CN111931819A - Machine fault prediction and classification method based on deep learning - Google Patents
Machine fault prediction and classification method based on deep learning Download PDFInfo
- Publication number
- CN111931819A CN111931819A CN202010668783.6A CN202010668783A CN111931819A CN 111931819 A CN111931819 A CN 111931819A CN 202010668783 A CN202010668783 A CN 202010668783A CN 111931819 A CN111931819 A CN 111931819A
- Authority
- CN
- China
- Prior art keywords
- model
- classification
- fault
- prediction
- stage
- 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
- 238000000034 method Methods 0.000 title claims abstract description 41
- 238000013135 deep learning Methods 0.000 title claims abstract description 18
- 238000013145 classification model Methods 0.000 claims abstract description 42
- 238000012549 training Methods 0.000 claims abstract description 37
- 238000000605 extraction Methods 0.000 claims abstract description 30
- 238000007781 pre-processing Methods 0.000 claims abstract description 15
- 230000002457 bidirectional effect Effects 0.000 claims abstract description 12
- 230000006870 function Effects 0.000 claims description 12
- 230000008569 process Effects 0.000 claims description 9
- 230000002159 abnormal effect Effects 0.000 claims description 6
- 238000013528 artificial neural network Methods 0.000 claims description 6
- 238000013527 convolutional neural network Methods 0.000 claims description 6
- 230000009467 reduction Effects 0.000 claims description 6
- 238000012360 testing method Methods 0.000 claims description 6
- 238000012417 linear regression Methods 0.000 claims description 4
- 238000011478 gradient descent method Methods 0.000 claims description 3
- 238000012545 processing Methods 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims description 3
- 238000004519 manufacturing process Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000012423 maintenance Methods 0.000 description 3
- 230000007547 defect Effects 0.000 description 2
- 230000036541 health Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 230000004913 activation Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000007797 corrosion Effects 0.000 description 1
- 238000005260 corrosion Methods 0.000 description 1
- 238000005336 cracking Methods 0.000 description 1
- 230000002950 deficient Effects 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000013508 migration Methods 0.000 description 1
- 230000005012 migration Effects 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 238000013021 overheating Methods 0.000 description 1
- 238000006116 polymerization reaction Methods 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000005096 rolling process Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
-
- 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)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention provides a machine fault prediction and classification method based on deep learning, which comprises the following steps: a data preprocessing stage and a feature extraction stage, wherein a feature extraction model of the CNN-based automatic encoder is established, and features of a preprocessed data set are extracted; a prediction model training stage, namely constructing a prediction model based on bidirectional LSTM, and training the prediction model by using the characteristics obtained in the characteristic extraction stage; a classification model training stage, which is to train a DNN classification model based on an automatic encoder by using the characteristics obtained in the characteristic extraction stage and adopting a supervised learning mode; and a model using stage, namely predicting the time of the fault occurrence by using the trained prediction model, and performing mode classification, fault category classification, fault source positioning and fault degree division on the impending fault by using the trained classification model.
Description
Technical Field
The invention relates to the field of machine health monitoring, in particular to a machine fault prediction and classification method based on deep learning.
Background
Manufacturing systems typically fail due to degraded performance or abnormal operation, resulting in overload, deformation, cracking, overheating, corrosion, and wear. Failure can result in higher operating costs, lower productivity, less defective parts waste, and even unexpected downtime. To implement smart manufacturing, it is important for smart factories to monitor machine conditions, identify incipient defects, diagnose the root cause of a fault, and then incorporate the information into manufacturing production and control.
The prior art discloses a fault prediction method for a migration convolutional neural network, which predicts the type of a fault by collecting signals after the fault occurs; the prior art also discloses a rolling bearing fault prediction method, fault time prediction can be realized through a depth long and short memory network, the research and the invention play a great promoting role in machine health monitoring, but only single fault time prediction or fault type prediction can be realized.
Disclosure of Invention
Aiming at the defects in the prior art, the invention provides a machine fault prediction and classification method based on deep learning, which can predict the occurrence of faults and classify the faults to be occurred while predicting the occurrence of the faults.
The present invention achieves the above-described object by the following technical means.
