CN112528414A - SOM-MQE-based aircraft engine fault early warning method - Google Patents
SOM-MQE-based aircraft engine fault early warning method Download PDFInfo
- Publication number
- CN112528414A CN112528414A CN202011501087.2A CN202011501087A CN112528414A CN 112528414 A CN112528414 A CN 112528414A CN 202011501087 A CN202011501087 A CN 202011501087A CN 112528414 A CN112528414 A CN 112528414A
- Authority
- CN
- China
- Prior art keywords
- aircraft engine
- data
- mqe
- early warning
- som
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 52
- 238000013139 quantization Methods 0.000 claims abstract description 76
- 238000012549 training Methods 0.000 claims abstract description 55
- 230000002159 abnormal effect Effects 0.000 claims abstract description 38
- 238000012360 testing method Methods 0.000 claims description 42
- 210000002569 neuron Anatomy 0.000 claims description 7
- 230000006872 improvement Effects 0.000 description 8
- 230000008569 process Effects 0.000 description 8
- 238000013528 artificial neural network Methods 0.000 description 6
- 238000004422 calculation algorithm Methods 0.000 description 6
- 238000004458 analytical method Methods 0.000 description 3
- 238000003745 diagnosis Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000013507 mapping Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 230000007423 decrease Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 230000036541 health Effects 0.000 description 2
- 238000012423 maintenance Methods 0.000 description 2
- 238000003860 storage Methods 0.000 description 2
- 238000012706 support-vector machine Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 210000004556 brain Anatomy 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000000446 fuel Substances 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000002194 synthesizing effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/15—Vehicle, aircraft or watercraft design
-
- 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
- G06N3/08—Learning methods
-
- 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
- G06F2119/02—Reliability analysis or reliability optimisation; Failure analysis, e.g. worst case scenario performance, failure mode and effects analysis [FMEA]
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computer Hardware Design (AREA)
- Biophysics (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Automation & Control Theory (AREA)
- Aviation & Aerospace Engineering (AREA)
- Computational Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Testing Of Engines (AREA)
Abstract
The invention relates to an aircraft engine fault early warning method based on SOM-MQE, belongs to the technical field of equipment fault early warning, and solves the problem that in the prior art, the accuracy of a fault prediction result is low due to individualized differences in the aspects of operating environments, service lives and the like of different aircraft engines. The method comprises the following steps: acquiring full life cycle data of a plurality of training aircraft engines; dividing the full life cycle data into normal data and abnormal data according to the using duration; the abnormal data comprises fault early warning data in an early warning interval in the whole life cycle; training an SOM model of the corresponding aircraft engine by using normal data in each training aircraft engine, and obtaining a minimum quantization error MQE of fault early warning data in the aircraft engine; determining MQE ranges corresponding to early warning intervals based on MQE average values of fault early warning data in a plurality of training aircraft engines; and based on the MQE range, performing fault early warning of the aircraft engine to be early warned.
Description
Technical Field
The invention relates to the technical field of equipment fault early warning, in particular to an aircraft engine fault early warning method based on SOM-MQE.
Background
In actual industrial production, the service life of complex systems such as a standard part and an engine is limited, and the complex systems need to be repaired (or replaced) after running for a period of time. Therefore, the accident can be effectively reduced by timely carrying out fault early warning on the equipment. At present, fault diagnosis technology of mechanical equipment mainly focuses on both signal analysis and intelligent diagnosis. In the aspect of signal analysis, the relevant practitioners are required to have good knowledge base and experience; in the aspect of intelligent diagnosis, the application of machine learning related algorithms has a remarkable effect, and mainly includes methods of pattern recognition such as Support Vector Machines (SVMs), Artificial Neural Networks (ANN), Kernel methods (Kernel methods), Convolutional Neural networks (Convolutional networks), and the like. In the training process of the model, a large amount of historical abnormal data is required to fit model parameters, but for an actual industrial scene, normal state data of equipment can be conveniently and accurately acquired, and fault samples are often difficult to acquire, which provides great challenges for common supervised learning models.
