CN110362065A - A kind of method for diagnosing status of aero-engine anti-asthma control system - Google Patents
A kind of method for diagnosing status of aero-engine anti-asthma control system Download PDFInfo
- Publication number
- CN110362065A CN110362065A CN201910646791.8A CN201910646791A CN110362065A CN 110362065 A CN110362065 A CN 110362065A CN 201910646791 A CN201910646791 A CN 201910646791A CN 110362065 A CN110362065 A CN 110362065A
- Authority
- CN
- China
- Prior art keywords
- control system
- aero
- asthma control
- engine anti
- engine
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B23/00—Testing or monitoring of control systems or parts thereof
- G05B23/02—Electric testing or monitoring
- G05B23/0205—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults
- G05B23/0218—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults
- G05B23/0243—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults model based detection method, e.g. first-principles knowledge model
- G05B23/0245—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults model based detection method, e.g. first-principles knowledge model based on a qualitative model, e.g. rule based; if-then decisions
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Automation & Control Theory (AREA)
- Testing And Monitoring For Control Systems (AREA)
- Control Of Positive-Displacement Air Blowers (AREA)
Abstract
The invention belongs to aero-engine fault diagnosis technique field more particularly to a kind of method for diagnosing status of the aero-engine anti-asthma control system based on hidden Markov model.This method comprises the following steps: A1, the sensing data for obtaining aero-engine anti-asthma control system;As input data, input hidden Markov model trained in advance obtains output result for A2, the sensing data that will acquire;A3, output result is matched with status information predetermined, the current state of output aero-engine anti-asthma control system;Wherein, the hidden Markov model trained in advance is the training set of the status information of the sensing data based on the aero-engine anti-asthma control system in default historical time section and corresponding aero-engine anti-asthma control system, the model after being trained.This method is based on hidden Markov model, and implementation complexity is low, and fault diagnosis levels of precision is higher.
Description
Technical field
The invention belongs to aero-engine fault diagnosis technique fields more particularly to a kind of based on hidden Markov model
The method for diagnosing status of aero-engine anti-asthma control system.
Background technique
In aero-engine type, there is a kind of gas-turbine unit, occasional enters its aerostatic press at work
A kind of unstable working condition, becomes surge.Surge once occur, it will cause CR Critical consequence, can damage aerostatic press,
Entire engine is damaged, flight safety accident is caused.
There is the more mature technology for avoiding surge from occurring at present, which is applied on engine, is known as
Aero-engine anti-asthma control system.But although anti-asthma control system can effectively inhibit the surge of aero-engine aerostatic press,
But because the system control planning is complicated, cause failure rate higher, it is same because of complicated control logic, cause to increase
The difficulty of the condition diagnosing of the system.
Summary of the invention
(1) technical problems to be solved
For existing technical problem, the present invention provides a kind of condition diagnosing of aero-engine anti-asthma control system
Method, this method are based on hidden Markov model, and implementation complexity is low, and fault diagnosis levels of precision is higher.
(2) technical solution
The present invention provides a kind of method for diagnosing status of aero-engine anti-asthma control system, includes the following steps:
A1, the sensing data for obtaining aero-engine anti-asthma control system;
As input data, input hidden Markov model trained in advance is obtained defeated for A2, the sensing data that will acquire
Result out;
A3, output result is matched with status information predetermined, exports aero-engine anti-asthma control system
Current state;
Wherein, the hidden Markov model trained in advance is anti-based on the aero-engine in default historical time section
The training set for breathing heavily the sensing data of control system and the status information of corresponding aero-engine anti-asthma control system, is trained
Model afterwards.
Further, the sensing data of the aero-engine anti-asthma control system includes: engine intake temperature, starts
Machine high-pressure compressor rotor revolving speed and engine variable guide vane angle.
Further, the status information predetermined includes: health status, degenerate state I, degenerate state II, moves back
Change state III, malfunction I and malfunction II.
Further, it before the step A1, further comprises the steps of:
A0, the state for initializing aero-engine anti-asthma control system,
Specifically, the status information of anti-asthma control system: health status 1 is set, other states are 0.
Further, in the hidden Markov model, using the sensing data of aero-engine anti-asthma control system as
Observation;
Using the status information of aero-engine anti-asthma control system as hidden state.
