CN111157894A - Motor fault diagnosis method, device and medium based on convolutional neural network - Google Patents
Motor fault diagnosis method, device and medium based on convolutional neural network Download PDFInfo
- Publication number
- CN111157894A CN111157894A CN202010037544.0A CN202010037544A CN111157894A CN 111157894 A CN111157894 A CN 111157894A CN 202010037544 A CN202010037544 A CN 202010037544A CN 111157894 A CN111157894 A CN 111157894A
- Authority
- CN
- China
- Prior art keywords
- neural network
- convolutional neural
- fault diagnosis
- parameter information
- electric parameter
- 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.)
- Pending
Links
Images
Classifications
-
- 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
- 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
Abstract
The invention relates to a motor fault diagnosis method, a device and a medium based on a convolutional neural network, which comprises the following steps: step 1, training a model; a, acquiring electric parameter information and non-electric parameter information of a motor; extracting statistic characteristics and/or frequency spectrum characteristics in the electric parameter information, and performing Park transformation on the electric parameter information and then performing Fourier transformation to obtain transformed frequency spectrum characteristics and artificial characteristics; b, pre-training by using a convolutional neural network, and inputting the non-electric parameter information into the pre-trained convolutional neural network to obtain a recessive characteristic; inputting a classification model for training; step 2, collecting real-time sample data; and 3, carrying out fault diagnosis. In order to solve the problems in the prior art, the invention adopts the convolutional neural network to extract the vibration characteristics and fuses the vibration characteristics with the current characteristics, thereby effectively improving the accuracy of fault diagnosis classification and reducing the diagnosis time.
Description
Technical Field
The invention relates to the field of fault diagnosis and protection of high-voltage and low-voltage motors, in particular to a fault diagnosis classification method and device based on a convolutional neural network and a computer readable storage medium.
Background
The protection of the motors (including the high-voltage motor and the low-voltage motor) needs to detect the states of the motors, judge possible faults of the motors and then execute corresponding protection strategies according to the types of the faults. Various sensors are used to detect electrical parameters (current, voltage, power factor, etc.) and non-electrical parameters (vibration signals, displacements, etc.) during operation of the motor.
Conventional data analysis and fault diagnosis methods include:
first, the electrical and non-electrical parameters are analyzed independently, giving analysis results, which are often not very efficient and accurate.
And secondly, directly inputting the electrical parameters and the non-electrical parameters into a deep neural network for feature classification. This approach enables feature classification, but is time consuming and less accurate.
Disclosure of Invention
The motor fault diagnosis method based on the convolutional neural network provided by the invention effectively improves the accuracy of fault diagnosis classification and reduces the diagnosis time. A convolutional neural network-based motor fault diagnosis apparatus and a computer-readable storage medium are also provided.
A motor fault diagnosis method based on a convolutional neural network comprises the following steps:
a, acquiring electric parameter information and non-electric parameter information of a motor; extracting statistic characteristics and/or frequency spectrum characteristics in the electric parameter information, and performing Park transformation on the electric parameter information and then performing Fourier transformation to obtain transformed frequency spectrum characteristics; the statistic characteristics and/or the frequency spectrum characteristics and the transformed frequency spectrum characteristics jointly form artificial characteristics;
b, pre-training by using a convolutional neural network, and inputting the non-electric parameter information into the pre-trained convolutional neural network to obtain a recessive characteristic;
inputting the artificial features and the implicit features into a classification model together for training;
and 3, processing the collected real-time sample data according to the mode of the step A, B, and inputting the processed real-time sample data into the trained classification model to obtain a fault diagnosis result.
Further, the electrical parameter information is voltage and current waveforms; the non-electrical parameter information is a three-axis acceleration waveform.
Further, step a includes the step of adding noise to the electrical parametric information and the non-electrical parametric information signals.
Further, the statistical characteristic of the electrical quantity includes: maximum, minimum, mean, variance, standard deviation, root mean square, skewness, kurtosis, form factor, peak factor, pulse factor, margin factor.
Further, the classification model is a LightGBM network.
