CN112464563A - Data mining method for steam turbine fault diagnosis - Google Patents
Data mining method for steam turbine fault diagnosis Download PDFInfo
- Publication number
- CN112464563A CN112464563A CN202011357884.8A CN202011357884A CN112464563A CN 112464563 A CN112464563 A CN 112464563A CN 202011357884 A CN202011357884 A CN 202011357884A CN 112464563 A CN112464563 A CN 112464563A
- Authority
- CN
- China
- Prior art keywords
- model
- steam turbine
- parameters
- data
- output
- 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
- 238000003745 diagnosis Methods 0.000 title claims abstract description 22
- 238000000034 method Methods 0.000 title claims abstract description 22
- 238000007418 data mining Methods 0.000 title claims abstract description 19
- 238000012545 processing Methods 0.000 claims abstract description 7
- 238000010276 construction Methods 0.000 claims abstract description 4
- 238000007781 pre-processing Methods 0.000 claims abstract description 4
- 238000012795 verification Methods 0.000 claims abstract description 4
- 210000002569 neuron Anatomy 0.000 claims description 16
- 230000006870 function Effects 0.000 claims description 9
- 238000012549 training Methods 0.000 claims description 9
- 238000012360 testing method Methods 0.000 claims description 8
- 230000009286 beneficial effect Effects 0.000 claims description 5
- 230000004913 activation Effects 0.000 claims description 3
- 238000012544 monitoring process Methods 0.000 claims description 3
- 238000013135 deep learning Methods 0.000 abstract description 4
- 238000012423 maintenance Methods 0.000 description 9
- 238000001514 detection method Methods 0.000 description 5
- 230000008901 benefit Effects 0.000 description 4
- 230000008439 repair process Effects 0.000 description 4
- 210000004027 cell Anatomy 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 238000012706 support-vector machine Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 230000006386 memory function Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000035945 sensitivity Effects 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000013145 classification model Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 210000002364 input neuron Anatomy 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 210000004205 output neuron Anatomy 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000000638 stimulation Effects 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- 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
- G06N3/048—Activation functions
-
- 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
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/20—Administration of product repair or maintenance
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2111/00—Details relating to CAD techniques
- G06F2111/10—Numerical modelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2113/00—Details relating to the application field
- G06F2113/28—Fuselage, exterior or interior
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/08—Thermal analysis or thermal optimisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/14—Force analysis or force optimisation, e.g. static or dynamic forces
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Business, Economics & Management (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Software Systems (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Water Supply & Treatment (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Primary Health Care (AREA)
- Geometry (AREA)
- Public Health (AREA)
- Computer Hardware Design (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Control Of Turbines (AREA)
Abstract
The invention designs a data mining method for steam turbine fault diagnosis, which is used for carrying out nonlinear processing on steam turbine operating parameters and comprises four steps of data preprocessing, model construction, model parameter selection and model verification. The method carries out data mining by means of the deep learning network, is used for fault identification and diagnosis work, realizes higher identification and diagnosis precision, and has extremely high popularization and application values.
Description
Technical Field
The application belongs to the technical field of thermal power equipment fault detection, and particularly relates to a data mining method for steam turbine fault diagnosis.
Background
The power industry is the foundation of economic development, and thermal power is the main force of development of the power industry in China. To improve efficiency, the equipment of thermal power plants is becoming larger and more complex. The widespread use of large-scale equipment has produced enormous economic benefits, but has also raised a number of problems. For example, due to lack of human and information resources, the diagnosis and repair of malfunctioning equipment often cannot be performed in a timely manner. Steam turbines, which are the core devices of steam power plants, are complex in structure and time-consuming to diagnose and repair faults. Moreover, the engineers of a typical thermal power plant can only handle routine or simple maintenance tasks, and are highly dependent on technical support of original equipment manufacturers to perform complex fault diagnosis and maintenance. In such maintenance cases, besides the expensive testing costs, the equipment is also subject to long downtimes and risks.
To reduce maintenance costs and the risk of significant experimentation, it is critical to introduce electronic maintenance in the steam turbine, which helps identify the root cause of component failure, reduces production system failures, eliminates expensive unplanned shutdown maintenance, and improves production efficiency.