A deep learning based machine fault prediction and classification method comprises the following steps:
a data preprocessing stage, which is used for preprocessing the acquired time sequence data;
a characteristic extraction stage, namely establishing a characteristic extraction model of the CNN-based automatic encoder, and extracting the characteristics of the preprocessed data set by using the characteristic extraction model;
a prediction model training stage, namely constructing a prediction model based on bidirectional LSTM, and training the prediction model by using the characteristics obtained in the characteristic extraction stage;
a classification model training stage, which is to train the DNNs classification model based on the automatic encoder by using the characteristics obtained in the characteristic extraction stage and adopting a supervised learning mode; and
and in the model using stage, the trained prediction model is used for predicting the time of the fault occurrence, and the trained classification model is used for carrying out mode classification, fault category classification, fault source positioning and fault degree division on the impending fault.
Furthermore, in the data preprocessing stage, time series data are acquired through a sensor and the like, and are preprocessed, wherein the preprocessing comprises data abnormal value processing, missing value filling and data dimension reduction.
Further, the feature extraction stage specifically includes:
the method comprises the steps of constructing a CNN-based automatic encoder comprising an input layer, an encoding layer, a decoding layer and an output layer, wherein the input N of the input layer is a data set obtained in a data preprocessing stage, the encoding layer carries out down-sampling and Feature extraction to obtain Feature, the decoding layer reduces the Feature into the input N through deconvolution, a loss function F (Y, N) is established, the loss function is MSE, and the loss function is reduced through a gradient descent method so that the input N is equal to the output Y to extract input features.
Further, the air conditioner is provided with a fan,
the specific steps of constructing the prediction model based on the bidirectional LSTM are as follows:
constructing two layers of bidirectional LSTM neural networks to encode a time mode;
stacking two fully connected dense layers together to process the output of the LSTM;
predicting by adopting a linear regression layer;
the training of the prediction model by using the features obtained in the feature extraction stage specifically comprises the following steps:
during training, given a predicted output and a real target, calculating the mean square error of training data, performing back propagation to update model parameters, introducing Dropout during model training, and randomly shielding part of hidden outputs through Dropout so that the hidden outputs cannot influence forward propagation in the training process, and closing Dropout during a testing stage to enable all hidden outputs to influence the testing of the prediction model.
Further, the classification model training phase:
the method comprises the steps of establishing a DNN model based on an automatic encoder, wherein the DNN is a multilayer neural network formed by overlapping a plurality of self-encoders, adding a softmax classifier at the top of the DNN model to obtain a classification model, extracting features layer by adopting a bottom-up unsupervised learning method for the classification model, and finely adjusting the classification model by utilizing a supervised learning method to obtain a trained classification model.
Further, the classification model is a hierarchical DNN model, the first level of the classification model is used for mode division, and the second level of the classification model is used for constructing different DNN classification models for each mode; the third level is used for fault severity classification for each fault category.
Further, the model using stage is specifically as follows:
inputting the features obtained in the feature extraction stage into a trained prediction model for prediction, when the predicted value exceeds a set threshold value, giving an early warning, inputting the predicted data into the trained classification model, and performing fault category classification, fault source positioning and fault degree division on the to-be-generated fault.
The invention has the beneficial effects that:
the invention provides a machine fault prediction and classification method based on deep learning, which can predict the occurrence time of a fault in advance through a prediction model based on bidirectional LSTM, classify prediction data through a classification model based on an autoencoder to predict the type, position and fault degree of the fault, detect the abnormal condition of the machine in early stage, reasonably arrange the shutdown time to perform planned maintenance, prolong the effective service life of the machine, improve the working efficiency of the machine and avoid the loss of time cost and economic cost caused by shutdown in peak period.
Drawings
FIG. 1 is a schematic flow chart of a method for predicting and classifying machine faults based on deep learning according to an embodiment of the present invention;
FIG. 2 is a block diagram of an auto encoder model according to an embodiment of the present invention;
FIG. 3 is a block diagram of a classification model according to an embodiment of the invention;
fig. 4 is a block diagram of a bi-directional LSTM in accordance with an embodiment of the present invention.
Detailed Description
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in detail below. In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention. This invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein, but rather should be construed as broadly as the present invention is capable of modification in various respects, all without departing from the spirit and scope of the present invention.
The following first describes a deep learning based machine fault prediction and classification method according to an embodiment of the present invention in detail with reference to the accompanying drawings.