Further, the convolutional neural network includes: the multilayer structure comprises a first convolution layer, a first pooling layer, a second convolution layer, a second pooling layer and a full-link layer.
The invention also provides a motor fault diagnosis device based on the convolutional neural network, which comprises a processor and a memory, wherein the processor is used for executing the computer program stored in the memory so as to realize the method.
The present invention also provides a computer-readable storage medium storing a computer program implementing the above method.
In order to solve the problems in the prior art, the invention adopts the convolutional neural network to extract the vibration characteristics and fuses the vibration characteristics with the current characteristics, thereby effectively improving the accuracy of fault diagnosis classification and reducing the diagnosis time.
Drawings
FIG. 1 is a schematic diagram of an embodiment of the present invention;
fig. 2 is a flow chart of an embodiment of the present invention.
Detailed Description
The following description is made with reference to the accompanying drawings.
Method embodiment
The invention belongs to an intelligent diagnosis and identification method in a whole view. The design idea of the embodiment is as follows: by taking advantage of the network structure of the convolutional neural network, the inherent implicit characteristic of the triaxial acceleration signal is automatically extracted through the network, rather than directly utilizing the network to obtain a result. And the method is combined with the artificial features extracted by the three-phase current, decision analysis is carried out on the two data sources together, and finally training is carried out by utilizing a LightGBM decision tree network to generate a training model. And simultaneously, acquiring three-phase current and triaxial acceleration data of the current motor in real time, and inputting the characteristics into a trained classification model for fault diagnosis after performing characteristic fusion according to the method to obtain a current motor fault classification result.
As shown in fig. 1 and 2, the method specifically comprises the following steps:
In the embodiment, the electric parameter is a three-phase current signal, and the non-electric parameter is a three-axis vibration acceleration signal. The three-phase current signals are phase A, phase B and phase C currents, each phase collects 10s continuous waveforms, and the sampling rate is 1 KHz. And simultaneously slicing three-phase current signals, wherein each phase of current is 10K points in total, 300 samples are sliced averagely in each phase, each sample in each phase is 100 points in total, and three-phase samples on the same time slice are combined into an array with the dimensionality of [3,100 ]. The three-axis acceleration signals are X-axis acceleration signals, Y-axis acceleration signals and Z-axis acceleration signals, each axis respectively collects 10s of continuous waveforms, and the sampling rate is 10 KHz. And simultaneously slicing the three-axis acceleration signals, wherein each axis of acceleration signals has 100K points, each axis averagely slices 300 samples, each axis has 1000 points, and the three-axis samples on the same time slice are combined into an array with a dimension [3,1000 ]. The three-phase current signals and the three-axis acceleration signals must be acquired simultaneously, the sampling rate is adjustable, the sampling duration is adjustable, the number of sliced samples is also adjustable, the sampling rate does not need to be kept consistent, and the sampling duration needs to be consistent.
The acquisition of the electrical parameters and the non-electrical parameters needs to be noticed, the sampling rate can be different, and the sampling time is kept consistent as much as possible. Meanwhile, when the slicing is carried out on various signals, the slicing is carried out according to the same time period, so that the consistency of the number of samples is ensured, and meanwhile, the fused characteristics are ensured to be various characteristics in the same time period during characteristic fusion.
And signal noise is added, and in the model training stage, the original data is subjected to signal noise addition, so that the data has certain errors and has experimental value. In this embodiment, the noise added by the signal plus noise is gaussian white noise, and gaussian noise with a signal-to-noise ratio of 13dB is added to the data, that is, power noise of 5% of the original data.
And calculating the statistic characteristics of each phase in each sample, including the maximum value, the minimum value, the mean value, the variance, the standard deviation, the root mean square, the skewness, the kurtosis, the form factor, the peak factor, the pulse factor and the margin factor, of the three-phase current samples. There were 12 per phase and 36 per sample. And performing fast Fourier transform on each phase current of each sample to obtain corresponding frequency spectrum arrays, wherein the frequency spectrum arrays comprise three groups, and each group comprises 51 characteristic points. After performing Park transformation on each sample three-phase current, performing fast Fourier transformation to obtain a group of frequency spectrum arrays, wherein 51 feature points are obtained in total. The sample feature points are combined, so that each sample of the three-phase current signal has 240 feature points. The embodiment extracts the frequency spectrum feature (which is not subjected to Pak transformation) and the frequency spectrum feature after transformation at the same time, and the frequency spectrum feature after transformation have the functions of comparison and further verification, so that the diagnosis accuracy can be further improved compared with the method for extracting only the frequency spectrum feature after transformation.