In an electronic maintenance system, an intelligent fault detection system is important for identifying faults, a data mining technology is the core of the intelligent fault detection system, and the performance of the intelligent fault detection system can be greatly improved through data mining. Applying these techniques to fault detection may eliminate additional tests or experimentation that typically involve high costs and high risks. In the prior literature or patent, the data mining techniques commonly used are: artificial neural networks, fuzzy logic systems, support vector machine classification, and the like. For example, some literature utilizes Support Vector Machines (SVMs) and dimension reduction schemes to build classification models to identify steam turbine faults in thermal power plants.
Although relevant documents research fault identification and diagnosis of steam turbines, most of the existing research aims at traditional feature identification and classification, the verified data set is very limited, and the expandability of the model is often questioned. In addition, given the diversity of steam turbine faults, it is difficult for a small data set to generalize over varying fault types and fault manifestations. With the development of deep learning networks, big data-based fault identification and diagnosis are now possible.
Disclosure of Invention
The technical problem to be solved by the invention is to provide a data mining method for steam turbine fault diagnosis, which is used for fault identification and diagnosis work by means of data mining through a deep learning network.
In order to solve the problems, the technical scheme adopted by the invention is as follows:
a data mining method for steam turbine fault diagnosis for non-linear processing of steam turbine operating parameters, comprising the steps of:
s1, preprocessing data: acquiring steam turbine operating parameters and processing the parameters into required standard requirements;
s2, model construction: for non-linearly analyzing steam turbine key parameters and making decisions based on current and historical information to diagnose the fault type;
s3, model parameter selection: performing parameter tuning on parameters required by the model to find an optimal parameter combination;
s4, model verification: acquiring original data, dividing the original data into training data and testing data, determining model parameters through the training data, and finishing model training; model performance is evaluated by testing the data and measuring the error between the model output and the expected output to minimize the model output error.
The technical scheme of the invention is further improved as follows: in step S2, the model structure adopted for the non-linear processing of the steam turbine operating parameters is an LSTM model, the LSTM model includes an input layer, a plurality of hidden layers, and an output layer, and data is input from the input layer, processed by the hidden layers, and then the fault type is output by the output layer;
an input layer: the size corresponds to the number of parameters monitoring the operation of the steam turbine;
hiding the layer: the method comprises the following steps of including one or more LSTM neurons, wherein the LSTM neurons are used for extracting nonlinear relations between input parameters and between past states so as to construct a group of parameter combinations beneficial to classification; each LSTM neuron is provided with weight and deviation, and the contribution degree of each LSTM neuron to the output is determined;
an output layer: and a decision layer of the model for outputting the fault type of the steam turbine.
The technical scheme of the invention is further improved as follows: the number of LSTM neurons in the hidden layer is between the number of input layer and output layer neurons, and the specific values are optimized by the model parameter selection process in step S3.
The technical scheme of the invention is further improved as follows: in step S3, the model parameters of the LSTM model include the degree of regularization, the number of hidden layers, the number of cells in each hidden layer, the type of activation function, the loss rate (droop rate), the learning rate, the batch size, and the number of runs, and the model parameters are optimized by applying random search.
The technical scheme of the invention is further improved as follows: the random search was performed by using a Hyperopt toolkit, setting the number of searches to 50, the tolerance to 10, and the minimum gain to 10-3。
The technical scheme of the invention is further improved as follows: in step S4, the error between the model output and the expected output is measured using a loss function.
The technical scheme of the invention is further improved as follows: the loss function is cross-entropy.
Due to the adoption of the technical scheme, the invention has the beneficial effects that:
the LSTM model of the invention is used for non-linearly analyzing key parameters of the steam turbine, such as the temperature and the pressure of primary steam, the temperature and the pressure of reheated steam, the vibration of a steam turbine generator, the rotating speed of turbine blades and the like. In addition, the LSTM model has a memory function, and can selectively memorize historical information for comparison, thereby facilitating current decision-making. And finally, a decision layer with full links makes decisions according to the current and historical information to diagnose the fault type. The engineer can perform targeted maintenance or repair of the steam turbine based on the decision.
Drawings
FIG. 1 is a schematic diagram of an LSTM model according to an embodiment of the present invention;
FIG. 2 is a graph of ROC curves for an embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to examples.
The invention discloses a data mining method for steam turbine fault diagnosis, which is described in detail by way of example below with reference to fig. 1-2.
