CN116028849B - Emulsion pump fault diagnosis method based on depth self-coding network - Google Patents
Emulsion pump fault diagnosis method based on depth self-coding network Download PDFInfo
- Publication number
- CN116028849B CN116028849B CN202211723315.XA CN202211723315A CN116028849B CN 116028849 B CN116028849 B CN 116028849B CN 202211723315 A CN202211723315 A CN 202211723315A CN 116028849 B CN116028849 B CN 116028849B
- Authority
- CN
- China
- Prior art keywords
- emulsion pump
- real
- depth self
- coding network
- time parameters
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 239000000839 emulsion Substances 0.000 title claims abstract description 61
- 238000003745 diagnosis Methods 0.000 title claims abstract description 56
- 238000000034 method Methods 0.000 title claims abstract description 38
- 238000012549 training Methods 0.000 claims abstract description 31
- 238000012544 monitoring process Methods 0.000 claims abstract description 24
- 230000002159 abnormal effect Effects 0.000 claims abstract description 15
- 230000008569 process Effects 0.000 claims abstract description 9
- 230000001960 triggered effect Effects 0.000 claims abstract description 4
- 230000006870 function Effects 0.000 claims description 11
- 238000012545 processing Methods 0.000 claims description 9
- 238000012360 testing method Methods 0.000 claims description 8
- 238000011478 gradient descent method Methods 0.000 claims description 5
- 210000002569 neuron Anatomy 0.000 claims description 4
- 238000007781 pre-processing Methods 0.000 claims description 4
- 238000010606 normalization Methods 0.000 claims description 3
- 238000004458 analytical method Methods 0.000 abstract description 5
- 238000013135 deep learning Methods 0.000 description 6
- 238000004364 calculation method Methods 0.000 description 4
- 230000008859 change Effects 0.000 description 4
- 238000011160 research Methods 0.000 description 3
- 230000005856 abnormality Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 239000003245 coal Substances 0.000 description 2
- 238000013136 deep learning model Methods 0.000 description 2
- 230000036541 health Effects 0.000 description 2
- 230000000670 limiting effect Effects 0.000 description 2
- 238000012423 maintenance Methods 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 208000019901 Anxiety disease Diseases 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 230000036506 anxiety Effects 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 238000010008 shearing Methods 0.000 description 1
Landscapes
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
Abstract
The invention belongs to the technical field of emulsion pump diagnosis, and relates to an emulsion pump fault diagnosis method based on a depth self-coding network, which comprises the following steps: 1) Monitoring real-time parameters in the running process of the emulsion pump; 2) Analyzing the real-time parameters, comparing the real-time parameters with a set threshold value, and if the real-time parameters are within the threshold value range, indicating that the emulsion pump is normal in operation, and continuing to monitor and analyze the real-time parameters; if the real-time parameter exceeds the threshold value, the operation of the emulsion pump is abnormal, and an alarm is triggered; 3) Inputting real-time parameters corresponding to the alarm time into a depth self-coding network model for training; 4) And acquiring actual data of the operation of the emulsion pump on line, and inputting the actual data into the trained depth self-coding network model to obtain an emulsion pump fault diagnosis result. According to the invention, through parameter analysis and a depth self-coding network model, historical data and monitoring data can be analyzed, the fault diagnosis precision and efficiency diagnosis are obviously improved, and the result is more reliable.
Description
Technical Field
The invention belongs to the technical field of emulsion pump diagnosis, and relates to an emulsion pump fault diagnosis method based on a depth self-coding network.
Background
The emulsion pump is a pump body which generates strong shearing force in high-speed rotation to realize mixing, homogenizing, dispersing and crushing, and the abnormal operation of operators can lead to the damage of the pump body and even cause the threat of personal safety, so the identification of the abnormal operation behavior of the emulsion pump has important significance.