The method comprises the steps of extracting characteristics of electric parameters (three-phase currents Ia, Ib and Ic), carrying out Park conversion on the three-phase currents, and extracting frequency spectrum characteristics of the three-phase currents. The Park transformation characteristics have good diagnosis effect on the analysis of the turn-to-turn short circuit fault of the motor stator winding. The Park transformation is a coordinate transformation which is most commonly used for analyzing the operation of the synchronous motor at present, the Park transformation projects three-phase currents of a, b and c of the stator to a direct axis (d axis) rotating along with the rotor, and a quadrature axis (q axis) and a zero axis (0 axis) vertical to a dq plane, so that the diagonalization of a stator inductance matrix is realized, and the operation analysis of the synchronous motor is simplified.
For the triaxial acceleration sample, feature extraction is not needed, and only the triaxial acceleration sample needs to be input into a Convolutional Neural Network (CNN). The corresponding convolutional neural network structure needs to be designed, the input is a matrix of [3,1000,1], 32 convolutional kernels are input, and the first convolutional layer is [1,900,32 ]. Then, pooling is performed, the pooling core [1,5,1] is performed with a step size of 5, and the first pooling layer is [1,180,32 ]. The convolution is continued with 64 second convolution kernels, 1,21,32, and 1,160,64 second convolution kernels. Then pooling is performed, pooling nuclei [1,5,1], step size 5, and a second pooling layer [1,32,64 ]. And then mapping the second pooling layer to a full-link layer, and multiplying by a relu activation function, namely converting the three-bit array into a one-dimensional array for output, wherein the output length is 1024. There are 1024 feature points of the convolutional neural network for each triaxial acceleration sample.
CNN pre-training processing, the sample characteristics extracted by the convolutional neural network are untrained networks, and the weights and deviations in the networks are random under the initial condition, so that the obtained sample characteristics are unstable after one round of network data processing, and the sample characteristics cannot be well represented. For this reason, the present embodiment first uses a pre-training process when extracting the sample features using the convolutional neural network. And training by using a convolutional neural network with the same network structure, and after a plurality of iterations, outputting each sample characteristic in a full connection layer. The number of iterations chosen here is 1000 rounds.
And (3) feature fusion, namely combining the features (240) of the three-phase current samples in the same time segment with the three-axis acceleration features (1024), and obtaining 1264 mixed features for each sample. And then inputting the mixed features into a LightGBM network, and obtaining a classification result through training and outputting the classification result. LightGBM is a fast, distributed, high-performance decision tree algorithm-based gradient boosting framework.
And 2, collecting a real-time sample.
Similar to the training step, the real-time samples also comprise electric parameter samples and non-electric parameter samples, and the electric parameter samples are used for extracting artificial features in the step 1; and (4) carrying out CNN pre-training on the non-electric parameters to obtain the output implicit characteristics.
And 3, diagnosing and classifying the faults.
And fusing the artificial features and the recessive features, namely sending the artificial features and the recessive features into the trained LightGBM network together to obtain a fault classification result.
In this embodiment, the electrical parameter extracts artificial features, but the non-electrical parameter does not extract features, and implicit features are automatically extracted through a convolutional neural network, because many empirical knowledge can be used to extract its main features for the electrical parameter as a simple signal. For non-electrical parameters (vibration), as a high-frequency complex signal, the intrinsic features are difficult to artificially represent and extract, so that it is desirable to extract features by means of a convolutional neural network. And after the features are fused, the features are not input into a convolutional neural network for training, and the LightGBM network is used for directly carrying out classification output.