Data pre-processing
First, the input and output variables of the model are determined based on the performance of the steam turbine and the associated influencing parameters. And secondly, clearing data and converting the data to improve the data quality, thereby being beneficial to improving the accuracy and efficiency of the data mining process. In this step, it is necessary to attempt to remove the missing data, outliers and correct inconsistencies in the data. Furthermore, different monitoring attributes have different scales. We need to normalize all attribute values to the same scale to avoid the impact of the different scales on the classification results. In general, all attribute values can be normalized to the interval [0,1] using a min-max normalization equation.
Model construction
The proposed model structure consists of one input layer, a plurality of hidden layers and one output layer. As shown in FIG. 1, the input level is the left-most column, and the input level size corresponds to the number of parameters that monitor the operation of the steam turbine. The hidden layer is an intermediate layer and is composed of one or more LSTM units, and neurons of the hidden layer are used for extracting nonlinear relations between input parameters and between past states, so that a new group of parameter combinations beneficial to classification is constructed. Each neuron has a weight and bias that determines its degree of contribution to the output. The number of hidden layer elements is determined empirically. The optimal size of the hidden neuron is typically between the input and output sizes. Furthermore, good performance can be expected using the average of the input and output neurons.
The LSTM processes the input signal using three gate control units (input gate, output gate, forgetting gate) and a memory unit to accomplish the extraction of the nonlinear features.
The forgetting gate mainly controls the capacity of the memory unit, and the processing of the input signals by the three gate control units is shown as the following formula:
ft=σ(Wxfxt+Whfht-1+bf) (1)
it=σ(Wxixt+Whiht-1+bi) (2)
ot=σ(Wxoxt+Whoht-1+bo) (3)
the memory unit of the LSTM calculates the storage information according to the formula (4) and the formula (5):
finally, the output of the LSTM is calculated by the following equation (6),
where W and b represent the weight and offset of the node, xtIs the input of the LSTM cell, ht-1Is the output at the previous moment, the operator ° is the Hadamard product, and the operation σ () refers to sigmoid calculation.
The output layer is the decision layer of the model, and is positioned in the column at the rightmost side and used for outputting the fault type of the steam turbine. The number of units of this layer thus corresponds to the number of fault types of the steam turbine. If the number of units is 2, it corresponds to two categories, i.e. the decision is made whether there is a fault or whether there is a certain fault. If the number of units is >2, then the multi-classification is corresponded, i.e. the probability that the model output is the plurality of fault types corresponding to the current state. The fault type with the highest corresponding probability is the most possible state of the current input.
The model uses cross-entropy as a loss function to measure the error between the model output and the expected output. The weights and biases for each layer of neurons are then updated using gradient descent and back propagation iterations to minimize the output error. The calculation formula of the loss function is shown in the following formula (7):
model parameter selection
The selection of model parameters is an important step for constructing a deep learning network, and directly determines the performance of the model. In the present model, the parameters to be selected are: the degree of regularization, the number of hidden layers, the number of cells per hidden layer, the type of activation function, the loss rate (droop rate), the learning rate, the batch size, the number of runs. These parameters, collectively referred to as hyper-parameters, must be adjusted so that the model can achieve optimal performance.
Currently, there are many ways to select the hyper-parameters. The most widely used strategies are grid search and manual search, which search one by one for a user's empirically specified set of hyper-parameters. These strategies suffer from cursing of dimensions, the number of searches growing exponentially with increasing parameters. Random search is a variation of grid search that performs random searches in a hyper-parameter set, rather than search one by one. Random searching has proven its superiority, especially if only a small number of hyper-parameters affect the performance of the model.
Therefore, in this patent, we apply a random search on the required parameters to find the optimal combination of parameters. The search is supported by the hyperopt toolkit. The number of searches was set to 50, the tolerance was set to 10, and the minimum gain was set to 10-3. The search space is shown in the following table.
Model validation
The data was first divided into training data and test data at a ratio of 3: 2. and determining model parameters by using the training data to complete model training. And test data is used to evaluate model performance. In order to fully verify the performance of the model, three standards of precision, kappa value and ROC curve can be adopted to evaluate the performance of the model.
In precision, the precision of the model can reach more than 95%.
kappa number: kappa stability says how well two objects fit together, and the higher the kappa value, the higher the mildness, and the higher the accuracy of the side-printed model.
ROC curve: as shown in fig. 2, a receiver operating characteristic curve (ROC curve) is also called a sensitivity curve. The reason for this is that each point on the curve reflects the same sensitivity, and they all respond to the same signal stimulus, but only the results obtained under two different criteria. The working characteristic curve of the subject is a graph formed by taking the False positive probability (False positive rate) as the horizontal axis and the True positive probability (True positive rate) as the vertical axis, and is drawn by different results of the subject under specific stimulation conditions due to different judgment standards.