The fault mechanism of the emulsion pump is complex, and the fault diagnosis research of the emulsion pump is slow to develop at present and is mostly limited to reliability analysis and health assessment. The traditional fault diagnosis method of the existing emulsion pump comprises the following steps: although the method for diagnosing individual faults such as a temperature monitoring method, a vibration analysis method, an oil analysis method and the like can diagnose an emulsion pump, the following problems exist: the traditional fault diagnosis method is too dependent on the technical level of maintenance personnel, and diagnosis precision and efficiency cause anxiety are not beneficial to intelligent development of the coal mine; the existing fault diagnosis method has low precision, needs a large amount of data, lacks data analysis of complex faults, and has unbalanced precision and efficiency.
Disclosure of Invention
The invention aims to provide an emulsion pump fault diagnosis method based on a depth self-coding network, which is characterized in that the accuracy and the efficiency of fault diagnosis are obviously improved by analyzing historical monitoring data, and the result is more reliable.
In order to achieve the above purpose, the technical scheme adopted by the invention is as follows:
an emulsion pump fault diagnosis method based on a depth self-coding network comprises the following steps:
1) Monitoring real-time parameters in the running process of the emulsion pump;
2) Analyzing the real-time parameters, comparing the real-time parameters with a set threshold value, and if the real-time parameters are within the threshold value range, indicating that the emulsion pump is normal in operation, and continuing to monitor and analyze the real-time parameters; if the data exceeds the threshold value, the operation of the emulsion pump is abnormal, and an alarm is triggered;
3) Inputting real-time parameters corresponding to the alarm time into a depth self-coding network model for training;
4) And acquiring actual data of the operation of the emulsion pump on line, and inputting the actual data into the trained depth self-coding network model to obtain an emulsion pump fault diagnosis result.
Further, in the step 1), real-time parameters in the running of the emulsion pump are collected by installing a sensor group.
Further, the specific process of the step 2) is as follows:
2.1 Processing the real-time data acquired in the step 1) and storing the processed real-time data into a database;
2.2 Setting a threshold value when the running parameters of the emulsion pump are abnormal according to the empirical value;
2.3 Comparing the real-time parameter corresponding to a certain moment with a threshold value, if the acquired real-time parameter exceeds the threshold value, indicating that the emulsion pump is abnormal in operation, continuously judging that the emulsion pump is abnormal three times, and triggering an alarm.
Further, the specific process of the step 3) is as follows:
3.1 Pre-processing the monitored real-time parameters, preparing a data set, and according to 8: dividing a training set and a testing set, and calibrating labels on part of the training set;
3.2 A depth self-coding network model is constructed, the number of neurons of each hidden layer is determined, and the weight is randomly initialized;
3.3 Inputting a label-free training set, carrying out training iteration on the depth self-coding network model, and updating weight and bias by minimizing a mean square error loss value between an input layer and an output layer;
3.4 Adding a classifier after the depth self-coding network model is trained in the step 3.3), inputting a training set sample containing a label, fine-tuning the weight by a gradient descent method and completing classifier learning;
3.5 Inputting a test set into the depth self-coding network model after learning in the step 3.4), comparing the output diagnosis result with the calibration label, and if the diagnosis result is the same as the calibration label, storing the training model.
Further, in the step 3.3), the mean square error loss value is calculated according to the following formula:
Wherein: l mse is a loss function, m is the number of samples in a batch, k represents the kth sample, j represents the number of dimensions of each sample, Representing the actual output value of the j-th dimension in the kth sample,/>Representing the predicted value of the j-th dimension in the kth sample.
Further, in the step 3.3), the iteration number is 100, and the learning rate of the parameter is 0.001.
Further, in the step 3.4), the weight fine adjustment is to ensure that the cross entropy L cel is minimum, and the cross entropy is calculated according to the following formula:
Wherein: l cel is the cross entropy function, m is the number of samples a batch contains, k represents the kth sample, j represents the number of dimensions of each sample, Representing the actual output value of the j-th dimension in the kth sample,/>Representing the predicted value of the j-th dimension in the kth sample.
Further, in the step 3.4), the classifier is Softmax, the iteration number is 100, and the learning rate of the parameter is 0.001.