In this embodiment, the non-electrical parameter (vibration) waveform is a triaxial acceleration waveform, so the input convolutional neural network is an array of three rows and N columns, and the convolutional neural network can well extract the intrinsic connection characteristics of waveforms in three directions by using the convolution advantages of the convolutional neural network. The traditional convolutional neural network reaches a full connection layer through internal network convolution and pooling, the full connection layer data is processed by utilizing a softmax layer, and the final fault classification result is output. Such multiple rounds of iterative training consume a large amount of training time, which is a drawback of all deep neural networks. For this reason, we want to extract features using convolutional neural networks, rather than training using convolutional neural networks. Certainly, in order to make the features extracted by the network better, a pre-training method is also used, that is, a better network structure is obtained by using hundreds of network training rounds (the number of rounds is less and adjustable), and then the original non-electrical parameter waveform is input to obtain the features output by the full connection layer (FC), that is, the implicit features of the input waveform.
As another embodiment, the sample structure and the selection of the artificial features may be different from the present embodiment, for example, the number, length, and the like of the sample points may be variable, and the number of the artificial features may be selectable.
As another embodiment, the structure of the convolutional neural network may be designed as needed, and is not limited to the structure and the number of layers in this embodiment.
As other embodiments, the LightGBM network may be replaced with other types of classification models.
As other embodiments, the electrical parameter information is not limited to current and voltage, but may include power, power parameters, and the like; the non-electrical parameter information is not limited to the vibration signal, but may include information such as temperature and displacement.
Computer-readable storage medium embodiments
The computer-readable storage medium referred to in this embodiment includes a semiconductor, a magnetic core, a magnetic drum, a magnetic tape, a laser disk, etc., on which a computer program is stored, which when executed, enables the method of the above-described method embodiments.
Device embodiment
The apparatus referred to in this embodiment includes a processor and a memory, the processor executes a computer program stored in the memory, and the computer program, when executed, can implement the method of the above-mentioned method embodiment.
Claims (10)
1. A motor fault diagnosis method based on a convolutional neural network is characterized in that: the method comprises the following steps:
step 1, training a model;
a, acquiring electric parameter information and non-electric parameter information of a motor; extracting statistic characteristics and/or frequency spectrum characteristics in the electric parameter information, and performing Park transformation on the electric parameter information and then performing Fourier transformation to obtain transformed frequency spectrum characteristics; the statistic characteristics and/or the frequency spectrum characteristics and the transformed frequency spectrum characteristics jointly form artificial characteristics;
b, pre-training by using a convolutional neural network, and inputting the non-electric parameter information into the pre-trained convolutional neural network to obtain a recessive characteristic;
inputting the artificial features and the implicit features into a classification model together for training;
step 2, collecting real-time sample data;
and 3, processing the collected real-time sample data according to the mode of the step A, B, and inputting the processed real-time sample data into the trained classification model to obtain a fault diagnosis result.
2. The motor fault diagnosis method based on the convolutional neural network as claimed in claim 1, wherein: the electrical parameter information is voltage and current waveforms; the non-electrical parameter information is a three-axis acceleration waveform.
3. The motor fault diagnosis method based on the convolutional neural network as claimed in claim 1, wherein: the electrical parameter and the non-electrical parameter have the same sampling time and the same or different sampling rate.
4. The motor fault diagnosis method based on the convolutional neural network as claimed in claim 1, wherein: the method also comprises the step of slicing the electric parameter signal and the non-electric parameter signal, and during slicing, slicing is carried out according to the same time period, so that the consistency of the sample number is ensured, and meanwhile, the fused characteristics are ensured to be various characteristics in the same time period during characteristic fusion.
5. The motor fault diagnosis method based on the convolutional neural network as claimed in claim 1, wherein: step a further includes the step of adding noise to the electrical parametric information and non-electrical parametric information signals.
6. The motor fault diagnosis method based on the convolutional neural network as claimed in claim 1, wherein: the statistical quantity characteristic of the electrical quantity comprises: maximum, minimum, mean, variance, standard deviation, root mean square, skewness, kurtosis, form factor, peak factor, pulse factor, margin factor.
7. The motor fault diagnosis method based on the convolutional neural network as claimed in claim 1, wherein: the classification model is a LightGBM network.
8. The motor fault diagnosis method based on the convolutional neural network as claimed in claim 1, wherein: the convolutional neural network includes: the multilayer structure comprises a first convolution layer, a first pooling layer, a second convolution layer, a second pooling layer and a full-link layer.
9. A convolutional neural network-based motor fault diagnosis apparatus comprising a processor and a memory, the processor being configured to execute a computer program stored in the memory to implement the method of any one of claims 1 to 8.
10. A computer-readable storage medium, characterized in that it stores a computer program to implement the method according to any one of claims 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010037544.0A CN111157894A (en) | 2020-01-14 | 2020-01-14 | Motor fault diagnosis method, device and medium based on convolutional neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010037544.0A CN111157894A (en) | 2020-01-14 | 2020-01-14 | Motor fault diagnosis method, device and medium based on convolutional neural network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111157894A true CN111157894A (en) | 2020-05-15 |
Family
ID=70563236
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010037544.0A Pending CN111157894A (en) | 2020-01-14 | 2020-01-14 | Motor fault diagnosis method, device and medium based on convolutional neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111157894A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112232212A (en) * | 2020-10-16 | 2021-01-15 | 广东石油化工学院 | Triple concurrent fault analysis method and system, large unit device and storage medium |
| CN112232414A (en) * | 2020-10-16 | 2021-01-15 | 广东石油化工学院 | Triple concurrency fault analysis method based on X and Y dual-measurement-point spectrum data |
| CN112270227A (en) * | 2020-10-16 | 2021-01-26 | 广东石油化工学院 | Oil film whirl and friction concurrent fault analysis method and analysis system |
| CN112284721A (en) * | 2020-10-16 | 2021-01-29 | 广东石油化工学院 | Double fault analysis method and system for friction and rotor imbalance of large unit |
| CN112329825A (en) * | 2020-10-23 | 2021-02-05 | 贵州电网有限责任公司 | Transformer mechanical fault diagnosis method based on information dimension division and decision tree lifting |
| CN113009338A (en) * | 2021-02-26 | 2021-06-22 | 中国长江三峡集团有限公司 | Offshore wind power variable pitch motor stator turn-to-turn short circuit fault diagnosis method |
| CN113159218A (en) * | 2021-05-12 | 2021-07-23 | 北京联合大学 | Radar HRRP multi-target identification method and system based on improved CNN |
| CN117436025A (en) * | 2023-12-21 | 2024-01-23 | 青岛鼎信通讯股份有限公司 | Fault indicator-based non-fault abnormal waveform screening method |
| CN117514885A (en) * | 2023-11-23 | 2024-02-06 | 德州隆达空调设备集团有限公司 | Fault detection method and device for axial flow fan |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101576604A (en) * | 2009-01-04 | 2009-11-11 | 湖南大学 | Method for diagnosing failures of analog circuit based on heterogeneous information fusion |
| CN102494899A (en) * | 2011-11-25 | 2012-06-13 | 华南理工大学 | Composite fault diagnosis method for diesel engine and diagnosis system |
| CN102954857A (en) * | 2012-10-17 | 2013-03-06 | 东南大学 | Vane unbalance fault diagnosis method of wind turbine generator set based on current signal |
| CN105069291A (en) * | 2015-08-06 | 2015-11-18 | 温州大学 | EMD (empirical mode decomposition) and BP (back propagation) neural network based motor bearing fault identification method |
| CN105910827A (en) * | 2016-04-25 | 2016-08-31 | 东南大学 | Induction motor fault diagnosis method based on discriminant convolutional feature learning |
| CN106054078A (en) * | 2016-07-26 | 2016-10-26 | 上海电力学院 | Fault identification method for inter-turn short circuit of stator windings in doubly-fed motor at sea |