Referring to fig. 1, a method for predicting and classifying machine faults based on deep learning according to an embodiment of the present invention includes the following steps
S1 data preprocessing stage:
acquiring time series data through a sensor and the like and preprocessing the time series data to obtain a data set { Xj(j ═ 1, 2, … m), where XjIs a k-dimensional vector. The preprocessing mainly comprises data abnormal value processing, missing value filling, data dimension reduction and the like, wherein the data dimension reduction specifically comprises the following steps:
converting the k-dimensional vector into a two-dimensional gray image, adopting a method of reducing the dimension at least twice when the k value is larger, firstly converting the k-dimensional vector into k by utilizing segmented polymerization in one dimension reduction1Dimension vector, k1<k, resulting in a data set { Xi(i ═ 1, 2, … m) in which X isiFor k1 dimensional vector, then normalizing the obtained data set to make all elements take values between 0 and 1, and making k1Dimension vector XiConversion into a grey scale map Mml iWherein
Ceil () represents taking the integer boundary, if the data dimension is still very large, adopting secondary dimension reduction, secondary dimension reduction pairGrey scale map Mml iScaling the equal proportion to obtain a scaled data set { N }iIn which N isiIs k2Dimension vector, k2<k1。
S2 feature extraction stage:
as shown in FIG. 2, a CNN-based automatic encoder is constructed, which comprises an input layer 1, an encoding layer 2, a decoding layer 4 and an output layer 5, wherein the input N of the input layer is a data set { X ] obtained in a data preprocessing stageiEither { N } or { N }iAnd (3) carrying out down-sampling and Feature extraction on the coding layer 2 to obtain Feature, reducing the Feature into input N by the decoding layer 4 through deconvolution, establishing a loss function F (Y, N), wherein the loss function is MSE, and reducing the loss function through a gradient descent method to enable the input to be equal to the output so as to extract input data NiThe characteristics of (1). The feature extraction model training stage uses the coding layer 2 and the decoding layer 4, and the feature extraction model using stage uses only the coding layer 2.
S3 predictive model training phase:
and constructing a prediction model based on the bidirectional LSTM, and training the prediction model by using the characteristics obtained in the characteristic extraction stage.
As shown in fig. 4, the specific steps of constructing the bidirectional LSTM-based prediction model are as follows:
the prediction model first constructs a two-layer bi-directional LSTM neural network to encode a time pattern, then stacks two fully connected dense layers together to process the LSTM output, and finally employs a linear regression layer for prediction. Bidirectional LSTM can process sequence data in both directions, including forward and reverse modes using two separate hidden layers, and then forward it to the same output layer. The following equations define the corresponding hidden layer functions, → and ← representing the forward and backward processes, respectively:
the complete LSTM hidden element representation ht is the concatenated vector of the forward and reverse process outputs, as follows:
wherein f istTo forget the gate output, itFor input to the gate output, otFor output of the output gate, htIs LSTM the mostFinal output, σ is the activation function ReLu, all W ∈ Rd×k,V∈Rd×d,b∈RdObtained in model training.
Two fully connected dense layers are stacked together, with the output of one layer being used as input to the next layer, and the formula for each layer being:
oi=g(Wi*hi+bi)
wherein o isiAnd hiRespectively representing the output and input of the ith fully-connected layer, WiAnd biRespectively representing the transformation matrix and the bias term in the ith fully-connected layer, with the function g () set to ReLu.
The feature training of the prediction model obtained in the feature extraction stage specifically comprises:
during training, given the predicted output and a real target, calculating the mean square error of training data, performing back propagation to update model parameters, introducing Dropout during model training, wherein partial hidden outputs are randomly shielded through Dropout so that the neurons cannot influence forward propagation in the training process, Dropout is closed in a testing stage, all hidden outputs influence the testing of the prediction model, and finally learned data represents that the prediction is input into a linear regression layer for prediction.
S4 classification model training phase:
establishing a DNN model based on an automatic encoder, wherein the structure of the DNN is as shown in FIG. 3, the DNN can be simply regarded as a multi-layer neural network formed by overlapping a plurality of self-encoders, then a softmax classifier is added to the top of the DNN model to obtain a classification model, the classification model is a DNN model with a hierarchical structure, the first level of the classification model is used for mode division, and the second level of the classification model is used for constructing different DNN classification models for each mode to obtain a more accurate fault classification result; to provide useful information for predictive maintenance, an additional DNN is constructed in the third hierarchy, further dividing a fault in a given mode into several classes of different fault severity.