Further, in the step 3.1), the preprocessing includes missing value and outlier processing; and the data in the training set and the testing set are subjected to normalization processing.
The beneficial effects of the invention are as follows:
1. the invention provides a fault diagnosis method combining parameter diagnosis and deep learning diagnosis, which can exert the advantages of high timeliness of the parameter diagnosis method and high diagnosis precision of the deep learning diagnosis method, and obviously improves the precision and efficiency of the fault diagnosis.
2. The invention applies deep learning to fault diagnosis of emulsion pump stations, self-adaptively excavates characteristics from equipment monitoring data, reduces dependence on expert experience, ensures more reliable results by analyzing historical monitoring data, improves the accuracy of fault diagnosis, optimizes maintenance and repair strategies of equipment, and enriches means of fault diagnosis and health management of coal mine machinery.
3. The emulsion pump station fault diagnosis model based on deep learning is constructed, the corresponding relation between the monitoring parameters and the fault state is reflected, and the intelligent degree of fault diagnosis is improved.
Drawings
FIG. 1 is a hierarchical fault diagnosis flow provided by the present invention;
FIG. 2 is a parameter diagnostic flow;
FIG. 3 is an automatic encoder configuration;
FIG. 4 is a depth self-encoding network architecture;
FIG. 5 is a fault diagnosis flow of an emulsion pump station based on a depth self-coding network.
Detailed Description
The invention will now be described in detail with reference to the drawings and examples.
Referring to fig. 1, the fault diagnosis method of the emulsion pump provided by the invention is a hierarchical fault diagnosis method combining parameter diagnosis and deep learning diagnosis.
First, the system monitors the running state of the emulsion pump by parameter diagnosis, and the real-time parameters in the running of the emulsion pump are acquired through a sensor group (such as a pressure sensor, a flow sensor, a temperature sensor and the like) arranged on the emulsion pump. If the real-time parameters are normal, monitoring is continued, and if the real-time parameters are abnormal, an alarm is triggered and the deep learning diagnosis is skipped.
The parameter diagnosis method can be established only by effectively monitoring the data for a long time, the parameter change of the mine emulsion pump during the past abnormality and the fault is recorded, the monitoring parameters of the machine in the running process always change within a certain range, and when the equipment is abnormal, one or more monitoring parameters are mutated. Therefore, on the premise of carrying out long-term data monitoring and analysis on the emulsion pump, setting a parameter abnormal threshold according to the variation range of the monitoring parameter of the ore emulsion pump, and alarming when the monitored real-time parameter exceeds the set threshold, wherein the method is the parameter diagnosis.
For parameter diagnosis, firstly, acquiring and storing real-time parameters, installing sensors in an emulsion pump, transmitting the sensor to a remote monitoring center through a special line (Modbus/TCP protocol), storing the sensor in a database after data processing, and storing multi-source data in the database at the same time of 5s interval. And then setting a threshold value, determining the maximum range of the monitoring real-time parameter by combining the technical specification, the technical protocol and the instruction of the emulsion pump station, analyzing the parameter change in actual production to determine the change range of the monitoring real-time parameter, and finishing the threshold value of the normal working range of the monitoring real-time parameter.
Referring to fig. 2, the flow of parameter diagnostics is: firstly, reading the whole row data at a certain moment from a database, comparing the whole row data with a set threshold value respectively, reading the whole row data at the next moment if a certain monitoring parameter or a plurality of monitoring parameters meet fault state conditions, repeating the continuous judgment for three times in order to avoid the problems of missing diagnosis, misdiagnosis and the like caused by data missing and data abnormality in the data acquisition and transmission process, and judging that diagnosis is correct if a certain monitoring real-time parameter or a plurality of monitoring real-time parameters meet the fault state conditions for three times continuously, triggering an alarm and storing a fault result.
