CN112418306B - Gas turbine compressor fault early warning method based on LSTM-SVM - Google Patents
Gas turbine compressor fault early warning method based on LSTM-SVM Download PDFInfo
- Publication number
- CN112418306B CN112418306B CN202011309245.4A CN202011309245A CN112418306B CN 112418306 B CN112418306 B CN 112418306B CN 202011309245 A CN202011309245 A CN 202011309245A CN 112418306 B CN112418306 B CN 112418306B
- Authority
- CN
- China
- Prior art keywords
- compressor
- fault
- alarm
- svm
- gas turbine
- 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
- 238000000034 method Methods 0.000 title claims abstract description 25
- 208000024891 symptom Diseases 0.000 claims abstract description 18
- 238000012544 monitoring process Methods 0.000 claims abstract description 15
- 238000012549 training Methods 0.000 claims abstract description 13
- 238000012545 processing Methods 0.000 claims abstract description 8
- 238000013135 deep learning Methods 0.000 claims abstract description 4
- 239000013598 vector Substances 0.000 claims description 16
- 238000013145 classification model Methods 0.000 claims description 13
- 238000012360 testing method Methods 0.000 claims description 6
- 230000006870 function Effects 0.000 claims description 2
- 238000011160 research Methods 0.000 description 6
- 238000005299 abrasion Methods 0.000 description 3
- 230000008859 change Effects 0.000 description 3
- 238000005260 corrosion Methods 0.000 description 3
- 230000007797 corrosion Effects 0.000 description 3
- 238000003745 diagnosis Methods 0.000 description 3
- 230000007423 decrease Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000010248 power generation Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 239000000428 dust Substances 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003746 surface roughness Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- 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
- G06N3/044—Recurrent networks, e.g. Hopfield 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
- G06N3/045—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
- G06N3/08—Learning methods
-
- 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
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
Abstract
The invention relates to a gas turbine compressor fault early warning method based on an LSTM-SVM, which comprises the following steps: and establishing a fault knowledge base of the gas turbine compressor, and excavating the relation between the fault type and the fault symptom of the compressor. Determining the symptom parameter types of the air compressor; training normal data of the sign parameters of the gas compressor by using a deep learning LSTM algorithm, and establishing a good prediction model; and monitoring a data curve output by the model, processing alarm data through positive and negative deviation degrees, performing fault classification as an input parameter of the SVM, and determining the fault type of the air compressor. The invention can quickly find the fault trend of the air compressor through the early warning information, and provides important decision support for early fault early warning of the air compressor.
Description
Technical Field
The invention relates to the field of equipment fault early warning of thermal power plants, in particular to a gas turbine compressor fault early warning method based on an LSTM-SVM.
Background
In the power industry, the gas turbine generator set becomes one of the mainstream power generation modes at present due to the advantages of quick start and stop, high thermal efficiency, less pollution and the like, but the key core technology of the gas turbine in China still depends on abroad. In order to change this current situation, the country has increased policy support for gas turbines, driving the gas turbine industry to develop rapidly. The compressor is one of the important components of a gas turbine, and its operating state directly affects the safety and reliability of the gas turbine. However, the compressor is operated in a high-rotation-speed and high-temperature environment for a long time, and faults such as blade fouling, abrasion corrosion and the like often occur. If the failure trend of the gas turbine compressor can be found early, the failures are repaired and protected in advance, and the risk of unstable operation or unplanned shutdown of the gas turbine caused by the failure of the compressor is reduced. Therefore, the early warning of the fault of the gas compressor has important significance for the stable operation of the gas turbine.