| CN106841949A (en) * | 2017-03-09 | 2017-06-13 | 杭州安脉盛智能技术有限公司 | Three-phase asynchronous Ac motor stator insulation on-line monitoring method and device |
| CN107421741A (en) * | 2017-08-25 | 2017-12-01 | 南京信息工程大学 | A kind of Fault Diagnosis of Roller Bearings based on convolutional neural networks |
| CN107957551A (en) * | 2017-12-12 | 2018-04-24 | 南京信息工程大学 | Stacking noise reduction own coding Method of Motor Fault Diagnosis based on vibration and current signal |
| CN108830326A (en) * | 2018-06-21 | 2018-11-16 | 河南工业大学 | A kind of automatic division method and device of MRI image |
| CN109145886A (en) * | 2018-10-12 | 2019-01-04 | 西安交通大学 | A kind of asynchronous machine method for diagnosing faults of Multi-source Information Fusion |
| CN109858503A (en) * | 2017-11-30 | 2019-06-07 | 株洲中车时代电气股份有限公司 | The traction converter failure diagnostic method of decision tree is promoted based on gradient |
| CN110132598A (en) * | 2019-05-13 | 2019-08-16 | 中国矿业大学 | Slewing rolling bearing fault noise diagnostics algorithm |
-
2020
- 2020-01-14 CN CN202010037544.0A patent/CN111157894A/en active Pending
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101576604A (en) * | 2009-01-04 | 2009-11-11 | 湖南大学 | Method for diagnosing failures of analog circuit based on heterogeneous information fusion |
| CN102494899A (en) * | 2011-11-25 | 2012-06-13 | 华南理工大学 | Composite fault diagnosis method for diesel engine and diagnosis system |
| CN102954857A (en) * | 2012-10-17 | 2013-03-06 | 东南大学 | Vane unbalance fault diagnosis method of wind turbine generator set based on current signal |
| CN105069291A (en) * | 2015-08-06 | 2015-11-18 | 温州大学 | EMD (empirical mode decomposition) and BP (back propagation) neural network based motor bearing fault identification method |
| CN105910827A (en) * | 2016-04-25 | 2016-08-31 | 东南大学 | Induction motor fault diagnosis method based on discriminant convolutional feature learning |
| CN106054078A (en) * | 2016-07-26 | 2016-10-26 | 上海电力学院 | Fault identification method for inter-turn short circuit of stator windings in doubly-fed motor at sea |
| CN106841949A (en) * | 2017-03-09 | 2017-06-13 | 杭州安脉盛智能技术有限公司 | Three-phase asynchronous Ac motor stator insulation on-line monitoring method and device |
| CN107421741A (en) * | 2017-08-25 | 2017-12-01 | 南京信息工程大学 | A kind of Fault Diagnosis of Roller Bearings based on convolutional neural networks |
| CN109858503A (en) * | 2017-11-30 | 2019-06-07 | 株洲中车时代电气股份有限公司 | The traction converter failure diagnostic method of decision tree is promoted based on gradient |
| CN107957551A (en) * | 2017-12-12 | 2018-04-24 | 南京信息工程大学 | Stacking noise reduction own coding Method of Motor Fault Diagnosis based on vibration and current signal |
| CN108830326A (en) * | 2018-06-21 | 2018-11-16 | 河南工业大学 | A kind of automatic division method and device of MRI image |
| CN109145886A (en) * | 2018-10-12 | 2019-01-04 | 西安交通大学 | A kind of asynchronous machine method for diagnosing faults of Multi-source Information Fusion |
| CN110132598A (en) * | 2019-05-13 | 2019-08-16 | 中国矿业大学 | Slewing rolling bearing fault noise diagnostics algorithm |
Non-Patent Citations (1)
| Title |
|---|
| 李俊瑶: ""交流电机电流特征分析与健康感知系统实现"", 《中国知网》 * |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112232212A (en) * | 2020-10-16 | 2021-01-15 | 广东石油化工学院 | Triple concurrent fault analysis method and system, large unit device and storage medium |
| CN112232414A (en) * | 2020-10-16 | 2021-01-15 | 广东石油化工学院 | Triple concurrency fault analysis method based on X and Y dual-measurement-point spectrum data |
| CN112270227A (en) * | 2020-10-16 | 2021-01-26 | 广东石油化工学院 | Oil film whirl and friction concurrent fault analysis method and analysis system |
| CN112284721A (en) * | 2020-10-16 | 2021-01-29 | 广东石油化工学院 | Double fault analysis method and system for friction and rotor imbalance of large unit |
| CN112232414B (en) * | 2020-10-16 | 2021-06-15 | 广东石油化工学院 | Triple concurrency fault analysis method based on X and Y dual-measurement-point spectrum data |
| CN112329825A (en) * | 2020-10-23 | 2021-02-05 | 贵州电网有限责任公司 | Transformer mechanical fault diagnosis method based on information dimension division and decision tree lifting |
| CN112329825B (en) * | 2020-10-23 | 2022-12-06 | 贵州电网有限责任公司 | Transformer mechanical fault diagnosis method based on information dimension division and decision tree lifting |
| CN113009338A (en) * | 2021-02-26 | 2021-06-22 | 中国长江三峡集团有限公司 | Offshore wind power variable pitch motor stator turn-to-turn short circuit fault diagnosis method |
| CN113159218A (en) * | 2021-05-12 | 2021-07-23 | 北京联合大学 | Radar HRRP multi-target identification method and system based on improved CNN |
| CN117514885A (en) * | 2023-11-23 | 2024-02-06 | 德州隆达空调设备集团有限公司 | Fault detection method and device for axial flow fan |
| CN117436025A (en) * | 2023-12-21 | 2024-01-23 | 青岛鼎信通讯股份有限公司 | Fault indicator-based non-fault abnormal waveform screening method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111157894A (en) | Motor fault diagnosis method, device and medium based on convolutional neural network | |
| WO2019090879A1 (en) | Analog circuit fault diagnosis method based on cross wavelet features | |
| WO2017128455A1 (en) | Analogue circuit fault diagnosis method based on generalized multiple kernel learning-support vector machine | |
| CN110108992B (en) | Cable partial discharge fault identification method and system based on improved random forest algorithm | |
| CN110726898B (en) | Power distribution network fault type identification method | |
| CN111650453A (en) | Power equipment diagnosis method and system based on windowing characteristic Hilbert imaging | |
| CN110705456A (en) | Micro motor abnormity detection method based on transfer learning | |
| Koley et al. | Wavelet-aided SVM tool for impulse fault identification in transformers | |
| CN110070102B (en) | Method for establishing sequence-to-sequence model for identifying power quality disturbance type | |
| CN110672905A (en) | CNN-based self-supervision voltage sag source identification method | |
| CN113159226B (en) | Inverter fault diagnosis method with integration of depth features and statistical features | |
| CN112083328A (en) | Fault diagnosis method, system and device for high-voltage circuit breaker | |
| CN111553112A (en) | Power system fault identification method and device based on deep belief network | |
| CN114325256A (en) | Power equipment partial discharge identification method, system, equipment and storage medium | |
| CN108828437B (en) | Analog circuit fault feature extraction method based on cloud correlation coefficient matrix | |
| CN109782158B (en) | Analog circuit diagnosis method based on multi-stage classification | |
| Sun et al. | Fault diagnosis of conventional circuit breaker accessories based on grayscale image of current signal and improved ZFNet-DRN | |
| CN112146880B (en) | Intelligent diagnosis method for internal structure faults of rolling bearing at different rotating speeds | |
| CN114330486A (en) | Power system bad data identification method based on improved Wasserstein GAN | |
| CN111666912B (en) | Partial discharge fusion feature extraction method considering electrical feature quantity and graphic feature | |
| CN115015683B (en) | Cable production performance test method, device, equipment and storage medium | |
| CN113341223B (en) | Method suitable for power grid harmonic analysis and harmonic source positioning | |
| CN114545147A (en) | Voltage sag source positioning method based on deep learning in consideration of time-varying topology | |
| Lewin et al. | Identification of PD defect typologies using a support vector machine | |
| CN114186590A (en) | Power distribution network single-phase earth fault identification method based on wavelet and deep learning |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200515 |
|
| RJ01 | Rejection of invention patent application after publication |