The deep learning model adopts a deep self-coding network, is a deep learning model commonly used for non-supervision learning, can directly cross fault mechanism research and signal processing algorithm research, constructs an end-to-end fault diagnosis model of a general part, and realizes fault classification through self-adaptive feature extraction, selection and classification. The depth self-coding network is formed by stacking a plurality of self-encoders.
Referring to fig. 3, the self-encoder includes three layers: the input layer, the hidden layer and the output layer form a coding network, the input layer and the hidden layer convert an input vector x containing high-dimensional characteristics into a low-dimensional characteristic vector y, the hidden layer and the output layer form a decoding network, and the low-dimensional characteristic vector y is converted into an output vector z containing high-order characteristics.
The mathematical expression of the self-encoder is as follows:
wherein x＝(x1,x2,…,xn),y＝(y1,y2,…ym),z＝(z1,z2,…zm) is the input, the value of hidden layer neuron, and the output, respectively. f is the activation function, W and W 'are weights, and b' are offsets.
And sequentially stacking a plurality of self-encoders, inputting the calculation results of the first-layer self-encoder into the second-layer self-encoder for calculation again, and inputting the calculation results into the third-layer self-encoder after calculation, namely the depth self-encoding network, wherein the structure is shown in fig. 4.
Training the depth self-coding network, firstly performing unsupervised pre-training, inputting training samples without labels, and updating weights and offsets by minimizing a mean square error loss value between an input layer and an output layer, namely reconstructing errors. And then, performing supervised fine tuning, adding a classifier after the network is encoded, inputting a certain proportion of samples containing labels, adjusting the weight and bias of an hidden layer, and completing the training of the classifier.
The model learning process is to update the weight and bias of the network to minimize the loss function, and the loss function has a mean square error (mse) and a cross entropy (cel) and has the following formula:
m is the number of samples contained in a batch, k represents the kth sample, j represents the number of dimensions of each sample, Representing the actual output value of the j-th dimension in the kth sample,/>Representing the predicted value of the j-th dimension in the kth sample.
Referring to fig. 5, the fault monitoring process for the emulsion pump station through the depth self-coding network is as follows:
(1) The method comprises the steps of collecting multi-source data of all monitoring parameters of an emulsion pump station at different moments, wherein 14 fault forms are covered, missing values and abnormal values are processed, a data set is manufactured in a standardized mode, a label is calibrated, and the method is characterized in that: and 2, dividing the training set and the testing set, and carrying out normalization processing on all data sets.
When the method is implemented, the partial data in the training set are labeled, and the training set is divided into a label-free training set and a label-containing training set.
(2) And constructing a depth self-coding network model, determining the number of neurons of each hidden layer, and randomly initializing the weight.
(3) And inputting a label-free training set, carrying out training iteration on the depth self-coding network model, and updating the weight and the bias by minimizing a mean square error loss value between an input layer and an output layer.
The weight is updated by a gradient descent method, the loss function is mean square error, 50 weight models Epochs weight models and the learning rate is 0.01 weight model.
The parameter Epochs is defined to set the number of iterations, 0.01 is defined to the parameter learning rate, and the function is to determine the step size of the weight iteration.
(4) And 3.3) adding a classifier after the trained depth self-coding network model, inputting a training set sample containing a label, fine-tuning the weight by a gradient descent method, and completing classifier learning.
And (3) inputting a training set containing labels, fine-tuning weights by a gradient descent method, and completing classifier learning, wherein the classifier is Softmax, the loss function is cross entropy, the learning rate is 0.001, and the number of the loss functions is 100 Epochs.
The limiting effect of the parameter Epochs is to set the iteration number, 0.001 is the limiting of the parameter learning rate, and the effect is to determine the step size of the weight iteration.
(5) The model inputs the test set, outputs the diagnosis result and compares with the calibration label, if the condition is satisfied, that is, the diagnosis result is the same as the calibration label, the accuracy is high enough, and the corresponding deep self-coding network training model is saved.
And finally, directly diagnosing faults by using the online data to drive the stored model.