And then, extracting features layer by adopting a bottom-up unsupervised learning method for the classification model, and then finely adjusting the classification model by utilizing a supervised learning method to obtain the trained classification model. The specific method for training the classification model comprises the following steps:
first, the DNN is pre-trained using an unsupervised greedy, layer-by-layer training algorithm, and then the first Automatic Encoder (AE) is trained by giving an unlabeled dataset x (features from the feature extraction stage) as input to the encoder network1) Coding feature h1Is AE1By making the input equal to the output to obtain the training parameter theta1Then, h is1Used as a second Automatic Encoder (AE)2) Input of (2), and training AE2To obtain a network training parameter theta2,h2Is AE2Can be regarded as AE2After that, h is selected2As a third Automatic Encoder (AE)3) Is repeated to obtain the hidden layer feature h of the "Auto Encoder (AE)"NAnd corresponding network training parameter thetaN。
S5 model use phase:
inputting the features obtained in the feature extraction stage into a trained prediction model for prediction, when the predicted value exceeds a set threshold value, giving an early warning, inputting the predicted data into the trained classification model, and performing fault category classification, fault source positioning and fault degree division on the to-be-generated fault.
The machine fault prediction and classification method based on deep learning provided by the embodiment of the invention can predict the occurrence time of the fault in advance through the prediction model based on the bidirectional LSTM, and classify the prediction data through the classification model based on the self-encoder to predict the type, position and fault degree of the fault, can detect the abnormal condition of the machine in early stage, reasonably arrange the shutdown time to carry out planned maintenance, thereby prolonging the effective service life of the machine, improving the working efficiency of the machine and avoiding the loss of time cost and economic cost caused by shutdown in the peak period.
The above-mentioned embodiments only express several embodiments of the present invention, and the description thereof is more specific and detailed, but not construed as limiting the scope of the present invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the inventive concept, which falls within the scope of the present invention. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (7)
1. A deep learning based machine fault prediction and classification method is characterized by comprising the following steps:
a data preprocessing stage, which is used for preprocessing the acquired time sequence data;
a characteristic extraction stage, namely establishing a characteristic extraction model of the CNN-based automatic encoder, and extracting the characteristics of the preprocessed data set by using the characteristic extraction model;
a prediction model training stage, namely constructing a prediction model based on bidirectional LSTM, and training the prediction model by using the characteristics obtained in the characteristic extraction stage;
a classification model training stage, which is to train the DNNs classification model based on the automatic encoder by using the characteristics obtained in the characteristic extraction stage and adopting a supervised learning mode; and
and in the model using stage, the trained prediction model is used for predicting the time of the fault occurrence, and the trained classification model is used for carrying out mode classification, fault category classification, fault source positioning and fault degree division on the impending fault.
2. The deep learning-based machine fault prediction and classification method according to claim 1, wherein the data preprocessing stage acquires time series data through a sensor and the like and performs preprocessing, including data abnormal value processing, missing value filling and data dimension reduction.
3. The deep learning-based machine fault prediction and classification method according to claim 1, wherein the feature extraction stage specifically comprises:
the method comprises the steps of constructing a CNN-based automatic encoder comprising an input layer, an encoding layer, a decoding layer and an output layer, wherein the input N of the input layer is a data set obtained in a data preprocessing stage, the encoding layer carries out down-sampling and Feature extraction to obtain Feature, the decoding layer reduces the Feature into the input N through deconvolution, a loss function F (Y, N) is established, the loss function is MSE, and the loss function is reduced through a gradient descent method so that the input N is equal to the output Y to extract input features.
4. The deep learning based machine fault prediction and classification method of claim 1,
the specific steps of constructing the prediction model based on the bidirectional LSTM are as follows:
constructing two layers of bidirectional LSTM neural networks to encode a time mode;
stacking two fully connected dense layers together to process the output of the LSTM;
predicting by adopting a linear regression layer;
the training of the prediction model by using the features obtained in the feature extraction stage specifically comprises the following steps:
during training, given a predicted output and a real target, calculating the mean square error of training data, performing back propagation to update model parameters, introducing Dropout during model training, and randomly shielding part of hidden outputs through Dropout so that the hidden outputs cannot influence forward propagation in the training process, and closing Dropout during a testing stage to enable all hidden outputs to influence the testing of the prediction model.
5. The deep learning-based machine fault prediction and classification method according to claim 1, wherein the classification model training phase specifically comprises:
the method comprises the steps of establishing a DNN model based on an automatic encoder, wherein the DNN is a multilayer neural network formed by overlapping a plurality of self-encoders, adding a softmax classifier at the top of the DNN model to obtain a classification model, extracting features layer by adopting a bottom-up unsupervised learning method for the classification model, and finely adjusting the classification model by utilizing a supervised learning method to obtain a trained classification model.
6. The deep learning-based machine fault prediction and classification method according to claim 5, wherein the classification model is a hierarchical DNN model, a first level of the classification model is used for mode division, and a second level is used for constructing a different DNN classification model for each mode; the third level is used for fault severity classification for each fault category.
7. The deep learning-based machine fault prediction and classification method according to claim 1, wherein the model usage phases are specifically:
inputting the features obtained in the feature extraction stage into a trained prediction model for prediction, when the predicted value exceeds a set threshold value, giving an early warning, inputting the predicted data into the trained classification model, and performing fault category classification, fault source positioning and fault degree division on the to-be-generated fault.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010668783.6A CN111931819A (en) | 2020-07-13 | 2020-07-13 | Machine fault prediction and classification method based on deep learning |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010668783.6A CN111931819A (en) | 2020-07-13 | 2020-07-13 | Machine fault prediction and classification method based on deep learning |
Publications (1)
Publication Number | Publication Date |
---|---|
CN111931819A true CN111931819A (en) | 2020-11-13 |
Family
ID=73312468
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010668783.6A Pending CN111931819A (en) | 2020-07-13 | 2020-07-13 | Machine fault prediction and classification method based on deep learning |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111931819A (en) |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112462198A (en) * | 2020-11-17 | 2021-03-09 | 国网四川省电力公司电力科学研究院 | Power grid fault line judgment method and system based on self-encoder |
CN112491468A (en) * | 2020-11-20 | 2021-03-12 | 福州大学 | FBG sensing network node fault positioning method based on twin node auxiliary sensing |
CN112989976A (en) * | 2021-03-03 | 2021-06-18 | 南京航空航天大学 | Equipment failure mode prediction method based on double-depth learning model |
CN113408371A (en) * | 2021-06-01 | 2021-09-17 | 武汉理工大学 | Early fault diagnosis method and device |
CN114932582A (en) * | 2022-06-16 | 2022-08-23 | 上海交通大学 | Robot small-probability failure prediction method based on Bi-GRU self-encoder |
CN115242680A (en) * | 2022-07-30 | 2022-10-25 | 北京理工大学 | Node classification method of graph neural network based on multi-stage training in communication network |
CN115906642A (en) * | 2022-11-28 | 2023-04-04 | 东莞科达五金制品有限公司 | Bearing production detection control method and device |
CN116188449A (en) * | 2023-03-13 | 2023-05-30 | 哈尔滨市科佳通用机电股份有限公司 | Rail wagon relief valve pull rod split pin loss fault identification method and equipment |
CN117009861A (en) * | 2023-10-08 | 2023-11-07 | 湖南国重智联工程机械研究院有限公司 | Hydraulic pump motor life prediction method and system based on deep learning |
CN117078232A (en) * | 2023-10-17 | 2023-11-17 | 山东沪金精工科技股份有限公司 | Processing equipment fault prevention system and method based on big data |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109033450A (en) * | 2018-08-22 | 2018-12-18 | 太原理工大学 | Lift facility failure prediction method based on deep learning |
CN110647891A (en) * | 2019-09-17 | 2020-01-03 | 上海仪电(集团)有限公司中央研究院 | CNN (convolutional neural network) -based automatic extraction method and system for time sequence data characteristics of self-encoder |
CN110779988A (en) * | 2019-10-30 | 2020-02-11 | 同济大学 | Bolt life prediction method based on deep learning |
-
2020
- 2020-07-13 CN CN202010668783.6A patent/CN111931819A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109033450A (en) * | 2018-08-22 | 2018-12-18 | 太原理工大学 | Lift facility failure prediction method based on deep learning |
CN110647891A (en) * | 2019-09-17 | 2020-01-03 | 上海仪电(集团)有限公司中央研究院 | CNN (convolutional neural network) -based automatic extraction method and system for time sequence data characteristics of self-encoder |
CN110779988A (en) * | 2019-10-30 | 2020-02-11 | 同济大学 | Bolt life prediction method based on deep learning |
Cited By (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112462198A (en) * | 2020-11-17 | 2021-03-09 | 国网四川省电力公司电力科学研究院 | Power grid fault line judgment method and system based on self-encoder |
CN112491468A (en) * | 2020-11-20 | 2021-03-12 | 福州大学 | FBG sensing network node fault positioning method based on twin node auxiliary sensing |
CN112989976A (en) * | 2021-03-03 | 2021-06-18 | 南京航空航天大学 | Equipment failure mode prediction method based on double-depth learning model |
CN113408371A (en) * | 2021-06-01 | 2021-09-17 | 武汉理工大学 | Early fault diagnosis method and device |
CN114932582B (en) * | 2022-06-16 | 2024-01-23 | 上海交通大学 | Robot small probability failure prediction method based on Bi-GRU self-encoder |
CN114932582A (en) * | 2022-06-16 | 2022-08-23 | 上海交通大学 | Robot small-probability failure prediction method based on Bi-GRU self-encoder |
CN115242680A (en) * | 2022-07-30 | 2022-10-25 | 北京理工大学 | Node classification method of graph neural network based on multi-stage training in communication network |
CN115906642A (en) * | 2022-11-28 | 2023-04-04 | 东莞科达五金制品有限公司 | Bearing production detection control method and device |
CN116188449B (en) * | 2023-03-13 | 2023-08-08 | 哈尔滨市科佳通用机电股份有限公司 | Rail wagon relief valve pull rod split pin loss fault identification method and equipment |
CN116188449A (en) * | 2023-03-13 | 2023-05-30 | 哈尔滨市科佳通用机电股份有限公司 | Rail wagon relief valve pull rod split pin loss fault identification method and equipment |
CN117009861A (en) * | 2023-10-08 | 2023-11-07 | 湖南国重智联工程机械研究院有限公司 | Hydraulic pump motor life prediction method and system based on deep learning |
CN117009861B (en) * | 2023-10-08 | 2023-12-15 | 湖南国重智联工程机械研究院有限公司 | Hydraulic pump motor life prediction method and system based on deep learning |
CN117078232A (en) * | 2023-10-17 | 2023-11-17 | 山东沪金精工科技股份有限公司 | Processing equipment fault prevention system and method based on big data |
CN117078232B (en) * | 2023-10-17 | 2024-01-09 | 山东沪金精工科技股份有限公司 | Processing equipment fault prevention system and method based on big data |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111931819A (en) | Machine fault prediction and classification method based on deep learning | |
CN106555788B (en) | Application based on the deep learning of Fuzzy Processing in hydraulic equipment fault diagnosis | |
CN107941537A (en) | A kind of mechanical equipment health state evaluation method | |
CN111178553A (en) | Industrial equipment health trend analysis method and system based on ARIMA and LSTM algorithms | |
Yacout | Fault detection and diagnosis for condition based maintenance using the logical analysis of data | |
CN113762329A (en) | Method and system for constructing state prediction model of large rolling mill | |
Pu et al. | A one-class generative adversarial detection framework for multifunctional fault diagnoses | |
CN112132394B (en) | Power plant circulating water pump predictive state evaluation method and system | |
Guo et al. | System operational reliability evaluation based on dynamic Bayesian network and XGBoost | |
Abou | Fuzzy-logic-based network for complex systems risk assessment: Application to ship performance analysis | |
CN114444582A (en) | Mechanical equipment fault diagnosis method based on convolutional neural network and Bayesian network | |
CN113485863A (en) | Method for generating heterogeneous unbalanced fault samples based on improved generation countermeasure network | |
Pariaman et al. | Anomaly detection using LSTM-Autoencoder to predict coal pulverizer condition on Coal-fired power plant | |
CN116821730B (en) | Fan fault detection method, control device and storage medium | |
CN117454771A (en) | Mechanical equipment dynamic maintenance decision-making method based on evaluation and prediction information | |
CN117493930A (en) | Multi-element time sequence anomaly detection method based on contrast generation countermeasure network | |
Abidi et al. | Predictive Maintenance Planning for Industry 4.0 Using Machine Learning for Sustainable Manufacturing. Sustainability 2022, 14, 3387 | |
CN114298413B (en) | Hydroelectric generating set runout trend prediction method | |
CN113191306A (en) | Equipment abnormal state prediction method based on edge calculation | |
CN116956089A (en) | Training method and detection method for temperature anomaly detection model of electrical equipment | |
CN114494766A (en) | Turnout fault diagnosis method based on hybrid deep learning model | |
Le et al. | An energy data-driven decision support system for high-performance manufacturing industries | |
CN108388232B (en) | Method for monitoring operation mode fault in crude oil desalting process | |
Chowdhury et al. | Control chart pattern recognition: A comparison between statistical correlation measure and support vector machine (svm) | |
Vachkov | Intelligent data analysis for performance evaluation and fault diagnosis in complex systems |
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 |