At present, the research on the gas compressor is mainly focused on the influence of the fault type of the gas compressor on the performance of the gas turbine unit, the research on the early warning of the fault of the gas compressor is less, and the early warning of the fault of power plant equipment is provided with a plurality of methods and results. The method for early warning the faults of the fans of the power plant by utilizing the multi-element state estimation and the deviation degree is adopted, a parameter model of the fans is established through the multi-element state estimation, the deviation degree monitoring model is utilized to output results, the fault development process is captured, and early warning is realized. The power generation equipment fault early warning system based on the similarity principle is used for establishing a model matrix through mathematical analysis on historical data, comparing an actual value with an estimated value output by the model, and alarming after exceeding a preset deviation. The dynamic early warning model of the steam turbine generator set is established by utilizing the gray theory and the similarity principle, and the abnormal state of the equipment can be found in time by adopting the hypersphere similarity analysis technology, so that a new method is provided for equipment fault early warning. Although the method is not applied to the field of gas turbines, the method has certain reference value, the structure of the gas compressor is complex, the correlation of the corresponding characteristic parameter changes among various fault types is strong, and the difficulty of early fault early warning of the gas turbine gas compressor is increased.
Disclosure of Invention
The invention aims to solve the early fault early warning problem of a gas turbine compressor.
In order to achieve the above purpose, the technical scheme of the invention provides a gas turbine compressor fault early warning method based on an LSTM-SVM, which is characterized in that an LSTM prediction model is built by using normal historical data of a gas compressor, and the output prediction error of the LSTM prediction model is within 0.5%. And substituting 7 characteristic parameters of the inlet flow, the efficiency, the inlet temperature, the outlet temperature, the inlet pressure, the outlet pressure and the class group pressure ratio of the air compressor into the LSTM prediction model to obtain a residual curve of a predicted value and an actual value, setting an early warning threshold value and giving out an overrun alarm. Finally, the alarm information of the alarm point is subjected to data processing through positive and negative deviation degrees and is used as an input parameter of an SVM to carry out fault classification, so that the fault type of the gas compressor is determined, and early fault early warning of the gas turbine gas compressor is realized, and the method specifically comprises the following steps:
s1, establishing a fault knowledge base of a gas turbine compressor, wherein the fault knowledge base comprises a relation between a compressor fault type and a fault compressor symptom parameter; the changes of the fault type of the air compressor and the sign parameters of the related air compressor are simply and intuitively found through a fault knowledge base, the changes are analyzed and inferred, the relation between the fault type of the air compressor and the sign parameters of the fault air compressor is excavated, and the influence on the sign parameters of the air compressor is determined to be larger when the fault of the air compressor occurs;
s2, determining the types of the compressor symptom parameters, training the determined normal data of the compressor symptom parameters by using a deep learning LSTM algorithm, and establishing a prediction model, so that the residual error range between the predicted value and the actual value output by the prediction model is controlled within 0.5%;
s3, substituting the actual value of the compressor symptom parameter into the prediction model established in the step S2 for monitoring, alarming when the residual error between the prediction value and the actual value output by the prediction model exceeds 0.9%, and processing positive and negative deviation degree of alarm information, wherein the alarm information is the data value of the compressor symptom parameter of an alarm point;
s4, classifying positive and negative deviation of the alarm information as input parameters of the SVM classification model, determining the fault type of the air compressor, and realizing fault early warning of the air compressor.
Preferably, the step S3 alarms by monitoring residual curves of characteristic parameters of the compressor, wherein the characteristic parameters of the compressor are air inlet flow, efficiency, inlet temperature, outlet temperature, inlet pressure, outlet pressure and stage group pressure ratio of the compressor, and are respectively marked as x 1 ,x 2 ,…,x 7 ；
Setting an alarm limit, wherein the alarm limit comprises a high alarm limit and a low alarm limit, and a residual error between a predicted value and an actual value output by a prediction model exceeds the limit for alarm and uploads alarm information;
the positive and negative deviation degree processing is carried out on the alarm information by using the following formula (1):
in formula (1), i=1, 2, …,7; m is m i The deviation degree of the alarm point of the characteristic parameter of the ith air compressor is; x is x i real Is the actual value of the characteristic parameter of the ith compressor; x is x i prediction Is the predicted value of the characteristic parameter of the ith compressor;
defining the exceeding high alarm limit as positive deviation degree, namely residual error monitoring exceeding upper limit; the exceeding of the low alarm limit is defined as negative deviation, namely, the residual error monitoring exceeds the lower limit; if no alarm information exists, defining as 0;
the LSTM prediction model is used for monitoring 7 characteristic parameters of the air compressors, so that a vector group M is formed by the deviation degree of alarm points corresponding to the 7 characteristic parameters of the air compressors k ，M k ＝{m 1 ,m 2 ,…,m 7 And k is the number of samples of the alarm information vector group and is used as the input of the SVM classification model, wherein for a training sample k, the number of samples is 20, and the number of test samples k is 80.
Preferably, in step S4, N sets of alarm vectors are obtained through a residual curve output by the prediction model and are used as input of an SVM classification model, and the SVM is used for classifying, so as to determine the fault type of the compressor, wherein the SVM classification model training set T is shown in the following formula (2):
T＝{(M 1 ,y 1 ),(M 2 ,y 2 ),…,(M k ,y k ),…,(M N ,y N )} (2)
in the formula (2), M k The k-th alarm deviation vector group; y is k For marking the fault type of the compressor, y k ＝{-1，1}；
The objective function and constraint conditions are represented by the following formulas (3), (4):
s.t.y k (wm k +b)≥1,k＝1,2,...,N (4)
in the formulas (3) and (4), w is a hyperplane normal vector dividing the fault type of the compressor, and b is a hyperplane offset term dividing the fault type of the compressor.
Preferably, in step S4, if there are 4 fault categories in the gas turbine compressor fault early warning, in the SVM classification model, a method of one category for the rest is adopted to solve the problem of classifying the support vector machine 4, that is, one category is firstly separated to determine the category, the other three categories are separated into another category to form a sample set, after four times of the above operations are completed, 4 sample sets are obtained, thereby forming 4 classification problems, each sample set has a classification result, and the largest value is selected as the final classification result.
Aiming at the early-stage fault early-warning problem of the gas turbine compressor, the invention provides a gas turbine compressor fault early-warning method based on an LSTM-SVM. An LSTM prediction model is established through historical data of the gas compressor, characteristic parameters of the gas compressor are substituted into the prediction model, a residual curve between a predicted value and an actual value output by the model is monitored, an alarm threshold is set, overrun alarm is carried out, alarm information of an alarm point is subjected to data processing through positive and negative deviation degrees and is used as an input parameter of a support vector machine, the fault type of the gas compressor is identified, and therefore gas turbine gas compressor fault early warning is achieved.
The invention can quickly and accurately find the early failure trend of the gas turbine compressor, maintain and protect the gas turbine compressor in time, reduce the economic loss caused by the failure of the gas turbine compressor and ensure that the gas turbine can reliably and safely run. Compared with the prior art, the invention has the following specific advantages:
(1) The prediction model established by the deep learning LSTM algorithm has a good prediction effect on time samples, is very suitable for the compressor operation data sampled in time, and has small prediction error.
(2) The research results of scholars literature and the analysis of the fault cases of the gas compressor are utilized to construct a gas compressor fault knowledge base, so that the relationship between the fault type and the symptom parameters of the gas compressor can be intuitively and simply reflected.
(3) The SVM is suitable for training small samples, has high classification speed, determines a final result by a few support vectors, can help us to grasp key samples, and has good robustness.
Drawings
FIG. 1 is a block diagram of a gas turbine compressor fault warning method of an LSTM-SVM of the present invention;
FIG. 2 is an overall flow of gas turbine compressor fault warning;
FIG. 3 is an actual and predicted value of the gas turbine compressor air intake;
FIG. 4 is a graph of predicted and actual compressor air intake;
FIG. 5 is a graph of compressor air intake residual percentage;
FIG. 6 is a predicted and actual compressor blade fouling intake;
FIG. 7 is a compressor blade fouling air inflow residual;
FIG. 8 is a predicted and actual blade fouling compressor efficiency value;
FIG. 9 is a blade fouling compressor efficiency residual;
FIG. 10 shows the LSTM-SVM training sample diagnosis;
FIG. 11 shows the LSTM-SVM test sample diagnosis results.
Detailed Description
The invention will be further illustrated with reference to specific examples. It is to be understood that these examples are illustrative of the present invention and are not intended to limit the scope of the present invention. Further, it is understood that various changes and modifications may be made by those skilled in the art after reading the teachings of the present invention, and such equivalents are intended to fall within the scope of the claims appended hereto.
As shown in FIG. 1, the gas turbine compressor fault early warning method is integrally an LSTM-SVM. And establishing a fault knowledge base of the gas turbine compressor according to expert knowledge, experience and compressor fault cases, wherein the knowledge base mainly comprises fault types and fault symptom parameters. The gas compressor fault knowledge base can simply and intuitively discover the change relation between the gas compressor fault type and related symptom parameters, so that 7 symptom parameters of gas compressor inlet flow, efficiency, inlet temperature, outlet temperature, inlet pressure, outlet pressure and stage group pressure ratio are determined, and then 7 characteristic parameters of gas compressor inlet flow, efficiency … stage group pressure ratio and the like are substituted into an LSTM prediction model to obtain a residual curve of a predicted value and an actual value, and an early warning threshold value is set and an overrun warning is performed. And finally, carrying out data processing on the alarm information of the alarm points through positive and negative deviation degrees, carrying out fault classification as input parameters of the SVM, and determining the fault type of the gas compressor, thereby realizing gas turbine gas compressor fault early warning. The specific flow chart is shown in fig. 2. The fault early warning method specifically comprises the following steps:
s1, firstly, determining a study object. The gas turbine PG9351FA of a certain power plant is taken as a research object, the atmospheric pressure of the performance parameter of a gas compressor is 101.3kpa, the output power is 255.6MW, the pressure ratio of the gas compressor is 15.4, and the like. The compressor frequently fails in the environment with high temperature and high rotation speed after long-term operation, and the common faults of the compressor are compressor blade fouling, blade abrasion corrosion, air inlet icing and compressor surge. Taking compressor blade fouling as an example. During the operation of the compressor, dust in the air can be sucked into the compressor, and the surface roughness of a rotor of the compressor can be increased along with long-time accumulation. Compressor fouling affects changes in compressor symptom parameters, which can reduce the operating efficiency of the gas turbine. Therefore, the construction of the gas compressor fault knowledge base can simply and intuitively discover the fault type of the gas compressor and the related sign parameter change relation thereof, and provides a basis for a rule reasoning fault diagnosis method. And constructing a fault knowledge base of the gas compressor according to the research results of expert scholars and the fault case analysis of the gas compressor, as shown in figure 3.
S2, selecting research data. The invention excavates the hidden relation between the fault type and the symptom parameter of the compressor according to the fault knowledge base of the compressor, selects 7 characteristic parameters of the inlet flow, the inlet temperature, the outlet temperature, the inlet pressure, the outlet pressure, the stage group pressure ratio and the compressor efficiency of the compressor as the state indexes reflecting the compressor of the gas turbine, and respectively marks as x 1 ,x 2 ,…,x 7 . And monitoring residual curves of the 7 characteristic parameter predicted values and the actual values, capturing early fault characteristics of the gas compressor, and providing basis for fault identification and early warning of the gas turbine gas compressor. By training the normal air inflow of the gas turbine compressor, the curve of the predicted value and the actual value output by the LSTM prediction model is observed, and the curve has good consistency, as shown in figure 4. Fig. 5 is a residual percentage curve of the predicted value and the actual value of the normal air inflow of the air compressor output by the model, and as seen from fig. 5, the residual percentage curve error of the predicted value and the actual value of the air inflow of the air compressor is within 0.5%, so that the predicted effect is very good.
S3, in order to capture the fault trend of the gas turbine compressor, early fault data of a section of compressor blade scale deposit air inflow are selected to be brought into an LSTM prediction model, and a predicted value and an actual value of the compressor blade scale deposit air inflow are obtained, as shown in FIG. 6, it can be seen that the actual value of the air inflow begins to deviate downwards when the compressor blade scales, and the performance of the compressor begins to decline. By monitoring the residual curve of the predicted value and the actual value of the air inflow of the air compressor, as shown in fig. 7, the fluctuation of the residual curve of the air inflow of the air compressor blade scale at the beginning is relatively uniform, the air is alarmed after exceeding the alarm limit to the 116 th point, the actual value of the alarm point is 616.48kg/s, the predicted value is 623.08kg/s, and the negative deviation is-1.05 according to the deviation formula (1) and the low alarm. For compressor blade fouling failure, the compressor efficiency also tends to decrease as shown in fig. 8. By monitoring the residual curve of the predicted and actual values of the compressor blade fouling efficiency, as shown in FIG. 9, the initial fluctuation of the residual of the compressor blade fouling efficiency is relatively normal, beyond the 148 th pointAnd (5) alarming. The actual value of the alarm point is 87.15%, and the predicted value is 87.92%. And obtaining negative deviation degree-0.87 according to the deviation degree formula (1) and the low alarm. The positive and negative deviation of other characteristic parameters of the compressor blade scale is obtained by the method in the same way, and an alarm vector group M is formed 1 { -1.05, -0.87,0,0.93,0, -0.97, -1.12}, as input parameters of the SVM classification model.
S4, aiming at the PG9351FA type gas turbine, the common faults of the gas compressor mainly comprise 4 types of blade fouling, blade abrasion corrosion, air inlet icing and surging, and the 4 types of faults are respectively marked as 1,2, 3 and 4. And acquiring an alarm deviation degree vector group from the LSTM prediction model as an input of the SVM classification model for classification recognition. Firstly, 5 alarm vector groups are selected for training for each fault type, 20 training samples are used in total, the training results are shown in fig. 10, and the accuracy of the classification results is high. And selecting test samples, wherein 20 alarm vector groups are selected for each fault type to test, and 80 test samples are taken in total, and the SVM classification result is shown in FIG. 11. From the SVM classification result in fig. 11, it can be seen that the accuracy reaches 98.7%.
An LSTM prediction model is established for the characteristic parameters of the gas turbine compressor, a residual curve of the predicted value and the actual value of the characteristic parameters is monitored, early failure trend of the gas turbine compressor is captured through fluctuation of residual to alarm, and positive and negative deviation is introduced to quantitatively process alarm information of an alarm point. Finally, the fault type of the gas turbine compressor is judged by using the SVM classification model, and a method reference is provided for early fault early warning work of the gas turbine compressor.
Claims (3)
1. The gas turbine compressor fault early warning method based on the LSTM-SVM is characterized by comprising the following steps of:
s1, establishing a fault knowledge base of a gas turbine compressor, wherein the fault knowledge base comprises a relation between a compressor fault type and a fault compressor symptom parameter; the changes of the fault type of the air compressor and the sign parameters of the related air compressor are simply and intuitively found through a fault knowledge base, the changes are analyzed and inferred, the relation between the fault type of the air compressor and the sign parameters of the fault air compressor is excavated, and the influence on the sign parameters of the air compressor is determined to be larger when the fault of the air compressor occurs;
s2, determining the types of the compressor symptom parameters, training the determined normal data of the compressor symptom parameters by using a deep learning LSTM algorithm, and establishing a prediction model, so that the residual error range between the predicted value and the actual value output by the prediction model is controlled within 0.5%;
s3, substituting the actual value of the compressor symptom parameter into the prediction model established in the step S2 for monitoring, alarming when the residual error between the prediction value and the actual value output by the prediction model exceeds 0.9%, and processing positive and negative deviation degree of alarm information, wherein the alarm information is the data value of the compressor symptom parameter of an alarm point, and when alarming is carried out by monitoring a residual error curve of the compressor characteristic parameter, the compressor characteristic parameter is the air inlet flow, efficiency, inlet temperature, outlet temperature, inlet pressure, outlet pressure and stage group pressure ratio of the compressor, and is respectively marked as x 1 ,x 2 ,…,x 7 ；
Setting an alarm limit, wherein the alarm limit comprises a high alarm limit and a low alarm limit, and a residual error between a predicted value and an actual value output by a prediction model exceeds the limit for alarm and uploads alarm information;
the positive and negative deviation degree processing is carried out on the alarm information by using the following formula (1):
in formula (1), i=1, 2, …,7; m is m i The deviation degree of the alarm point of the characteristic parameter of the ith air compressor is; x is x i real Is the actual value of the characteristic parameter of the ith compressor; x is x i prediction Is the predicted value of the characteristic parameter of the ith compressor;
defining the exceeding high alarm limit as positive deviation degree, namely residual error monitoring exceeding upper limit; the exceeding of the low alarm limit is defined as negative deviation, namely, the residual error monitoring exceeds the lower limit; if no alarm information exists, defining as 0;
monitoring 7 compressor characteristic parameters through LSTM prediction modelThereby forming a vector group M by the deviation degree of the alarm points corresponding to the 7 characteristic parameters of the air compressor k ，M k ＝{m 1 ,m 2 ,…,m 7 And k is the number of samples of the alarm information vector group and is used as the input of the SVM classification model, wherein for a training sample k of 20, a test sample k of 80
S4, classifying positive and negative deviation of the alarm information as input parameters of the SVM classification model, determining the fault type of the air compressor, and realizing fault early warning of the air compressor.
2. The gas turbine compressor fault early warning method based on the LSTM-SVM according to claim 1, wherein in step S4, N sets of alarm vector sets are obtained through a residual curve output by a prediction model as input of an SVM classification model, classification is performed by using SVM, and a fault type of the compressor is determined, wherein a training set T of the SVM classification model is represented by the following formula (2):
T＝{(M 1 ,y 1 ),(M 2 ,y 2 ),…,(M k ,y k ),…,(M N ,y N )} (2)
in the formula (2), M k The k-th alarm deviation vector group; y is k For marking the fault type of the compressor, y k ＝{-1，1}；
The objective function and constraint conditions are represented by the following formulas (3), (4):
s.t.y k (wm k +b)≥1,k＝1,2,...,N (4)
in the formulas (3) and (4), w is a hyperplane normal vector dividing the fault type of the compressor, and b is a hyperplane offset term dividing the fault type of the compressor.
3. The gas turbine compressor fault early warning method based on the LSTM-SVM according to claim 1, wherein in the step S4, the gas turbine compressor fault early warning has 4 fault categories, then in the SVM classification model, a method of one category to the other category is adopted for solving the problem of supporting vector machine 4 classification, namely, one category is firstly divided into a certain category, the other three categories are divided into another category to form a sample set, 4 sample sets are obtained after four times of the operations are completed, thus 4 classification problems are formed, each sample set has a classification result, and the largest value is selected as a final classification result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011309245.4A CN112418306B (en) | 2020-11-20 | 2020-11-20 | Gas turbine compressor fault early warning method based on LSTM-SVM |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011309245.4A CN112418306B (en) | 2020-11-20 | 2020-11-20 | Gas turbine compressor fault early warning method based on LSTM-SVM |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112418306A CN112418306A (en) | 2021-02-26 |
| CN112418306B true CN112418306B (en) | 2024-03-29 |
Family
ID=74773811
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011309245.4A Active CN112418306B (en) | 2020-11-20 | 2020-11-20 | Gas turbine compressor fault early warning method based on LSTM-SVM |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112418306B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112632805B (en) * | 2021-03-15 | 2021-06-01 | 国能大渡河大数据服务有限公司 | Analysis early warning method, system, terminal and medium for crossing vibration area of unit |
| CN117516927B (en) * | 2024-01-05 | 2024-04-05 | 四川省机械研究设计院(集团)有限公司 | Gearbox fault detection method, system, equipment and storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109814527A (en) * | 2019-01-11 | 2019-05-28 | 清华大学 | Based on LSTM Recognition with Recurrent Neural Network industrial equipment failure prediction method and device |
| CN110161343A (en) * | 2019-06-12 | 2019-08-23 | 中南大学 | A kind of non-intrusion type real-time dynamic monitoring method of intelligence train exterior power receiving device |
| CN111237134A (en) * | 2020-01-14 | 2020-06-05 | 上海电力大学 | Offshore double-fed wind driven generator fault diagnosis method based on GRA-LSTM-stacking model |
| CN111340110A (en) * | 2020-02-25 | 2020-06-26 | 湖南师范大学 | Fault early warning method based on industrial process running state trend analysis |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2019028269A2 (en) * | 2017-08-02 | 2019-02-07 | Strong Force Iot Portfolio 2016, Llc | Methods and systems for detection in an industrial internet of things data collection environment with large data sets |
-
2020
- 2020-11-20 CN CN202011309245.4A patent/CN112418306B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109814527A (en) * | 2019-01-11 | 2019-05-28 | 清华大学 | Based on LSTM Recognition with Recurrent Neural Network industrial equipment failure prediction method and device |
| CN110161343A (en) * | 2019-06-12 | 2019-08-23 | 中南大学 | A kind of non-intrusion type real-time dynamic monitoring method of intelligence train exterior power receiving device |
| CN111237134A (en) * | 2020-01-14 | 2020-06-05 | 上海电力大学 | Offshore double-fed wind driven generator fault diagnosis method based on GRA-LSTM-stacking model |
| CN111340110A (en) * | 2020-02-25 | 2020-06-26 | 湖南师范大学 | Fault early warning method based on industrial process running state trend analysis |
Non-Patent Citations (2)
| Title |
|---|
| 基于多维时间序列状态估计的压气机故障预警模型建立;黄佳林;茅大钧;;上海电机学院学报;20170825(04);全文 * |
| 基于长短期记忆网络的风电机组齿轮箱故障预测;何群;尹飞飞;武鑫;谢平;江国乾;;计量学报;20201026(10);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112418306A (en) | 2021-02-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR101955305B1 (en) | Gas turbine sensor failure detection utilizing a sparse coding methodology | |
| US6587737B2 (en) | Method for the monitoring of a plant | |
| US8566070B2 (en) | Apparatus abnormality monitoring method and system | |
| CN112418306B (en) | Gas turbine compressor fault early warning method based on LSTM-SVM | |
| CN104035331B (en) | Unit running optimization instructs system and equipment thereof | |
| CN115034248A (en) | Automatic diagnostic method, system and storage medium for equipment | |
| CN112067335A (en) | Power plant blower fault early warning method based on multivariate state estimation | |
| US10401847B2 (en) | System and method for issue detection of industrial processes | |
| WO2016208315A1 (en) | Plant diagnosis device and plant diagnosis method | |
| KR20190082715A (en) | Data classification method based on correlation, and a computer-readable storege medium having program to perform the same | |
| CN111648992B (en) | Gas turbine compressor fault identification early warning method | |
| CN110007660B (en) | Online soft measurement method for transient equivalent thermal stress of steam turbine set of thermal power plant | |
| JP6830414B2 (en) | Diagnostic device and diagnostic method | |
| CN110826600B (en) | Engine surge prediction method based on adaptive resonance network online incremental learning | |
| CN110276115B (en) | Gas path fault diagnosis method based on gas turbine blade profile parameters | |
| CN112067289A (en) | Motor shaft and transmission shaft abnormal vibration early warning algorithm based on neural network | |
| CN114936520A (en) | Steam turbine fault early warning analysis method based on improved MSET | |
| Mi et al. | A nonparametric cumulative sum-based fault detection method for rolling bearings using high-level extended isolated forest | |
| Tingting et al. | Early warning method for power station auxiliary failure considering large-scale operating conditions | |
| Sun et al. | Anomaly detection of marine diesel engine valve system based on Mahalanobis distance | |
| Koukoura et al. | An insight into wind turbine planet bearing fault prediction using SCADA data | |
| Hamid et al. | Bearing fault identification for high-speed wind turbines using cnn | |
| CN116975755A (en) | Noise-tag-scene-oriented stepping robust semi-supervised gas turbine anomaly detection method and system | |
| Qosim et al. | Fault Detection System Using Machine Learning on Synthesis Loop Ammonia Plant | |
| CN116089877A (en) | Fault diagnosis method and system for turbine through-flow part of power plant |
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 |