The stored model is a trained model, and the fault can be diagnosed only by inputting data to be diagnosed, so that the actual data of the emulsion pump is monitored on line, and the actual data is input into the trained model to diagnose the fault result of the emulsion pump.
Claims (8)
1. The emulsion pump fault diagnosis method based on the depth self-coding network is characterized by comprising the following steps of:
1) Monitoring real-time parameters in the running process of the emulsion pump;
2) Analyzing the real-time parameters, comparing the real-time parameters with a set threshold value, and if the real-time parameters are within the threshold value range, indicating that the emulsion pump is normal in operation, and continuing to monitor and analyze the real-time parameters; if the data exceeds the threshold value, the emulsion pump is abnormal in operation, and an alarm is triggered;
3) Inputting real-time parameters corresponding to the alarm time into a depth self-coding network model for training;
4) Acquiring actual data of the operation of the emulsion pump on line, and inputting the actual data into the trained depth self-coding network model to obtain an emulsion pump fault diagnosis result;
the specific process of the step 3) is as follows:
3.1 Pre-processing the monitored real-time parameters, preparing a data set, and according to 8: dividing a training set and a testing set, and calibrating labels on part of the training set;
3.2 A depth self-coding network model is constructed, the number of neurons of each hidden layer is determined, and the weight is randomly initialized;
3.3 Inputting a label-free training set, carrying out training iteration on the depth self-coding network model, and updating weight and bias by minimizing a mean square error loss value between an input layer and an output layer;
3.4 Adding a classifier after the depth self-coding network model is trained in the step 3.3), inputting a training set sample containing a label, fine-tuning the weight by a gradient descent method and completing classifier learning;
3.5 Inputting a test set into the depth self-coding network model after learning in the step 3.4), comparing the output diagnosis result with the calibration label, and if the diagnosis result is the same as the calibration label, storing the training model.
2. The method for diagnosing a fault in an emulsion pump based on a depth self-encoding network according to claim 1, wherein in said step 1), real-time parameters during operation of the emulsion pump are collected by installing a sensor group.
3. The emulsion pump fault diagnosis method based on the depth self-coding network according to claim 2, wherein the specific process of the step 2) is as follows:
2.1 Processing the real-time data acquired in the step 1) and storing the processed real-time data into a database;
2.2 Setting a threshold value when the running parameters of the emulsion pump are abnormal according to the empirical value;
2.3 Comparing the real-time parameter corresponding to a certain moment with a threshold value, if the acquired real-time parameter exceeds the threshold value, indicating that the emulsion pump is abnormal in operation, continuously judging that the emulsion pump is abnormal three times, and triggering an alarm.
4. The emulsion pump failure diagnosis method based on depth self-encoding network according to claim 3, wherein in the step 3.3), the mean square error loss value is calculated according to the following formula:
Wherein: l mse is a loss function, m is the number of samples contained in a batch, k represents the kth sample, j represents the number of dimensions of each sample, y k j represents the actual output value of the jth dimension in the kth sample, and x k j represents the predicted value of the jth dimension in the kth sample.
5. The method for diagnosing a fault in an emulsion pump based on a depth self-encoding network as recited in claim 4, wherein in said step 3.3), the iteration number is 100, and the learning rate of the parameter is 0.001.
6. The emulsion pump fault diagnosis method based on the depth self-encoding network according to claim 5, wherein in the step 3.4), the weight fine adjustment is to ensure that the cross entropy L cel is minimum, and the cross entropy is calculated according to the following formula:
Wherein: l cel is a cross entropy function, m is the number of samples contained in a batch, k represents the kth sample, j represents the number of dimensions of each sample, y k j represents the actual output value of the jth dimension in the kth sample, and x k j represents the predicted value of the jth dimension in the kth sample.
7. The method for diagnosing a fault in an emulsion pump based on a depth self-encoding network as claimed in claim 6, wherein in the step 3.4), the classifier is Softmax, the iteration number is 100, and the learning rate of the parameters is 0.001.