The self-organizing mapping network (SOM) is an unsupervised artificial neural network, and gradually optimizes the network by means of mutual competition among neurons through simulating the characteristics of human brain on signal processing. The SOM algorithm has a simple structure and a small quantity of parameters, so that the SOM algorithm is widely used for speech recognition, clustering, abnormal value elimination and the like. In the prior art, the SOM algorithm and the immune neural network are combined to predict and diagnose the fault of a fuel system in an aircraft engine, and the method classifies the fault by using the BP neural network on the basis of using the SOM algorithm to process data, and still requires sufficient abnormal data samples. However, the research does not consider individual differences in the operating environments, the service lives and the like of different aircraft engines, so that the accuracy of the fault prediction result is low.
Disclosure of Invention
In view of the foregoing analysis, the embodiments of the present invention are directed to providing an airplane engine fault early warning method based on SOM-MQE, so as to solve the problem in the prior art that the accuracy of a fault prediction result is low due to individual differences in operating environments, service lives, and the like of different airplane engines.
An aircraft engine fault early warning method based on SOM-MQE comprises the following steps:
acquiring full life cycle data of a plurality of training aircraft engines; the full life cycle data is divided into normal data and abnormal data according to the using duration; the abnormal data comprises fault early warning data in an early warning interval in the whole life cycle;
training an SOM model of a corresponding aircraft engine by using normal data in each training aircraft engine, and obtaining a minimum quantization error MQE of fault early warning data in the aircraft engine;
determining a minimum quantization error MQE range corresponding to an early warning interval based on an average value of the minimum quantization errors MQE of the fault early warning data in the plurality of training aircraft engines;
and performing fault early warning on the aircraft engine to be early warned based on the minimum quantization error MQE range.
On the basis of the above scheme, the present embodiment further makes the following improvements:
based on the further improvement of the above method, after determining the minimum quantization error MQE range, the method further includes:
training an SOM model of a test aircraft engine by using normal data in the test aircraft engine, and acquiring a minimum quantization error MQE of abnormal data in the test aircraft engine;
acquiring an abnormal data interval of which the minimum quantization error MQE of the abnormal data is within the range of the minimum quantization error MQE;
obtaining an early warning accuracy rate based on the ratio of the intersection of the abnormal data interval and the early warning interval to the early warning interval duration;
and if the accuracy is equal to or higher than a set accuracy threshold, performing fault early warning on the aircraft engine to be early warned based on the minimum quantization error MQE range.
Based on the further improvement of the method, the method further comprises the following steps:
if the accuracy is lower than the set accuracy threshold, adjusting the SOM model parameters of the test aircraft engine or the proportion of normal data in the test aircraft engine, retraining the SOM model of the test aircraft engine, and reacquiring the early warning accuracy.
Based on the further improvement of the method, the method further comprises the following steps:
if the accuracy is lower than the set accuracy threshold, adjusting one or more of the parameters of the SOM model of the training aircraft engine, the proportion of normal data in the full life cycle data and the range of the early warning interval, and re-determining the range of the minimum quantization error MQE and re-obtaining the early warning accuracy.
Based on further improvement of the method, the parameters of the SOM model for training the aircraft engine and testing the aircraft engine are adjusted, and the parameters comprise one or more of a learning rate initialization parameter, an initial win neighborhood and a neuron weight initial value in the SOM model.
Based on the further improvement of the method, the performing fault early warning of the aircraft engine to be early warned based on the minimum quantization error MQE range includes:
when normal data of an aircraft engine to be early warned are collected, training an SOM (self-organizing map) model of the aircraft engine to be early warned by utilizing the collected normal data of the aircraft engine to be early warned;
acquiring a minimum quantization error MQE of abnormal data in an aircraft engine to be early warned every time the abnormal data of the aircraft engine to be early warned is acquired;
and if the minimum quantization error MQE of the abnormal data in the aircraft engine to be early-warned is within the range of the minimum quantization error MQE, performing fault early warning on the aircraft engine to be early-warned.
Based on a further improvement of the above method, the minimum quantization error MQE range has a lower limit of 80% and an upper limit of 120% of the average value of the minimum quantization error MQE and the minimum quantization error MQE.
On the basis of a further development of the method described above,
the normal data is the data of the first 60% of the service life of the full life cycle data of the aircraft engine;
the abnormal data is data of later 40% of service life in the full life cycle data of the aircraft engine;
the early warning interval is in the range of 88% -92% of the service life of the full life cycle data of the aircraft engine.
In a further development of the above method, the full life cycle data of the aircraft engine comprises: engine intake temperature, exhaust temperature, oil pressure, compressed air pressure, and vibration speed parameters.
Based on the further improvement of the method, the training the SOM model of the current aircraft engine by using the normal data in each aircraft engine and obtaining the minimum quantization error MQE of the fault early warning data in the current aircraft engine includes:
training an SOM (model of mass) model of the current aircraft engine by using normal data in each aircraft engine to obtain an optimal matching unit;
and taking the minimum value of the distance between the fault early warning data and the optimal matching unit as the minimum quantization error MQE of the fault early warning data.
Compared with the prior art, the invention can realize at least one of the following beneficial effects:
firstly, the SOM-MQE-based aircraft engine fault early warning method provided by the invention fully considers the individual differences in the aspects of operating environment, service life, factory settings, wear degree and the like of different aircraft engines, and trains a plurality of SOM models for training the aircraft engines respectively to obtain accurate SOM model training results corresponding to the aircraft engines and corresponding minimum quantization errors MQE;
secondly, the SOM-MQE-based aircraft engine fault early warning method provided by the invention considers that the time for easily generating fault early warning of different aircraft engines of the same type is approximately the same (namely, the early warning interval corresponding to each training aircraft engine), so that the minimum quantization error MQE range corresponding to the early warning interval can be determined by calculating the average value of the minimum quantization errors MQE of fault early warning data in a plurality of training aircraft engines, the minimum quantization error MQE range determined by adopting the method can furthest integrate individualized differences of different aircraft engines, and can be used for fault early warning of aircraft engines of the same type;
thirdly, the fault early warning method for the aircraft engine based on the SOM-MQE provided by the invention carries out fault early warning on the aircraft engine to be early warned based on the minimum quantization error MQE range, can remind a user of repairing or replacing the aircraft engine in time, effectively reduces the accidental repairing times of the aircraft engine and ensures the use safety of the aircraft engine.
In the invention, the technical schemes can be combined with each other to realize more preferable combination schemes. Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and drawings.
Drawings
The drawings are only for purposes of illustrating particular embodiments and are not to be construed as limiting the invention, wherein like reference numerals are used to designate like parts throughout.
FIG. 1 is a flow chart of a SOM-MQE-based aircraft engine fault warning method provided in embodiment 1 of the present invention;
fig. 2 is an MQE distribution diagram of the fault warning data provided in embodiment 2 of the present invention;
fig. 3 is a diagram illustrating a variation trend of MQE values of measured data of test equipment No. 1 according to embodiment 2 of the present invention.
Detailed Description
The accompanying drawings, which are incorporated in and constitute a part of this application, illustrate preferred embodiments of the invention and together with the description, serve to explain the principles of the invention and not to limit the scope of the invention.
Example 1
The embodiment 1 of the invention discloses an aircraft engine fault early warning method based on SOM-MQE, and the flow chart is shown in figure 1 and comprises the following steps:
step S1: acquiring full life cycle data of a plurality of training aircraft engines; the full life cycle data is divided into normal data and abnormal data according to the using duration; the abnormal data comprises fault early warning data in an early warning interval in the whole life cycle;
step S2: training an SOM model of a corresponding aircraft engine by using normal data in each training aircraft engine, and obtaining a minimum quantization error MQE of fault early warning data in the aircraft engine;
step S3: determining a minimum quantization error MQE range corresponding to an early warning interval based on an average value of the minimum quantization errors MQE of the fault early warning data in the plurality of training aircraft engines;
step S4: and performing fault early warning on the aircraft engine to be early warned based on the minimum quantization error MQE range.
Compared with the prior art, the SOM-MQE-based aircraft engine fault early warning method provided by the embodiment of the invention fully considers the individual differences in the aspects of the operating environment, the service life, the factory setting, the wear degree and the like of different aircraft engines, and respectively trains a plurality of SOM models for training the aircraft engines to obtain the accurate SOM model training result corresponding to the aircraft engine and the corresponding minimum quantization error MQE; meanwhile, considering that the time for easily generating fault early warning of different aircraft engines of the same type is approximately the same (namely, the early warning interval corresponding to each training aircraft engine), the minimum quantization error MQE range corresponding to the early warning interval can be determined by calculating the average value of the minimum quantization errors MQE of the fault early warning data in a plurality of training aircraft engines, and the minimum quantization error MQE range determined by adopting the method can be used for maximally synthesizing the individualized differences of the different aircraft engines and can be used for fault early warning of the aircraft engines of the same type; in addition, the fault early warning of the aircraft engine to be early warned is carried out based on the minimum quantization error MQE range, a user can be reminded to repair or replace the aircraft engine in time, the accidental maintenance frequency of the aircraft engine is effectively reduced, and the use safety of the aircraft engine is guaranteed.
It should be noted that the training aircraft engine, the aircraft engine to be warned in the embodiment, and the test aircraft engine described later are all the same type of aircraft engine. Meanwhile, the training aircraft engine and the testing aircraft engine both contain full life cycle data of the aircraft engine from a healthy state, a decline state to a fault state.
Preferably, after determining the minimum quantization error MQE range, the accuracy of the determined minimum quantization error MQE range may be verified by performing the following operations:
training an SOM model of a test aircraft engine by using normal data in the test aircraft engine, and acquiring a minimum quantization error MQE of abnormal data in the test aircraft engine;
acquiring an abnormal data interval of which the minimum quantization error MQE of the abnormal data is within the range of the minimum quantization error MQE;
obtaining an early warning accuracy rate based on the ratio of the intersection of the abnormal data interval and the early warning interval to the early warning interval duration; it should be noted that the ratio of the intersection of the abnormal data interval and the early warning interval (i.e., the early warning interval determined according to the tested aircraft engine) to the duration of the early warning interval represents the coincidence degree of the early warning result corresponding to the tested aircraft engine and the early warning interval set in the training process, and can be used for representing the early warning accuracy.
If the accuracy is equal to or higher than the set accuracy threshold, the determined minimum quantization error MQE range can meet the early warning requirement, and at this time, fault early warning of the aircraft engine to be early warned can be performed based on the minimum quantization error MQE range. The accuracy threshold value can be determined according to the test accuracy requirement. Illustratively, in the absence of explicit fault tag data, the accuracy threshold may be set to 70% according to the general requirements of the industry.
In the embodiment, in order to avoid reduction of the early warning accuracy rate caused by individual differences of a single test aircraft engine, the early warning accuracy rates corresponding to a plurality of test aircraft engines can be calculated simultaneously; if the early warning accuracy rates corresponding to the tested aircraft engines above 90% (the parameter can be adjusted according to the test accuracy requirement) are all equal to or higher than the set accuracy rate threshold, it indicates that the determined minimum quantization error MQE range can meet the early warning requirement. Otherwise, the reason for this can be analyzed from two aspects:
(1) problems associated with testing aircraft engines:
1) testing improper setting of SOM model parameters of an aircraft engine:
adjusting SOM model parameters of the test aircraft engine, retraining the SOM model of the test aircraft engine, and reacquiring the early warning accuracy; if the reacquired early warning accuracy is equal to or higher than the accuracy threshold, it is indicated that the prediction accuracy possibly caused by improper setting of the parameters of the SOM model is low;
2) the proportion of normal data of the tested aircraft engine is not properly selected:
and adjusting the proportion of normal data in the tested aircraft engine, retraining the SOM model of the tested aircraft engine, and reacquiring the early warning accuracy. If the reacquired early warning accuracy is equal to or higher than the accuracy threshold, it is indicated that the prediction accuracy is low probably due to improper selection of normal data of the tested aircraft engine;
if the predicted accuracy is still lower than the accuracy threshold, the SOM model parameters of the aircraft engine and the normal data ratio of the tested aircraft engine can be adjusted at the same time, and the early warning accuracy is calculated again.
If the set accuracy threshold requirement cannot be met through repeated tests, whether the determined minimum quantization error MQE range has a problem or not is considered, at the moment, one or more of the SOM model parameters of the training aircraft engine, the proportion of normal data in the full life cycle data and the range of the early warning interval can be adjusted, the minimum quantization error MQE range is determined again, and the early warning accuracy is obtained again.
The above process of testing and re-determining the minimum quantization error MQE range is complicated and may require trial and error to obtain the optimal minimum quantization error MQE range.
It should be noted that, adjusting the SOM model parameters of the training aircraft engine and the testing aircraft engine includes adjusting one or more parameters of the learning rate initialization parameter, the initial win neighborhood, and the initial neuron weight value in the SOM model.
Preferably, step S2 includes:
step S21: training an SOM (model of mass) model of the current aircraft engine by using normal data in each training aircraft engine to obtain an optimal matching unit;
it should be noted that the SOM belongs to an algorithm of a neural network, which can map data of a high dimension to a low dimension. In the training phase, the distance between the features of each input sample and the neurons of the mapping layer is calculated, the neuron closest to the input sample in the mapping layer is found, and this neuron is defined as the Best Matching Unit (BMU).
Step S22: and taking the minimum value of the distance between the fault early warning data and the optimal matching unit as the minimum quantization error MQE of the fault early warning data. The distance between the fault early warning data and the optimal matching unit can be obtained by a Euclidean distance calculation method.
Preferably, step S4 includes:
step S41: when normal data of an aircraft engine to be early warned are collected, training an SOM (self-organizing map) model of the aircraft engine to be early warned by utilizing the collected normal data of the aircraft engine to be early warned;
step S42: acquiring a minimum quantization error MQE of abnormal data in an aircraft engine to be early warned every time the abnormal data of the aircraft engine to be early warned is acquired;
step S43: and if the minimum quantization error MQE of the abnormal data in the aircraft engine to be early-warned is within the range of the minimum quantization error MQE, performing fault early warning on the aircraft engine to be early-warned.
Preferably, the minimum quantization error MQE range has a lower limit of 80% and an upper limit of 120% of the average value of the minimum quantization error MQE and the average value of the minimum quantization error MQE.
In the implementation process of the embodiment, the normal data, the abnormal data and the early warning interval can be determined in the following ways: the normal data is the data of the first 60% of the service life of the full life cycle data of the aircraft engine; the abnormal data is data of later 40% of service life in the full life cycle data of the aircraft engine; the early warning interval is in the range of 88% -92% of the service life of the full life cycle data of the aircraft engine.
Preferably, the full lifecycle data of the aircraft engine comprises: the full lifecycle data for the aircraft engine comprises: engine intake temperature, exhaust temperature, oil pressure, compressed air pressure, vibration speed parameters, and the like.
Example 2
The embodiment 2 of the invention discloses a training and testing process of the SOM-MQE-based aircraft engine fault early warning method in the embodiment 1, and the specific introduction is as follows:
the full life cycle data of 218 aircraft engines of the same type are selected, and the full life cycle data of each aircraft engine comprises the whole process from a healthy state, a decline state to a failure (scrapping) state of the aircraft engine. The specific contents included in the full life cycle data may be set with reference to embodiment 1. Assuming that the rated service life of such an aircraft engine is 357 cycles, the full life cycle of different aircraft engines does not necessarily correspond to different durations due to factory variations and different wear levels.
In the implementation process, a training set and a test set are divided firstly, full life cycle data of the first 100 aircraft engines are selected as the training set (the 100 aircraft engines are all used as training aircraft engines), and full life cycle data of the last 118 aircraft engines are selected as the test set (the 118 aircraft engines are all used as test aircraft engines). And normal data, abnormal data and early warning intervals are divided according to the mode in the embodiment 1.
The steps S2 and S3 in embodiment 1 (the step of normalizing the full life cycle data may be added before execution) are executed, and a MQE distribution diagram of the fault warning data is obtained, as shown in fig. 2. The average value of the minimum quantization errors MQE of the data (i.e., the fault warning data) in the warning intervals in all the training sets is calculated to be 2.64, and the average value is multiplied by 0.8 to obtain the lower limit of the range of the minimum quantization error MQE (i.e., the lower limit of the threshold of the fault warning data MQE in fig. 2) 2.11, and multiplied by 1.2 to obtain the upper limit of the range of the minimum quantization error MQE (i.e., the upper limit of the threshold of the fault warning data MQE in fig. 2) 3.16.
After the minimum quantization error MQE range is determined, the accuracy of the determined minimum quantization error MQE range needs to be tested. Similarly, the full life cycle data in the test set is normalized, the first 60% of the service life of the engines of the test airplane in the test set is used as normal data to train the SOM, and after the training of the model is finished, the MQE value of the model and the trained SOM is calculated by using the residual data of each engine in the test set. And then calculating the early warning accuracy according to the corresponding content in the embodiment 1.
Taking an engine in the test set as an example (called test equipment No. 1), the first 60% period data of the aircraft engine is selected as normal data for initializing the SOM model, and the remaining period data is used for verifying, as shown in FIG. 3, the broken line is that the remaining period data of the aircraft engine deviates from MQE values of the SOM model. As is evident from fig. 3, the MQE value becomes larger as the operation time of the aircraft engine increases, and the risk of the aircraft engine failing is higher and higher as the operation time deviates from the normal MQE model. The two horizontal lines in fig. 3 are the upper and lower limits of the minimum quantization error MQE range obtained by training according to the training set, that is, when the MQE value in the measured data is within this interval, the device is considered to be in a fault warning state. MQE values of the measured data in all the test set data are calculated, and according to the calculation mode in the embodiment 1, the early warning accuracy of the early warning model can be calculated to be 74.81%, and the requirement of 70% of the early warning accuracy of the equipment fault in the industry under the condition of lacking clear fault label data is met.
In summary, the method for early warning of the fault of the aircraft engine based on the SOM-MQE provided by the embodiment can make full use of massive health data in historical data and use experience to construct a reference model under the condition of lacking a clear fault label, and can judge the health state of the aircraft engine in real time based on the deviation between the observation data and a normal reference model (SOM model), so as to quickly and effectively perform fault early warning on the aircraft engine, reduce the maintenance times caused by governing the unexpected fault of the aircraft engine, and have a significant reference meaning for practical application in the industry.
Those skilled in the art will appreciate that all or part of the flow of the method implementing the above embodiments may be implemented by a computer program, which is stored in a computer readable storage medium, to instruct related hardware. The computer readable storage medium is a magnetic disk, an optical disk, a read-only memory or a random access memory.
The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present invention are included in the scope of the present invention.
Claims (10)
1. An aircraft engine fault early warning method based on SOM-MQE is characterized by comprising the following steps:
acquiring full life cycle data of a plurality of training aircraft engines; the full life cycle data is divided into normal data and abnormal data according to the using duration; the abnormal data comprises fault early warning data in an early warning interval in the whole life cycle;
training an SOM model of a corresponding aircraft engine by using normal data in each training aircraft engine, and obtaining a minimum quantization error MQE of fault early warning data in the aircraft engine;
determining a minimum quantization error MQE range corresponding to an early warning interval based on an average value of the minimum quantization errors MQE of the fault early warning data in the plurality of training aircraft engines;
and performing fault early warning on the aircraft engine to be early warned based on the minimum quantization error MQE range.
2. The SOM-MQE-based aircraft engine fault warning method of claim 1, further comprising, after determining the minimum quantization error MQE range:
training an SOM model of a test aircraft engine by using normal data in the test aircraft engine, and acquiring a minimum quantization error MQE of abnormal data in the test aircraft engine;
acquiring an abnormal data interval of which the minimum quantization error MQE of the abnormal data is within the range of the minimum quantization error MQE;
obtaining an early warning accuracy rate based on the ratio of the intersection of the abnormal data interval and the early warning interval to the early warning interval duration;
and if the accuracy is equal to or higher than a set accuracy threshold, performing fault early warning on the aircraft engine to be early warned based on the minimum quantization error MQE range.
3. The SOM-MQE-based aircraft engine fault warning method of claim 2, further comprising:
if the accuracy is lower than the set accuracy threshold, adjusting the SOM model parameters of the test aircraft engine or the proportion of normal data in the test aircraft engine, retraining the SOM model of the test aircraft engine, and reacquiring the early warning accuracy.
4. An SOM-MQE-based aircraft engine fault early warning method according to claim 2 or 3, further comprising:
if the accuracy is lower than the set accuracy threshold, adjusting one or more of the parameters of the SOM model of the training aircraft engine, the proportion of normal data in the full life cycle data and the range of the early warning interval, and re-determining the range of the minimum quantization error MQE and re-obtaining the early warning accuracy.
5. The SOM-MQE-based aircraft engine fault warning method according to claim 4, wherein adjusting the SOM model parameters of the training aircraft engine and the testing aircraft engine comprises adjusting one or more of a learning rate initialization parameter, an initial win neighborhood, and an initial neuron weight value in the SOM model.
6. An SOM-MQE-based aircraft engine fault early warning method according to claim 1 or 2, wherein the fault early warning of the aircraft engine to be early warned based on the minimum quantitative error MQE range comprises:
when normal data of an aircraft engine to be early warned are collected, training an SOM (self-organizing map) model of the aircraft engine to be early warned by utilizing the collected normal data of the aircraft engine to be early warned;
acquiring a minimum quantization error MQE of abnormal data in an aircraft engine to be early warned every time the abnormal data of the aircraft engine to be early warned is acquired;
and if the minimum quantization error MQE of the abnormal data in the aircraft engine to be early-warned is within the range of the minimum quantization error MQE, performing fault early warning on the aircraft engine to be early-warned.
7. An aircraft engine fault warning method based on SOM-MQE as claimed in claim 1, wherein the minimum quantization error MQE has a lower limit of 80% and an upper limit of 120% of the average value of the minimum quantization error MQE.
8. The SOM-MQE-based aircraft engine fault warning method of claim 1,
the normal data is the data of the first 60% of the service life of the full life cycle data of the aircraft engine;
the abnormal data is data of later 40% of service life in the full life cycle data of the aircraft engine;
the early warning interval is in the range of 88% -92% of the service life of the full life cycle data of the aircraft engine.
9. The SOM-MQE-based aircraft engine fault warning method of claim 1, wherein the full lifecycle data of the aircraft engine comprises: engine intake temperature, exhaust temperature, oil pressure, compressed air pressure, and vibration speed parameters.
10. The SOM-MQE-based aircraft engine fault pre-warning method according to claim 1, wherein the training of the SOM model of the current aircraft engine using the normal data in each aircraft engine and obtaining the minimum quantization error MQE of the fault pre-warning data in the current aircraft engine comprises:
training an SOM (model of mass) model of the current aircraft engine by using normal data in each aircraft engine to obtain an optimal matching unit;
and taking the minimum value of the distance between the fault early warning data and the optimal matching unit as the minimum quantization error MQE of the fault early warning data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011501087.2A CN112528414A (en) | 2020-12-17 | 2020-12-17 | SOM-MQE-based aircraft engine fault early warning method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011501087.2A CN112528414A (en) | 2020-12-17 | 2020-12-17 | SOM-MQE-based aircraft engine fault early warning method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112528414A true CN112528414A (en) | 2021-03-19 |
Family
ID=75001303
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011501087.2A Pending CN112528414A (en) | 2020-12-17 | 2020-12-17 | SOM-MQE-based aircraft engine fault early warning method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112528414A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114154779A (en) * | 2021-10-19 | 2022-03-08 | 北京电子工程总体研究所 | Health assessment and fault diagnosis method and system based on EMD-SOM |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105129109A (en) * | 2015-09-30 | 2015-12-09 | 北京航空航天大学 | Method for evaluating health of aircraft aileron actuator system based on multi-fractal theory and self-organizing map (SOM) network |
| CN106408687A (en) * | 2016-11-24 | 2017-02-15 | 沈阳航空航天大学 | Automobile engine fault early warning method based on machine learning method |
| CN109086804A (en) * | 2018-07-12 | 2018-12-25 | 中石化石油机械股份有限公司 | A kind of hydraulic device fault forecast method merged based on multi source status monitoring information and reliability characteristic |
| CN110410282A (en) * | 2019-07-24 | 2019-11-05 | 河北工业大学 | Wind turbines health status on-line monitoring and method for diagnosing faults based on SOM-MQE and SFCM |
-
2020
- 2020-12-17 CN CN202011501087.2A patent/CN112528414A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105129109A (en) * | 2015-09-30 | 2015-12-09 | 北京航空航天大学 | Method for evaluating health of aircraft aileron actuator system based on multi-fractal theory and self-organizing map (SOM) network |
| CN106408687A (en) * | 2016-11-24 | 2017-02-15 | 沈阳航空航天大学 | Automobile engine fault early warning method based on machine learning method |
| CN109086804A (en) * | 2018-07-12 | 2018-12-25 | 中石化石油机械股份有限公司 | A kind of hydraulic device fault forecast method merged based on multi source status monitoring information and reliability characteristic |
| CN110410282A (en) * | 2019-07-24 | 2019-11-05 | 河北工业大学 | Wind turbines health status on-line monitoring and method for diagnosing faults based on SOM-MQE and SFCM |
Non-Patent Citations (1)
| Title |
|---|
| 董玉亮: "基于保局投影与自组织映射的风电机组故障预警方法", 太阳能学报, vol. 36, no. 5, 31 May 2015 (2015-05-31), pages 1123 - 1129 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114154779A (en) * | 2021-10-19 | 2022-03-08 | 北京电子工程总体研究所 | Health assessment and fault diagnosis method and system based on EMD-SOM |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109947088A (en) | Equipment fault early-warning system based on model lifecycle management | |
| EP1982301B1 (en) | Method of condition monitoring | |
| JP6890527B2 (en) | Systems and methods for obtaining models of predictive inference about behavior, and non-temporary computer-readable storage media for them | |
| AU2013245998A1 (en) | Efficient health management, diagnosis and prognosis of a machine | |
| CN114386312A (en) | Equipment fault diagnosis method | |
| CN112668415A (en) | Aircraft engine fault prediction method | |
| CN111190349A (en) | Method, system and medium for monitoring state and diagnosing fault of ship engine room equipment | |
| CN111079861A (en) | Power distribution network voltage abnormity diagnosis method based on image rapid processing technology | |
| CN114994543A (en) | Energy storage power station battery fault diagnosis method and device and storage medium | |
| CN116720048A (en) | Power station auxiliary machine fault diagnosis method and system based on machine learning model | |
| CN116453438A (en) | Display screen parameter detection method, device, equipment and storage medium | |
| CN115758083A (en) | Motor bearing fault diagnosis method based on time domain and time-frequency domain fusion | |
| CN112528414A (en) | SOM-MQE-based aircraft engine fault early warning method | |
| CN115392782A (en) | Method and system for monitoring and diagnosing health state of process system of nuclear power plant | |
| CN116997867A (en) | Method and system for predicting the operation of a technical installation | |
| US20210080924A1 (en) | Diagnosis Method and Diagnosis System for a Processing Engineering Plant and Training Method | |
| CN117194163A (en) | Computer equipment, fault detection system, method and readable storage medium | |
| CN117052970A (en) | Intelligent control system and method for pneumatic ball valve assembly | |
| CN112286088A (en) | Method and application system for online application of power equipment fault prediction model | |
| CN116204825A (en) | Production line equipment fault detection method based on data driving | |
| TWI794907B (en) | Prognostic and health management system for system management and method thereof | |
| JP7490807B2 (en) | Training artificial intelligence modules for industrial applications | |
| CN112936342B (en) | Physical robot action evaluation system and method based on human body gesture recognition algorithm | |
| Won et al. | Prediction of remaining useful lifetime of membrane using machine learning | |
| KR20230075150A (en) | Method for managing system health |
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 | ||
| CB03 | Change of inventor or designer information | ||
| CB03 | Change of inventor or designer information |
Inventor after: Zhang Rui Inventor after: Qiu Bohua Inventor after: Wei Muheng Inventor after: Zhu Huimin Inventor before: Zhang Yu Inventor before: Qiu Bohua Inventor before: Wei Muheng Inventor before: Zhu Huimin Inventor before: Zhang Rui Inventor before: Tan Xiao |