In order to embody the advantages of the proposed model in the aspects of fault diagnosis accuracy and model expandability, the proposed deep learning model can be compared with the traditional SVM-based model, and the advantages of the invention are obvious compared with the advantages of the SVM model.
The invention provides a fault diagnosis system of a steam turbine based on an LSTM model. In this model, the LSTM model is used to nonlinearly analyze steam turbine key parameters such as the temperature and pressure of primary steam, the temperature and pressure of reheat steam, the vibration of the turbine generator, and the rotational speed of turbine blades. In addition, the LSTM model has a memory function, and can selectively memorize historical information for comparison, thereby facilitating current decision-making. And finally, making a decision by a fully-linked decision layer according to the current and historical information, and diagnosing the fault type. The engineer can perform targeted maintenance or repair of the steam turbine based on the decision. The scheme provided by the patent utilizes real data provided by conspiracy power companies, and the evaluation and verification of the validity of the proposed model are worthy of affirmation and dependence.
Claims (7)
1. A data mining method for steam turbine fault diagnosis for non-linear processing of steam turbine operating parameters, comprising the steps of:
s1, preprocessing data: acquiring steam turbine operating parameters and processing the parameters into required standard requirements;
s2, model construction: for non-linearly analyzing steam turbine key parameters and making decisions based on current and historical information to diagnose the fault type;
s3, model parameter selection: performing parameter tuning on parameters required by the model to find an optimal parameter combination;
s4, model verification: acquiring original data, dividing the original data into training data and testing data, determining model parameters through the training data, and finishing model training; model performance is evaluated by testing the data and measuring the error between the model output and the expected output to minimize the model output error.
2. The data mining method for steam turbine fault diagnosis according to claim 1, wherein in step S2, the steam turbine operating parameters are non-linearly processed by using an LSTM model having an input layer, a plurality of hidden layers and an output layer, and the data is input from the input layer, processed by the hidden layers, and the fault type is output from the output layer;
an input layer: the size corresponds to the number of parameters monitoring the operation of the steam turbine;
hiding the layer: the method comprises the following steps of including one or more LSTM neurons, wherein the LSTM neurons are used for extracting nonlinear relations between input parameters and between past states so as to construct a group of parameter combinations beneficial to classification; each LSTM neuron is provided with weight and deviation, and the contribution degree of each LSTM neuron to the output is determined;
an output layer: and a decision layer of the model for outputting the fault type of the steam turbine.
3. The method of data mining for steam turbine fault diagnosis of claim 2, wherein the number of LSTM neurons in the hidden layer is between the number of input layer and output layer neurons.
4. The data mining method for steam turbine fault diagnosis according to claim 1, wherein in step S3, the model parameters of the LSTM model include regularization degree, number of hidden layers, number of units per hidden layer, activation function type, loss rate, learning rate, batch size, and operation times, and the model parameters are optimized by applying random search.
5. The method of claim 4, wherein the random search is performed using a hyper pt kit, and wherein the number of searches is set to 50, the tolerance is set to 10, and the minimum gain is set to 10-3。
6. The data mining method for steam turbine fault diagnosis according to claim 5, wherein in step S4, the error between the model output and the expected output is measured by using a loss function.
7. The data mining method for steam turbine fault diagnosis according to claim 6, wherein the loss function is cross-error.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011357884.8A CN112464563A (en) | 2020-11-27 | 2020-11-27 | Data mining method for steam turbine fault diagnosis |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011357884.8A CN112464563A (en) | 2020-11-27 | 2020-11-27 | Data mining method for steam turbine fault diagnosis |
Publications (1)
Publication Number | Publication Date |
---|---|
CN112464563A true CN112464563A (en) | 2021-03-09 |
Family
ID=74809537
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202011357884.8A Pending CN112464563A (en) | 2020-11-27 | 2020-11-27 | Data mining method for steam turbine fault diagnosis |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112464563A (en) |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109555566A (en) * | 2018-12-20 | 2019-04-02 | 西安交通大学 | A kind of turbine rotor method for diagnosing faults based on LSTM |
CN109931678A (en) * | 2019-03-13 | 2019-06-25 | 中国计量大学 | Air-conditioning fault diagnosis method based on deep learning LSTM |
CN110110862A (en) * | 2019-05-10 | 2019-08-09 | 电子科技大学 | A kind of hyperparameter optimization method based on adaptability model |
CN110689171A (en) * | 2019-09-05 | 2020-01-14 | 哈尔滨工程大学 | Turbine health state prediction method based on E-LSTM |
US20200284265A1 (en) * | 2019-03-06 | 2020-09-10 | General Electric Company | Application of Machine Learning to Process High-Frequency Sensor Signals of a Turbine Engine |
CN111783487A (en) * | 2020-06-24 | 2020-10-16 | 中国工商银行股份有限公司 | Fault early warning method and device for card reader equipment |
-
2020
- 2020-11-27 CN CN202011357884.8A patent/CN112464563A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109555566A (en) * | 2018-12-20 | 2019-04-02 | 西安交通大学 | A kind of turbine rotor method for diagnosing faults based on LSTM |
US20200284265A1 (en) * | 2019-03-06 | 2020-09-10 | General Electric Company | Application of Machine Learning to Process High-Frequency Sensor Signals of a Turbine Engine |
CN109931678A (en) * | 2019-03-13 | 2019-06-25 | 中国计量大学 | Air-conditioning fault diagnosis method based on deep learning LSTM |
CN110110862A (en) * | 2019-05-10 | 2019-08-09 | 电子科技大学 | A kind of hyperparameter optimization method based on adaptability model |
CN110689171A (en) * | 2019-09-05 | 2020-01-14 | 哈尔滨工程大学 | Turbine health state prediction method based on E-LSTM |
CN111783487A (en) * | 2020-06-24 | 2020-10-16 | 中国工商银行股份有限公司 | Fault early warning method and device for card reader equipment |
Non-Patent Citations (1)
Title |
---|
陈伟: "深度学习在滚动轴承故障诊断中的应用研究", 《中国优秀硕士论文全文数据库》 * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Zhang et al. | Analog circuit incipient fault diagnosis method using DBN based features extraction | |
CN111237134B (en) | Offshore double-fed wind driven generator fault diagnosis method based on GRA-LSTM-stacking model | |
CN110738274A (en) | nuclear power device fault diagnosis method based on data driving | |
CN111539553B (en) | Wind turbine generator fault early warning method based on SVR algorithm and off-peak degree | |
US11840998B2 (en) | Hydraulic turbine cavitation acoustic signal identification method based on big data machine learning | |
US20210334658A1 (en) | Method for performing clustering on power system operation modes based on sparse autoencoder | |
CN105628425A (en) | Rotation machinery early stage fault diagnosis method based on heredity annealing optimization multi-core support vector machine | |
US11860608B2 (en) | Industrial equipment operation, maintenance and optimization method and system based on complex network model | |
CN110647911A (en) | Bearing fault diagnosis method based on principal component analysis and deep belief network | |
CN110991737A (en) | Ultra-short-term wind power prediction method based on deep belief network | |
CN115329986A (en) | Wind turbine generator anomaly detection and positioning method based on interpretable graph neural network | |
CN115496108A (en) | Fault monitoring method and system based on manifold learning and big data analysis | |
Qin et al. | High-efficiency generative adversarial network model for chemical process fault diagnosis | |
CN112085108A (en) | Photovoltaic power station fault diagnosis algorithm based on automatic encoder and K-means clustering | |
CN115163424A (en) | Wind turbine generator gearbox oil temperature fault detection method and system based on neural network | |
CN115146718A (en) | Depth representation-based wind turbine generator anomaly detection method | |
CN115238573A (en) | Hydroelectric generating set performance degradation trend prediction method and system considering working condition parameters | |
CN117408162A (en) | Power grid fault control method based on digital twin | |
CN117520809A (en) | Transformer fault diagnosis method based on EEMD-KPCA-CNN-BiLSTM | |
Li et al. | A novel fault early warning method for mechanical equipment based on improved MSET and CCPR | |
CN116578833A (en) | IGBT module aging fault diagnosis system based on optimized random forest model | |
CN112163474A (en) | Intelligent gearbox diagnosis method based on model fusion | |
CN112464563A (en) | Data mining method for steam turbine fault diagnosis | |
CN116317937A (en) | Distributed photovoltaic power station operation fault diagnosis method | |
CN114046816B (en) | Sensor signal fault diagnosis method based on lightweight gradient lifting decision tree |
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: 20210309 |