8. The emulsion pump failure diagnosis method based on depth self-encoding network according to claim 7, wherein in step 3.1), the preprocessing includes missing value and outlier processing; and the data in the training set and the testing set are subjected to normalization processing.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211723315.XA CN116028849B (en) | 2022-12-30 | 2022-12-30 | Emulsion pump fault diagnosis method based on depth self-coding network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211723315.XA CN116028849B (en) | 2022-12-30 | 2022-12-30 | Emulsion pump fault diagnosis method based on depth self-coding network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116028849A CN116028849A (en) | 2023-04-28 |
| CN116028849B true CN116028849B (en) | 2024-05-14 |
Family
ID=86077674
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211723315.XA Active CN116028849B (en) | 2022-12-30 | 2022-12-30 | Emulsion pump fault diagnosis method based on depth self-coding network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116028849B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116517921B (en) * | 2023-07-03 | 2023-11-10 | 成都飞机工业（集团）有限责任公司 | On-line detection method and system for aviation hydraulic oil vehicle state |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111290874A (en) * | 2020-02-29 | 2020-06-16 | 苏州浪潮智能科技有限公司 | Server fault monitoring and diagnosing method and device |
| CN112306036A (en) * | 2019-08-02 | 2021-02-02 | 中国石油化工股份有限公司 | Method for diagnosing operation fault of chemical process |
| CN113323699A (en) * | 2021-06-08 | 2021-08-31 | 中国矿业大学 | Method for accurately identifying fault source of hydraulic support system based on data driving |
| CN113378887A (en) * | 2021-05-14 | 2021-09-10 | 太原理工大学 | Emulsion pump fault grading diagnosis method |
| CN113567159A (en) * | 2021-06-17 | 2021-10-29 | 太原理工大学 | Scraper conveyor state monitoring and fault diagnosis method based on edge cloud cooperation |
| CN113762401A (en) * | 2021-09-13 | 2021-12-07 | 唯品会(广州)软件有限公司 | Self-adaptive classification task threshold adjusting method, device, equipment and storage medium |
| CN114048811A (en) * | 2021-11-10 | 2022-02-15 | 中国核动力研究设计院 | Wireless sensor node fault diagnosis method and device based on deep learning |
| CN114065613A (en) * | 2021-10-27 | 2022-02-18 | 中国华能集团清洁能源技术研究院有限公司 | Multi-working-condition process industrial fault detection and diagnosis method based on deep migration learning |
| CN114444582A (en) * | 2022-01-07 | 2022-05-06 | 太原理工大学 | Mechanical equipment fault diagnosis method based on convolutional neural network and Bayesian network |
| CN114741945A (en) * | 2022-02-17 | 2022-07-12 | 成都飞机工业（集团）有限责任公司 | Diagnosis method for online faults of aircraft engine |
| CN115219883A (en) * | 2022-08-19 | 2022-10-21 | 重庆交通职业学院 | Automatic detection system and detection method for electric automation equipment |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6917839B2 (en) * | 2000-06-09 | 2005-07-12 | Intellectual Assets Llc | Surveillance system and method having an operating mode partitioned fault classification model |
-
2022
- 2022-12-30 CN CN202211723315.XA patent/CN116028849B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112306036A (en) * | 2019-08-02 | 2021-02-02 | 中国石油化工股份有限公司 | Method for diagnosing operation fault of chemical process |
| CN111290874A (en) * | 2020-02-29 | 2020-06-16 | 苏州浪潮智能科技有限公司 | Server fault monitoring and diagnosing method and device |
| CN113378887A (en) * | 2021-05-14 | 2021-09-10 | 太原理工大学 | Emulsion pump fault grading diagnosis method |
| CN113323699A (en) * | 2021-06-08 | 2021-08-31 | 中国矿业大学 | Method for accurately identifying fault source of hydraulic support system based on data driving |
| CN113567159A (en) * | 2021-06-17 | 2021-10-29 | 太原理工大学 | Scraper conveyor state monitoring and fault diagnosis method based on edge cloud cooperation |