Further, the structural parameters of hidden Markov model include state-transition matrix A and observation probability matrix B;Shape
State shift-matrix A indicates the transition probability between two hidden states, and observation probability matrix B indicates the control of aero-engine anti-asthma
System, which is presently in hidden state, leads to the probability of each observation.
Further, the training formula of the state-transition matrix A are as follows:
A=[aij]
Wherein, aijIndicate that aero-engine anti-asthma control system is transferred to the probability of hidden state j from hidden state i;Aij
Aero-engine anti-asthma control system is transferred to the number of hidden state j from hidden state i in expression training set;I, j is indicated
The status information of aero-engine anti-asthma control system, value range 1,2...n, n indicate number of states, n=6.
Further, the training formula of the observation probability matrix B are as follows:
B=[bij]
In formula, bijWhen indicating that aero-engine anti-asthma control system is hidden state i, the probability that observation j occurs,
In,
In formula, b'iIndicate that aero-engine anti-asthma control system is described using gauss hybrid models to be occurred in hidden state i
When observation probability density function, m indicate observation dimension;wijIndicate that when aero-engine anti-asthma control system be hidden
When hiding state i, the weight of jth dimension observation;oijFor independent variable, jth dimension observation is indicated;N(oij|uij,σij) indicate
When aero-engine anti-asthma control system is in hidden state i, the gauss of distribution function of jth dimension, uijAnd σijIt respectively refers to for mean value
With standard deviation.
Further,
In formula, l indicates training samples number;It indicates when aero-engine anti-asthma control system is in hidden state i,
Jth ties up observation in the value of s-th of sample;LiIndicate maximum likelihood function.
Further, the judgement and prediction of the hidden state are completed by viterbi algorithm.
(3) beneficial effect
The method for diagnosing status of aero-engine anti-asthma control system provided by the invention based on hidden Markov model,
The malfunction of current aerospace engine anti-asthma control system can not only more accurately be diagnosed, additionally it is possible to accurately estimate aviation
The state change of engine anti-asthma control system.The control logic of aero-engine anti-asthma control system is simple, and failure rate is low, drop
The low difficulty of the condition diagnosing of the system.
Detailed description of the invention
Fig. 1 is the flow chart of the method for diagnosing status of aero-engine anti-asthma control system in the present invention;
Fig. 2 is the schematic diagram of hidden Markov model in the present invention.
Specific embodiment
In order to preferably explain the present invention, in order to understand, with reference to the accompanying drawing, by specific embodiment, to this hair
It is bright to be described in detail.
As shown in Figure 1, 2, the present invention provides a kind of method for diagnosing status of aero-engine anti-asthma control system, including instruction
Practice stage and implementation phase.
Training stage includes:
S1, the sensing data for presetting the aero-engine anti-asthma control system in historical time section and corresponding aviation are obtained
The status information of engine anti-asthma control system forms training set as training data;
S2, training data is inputted to hidden Markov model, training obtains the structural parameters of hidden Markov model: state
Shift-matrix A, observation probability matrix B, the hidden Markov model after being trained with this.
Implementation phase includes:
A0, the state for initializing aero-engine anti-asthma control system;
A1, the sensing data for obtaining aero-engine anti-asthma control system in real time;
A2, the sensing data that will acquire input the hidden Markov model after training, are exported as input data
As a result;
A3, output result is matched with status information predetermined, exports aero-engine anti-asthma control system
Current state.
Wherein, the sensing data of aero-engine anti-asthma control system includes: engine intake temperature, engine high pressure pressure
Mechanism of qi rotor speed and engine variable guide vane angle;
Status information predetermined includes: health status, degenerate state I, degenerate state II, degenerate state III, failure
State I and malfunction II.
The state for initializing aero-engine anti-asthma control system, specifically includes:
The status information of anti-asthma control system: health status 1 is set, other states are 0.
Further, using three sensing datas of aero-engine anti-asthma control system as in hidden Markov model
Observation indicates the set of all possible sensing datas with O.The observation of any time is a three-dimensional vector, uses O=
{o1,o2,o3Indicate, wherein o1,o2,o3Respectively indicate one in three supplemental characteristics, all supplemental characteristics are all successional;
Using the status information of aero-engine anti-asthma control system as the hidden state in hidden Markov model, Q is used
Indicate the set of all possible hidden states.Hidden Markov model is in use, hidden state is with the mode of a sequence
It shows, uses { siIndicate, wherein [0, t] i ∈, then { siRefer to the hidden state sequence of the t from the moment 0 to the moment, si∈Q。
Then the representation of training set is { s1,o1},{s2,o2},...{sl,ol, wherein l indicates training samples number.
State-transition matrix A indicates the transition probability between two hidden states, and calculation method is as follows:
A=[aij]
In formula, aijIndicate that aero-engine anti-asthma control system is transferred to the probability of hidden state j from hidden state i;Aij
Indicate the number for being transferred to hidden state j in training set from hidden state i;I, j indicates aero-engine anti-asthma control system
Status information, value range 1,2...n, n indicate number of states, n=6.
Observation probability matrix B indicates that hidden state locating for aero-engine anti-asthma control system can result in each observation
Probability, calculation method is as follows:
B=[bij]
In formula, bijWhen indicating hidden state i, the probability of observation j generation.
Since observation belongs to three-dimensional continuous variable, observation j has been extended to a range from a point by the present invention, tool
Body is as follows:
Firstly, describing sight of the aero-engine anti-asthma control system when hidden state i occurs using gauss hybrid models
The probability density function b' of measured valuei:
In formula, m indicates the dimension of observation, and m takes 3 in the present invention;wijIt indicates to work as aero-engine anti-asthma control system
When for hidden state i, the weight of jth dimension observation;oijFor independent variable, jth dimension observation is indicated;N(oij|uij,σij) table
Show when aero-engine anti-asthma control system is in hidden state i, the gauss of distribution function of jth dimension, uijAnd σijRespectively refer to generation
Mean value and standard deviation.
Gauss of distribution function are as follows:
As probability density function, b'iIt is o in observationijWhen it is meaningless, therefore, which is extended for by the present invention
One standard deviation range, i.e. calculating oijIn sectionIntegral.Then, it is when aero-engine anti-asthma controls
When system is in hidden state i, the probability of observation j generation are as follows:
Further, about uij、σijAnd wijTraining:
In formula, l indicates training samples number,It indicates when aero-engine anti-asthma control system is in hidden state i,
Jth ties up observation in the value of s-th of sample.
For wijTraining, using Maximum Likelihood Estimation:
In formula, LiIndicate maximum likelihood function, training process is to solve for wijEnable LiValue it is maximum.
Further, specific as follows using viterbi algorithm when judging hidden state:
T1, in t moment, hidden state be i all state transition paths in, the maximum value of each path probability, table
It is shown as δt(i)。
T2, according to step T1, release the maximum probability value at the t+1 moment:
In formula, δt+1(i) it indicates at the t+1 moment, hidden state is in all state transition paths of i, and each path is general
The maximum value of rate;When indicating that anti-asthma control system is hidden state i, observation o(t+1)Probability.When implementing, because
It is continuous function for state parameter, what is actually calculated is in O(t+1)The probability of one minizone of surrounding.
According to above-mentioned, can from 0 moment, recursion obtains the hidden state chain of maximum probability to current time, record always,
The as judging result of hidden state.
The technical principle of the invention is described above in combination with a specific embodiment, these descriptions are intended merely to explain of the invention
Principle shall not be construed in any way as a limitation of the scope of protection of the invention.Based on explaining herein, those skilled in the art
It can associate with other specific embodiments of the invention without creative labor, these modes fall within this hair
Within bright protection scope.
Claims (10)
1. a kind of method for diagnosing status of aero-engine anti-asthma control system, which comprises the steps of:
A1, the sensing data for obtaining aero-engine anti-asthma control system;
As input data, input hidden Markov model trained in advance obtains output knot for A2, the sensing data that will acquire
Fruit;
A3, output result is matched with status information predetermined, output aero-engine anti-asthma control system is current
State;
Wherein, the hidden Markov model trained in advance is based on the aero-engine anti-asthma control in default historical time section
The training set of the status information of the sensing data of system processed and corresponding aero-engine anti-asthma control system, after being trained
Model.
2. method for diagnosing status according to claim 1, which is characterized in that the aero-engine anti-asthma control system
Sensing data includes: engine intake temperature, engine high-pressure compressor rotor revolving speed and engine variable guide vane angle.
3. method for diagnosing status according to claim 2, which is characterized in that the status information predetermined includes:
Health status, degenerate state I, degenerate state II, degenerate state III, malfunction I and malfunction II.
4. method for diagnosing status according to claim 3, which is characterized in that before the step A1, further comprise the steps of:
A0, the state for initializing aero-engine anti-asthma control system,
Specifically, the status information of anti-asthma control system: health status 1 is set, other states are 0.
5. method for diagnosing status according to claim 4, which is characterized in that in the hidden Markov model, by aviation
The sensing data of engine anti-asthma control system is as observation;
Using the status information of aero-engine anti-asthma control system as hidden state.
6. method for diagnosing status according to claim 5, which is characterized in that the structural parameters of hidden Markov model include
State-transition matrix A and observation probability matrix B;
State-transition matrix A indicates the transition probability between two hidden states, and observation probability matrix B indicates that aero-engine is anti-
Asthma control system, which is presently in hidden state, leads to the probability of each observation.
7. method for diagnosing status according to claim 6, which is characterized in that the training formula of the state-transition matrix A
Are as follows:
A=[aij]
Wherein, aijIndicate that aero-engine anti-asthma control system is transferred to the probability of hidden state j from hidden state i;AijIt indicates
Aero-engine anti-asthma control system is transferred to the number of hidden state j from hidden state i in training set;I, j indicates aviation
The status information of engine anti-asthma control system, value range 1,2...n, n indicate number of states, n=6.
8. method for diagnosing status according to claim 7, which is characterized in that the training formula of the observation probability matrix B
Are as follows:
B=[bij]
In formula, bijWhen indicating that aero-engine anti-asthma control system is hidden state i, the probability of observation j generation, wherein
In formula, b'iIndicate using gauss hybrid models describe aero-engine anti-asthma control system hidden state i occur when
The probability density function of observation, m indicate the dimension of observation;wijIndicate that when aero-engine anti-asthma control system be to hide shape
When state i, the weight of jth dimension observation;oijFor independent variable, jth dimension observation is indicated;N(oij|uij,σij) indicate in aviation
When engine anti-asthma control system is in hidden state i, the gauss of distribution function of jth dimension, uijAnd σijIt respectively refers to for mean value and mark
It is quasi- poor.
9. method for diagnosing status according to claim 8, which is characterized in that
In formula, l indicates training samples number;Indicate the jth when aero-engine anti-asthma control system is in hidden state i
Observation is tieed up in the value of s-th of sample;LiIndicate maximum likelihood function.
10. method for diagnosing status according to claim 9, which is characterized in that the judgement and prediction of the hidden state are logical
Cross viterbi algorithm completion.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910646791.8A CN110362065B (en) | 2019-07-17 | 2019-07-17 | State diagnosis method of anti-surge control system of aircraft engine |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910646791.8A CN110362065B (en) | 2019-07-17 | 2019-07-17 | State diagnosis method of anti-surge control system of aircraft engine |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110362065A true CN110362065A (en) | 2019-10-22 |
CN110362065B CN110362065B (en) | 2022-07-19 |
Family
ID=68221057
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910646791.8A Active CN110362065B (en) | 2019-07-17 | 2019-07-17 | State diagnosis method of anti-surge control system of aircraft engine |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110362065B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112809310A (en) * | 2020-12-30 | 2021-05-18 | 中国人民解放军第五七一九工厂 | Method for discriminating and repairing manufacturing defects of anti-surge system of aircraft engine |
CN113571092A (en) * | 2021-07-14 | 2021-10-29 | 东软集团股份有限公司 | Method for identifying abnormal sound of engine and related equipment thereof |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CA2568572A1 (en) * | 2005-11-23 | 2007-05-23 | General Electric Company | System and method for generating closed captions |
CN105023011A (en) * | 2015-08-19 | 2015-11-04 | 苏州奥诺遥感科技有限公司 | HMM based crop phenology dynamic estimation method |
CN106339322A (en) * | 2016-09-13 | 2017-01-18 | 哈尔滨工程大学 | Method for software behavior prediction based on HMM-ACO |
CN106919759A (en) * | 2017-03-03 | 2017-07-04 | 哈尔滨工业大学 | The generalized approximate modeling method of the aero-engine performance based on fit sensitivity and model application |
CN109344195A (en) * | 2018-10-25 | 2019-02-15 | 电子科技大学 | Pipe safety event recognition and Knowledge Discovery Method based on HMM model |
-
2019
- 2019-07-17 CN CN201910646791.8A patent/CN110362065B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CA2568572A1 (en) * | 2005-11-23 | 2007-05-23 | General Electric Company | System and method for generating closed captions |
CN105023011A (en) * | 2015-08-19 | 2015-11-04 | 苏州奥诺遥感科技有限公司 | HMM based crop phenology dynamic estimation method |
CN106339322A (en) * | 2016-09-13 | 2017-01-18 | 哈尔滨工程大学 | Method for software behavior prediction based on HMM-ACO |
CN106919759A (en) * | 2017-03-03 | 2017-07-04 | 哈尔滨工业大学 | The generalized approximate modeling method of the aero-engine performance based on fit sensitivity and model application |
CN109344195A (en) * | 2018-10-25 | 2019-02-15 | 电子科技大学 | Pipe safety event recognition and Knowledge Discovery Method based on HMM model |
Non-Patent Citations (1)
Title |
---|
李晓明: "某航空发动机防喘控制系统故障预测与健康管理技术应用研究", 《中国博士学位论文全文数据库 工程科技II辑》 * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112809310A (en) * | 2020-12-30 | 2021-05-18 | 中国人民解放军第五七一九工厂 | Method for discriminating and repairing manufacturing defects of anti-surge system of aircraft engine |
CN112809310B (en) * | 2020-12-30 | 2022-01-25 | 中国人民解放军第五七一九工厂 | Method for discriminating and repairing manufacturing defects of anti-surge system of aircraft engine |
CN113571092A (en) * | 2021-07-14 | 2021-10-29 | 东软集团股份有限公司 | Method for identifying abnormal sound of engine and related equipment thereof |
CN113571092B (en) * | 2021-07-14 | 2024-05-17 | 东软集团股份有限公司 | Engine abnormal sound identification method and related equipment thereof |
Also Published As
Publication number | Publication date |
---|---|
CN110362065B (en) | 2022-07-19 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106021826B (en) | One kind is based on aero-engine complete machine method for predicting residual useful life under operating mode's switch and the matched variable working condition of similitude | |
CN108256173A (en) | A kind of Gas path fault diagnosis method and system of aero-engine dynamic process | |
CN112131760A (en) | CBAM model-based prediction method for residual life of aircraft engine | |
EP1443375A1 (en) | Fault diagnosis | |
CN106055770A (en) | Diagnostic method for gas path faults of aero-engine based on sliding mode theory | |
CN109162813B (en) | One kind being based on the modified Aeroengine Smart method for controlling number of revolution of iterative learning | |
CN110362065A (en) | A kind of method for diagnosing status of aero-engine anti-asthma control system | |
CN110516394A (en) | Aero-engine steady-state model modeling method based on deep neural network | |
CN109472062A (en) | A kind of variable cycle engine self-adaptive component grade simulation model construction method | |
CN109635318A (en) | A kind of aero-engine sensor intelligent analytic redundancy design method based on KEOS-ELM algorithm | |
CN109489987A (en) | Fanjet measurement biases fault-tolerant gas circuit performance distributed and filters estimation method | |
Li et al. | Study on gas turbine gas-path fault diagnosis method based on quadratic entropy feature extraction | |
CN111860791A (en) | Aero-engine thrust estimation method and device based on similarity transformation | |
CN110334383A (en) | Gas turbine fault diagnosis expert system method based on GA and L-M Combinatorial Optimization | |
CN106257253A (en) | Temperature sensor signal modification method and system | |
Reitz et al. | Design of experiments and numerical simulation of deteriorated high pressure compressor airfoils | |
Venkatachari et al. | Assessment of RANS-based transition models based on experimental data of the common research model with natural laminar flow | |
CN110516391A (en) | A kind of aero-engine dynamic model modeling method neural network based | |
CN109308484A (en) | Aero-engine multiclass failure minimum risk diagnostic method and device | |
CN112116101B (en) | Aeroengine fault diagnosis method based on group reduction kernel extreme learning machine | |
CN114154234A (en) | Modeling method, system and storage medium for aircraft engine | |
CN105785791A (en) | Modeling method of airborne propulsion system under supersonic speed state | |
Kong et al. | Study on condition monitoring of 2-Spool turbofan engine using non-linear gas path analysis method and genetic algorithms | |
CN116702380A (en) | Aeroengine performance degradation monitoring and model correction method based on digital twin | |
CN112733880B (en) | Aircraft engine fault diagnosis method and system and electronic equipment |
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 |