CN112478975A - Elevator door fault detection method based on audio features - Google Patents
Elevator door fault detection method based on audio features Download PDFInfo
- Publication number
- CN112478975A CN112478975A CN202011447416.XA CN202011447416A CN112478975A CN 112478975 A CN112478975 A CN 112478975A CN 202011447416 A CN202011447416 A CN 202011447416A CN 112478975 A CN112478975 A CN 112478975A
- Authority
- CN
- China
- Prior art keywords
- detection model
- fault
- data
- model
- features
- 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
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B66—HOISTING; LIFTING; HAULING
- B66B—ELEVATORS; ESCALATORS OR MOVING WALKWAYS
- B66B5/00—Applications of checking, fault-correcting, or safety devices in elevators
- B66B5/02—Applications of checking, fault-correcting, or safety devices in elevators responsive to abnormal operating conditions
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B66—HOISTING; LIFTING; HAULING
- B66B—ELEVATORS; ESCALATORS OR MOVING WALKWAYS
- B66B5/00—Applications of checking, fault-correcting, or safety devices in elevators
- B66B5/0006—Monitoring devices or performance analysers
- B66B5/0018—Devices monitoring the operating condition of the elevator system
- B66B5/0031—Devices monitoring the operating condition of the elevator system for safety reasons
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- 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
-
- 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
- G06N3/088—Non-supervised learning, e.g. competitive learning
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/08—Speech classification or search
- G10L15/16—Speech classification or search using artificial neural networks
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L21/00—Processing of the speech or voice signal to produce another audible or non-audible signal, e.g. visual or tactile, in order to modify its quality or its intelligibility
- G10L21/02—Speech enhancement, e.g. noise reduction or echo cancellation
- G10L21/0208—Noise filtering
Abstract
The invention relates to an elevator door fault detection method based on audio features, which comprises the following steps: s1, collecting the video stream of the elevator running time period, and acquiring the audio signal therein; s2, preprocessing the audio signal; s3, receiving the preprocessed data, and extracting data features; s4, taking the data features as learning targets, adaptively selecting an abnormal detection model or a supervised learning model according to the size of the existing data volume for training, and updating the detection model in real time; and S5, inputting the data characteristics into the updated detection model, giving a fault probability threshold value by the detection model, and determining whether to send out a fault alarm according to the fault probability threshold value. The elevator door fault detection method based on the audio frequency characteristics is low in cost and wide in application range.
Description
Technical Field
The invention belongs to the technical field of elevator door fault detection.
Background
The passenger elevator is most frequently contacted in the work and life of the door, and whether the passenger elevator is safely operated or not is directly related to the personal safety of passengers. Statistically, more than 80% of elevator failures and more than 70% of elevator accidents are caused by problems in the elevator door system. Once the door system breaks down, the normal operation of the whole machine is affected, and even the passengers are injured and killed. Therefore, the accurate identification of the door state in the elevator running process is very important for ensuring the front safe running of the elevator, and the technical support can be provided for identifying potential safety hazards of the elevator and diagnosing faults.
According to the prior art, whether the door state is abnormal or not in the elevator running process can be detected to a certain extent, but the prior art seriously depends on a large number of hardware devices, and is not a good choice in use cost and implementation effect. For example, chinese patent CN201910763678.8 discloses a system for elevator door sensor fusion, fault detection and service notification, which identifies the door trim through the output of sensors, through a processor, a memory and a plurality of sensors. For another example, chinese patent CN201711495102.5 discloses a system for detecting elevator door faults based on large-scale data. The system consists of a data distribution layer, a data transmission layer, a data processing layer and a distributed coordination service component, can distribute, transmit, analyze and fragment large-scale flow data of the elevator, relates to various sensors and complex judgment algorithms, has high cost and small application range, and judges rule load.
Disclosure of Invention
The invention aims to solve the problems and provides an elevator door fault detection method based on audio frequency characteristics, which is low in cost and wide in application range.
In order to achieve the above object, the present invention provides an elevator door fault detection method based on audio frequency characteristics, comprising:
s1, collecting the video stream of the elevator running time period, and acquiring the audio signal therein;
s2, preprocessing the audio signal;
s3, receiving the preprocessed data, and extracting data features;
s4, taking the data features as learning targets, adaptively selecting an abnormal detection model or a supervised learning model according to the size of the existing data volume for training, and updating the detection model in real time;
and S5, inputting the data characteristics into the updated detection model, giving a fault probability threshold value by the detection model, and determining whether to send out a fault alarm according to the fault probability threshold value.
According to an embodiment of the present invention, in step S2, the pre-processing on the audio signal includes, but is not limited to, filtering, denoising, and removing singular values.
According to an embodiment of the present invention, a wavelet threshold denoising method is used for denoising the audio signal:
setting orthogonal wavelets and a decomposition layer number N, and performing N-layer wavelet decomposition on the signal f (t), wherein N is 6;
carrying out nonlinear threshold processing on the wavelet transform coefficient of the measurement signal;
and reconstructing the processed wavelet coefficient.
According to an embodiment of the present invention, in step S3, the data features extracted include, but are not limited to, zero crossing rate, frequency domain energy, spectral centroid, mel-frequency-ratio coefficient.
According to one embodiment of the present invention, the adaptively selecting an anomaly detection model or a supervised learning model for training according to the size of the existing data volume comprises:
under the condition that door fault data is insufficient, namely under the condition that positive and negative samples are seriously unbalanced, an abnormal detection model is selected for training, in the feature selection process, an algorithm can be selected according to the matching features of feature dimension in a self-adaptive mode, wherein the algorithm comprises but is not limited to an exhaustion method, a variance selection method, a correlation coefficient method and unsupervised clustering, and the abnormal detection model comprises but is not limited to Isolation free, One class SVM, Autoencoder and GANormally.
According to an embodiment of the present invention, the adaptively selecting the anomaly detection model or the supervised learning model for training according to the size of the existing data volume further comprises:
in case of sufficient elevator samples, a supervised learning model is adaptively selected, including but not limited to Random forms, LightGBM, Xgboost, multiple model integration GRU, CNN, GCN.
According to one embodiment of the invention, in step S5, it is set that a fault alarm is issued when the fault probability threshold given by the detection model is greater than or equal to 0.6.
Drawings
Fig. 1 schematically shows a flow chart of an audio-based elevator door fault detection method according to the invention.
Detailed Description
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the embodiments will be briefly described below. It is obvious that the drawings in the following description are only some embodiments of the invention, and that for a person skilled in the art, other drawings can be derived from them without inventive effort.
The present invention is described in detail below with reference to the drawings and the specific embodiments, which are not repeated herein, but the embodiments of the present invention are not limited to the following embodiments.
As shown in fig. 1, the present invention provides an elevator door fault detection method based on audio characteristics, which includes S1, collecting a video stream of an elevator operation time period, and acquiring an audio signal therein; s2, preprocessing the audio signal; s3, receiving the preprocessed data, and extracting data features; s4, taking the data features as learning targets, adaptively selecting an abnormal detection model or a supervised learning model according to the size of the existing data volume for training, and updating the detection model in real time; and S5, inputting the data characteristics into the updated detection model, giving a fault probability threshold value by the detection model, and determining whether to send out a fault alarm according to the fault probability threshold value.
Specifically, in step S1, video monitoring audio data in the elevator may be called, an audio signal of an elevator operation time period is obtained from the monitoring data, and then the audio signal is preprocessed in step S2 to obtain data with better quality, so as to ensure the effect of the finally formed detection model. Preprocessing for the audio data includes, but is not limited to, filtering, noise reduction, and singular value removal.
When the audio signal is subjected to denoising processing, a wavelet threshold denoising method is used:
a noisy signal f (t) is set, which can be represented by the following model:
F(t)=s(t)+e(t) (1)
where s (t) is the true signal and e (t) is white noise. Wavelet transformation is simultaneously carried out on two sides of the above formula:
WTf(a,b)=WTs(a,b)+WTe(a,b) (2)
i.e. the wavelet transform of the actual measurement signal is equal to the sum of the wavelet transforms of the plurality of signals.
After wavelet transformation, the correlation of the signal f (t) can be removed to the maximum extent, and most energy is concentrated on a few wavelet coefficients with relatively large amplitude. And the noise e (t) is distributed on all time axes under various scales after wavelet transformation, and the amplitude is not very large. By using the principle, the wavelet coefficient of the noise is reduced to the maximum degree on each scale of the wavelet transformation, and then the processed wavelet coefficient is used for signal reconstruction, thereby achieving the purpose of suppressing the noise.
The method comprises the following specific steps:
the orthogonal wavelet and the number N of decomposition layers are set, and N-layer wavelet decomposition (N ═ 6) is performed on the signal f (t).
And carrying out nonlinear threshold processing on the wavelet transformation coefficient of the measurement signal.
And reconstructing the processed wavelet coefficient.
Next, in step S3, a feature construction is performed to extract important features from the data processed in step S2, so as to improve the effect of the detection model. Important features include, but are not limited to, audio features such as zero-crossing rate, frequency domain energy, spectral centroid, mel-to-cepstrum coefficient, etc.
Then, in S4, the data feature is used as a learning target, and an abnormality detection model or a supervised learning model is selected and trained adaptively according to the size of the existing data amount, so that the detection model is updated in real time. In particular in case of insufficient gate fault data, i.e. a severe imbalance of positive and negative samples. The fitting module automatically adopts an unsupervised anomaly detection model as a training target. The part comprises feature selection, model training, model testing, model evaluation and verification.
Specifically, in the feature selection process, the adaptive matching feature selection algorithm may be based on the feature dimension, and includes, but is not limited to, an exhaustive method, a variance selection method, a correlation coefficient method, and the like. Similarly, the model can also adaptively select a traditional anomaly detection method or a deep learning method, including but not limited to unsupervised clustering, Isolation cluster, One class SVM, Autoencoder, GANormaly, and the like.
Secondly, in the case of sufficient elevator samples, a supervised learning classification model is selected in an adaptive mode, and the part also comprises feature selection, model training, model testing, model evaluation and verification.
In particular, supervised learning classification models include, but are not limited to, conventional machine learning models. Such as Random forms, LightGBM, Xgboost, etc., or a method that employs integration of multiple models. Also, a deep learning model, such as GRU, CNN, GCN, etc., can be selected adaptively, thereby improving the detection effect. The detection effectiveness evaluation index includes, but is not limited to, accuracy, recall, or a weighted f1-score of both, as desired.
Taking a gate Unit (Gated current Unit, GRU) network model as an example:
the inputs to the GRU model are two: input x of the network at the present momentt(i.e., speech signal), output value h of last time GRUt-1. The outputs of the GRU are: and outputting the value GRU at the current moment. Where x, h are both vectors. Two gates are introduced into the GRU: a reset gate and an update gate.
Resetting a gate: determines how to combine the new input information with the previous memory;
and (4) updating the door: the amount of previous memory saved to the current time step is defined.
Update gate calculation:
first at time t the update gate z is calculated using the following formulat:
zt=σ(W(z)xt+U(z)ht-1),
Wherein x istThe output value of the previous time is h for the input vector of the t time stept-1;
Reset gate calculation:
essentially, the reset gate mainly determines how much past information needs to be forgotten, and we can calculate using the following expression:
rt=σ(W(r)xt+U(r)ht-1)
this expression is the same as the expression for the update gate, except that the parameters and uses of the linear transformation are different.
Current memory content
In the use of the reset gate, the new memory will use the reset gate to store the past related information, which is calculated by the expression:
final memory of current time step
In the last step, the network needs to calculate htThe vector will retain the information of the current unit and pass on to the next unit.
Updating the current sequence output:
v is a weight matrix, byIs an offset.
Calculating a full connection layer:
Y=X·W+B
x is input, W is the weight matrix, and B is the offset.
Sof tMax probability value output:
i∈[1,k]wherein z isiIs the output value of the ith node of the neural network model, and C is the number of output nodes, namely the number of classified categories.
Finally, in step S5, the data characteristics are input into the updated detection model, the detection model gives a failure probability threshold, and whether to send out a failure alarm is determined according to the failure probability threshold. According to one embodiment of the invention, the fault alarm is set to be sent when the fault probability threshold value given by the detection model is greater than or equal to 0.6.
Firstly, the method does not need to install sensor equipment on each part, and only needs to call video monitoring audio data in the elevator. The method obtains the audio signal of the elevator running time period from the monitoring data, takes the audio signal as the learning target of the detection model, and can self-adaptively select the detection model according to the size of the existing data volume, namely, the method has low cost and wide application range.
The above description is only one embodiment of the present invention, and is not intended to limit the present invention, and it is apparent to those skilled in the art that various modifications and variations can be made in the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (7)
1. An elevator door fault detection method based on audio features comprises the following steps:
s1, collecting the video stream of the elevator running time period, and acquiring the audio signal therein;
s2, preprocessing the audio signal;
s3, receiving the preprocessed data, and extracting data features;
s4, taking the data features as learning targets, adaptively selecting an abnormal detection model or a supervised learning model according to the size of the existing data volume for training, and updating the detection model in real time;
and S5, inputting the data characteristics into the updated detection model, giving a fault probability threshold value by the detection model, and determining whether to send out a fault alarm according to the fault probability threshold value.
2. The method for detecting elevator door malfunction as claimed in claim 1, wherein the preprocessing of the audio signal in step S2 includes but is not limited to filtering, noise reduction and singular value removal.
3. The method of claim 2, wherein the noise reduction processing is performed on the audio signal using a wavelet threshold denoising method:
setting orthogonal wavelets and a decomposition layer number N, and performing N-layer wavelet decomposition on the signal f (t), wherein N is 6;
carrying out nonlinear threshold processing on the wavelet transform coefficient of the measurement signal;
and reconstructing the processed wavelet coefficient.
4. The method for detecting elevator door fault based on audio features of claim 1, wherein in the step S3, the extracted data features include but are not limited to zero crossing rate, frequency domain energy, spectral centroid, mel-to-common coefficient.
5. The method of claim 1, wherein adaptively selecting an anomaly detection model or a supervised learning model for training based on the magnitude of the existing data volume comprises:
under the condition that door fault data is insufficient, namely under the condition that positive and negative samples are seriously unbalanced, an abnormal detection model is selected for training, in the feature selection process, an algorithm can be selected according to the matching features of feature dimension in a self-adaptive mode, wherein the algorithm comprises but is not limited to an exhaustion method, a variance selection method, a correlation coefficient method and unsupervised clustering, and the abnormal detection model comprises but is not limited to Isolation free, One class SVM, Autoencoder and GANormally.
6. The method of claim 4, wherein adaptively selecting an anomaly detection model or a supervised learning model for training based on the magnitude of the existing data volume further comprises:
in case of sufficient elevator samples, a supervised learning model is adaptively selected, including but not limited to Random forms, LightGBM, Xgboost, multiple model integration GRU, CNN, GCN.
7. The method for detecting elevator door fault based on audio frequency characteristics according to claim 1, wherein in step S5, it is set that when the fault probability threshold value given by the detection model is greater than or equal to 0.6, a fault alarm is issued.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011447416.XA CN112478975A (en) | 2020-12-09 | 2020-12-09 | Elevator door fault detection method based on audio features |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011447416.XA CN112478975A (en) | 2020-12-09 | 2020-12-09 | Elevator door fault detection method based on audio features |
Publications (1)
Publication Number | Publication Date |
---|---|
CN112478975A true CN112478975A (en) | 2021-03-12 |
Family
ID=74941374
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202011447416.XA Pending CN112478975A (en) | 2020-12-09 | 2020-12-09 | Elevator door fault detection method based on audio features |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112478975A (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111170108A (en) * | 2020-02-13 | 2020-05-19 | 浙江新再灵科技股份有限公司 | Method for detecting out-of-door elevator stopping based on multi-source sensor fault label sample data gate |
CN113086799A (en) * | 2021-04-09 | 2021-07-09 | 新沂慧科智能科技有限公司 | Elevator fault detection and early warning method based on image recognition |
CN113581961A (en) * | 2021-08-10 | 2021-11-02 | 江苏省特种设备安全监督检验研究院 | Automatic fault identification method for elevator hall door |
CN114002597A (en) * | 2021-10-25 | 2022-02-01 | 浙江理工大学 | Motor fault diagnosis method and system based on GRU network stator current analysis |
CN117361256A (en) * | 2023-10-10 | 2024-01-09 | 广东全联富士电梯有限公司 | Elevator safety management method and system based on artificial intelligence |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102190224A (en) * | 2010-03-19 | 2011-09-21 | 东芝电梯株式会社 | Automatic diagnosis device of elevator |
JP2013049546A (en) * | 2011-08-31 | 2013-03-14 | Hitachi Building Systems Co Ltd | Method for diagnosing abnormal sound for elevator and system |
CN205204106U (en) * | 2015-11-30 | 2016-05-04 | 华电重工股份有限公司 | Belt conveyor bearing roller operation noise monitoring predicts system with damage |
CN107522053A (en) * | 2017-07-11 | 2017-12-29 | 浙江新再灵科技股份有限公司 | A kind of elevator platform detecting system and method based on audio analysis |
CN108124065A (en) * | 2017-12-05 | 2018-06-05 | 浙江鹏信信息科技股份有限公司 | A kind of method junk call content being identified with disposal |
CN109748170A (en) * | 2019-01-18 | 2019-05-14 | 西人马(西安)测控科技有限公司 | Elevator faults alarm method, device, equipment and medium |
CN111717753A (en) * | 2020-06-29 | 2020-09-29 | 浙江新再灵科技股份有限公司 | Self-adaptive elevator fault early warning system and method based on multi-dimensional fault characteristics |
CN111860267A (en) * | 2020-07-13 | 2020-10-30 | 浙大城市学院 | Multichannel body-building movement identification method based on human body bone joint point positions |
-
2020
- 2020-12-09 CN CN202011447416.XA patent/CN112478975A/en active Pending
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102190224A (en) * | 2010-03-19 | 2011-09-21 | 东芝电梯株式会社 | Automatic diagnosis device of elevator |
JP2013049546A (en) * | 2011-08-31 | 2013-03-14 | Hitachi Building Systems Co Ltd | Method for diagnosing abnormal sound for elevator and system |
CN205204106U (en) * | 2015-11-30 | 2016-05-04 | 华电重工股份有限公司 | Belt conveyor bearing roller operation noise monitoring predicts system with damage |
CN107522053A (en) * | 2017-07-11 | 2017-12-29 | 浙江新再灵科技股份有限公司 | A kind of elevator platform detecting system and method based on audio analysis |
CN108124065A (en) * | 2017-12-05 | 2018-06-05 | 浙江鹏信信息科技股份有限公司 | A kind of method junk call content being identified with disposal |
CN109748170A (en) * | 2019-01-18 | 2019-05-14 | 西人马(西安)测控科技有限公司 | Elevator faults alarm method, device, equipment and medium |
CN111717753A (en) * | 2020-06-29 | 2020-09-29 | 浙江新再灵科技股份有限公司 | Self-adaptive elevator fault early warning system and method based on multi-dimensional fault characteristics |
CN111860267A (en) * | 2020-07-13 | 2020-10-30 | 浙大城市学院 | Multichannel body-building movement identification method based on human body bone joint point positions |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111170108A (en) * | 2020-02-13 | 2020-05-19 | 浙江新再灵科技股份有限公司 | Method for detecting out-of-door elevator stopping based on multi-source sensor fault label sample data gate |
CN113086799A (en) * | 2021-04-09 | 2021-07-09 | 新沂慧科智能科技有限公司 | Elevator fault detection and early warning method based on image recognition |
CN113581961A (en) * | 2021-08-10 | 2021-11-02 | 江苏省特种设备安全监督检验研究院 | Automatic fault identification method for elevator hall door |
CN114002597A (en) * | 2021-10-25 | 2022-02-01 | 浙江理工大学 | Motor fault diagnosis method and system based on GRU network stator current analysis |
CN117361256A (en) * | 2023-10-10 | 2024-01-09 | 广东全联富士电梯有限公司 | Elevator safety management method and system based on artificial intelligence |
CN117361256B (en) * | 2023-10-10 | 2024-03-12 | 广东全联富士电梯有限公司 | Elevator safety management method and system based on artificial intelligence |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112478975A (en) | Elevator door fault detection method based on audio features | |
CN109858104B (en) | Rolling bearing health assessment and fault diagnosis method and monitoring system | |
CN111562108A (en) | Rolling bearing intelligent fault diagnosis method based on CNN and FCMC | |
CN110040594B (en) | Convolutional neural network-based elevator operation detection system and method | |
CN112193959A (en) | Method and system for detecting abnormal sound of elevator | |
CN111678699B (en) | Early fault monitoring and diagnosing method and system for rolling bearing | |
CN109443752B (en) | Gear vibration signal noise reduction and fault diagnosis method based on VMD | |
CN111717753A (en) | Self-adaptive elevator fault early warning system and method based on multi-dimensional fault characteristics | |
CN110333078B (en) | Rolling bearing degradation state stage determination method | |
CN113780432B (en) | Intelligent detection method for operation and maintenance abnormity of network information system based on reinforcement learning | |
CN108508297B (en) | Fault arc detection method based on mutation coefficient and SVM | |
CN115984767A (en) | Abnormity early warning method and system based on real-time analysis of monitoring picture | |
CN114850968B (en) | Cutter abrasion monitoring method, device, terminal and medium based on vibration model | |
CN114785573A (en) | Intelligent substation process layer network abnormal flow detection method based on deep learning | |
CN112664410B (en) | Big data-based modeling method for unit online monitoring system | |
CN106175822A (en) | A kind of elctrocolonogram system based on sound transducer | |
CN112850408B (en) | Elevator emergency stop trapped person fault detection method based on multi-model fusion | |
CN114615086B (en) | Vehicle-mounted CAN network intrusion detection method | |
CN115758086A (en) | Method, device and equipment for detecting faults of cigarette cut-tobacco drier and readable storage medium | |
CN113086798B (en) | Elevator fault detection method based on gated cyclic network and typical correlation analysis | |
CN115632808A (en) | Threat perception method and device for intrusion detection model | |
Chin et al. | A method of fault signature extraction for improved diagnosis | |
CN113505817A (en) | Self-adaptive weighting training method for bearing fault diagnosis model samples under unbalanced data | |
CN112595515A (en) | Power shafting bearing fault detection method and system | |
CN110348481A (en) | One kind being based on the gravitational network inbreak detection method of neighbour's sample |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210312 |