| CN113762401A (en) * | 2021-09-13 | 2021-12-07 | 唯品会(广州)软件有限公司 | Self-adaptive classification task threshold adjusting method, device, equipment and storage medium |
| CN114065613A (en) * | 2021-10-27 | 2022-02-18 | 中国华能集团清洁能源技术研究院有限公司 | Multi-working-condition process industrial fault detection and diagnosis method based on deep migration learning |
| CN114048811A (en) * | 2021-11-10 | 2022-02-15 | 中国核动力研究设计院 | Wireless sensor node fault diagnosis method and device based on deep learning |
| CN114444582A (en) * | 2022-01-07 | 2022-05-06 | 太原理工大学 | Mechanical equipment fault diagnosis method based on convolutional neural network and Bayesian network |
| CN114741945A (en) * | 2022-02-17 | 2022-07-12 | 成都飞机工业（集团）有限责任公司 | Diagnosis method for online faults of aircraft engine |
| CN115219883A (en) * | 2022-08-19 | 2022-10-21 | 重庆交通职业学院 | Automatic detection system and detection method for electric automation equipment |
Non-Patent Citations (2)
| Title |
|---|
| 应用深度自编码网络和XGBoost的风电机组发电机故障诊断;赵洪山等;《电力系统自动化》;20190110;第43卷(第1期);81-86 * |
| 数据驱动的乳化液泵站故障诊断;牛锐祥;《中国优秀硕士学位论文全文数据库 工程科技I辑》;20230215(第2期);B021-472 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116028849A (en) | 2023-04-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110414155B (en) | Fan component temperature abnormity detection and alarm method with single measuring point | |
| CN111460728B (en) | Method and device for predicting residual life of industrial equipment, storage medium and equipment | |
| CN113255848B (en) | Water turbine cavitation sound signal identification method based on big data learning | |
| CN110738274A (en) | nuclear power device fault diagnosis method based on data driving | |
| CN110348150B (en) | Fault detection method based on correlation probability model | |
| CN113188794B (en) | Gearbox fault diagnosis method and device based on improved PSO-BP neural network | |
| CN113657454B (en) | Nuclear power rotating machinery state monitoring method based on autoregressive BiGRU | |
| CN112132394B (en) | Power plant circulating water pump predictive state evaluation method and system | |
| CN116028849B (en) | Emulsion pump fault diagnosis method based on depth self-coding network | |
| CN113343591B (en) | Product key part life end-to-end prediction method based on self-attention network | |
| CN116383636A (en) | Coal mill fault early warning method based on PCA and LSTM fusion algorithm | |
| CN115099260A (en) | Online monitoring mechanical fault real-time diagnosis method for double-screw oil transfer pump | |
| CN114444582A (en) | Mechanical equipment fault diagnosis method based on convolutional neural network and Bayesian network | |
| CN115409131A (en) | Production line abnormity detection method based on SPC process control system | |
| CN112036087A (en) | Multi-strategy fused nuclear power key equipment fault diagnosis method and system | |
| CN116383722A (en) | Fracturing measure process monitoring method based on gate control circulation unit neural network | |
| CN116109039A (en) | Data-driven anomaly detection and early warning system | |
| CN108416386A (en) | A kind of method and system judged extremely for Hydropower Unit bearing temperature | |
| CN110779988A (en) | Bolt life prediction method based on deep learning | |
| CN117556347A (en) | Power equipment fault prediction and health management method based on industrial big data | |
| CN117768235A (en) | Real-time flow monitoring alarm system based on Internet of things | |
| CN115659249B (en) | Abnormality detection method for intelligent station measurement control system | |
| CN116793666A (en) | Wind turbine generator system gearbox fault diagnosis method based on LSTM-MLP-LSGAN model | |
| CN116226679A (en) | Wind turbine generator gearbox abnormality detection method considering similarity of running states of multiple units | |
| CN113551904A (en) | Gear box multi-type concurrent fault diagnosis method based on hierarchical machine 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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |