimensional space mapping on the data set consisting of the binary data of 15 bits by using a conversion matrix 32768 × 10 to obtain a mapped 10-bit space mapping matrix L × 10, wherein L is the length of the original data, namely the number of the data sets consisting of the binary data of 15 bits.
As an optional implementation manner, in the first aspect of the embodiment of the present invention, the pre-trained network model includes:
acquiring multiple groups of motor sample data with different fault types and motor sample data without faults, and constructing a sample set;
and creating a network initial model, and training the network initial model by using the sample data of each motor in the sample set to obtain the trained network model.
As an optional implementation manner, in the first aspect of the embodiment of the present invention, the length L of the original data satisfies:
wherein, beta is the sampling frequency of the sensor, and n is the rotating speed of the motor.
A second aspect of an embodiment of the present invention discloses an electronic device, including: a memory storing executable program code; a processor coupled with the memory; the processor calls the executable program code stored in the memory for executing the motor fault diagnosis method based on the CN and the PCA disclosed by the first aspect of the embodiment of the invention.
A third aspect of the embodiments of the present invention discloses a computer-readable storage medium storing a computer program, where the computer program enables a computer to execute a method for diagnosing a motor fault based on CN and PCA disclosed in the first aspect of the embodiments of the present invention.
A fourth aspect of the embodiments of the present invention discloses a computer program product, which, when running on a computer, causes the computer to execute a method for diagnosing a fault of a motor based on CN and PCA as disclosed in the first aspect of the embodiments of the present invention.
A fifth aspect of the embodiments of the present invention discloses an application publishing platform, where the application publishing platform is used to publish a computer program product, where when the computer program product runs on a computer, the computer is caused to execute the method for diagnosing a motor fault based on CN and PCA disclosed in the first aspect of the embodiments of the present invention.
Compared with the prior art, the embodiment of the invention has the following beneficial effects:
the fault diagnosis method based on the CN and the PCA has high diagnosis precision, and meanwhile, the model convergence is fast, so that the separability of different types of data can be improved, and the fault diagnosis efficiency of the algorithm is further improved.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings without creative efforts.
FIG. 1 is a flow chart illustrating a motor fault diagnosis method based on CN and PCA according to an embodiment of the present invention;
FIG. 2 is a schematic structural diagram of a motor fault diagnosis device based on CN and PCA according to the embodiment of the invention;
fig. 3 is a schematic structural diagram of an electronic device disclosed in the embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
It should be noted that the terms "first", "second", "third", "fourth", and the like in the description and the claims of the present invention are used for distinguishing different objects, and are not used for describing a specific order. The terms "comprises," "comprising," and any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
A motor fault diagnosis method based on CN and PCA, an electronic device and a medium are used for judging whether a motor is in fault and fault type according to detection data.
Example one
Referring to fig. 1, fig. 1 is a schematic flow chart of a motor fault diagnosis method based on CN and PCA according to an embodiment of the present invention. The execution main body of the method described in the embodiment of the invention is built by software/hardware, and can use equipment with processing and storage functions such as a computer and a server, and can also use a mobile phone, a tablet computer and the like under the condition of small data volume. As shown in fig. 1, the motor fault diagnosis method based on CN and PCA includes the following steps:
and 110, receiving raw data of the motor collected by the sensor.
The sensor can collect internal parameter data of the motor, such as voltage, current, acceleration and the like, and can also collect vibration information of the base, the driving end and the like of the motor in the running process. The number of the sensors may be one or more, and when there are a plurality of sensors, different types of sensors are preferably used, for example, a vibration sensor is used to collect vibration information of the base, a voltage transformer is used to collect voltage information of the motor, and the like.
The raw data of the motor collected by each sensor is called raw data of one channel. It can be understood that: when one sensor is used for collecting motor data, the motor data can be called single-channel original data; when two sensors are used to collect motor data, it can be called dual channel raw data, etc.
The length (or quantity) of the original data should satisfy:
wherein, L is the length of the raw data of each channel, beta is the sampling frequency of the sensor, and n is the rotating speed of the motor. In a preferred embodiment of the present invention, if a plurality of sensors are used to collect raw data of the motor, sampling frequencies of the plurality of sensors are the same, and collecting time is equal, so that the lengths of the raw data of the channels are guaranteed to be equal.
And 120, normalizing the original data and converting the normalized original data into binary data with 15 bits.
In the CN transformation, the transformation matrix is 32768 × 10, i.e. 215X 10, therefore, constructing the original data as 15-bit binary data, on the one hand, with subsequent transformation, and on the other hand, because the data is not isolated, there is necessarily a certain correlation with one or more data adjacent to each other before and after or data in the same time of different channels, and if the original data within or between channels can be correlated, the diagnosis can be made more accurate.
The step 120 is used for the above-mentioned conversion operation, and performs normalization processing on the raw data, and correlates the raw data in the channels or between the channels.
Specifically, the method comprises the following steps:
for single channel raw data, i.e. raw data of the motor acquired by a single sensor, the correlation of the raw data within the channel can be achieved using the following means.
Normalizing the single-channel original data to be 0-32, namely each single-channel original data can be represented by 5-bit binary data; adjacent 3 binary data with 5 bits are selected to be connected in sequence, so that the binary data forming 15 bits can be spliced:
Si＝{Mi-1,Mi,Mi+1}
wherein S isiFor the constituent i-th 15-bit binary data, Mi5-bit binary data normalized for the ith original data of a single channel; i is more than or equal to 0 and less than or equal to L, and L is the length of single-channel original data; when i is 0, obtaining M by using a linear difference mode-1When i is equal to L, M is obtained by using a linear difference modeL+1。
For example, if the data obtained by the normalization processing of the i-1 th, i +1 th single-channel raw data is converted into binary data of 10001, 01110, 11010, respectively, the obtained binary data of the i-th 15 bits is 100010111011010.
It can be understood that: the number of 15-bit binary data obtained by normalizing and splicing single-channel original data is equal to that of the single-channel original data, and the single-channel original data has L data. The normalization process may use a Z-score normalization method.
Similarly, for the dual-channel original data, the following method can be used to correlate the original data between the channels:
normalizing the two-channel original data to be 0-127 and 0-255 respectively, namely expressing the original data of one channel by adopting 7-bit binary data, and expressing the original data of the other channel by adopting 8-bit binary data;
selecting 7-bit binary data after one channel is specified to be connected with 8-bit binary data after the other channel is specified to form 15-bit binary data:
Si＝{Mi,Ni}
wherein S isiFor the constituent i-th 15-bit binary data, MiNormalized 7-bit binary data of the ith original data of one channel; n is a radical ofiNormalized 8-bit binary data of the ith original data of the other channel; i is more than or equal to 0 and less than or equal to L, and L is the length of the original data of each channel.
For example, if the data obtained by normalizing the ith original data of the 1 st channel original data and the ith original data of the 2 nd channel original data is converted into binary data of 1000110 and 11011100, respectively, the obtained ith binary data of 15 bits is 100011011011100.
It is also understood that: the quantity of the binary data of 15 bits obtained by normalizing and splicing the two-channel original data is equal to the quantity of the original data of each channel of the two channels, and the binary data has L data. The normalization process may use a Z-score normalization method.
Similarly, for three channels of original data, the following method can be used to associate the original data between the channels.
Normalizing the original data of each channel of the three channels to be 0-32, namely, the original data of each channel in the three channels of original data can be represented by 5-bit binary data;
selecting the corresponding original data of three channels to connect, and forming 15-bit binary data:
Si＝{Mi,Ni,Oi}
wherein S isiFor the constituent i-th 15-bit binary data, Mi、Ni、OiRespectively 5-bit binary data of normalized ith original data of the three channels; i is more than or equal to 0 and less than or equal to L, and L is the length of the original data of each channel.
For example, if data obtained by normalizing ith original data of 1 st channel original data, ith original data of 2 nd channel and ith original data of 3 rd channel is converted into binary data of 10001, 01110 and 11010 respectively, the obtained ith binary data of 15 bits is 100010111011010.
It can be understood that: the number of 15-bit binary data obtained by normalizing and splicing three-channel original data is equal to that of single-channel original data, and the three-channel original data all have L data. The normalization process may use a Z-score normalization method.
For four-channel or above original data, a processing mode similar to that of two-channel or three-channel original data may be adopted, for example, for four-channel, the original data of each channel may be respectively normalized to 0-9, and 0-16, that is, three-bit binary, and four-bit binary representations may be respectively adopted, and then the normalized data at the same position is spliced to obtain 15-bit binary data. All five channels can be normalized to 0-9.
If more than fifteen channels are used, the present invention is no longer within the scope of protection.
And 130, performing space mapping on the binary data by using a CN algorithm to obtain a 10-dimensional space mapping matrix.
In recent years, CN (Color Names) has been widely used in the fields of object detection, image recognition, and motion recognition. They are language color labels, and color naming, according to the conclusions drawn by the studies of Berlin and Kay, consists of 11 basic components: black, blue, brown, gray, green, orange, pink, purple, red, white, and yellow. Color names are assigned artificially for rendering colors in the real world. Based on this color attribute method, the mapping of image learning for Google image search retrieval herein, the original RGB image can be converted into 11-dimensional color space with probability sum of 1 after mapping:
X0(i,j)＝W2C32768×10·X1(i,j)
in the field of computer vision, the gray values in conventional object trackers always need to be normalized within [ -0.5,0.5] [6 ]. The color names are normalized by a technique to achieve better performance. This technique is a normalization process by projecting color names to the orthogonal standard of this 10-dimensional subspace. Thus, the 11-dimensional space can be reduced to 10-dimensional. Meanwhile, the projection can enable the color feature to have the characteristic of centralization.
When the CN algorithm is applied to the embodiment of the invention, the decimal value corresponding to the acquired 15-bit binary data is calculated, the decimal value corresponding to the index value in the CN conversion matrix 32768 multiplied by 10 is determined, and then the high-dimensional space mapping matrix corresponding to the decimal value is constructed, and finally the result mapping matrix MR [0: L, 0:10] is obtained.
140, using PCA algorithm to perform dimensionality reduction processing on the space mapping matrix to obtain a 3-dimensional feature matrix.
In most research areas of research, a large amount of data is required to find rules between them. There is no doubt that a large amount of data will provide a large amount of information and make it easier to analyze. However, there may be correlation between many variables, so it is necessary to discard useless data and reduce the amount of data to reduce the variables and thus the amount of computation. If the data to be analyzed, or its dimensionality, is randomly reduced, a missing PCA, which will inevitably result in useful information, is used to solve the above-described problem.
The Principle of PCA (principal Component Analysis) is to project the original sample data into a new space, i.e. to map a set of matrices to another coordinate system. In the new space or coordinates, not all of the original samples are needed, but only the spatial coordinates corresponding to the eigenvalues of the largest linearly independent set of original samples. The calculation of eigenvalues and their corresponding eigenvectors is a key part of the PCA algorithm, which refers to the eigenvalues of the covariance matrix corresponding to the raw data here.
This document becomes L × 10-dimensional data after CN operation. If PCA is to be used for size reduction, a 10 x 10 covariance matrix needs to be calculated first. And secondly, calculating to obtain the eigenvalue of the covariance matrix and the corresponding eigenvector thereof. If the first 3 eigenvalues already account for more than 99% of all eigenvalues, only the eigenvectors corresponding to the first 9 eigenvalues are extracted, the selected eigenvectors constituting a 10 × 3 transformation matrix. And finally, multiplying the L multiplied by 10 by 3 by the CN projected data by a transformation matrix to obtain the corresponding coordinates of the original sample data in the new feature space, wherein the 10-dimensional data is successfully reduced to 3-dimensional on the basis of not losing useful information.
And 150, inputting the feature matrix into a pre-trained network model, and outputting a diagnosis result.
Because a general network model is directed to processing image data, the feature matrix may be converted into a three-channel image, for example, features in each dimension of the 3-dimensional feature matrix are normalized to 0-255, and then the 3-dimensional feature matrix represents pixel values of images of R, G, B three channels, so as to construct a R, G, B pixel point value matrix of three channels, and then a preset number of pixels satisfying the network model, for example, 100 pixels, are selected to construct an image of 10 × 10, so as to obtain a R, G, B image of three channels.
The network model can preferably adopt a network structure of MobileNet V3 small.
And selecting a sample set consisting of sample data of a plurality of motors to train the model. Preferably, the sample data comprises a plurality of data sources, for example, a plurality of single-channel sample data, a plurality of double-channel sample data, a plurality of three-channel sample data, a plurality of four-channel sample data, and the like, and the sample data should comprise a plurality of sample data of different fault types, such as loss of field, air gap variation, stator winding fault, and the like, and should also comprise a plurality of sample data of the motor when the fault does not exist.
The length of each sample data in the sample data set should also satisfy:
the sample data also needs the processing process of the step 120-140, the obtained 3-dimensional feature matrix is converted into a three-channel image, the three-channel image is input into the network initial model for training, the trained label is the fault type of the sample, for example, the label of the sample data without fault is defined as 0, the label of the sample data with loss of magnetic fault is defined as 1, the label of the gap variation fault is defined as 2 … …, and the label is used for performing back propagation on the output result of the network initial model, so that a proper function or value of each parameter of the network initial model is determined, and a final network model is obtained.
And inputting the processed three-channel image of the original data into the trained network model, so that the state of the motor corresponding to the original data can be obtained, and if the state is a fault state, the fault type can be determined.
In order to verify that CN + PCA can improve the spatial separability of various types of original data, the embodiment of the invention applies a CN + PCA method to process rolling bearing fault data of the university of Western storage, and a training and testing sample set is constructed. And meanwhile, a training sample and a test sample data set are directly constructed according to the original data of the university of western storage. Simulation verification is carried out on two different data sets on a lightweight CNN network (the overall structure is five layers), and results of simulation experiments show that the fault diagnosis method based on CN + PCA fault data processing has high diagnosis precision and rapid model convergence. The identification accuracy and error iteration relationship is shown in table 1. The western storage data experiment result proves that the separability between different types of sample data can be improved by applying the high-dimensional space mapping realized by the CN and applying the PCA to carry out the dimension reduction and the construction of the sample data. And further, the fault diagnosis efficiency of the algorithm is improved.
Table 1: loss and accuracy of fault diagnosis training and testing based on CN and PCA
| Number of iterations | Loss of training | Training accuracy | Testing for loss | Test accuracy |
| 1 | 0.0217 | 99.5438 | 0.0004 | 100.000 |
| 2 | 0.0011 | 99.9983 | 0.0001 | 100.000 |
| 3 | 0.0007 | 99.9983 | 0.0006 | 100.000 |
| 4 | 0.0005 | 100.000 | 0.0000 | 100.000 |
| 5 | 0.0004 | 100.000 | 0.0000 | 100.000 |
| 6 | 0.0005 | 99.9983 | 0.0000 | 100.000 |
| 7 | 0.0003 | 99.9983 | 0.0000 | 100.000 |
| 8 | 0.0003 | 100.000 | 0.0000 | 100.000 |
| 9 | 0.0002 | 99.9983 | 0.0000 | 100.000 |
| 10 | 0.0002 | 100.000 | 0.0000 | 100.000 |
| 11 | 0.0001 | 100.000 | 0.0000 | 100.000 |
| 12 | 0.0001 | 100.000 | 0.0000 | 100.000 |
| 13 | 0.0001 | 100.000 | 0.0000 | 100.000 |
| 14 | 0.0001 | 100.000 | 0.0000 | 100.000 |
| 15 | 0.0001 | 100.000 | 0.0000 | 100.000 |
| 16 | 0.0001 | 100.000 | 0.0000 | 100.000 |
| 17 | 0.0001 | 100.000 | 0.0000 | 100.000 |
| 18 | 0.0001 | 100.000 | 0.0000 | 100.000 |
| 19 | 0.0001 | 100.000 | 0.0000 | 100.000 |
| 20 | 0.0001 | 100.000 | 0.0000 | 100.000 |
Example two
Referring to fig. 2, fig. 2 is a schematic structural diagram of a motor fault diagnosis device based on CN and PCA according to an embodiment of the present invention. As shown in fig. 2, the apparatus may include:
a receiving unit 210, configured to receive raw data of the motor collected by the sensor;
a normalization unit 220, configured to perform normalization processing on the original data, and convert the original data into binary data of 15 bits;
a mapping unit 230, configured to perform spatial mapping on the binary data by using a CN algorithm to obtain a 10-dimensional spatial mapping matrix;
a dimension reduction unit 240, configured to perform dimension reduction processing on the spatial mapping matrix by using a PCA algorithm, to obtain a 3-dimensional feature matrix;
and the diagnosis unit 250 is used for inputting the feature matrix into a pre-trained network model and outputting a diagnosis result.
As an alternative embodiment, the receiving of raw data of the motor collected by the sensor includes:
receiving single-channel or multi-channel raw data of the motor collected by one or more types of sensors;
the sensor is any one or more of a vibration sensor, a voltage transformer, a current transformer and an acceleration sensor.
As an optional implementation manner, the original data is single-channel original data;
the normalizing the original data and converting the normalized original data into binary data with 15 bits includes:
normalizing the single-channel original data to be 0-32, namely expressing each single-channel original data by using 5-bit binary data;
selecting adjacent 3 binary data with 5 bits to be connected in sequence to form binary data with 15 bits:
Si＝{Mi-1,Mi,Mi+1}
wherein S isiFor the constituent i-th 15-bit binary data, Mi5-bit binary data normalized for the ith original data of a single channel; i is more than or equal to 0 and less than or equal to L, and L is the length of single-channel original data; when i is 0, obtaining M by using a linear difference modei-1When i is equal to L, linear difference mode is used for obtainingMi+1。
As an optional implementation manner, the original data is dual-channel original data;
the normalizing the original data and converting the normalized original data into binary data with 15 bits includes:
normalizing the two-channel original data to be 0-127 and 0-255 respectively, namely expressing the original data of one channel by adopting 7-bit binary data, and expressing the original data of the other channel by adopting 8-bit binary data;
selecting 7-bit binary data after one channel is specified to be connected with 8-bit binary data after the other channel is specified to form 15-bit binary data:
Si＝{Mi,Ni}
wherein S isiFor the constituent i-th 15-bit binary data, MiNormalized 7-bit binary data of the ith original data of one channel; n is a radical ofiNormalized 8-bit binary data of the ith original data of the other channel; i is more than or equal to 0 and less than or equal to L, and L is the length of the original data of each channel.
As an optional implementation manner, the raw data is three-channel raw data;
the normalizing the original data and converting the normalized original data into binary data with 15 bits includes:
normalizing the three-channel original data to be 0-32, namely expressing the original data of each channel in the three-channel original data by using 5-bit binary data;
selecting the corresponding original data of three channels to connect, and forming 15-bit binary data:
Si＝{Mi,Ni,Oi}
wherein S isiFor the constituent i-th 15-bit binary data, Mi、Ni、OiRespectively 5-bit binary data of normalized ith original data of the three channels; i is more than or equal to 0 and less than or equal to L, and L is the original data of each channelLength of (d).
As an optional implementation, the performing spatial mapping on the binary data by using a CN algorithm to obtain a 10-dimensional spatial mapping matrix includes:
and performing high-dimensional space mapping on the data set consisting of the binary data of 15 bits by using a conversion matrix 32768 × 10 to obtain a mapped 10-bit space mapping matrix L × 10, wherein L is the length of the original data, namely the number of the data sets consisting of the binary data of 15 bits.
As an alternative embodiment, the pre-trained network model includes:
acquiring multiple groups of motor sample data with different fault types and motor sample data without faults, and constructing a sample set;
and creating a network initial model, and training the network initial model by using the sample data of each motor in the sample set to obtain the trained network model.
As an optional implementation manner, the length L of the original data satisfies:
wherein, beta is the sampling frequency of the sensor, and n is the rotating speed of the motor.
EXAMPLE III
Referring to fig. 3, fig. 3 is a schematic structural diagram of an electronic device according to an embodiment of the disclosure. As shown in fig. 3, the electronic device may include:
a memory 310 storing executable program code;
a processor 320 coupled to the memory 310;
the processor 320 calls the executable program code stored in the memory 310 to execute some or all of the steps in the CN and PCA based motor fault diagnosis method in the first embodiment.
The embodiment of the invention discloses a computer-readable storage medium which stores a computer program, wherein the computer program enables a computer to execute part or all of the steps in the motor fault diagnosis method based on CN and PCA in the first embodiment.
The embodiment of the invention also discloses a computer program product, wherein when the computer program product runs on a computer, the computer is enabled to execute part or all of the steps in the motor fault diagnosis method based on the CN and the PCA in the first embodiment.
The embodiment of the invention also discloses an application publishing platform, wherein the application publishing platform is used for publishing the computer program product, and when the computer program product runs on a computer, the computer is enabled to execute part or all of the steps in the motor fault diagnosis method based on the CN and the PCA in the first embodiment.
In various embodiments of the present invention, it should be understood that the sequence numbers of the processes do not imply a necessary sequence of execution, and the execution sequence of the processes should be determined by their functions and inherent logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.
The integrated units, if implemented as software functional units and sold or used as a stand-alone product, may be stored in a computer accessible memory. Based on such understanding, the technical solution of the present invention, which is a part of or contributes to the prior art in essence, or all or part of the technical solution, can be embodied in the form of a software product, which is stored in a memory and includes several requests for causing a computer device (which may be a personal computer, a server, a network device, or the like, and may specifically be a processor in the computer device) to execute part or all of the steps of the method described in the embodiments of the present invention.
In the embodiments provided herein, it should be understood that "B corresponding to a" means that B is associated with a, from which B can be determined. It should also be understood, however, that determining B from a does not mean determining B from a alone, but may also be determined from a and/or other information.
Those of ordinary skill in the art will appreciate that some or all of the steps in the various methods of the embodiments described herein may be implemented by programming and instructing associated hardware, the program may be stored in a computer readable storage medium including Read-Only Memory (ROM), Random Access Memory (RAM), Programmable Read-Only Memory (PROM), Erasable Programmable Read-Only Memory (EPROM), One-time Programmable Read-Only Memory (OTPROM), Electrically Erasable rewritable Read-Only Memory (EEPROM), Compact Disc Read-Only Memory (CD-ROM) or other optical disk storage, magnetic disk storage, tape storage, or any other medium capable of being used to carry or store data.
The motor fault diagnosis method based on CN and PCA, the electronic device and the medium disclosed in the embodiments of the present invention are introduced in detail, and a specific example is applied in the text to explain the principle and the implementation of the present invention, and the description of the embodiments is only used to help understanding the method and the core idea of the present invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, there may be variations in the specific implementation manners and application ranges, and in summary, the content of the present specification should not be construed as a limitation to the present invention.
Claims (10)
1. A motor fault diagnosis method based on CN and PCA is characterized by comprising the following steps:
receiving raw data of a motor collected by a sensor;
normalizing the original data and converting the normalized original data into binary data with 15 bits;
performing space mapping on the binary data by using a CN algorithm to obtain a 10-dimensional space mapping matrix;
performing dimensionality reduction on the space mapping matrix by using a PCA algorithm to obtain a 3-dimensional feature matrix;
and inputting the characteristic matrix into a pre-trained network model, and outputting a diagnosis result.
2. The CN and PCA based motor fault diagnosis method of claim 1 wherein receiving raw data of the motor collected by the sensor comprises:
receiving single-channel or multi-channel raw data of the motor collected by one or more sensors;
the sensor is any one or more of a vibration sensor, a voltage transformer, a current transformer and an acceleration sensor.
3. The CN and PCA based motor fault diagnosis method of claim 2 wherein the raw data is single channel raw data;
the normalizing the original data and converting the normalized original data into binary data with 15 bits includes:
normalizing the single-channel original data to be 0-32, namely expressing each single-channel original data by using 5-bit binary data;
selecting adjacent 3 binary data with 5 bits to be connected in sequence to form binary data with 15 bits:
Si＝{Mi-1,Mi,Mi+1}
wherein S isiFor the constituent i-th 15-bit binary data, Mi5-bit binary data normalized for the ith original data of a single channel; i is more than or equal to 0 and less than or equal to L, and L is the length of single-channel original data; when i is 0, obtaining M by using a linear difference modei-1When i is equal to L, M is obtained by using a linear difference modei+1。
4. The CN and PCA based motor fault diagnosis method of claim 2, wherein the raw data is dual channel raw data;
the normalizing the original data and converting the normalized original data into binary data with 15 bits includes:
normalizing the two-channel original data to be 0-127 and 0-255 respectively, namely, the original data of one channel is represented by 7-bit binary data, and the original data of the other channel is represented by 8-bit binary data;
selecting 7-bit binary data after one channel is specified to be connected with 8-bit binary data after the other channel is specified to form 15-bit binary data:
Si＝{Mi,Ni}
wherein S isiFor the constituent i-th 15-bit binary data, MiNormalized 7-bit binary data of the ith original data of one channel; n is a radical ofiNormalized 8-bit binary data of the ith original data of the other channel; i is more than or equal to 0 and less than or equal to L, and L is the length of the original data of each channel.
5. The CN and PCA based motor fault diagnosis method of claim 2, wherein the raw data is three-channel raw data;
the normalizing the original data and converting the normalized original data into binary data with 15 bits includes:
normalizing the three-channel original data to be 0-32, namely expressing the original data of each channel in the three-channel original data by using 5-bit binary data;
selecting the corresponding original data of three channels to connect, and forming 15-bit binary data:
Si＝{Mi,Ni,Oi}
wherein S isiFor the constituent i-th 15-bit binary data, Mi、Ni、OiRespectively 5-bit binary data of the ith original data of the three channels after normalization; i is more than or equal to 0 and less than or equal to L, and L is the length of the original data of each channel.
6. The CN and PCA based motor fault diagnosis method of any one of claims 1 to 5, wherein the binary data is spatially mapped by a CN algorithm to obtain a 10-dimensional spatial mapping matrix, comprising:
and performing high-dimensional space mapping on the data set consisting of the binary data of 15 bits by using a conversion matrix 32768 × 10 to obtain a mapped 10-bit space mapping matrix L × 10, wherein L is the length of the original data, namely the number of the data sets consisting of the binary data of 15 bits.
7. The CN and PCA based motor fault diagnosis method of any of claims 1-5, wherein the pre-trained network model comprises:
acquiring multiple groups of motor sample data with different fault types and motor sample data without faults, and constructing a sample set;
and creating a network initial model, and training the network initial model by using the sample data of each motor in the sample set to obtain the trained network model.
8. The CN and PCA based motor fault diagnosis method according to any of claims 1-5, wherein the length L of the raw data satisfies:
wherein, beta is the sampling frequency of the sensor, and n is the rotating speed of the motor.
9. An electronic device, comprising: a memory storing executable program code; a processor coupled with the memory; the processor invokes the executable program code stored in the memory for performing a CN and PCA based motor fault diagnosis method of any of claims 1-8.
10. A computer readable storage medium storing a computer program, wherein the computer program causes a computer to execute a CN and PCA based motor fault diagnosis method according to any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110020431.4A CN112836346B (en) | 2021-01-07 | 2021-01-07 | Motor fault diagnosis method based on CN and PCA, electronic equipment and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110020431.4A CN112836346B (en) | 2021-01-07 | 2021-01-07 | Motor fault diagnosis method based on CN and PCA, electronic equipment and medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112836346A true CN112836346A (en) | 2021-05-25 |
| CN112836346B CN112836346B (en) | 2024-04-23 |
Family
ID=75928342
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110020431.4A Active CN112836346B (en) | 2021-01-07 | 2021-01-07 | Motor fault diagnosis method based on CN and PCA, electronic equipment and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112836346B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114236386A (en) * | 2021-12-16 | 2022-03-25 | 河北工业大学 | Fault diagnosis method for power converter of switched reluctance motor |
| CN117849617A (en) * | 2024-01-10 | 2024-04-09 | 关兴 | Star connection motor resistance sensing fault detection method, system, equipment and medium |
| CN117991093A (en) * | 2024-04-03 | 2024-05-07 | 成都航天凯特机电科技有限公司 | Permanent magnet synchronous motor fault diagnosis method |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108994088A (en) * | 2018-06-12 | 2018-12-14 | 武汉科技大学 | HAGC cylinder method for diagnosing faults and device based on PCA dimensionality reduction Yu DBN network |
| CN109272044A (en) * | 2018-09-19 | 2019-01-25 | 郑州云海信息技术有限公司 | A kind of image similarity determines method, apparatus, equipment and storage medium |
| CN110895701A (en) * | 2019-06-12 | 2020-03-20 | 河南理工大学 | Forest fire online identification method and device based on CN and FHOG |
| CN111191725A (en) * | 2019-12-31 | 2020-05-22 | 北京市燃气集团有限责任公司 | RBF fault diagnosis method, system, terminal and computer storage medium based on PCA data processing |
-
2021
- 2021-01-07 CN CN202110020431.4A patent/CN112836346B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108994088A (en) * | 2018-06-12 | 2018-12-14 | 武汉科技大学 | HAGC cylinder method for diagnosing faults and device based on PCA dimensionality reduction Yu DBN network |
| CN109272044A (en) * | 2018-09-19 | 2019-01-25 | 郑州云海信息技术有限公司 | A kind of image similarity determines method, apparatus, equipment and storage medium |
| CN110895701A (en) * | 2019-06-12 | 2020-03-20 | 河南理工大学 | Forest fire online identification method and device based on CN and FHOG |
| CN111191725A (en) * | 2019-12-31 | 2020-05-22 | 北京市燃气集团有限责任公司 | RBF fault diagnosis method, system, terminal and computer storage medium based on PCA data processing |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114236386A (en) * | 2021-12-16 | 2022-03-25 | 河北工业大学 | Fault diagnosis method for power converter of switched reluctance motor |
| CN114236386B (en) * | 2021-12-16 | 2023-10-31 | 河北工业大学 | Fault diagnosis method for power converter of switched reluctance motor |
| CN117849617A (en) * | 2024-01-10 | 2024-04-09 | 关兴 | Star connection motor resistance sensing fault detection method, system, equipment and medium |
| CN117991093A (en) * | 2024-04-03 | 2024-05-07 | 成都航天凯特机电科技有限公司 | Permanent magnet synchronous motor fault diagnosis method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112836346B (en) | 2024-04-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112836346B (en) | Motor fault diagnosis method based on CN and PCA, electronic equipment and medium | |
| CN110209824B (en) | Text emotion analysis method, system and device based on combined model | |
| Dias et al. | Using the Choquet integral in the pooling layer in deep learning networks | |
| CN109299320A (en) | A kind of information interacting method, device, computer equipment and storage medium | |
| CN111078847A (en) | Power consumer intention identification method and device, computer equipment and storage medium | |
| CN109409504A (en) | A kind of data processing method, device, computer and storage medium | |
| CN111970400B (en) | Crank call identification method and device | |
| CN116414990B (en) | Vehicle fault diagnosis and prevention method | |
| CN116503399B (en) | Insulator pollution flashover detection method based on YOLO-AFPS | |
| WO2021127982A1 (en) | Speech emotion recognition method, smart device, and computer-readable storage medium | |
| CN114782054A (en) | Customer service quality detection method based on deep learning algorithm and related equipment | |
| CN115392301A (en) | Converter transformer state identification method, converter transformer state identification device, converter transformer state identification equipment, converter transformer state identification medium and program product | |
| CN113850387A (en) | Expert system knowledge base construction method, question and answer method, system, device and medium | |
| CN114581789A (en) | Hyperspectral image classification method and system | |
| CN112163447B (en) | Multi-task real-time gesture detection and recognition method based on Attention and Squeezenet | |
| CN112733706B (en) | Motor fault diagnosis method based on bilinear LBP, electronic equipment and medium | |
| CN116883740A (en) | Similar picture identification method, device, electronic equipment and storage medium | |
| CN115827878A (en) | Statement emotion analysis method, device and equipment | |
| CN115982621A (en) | Rotary machine residual service life prediction method based on time convolution network | |
| CN117036843A (en) | Target detection model training method, target detection method and device | |
| CN111813830B (en) | Industrial time sequence data retrieval method based on rail transit industrial Internet | |
| CN114926702A (en) | Small sample image classification method based on depth attention measurement | |
| CN114220012A (en) | Textile cotton and linen identification method based on deep self-attention network | |
| CN116933445B (en) | Digital twin model construction method and system for die forging equipment | |
| Xie et al. | S2H-GNN: Learning Soft to Hard Feature Matching with Sparsified Graph Neural Network |
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 |