CN115526241A - Real-time abnormity detection method, device, equipment and medium for aviation hydraulic pump station - Google Patents
Real-time abnormity detection method, device, equipment and medium for aviation hydraulic pump station Download PDFInfo
- Publication number
- CN115526241A CN115526241A CN202211112811.1A CN202211112811A CN115526241A CN 115526241 A CN115526241 A CN 115526241A CN 202211112811 A CN202211112811 A CN 202211112811A CN 115526241 A CN115526241 A CN 115526241A
- Authority
- CN
- China
- Prior art keywords
- hydraulic pump
- pump station
- data
- aviation hydraulic
- sequence data
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 47
- 238000012549 training Methods 0.000 claims abstract description 75
- 230000011218 segmentation Effects 0.000 claims abstract description 63
- 238000000034 method Methods 0.000 claims abstract description 60
- 238000003062 neural network model Methods 0.000 claims abstract description 38
- 230000002159 abnormal effect Effects 0.000 claims description 31
- 230000015654 memory Effects 0.000 claims description 19
- 238000004422 calculation algorithm Methods 0.000 claims description 16
- 230000006870 function Effects 0.000 claims description 13
- 238000010606 normalization Methods 0.000 claims description 11
- 238000004590 computer program Methods 0.000 claims description 10
- 230000005856 abnormality Effects 0.000 claims description 8
- 238000012545 processing Methods 0.000 claims description 8
- 230000004913 activation Effects 0.000 claims description 7
- 238000003860 storage Methods 0.000 claims description 7
- 238000007781 pre-processing Methods 0.000 claims description 6
- 238000011897 real-time detection Methods 0.000 claims description 4
- 238000004140 cleaning Methods 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 claims description 3
- 238000004519 manufacturing process Methods 0.000 abstract description 11
- 230000008569 process Effects 0.000 description 28
- 238000012360 testing method Methods 0.000 description 15
- 238000004891 communication Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 238000013528 artificial neural network Methods 0.000 description 4
- 230000009471 action Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 239000010720 hydraulic oil Substances 0.000 description 2
- 238000012886 linear function Methods 0.000 description 2
- 230000007774 longterm Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 238000013515 script Methods 0.000 description 2
- 239000000725 suspension Substances 0.000 description 2
- 230000006399 behavior Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 230000000630 rising effect Effects 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 238000011895 specific detection Methods 0.000 description 1
- 230000000087 stabilizing effect Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
- 238000013024 troubleshooting Methods 0.000 description 1
Images
Classifications
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F04—POSITIVE - DISPLACEMENT MACHINES FOR LIQUIDS; PUMPS FOR LIQUIDS OR ELASTIC FLUIDS
- F04B—POSITIVE-DISPLACEMENT MACHINES FOR LIQUIDS; PUMPS
- F04B51/00—Testing machines, pumps, or pumping installations
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Mechanical Engineering (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The application discloses a real-time abnormity detection method, device, equipment and medium for an aviation hydraulic pump station, and relates to the technical field of aviation hydraulic pump station detection. The method comprises the steps of obtaining historical time sequence data of a target aviation hydraulic pump station; the historical time sequence data is used for representing the performance of the target aviation hydraulic pump station; performing segmentation fitting on the historical time sequence data to obtain segmentation points; classifying the historical time sequence data after segmentation fitting based on the segmentation points to obtain a training set; training the constructed initial neural network model based on the training set to obtain a data prediction model; and detecting real-time sequence data of the target aviation hydraulic pump station based on the data prediction model. By the aid of the technical scheme, the fault position of the target aviation pump station can be quickly and accurately detected, so that related faults can be more efficiently eliminated by workers, and production efficiency of airplane assembly can be improved.
Description
Technical Field
The application relates to the technical field, in particular to a real-time abnormity detection method, a real-time abnormity detection device, equipment and a medium for an aviation hydraulic pump station.
Background
The hydraulic pump station is widely applied to the industrial manufacturing industry, and the common hydraulic pump station mainly comprises a hydraulic oil pump, a control valve, an actuating cylinder, a hydraulic motor, an auxiliary device and the like. In the field of aircraft manufacturing, a hydraulic pump station is an important special device in the aircraft final assembly link, and is mainly used for providing hydraulic energy for testing actions such as undercarriage, speed reduction plate retraction and release in the aircraft final assembly experiment process.
Because aviation hydraulic power unit system structure is complicated and the part is many, and is higher to operational environment requirement simultaneously, therefore aviation hydraulic power unit breaks down the frequency higher in the use, and most fault conditions have proruption and contingency moreover, and the fault conditions are difficult to predict. Once a hydraulic pump station fails in the using process, the conditions of task suspension, planned delay and the like are caused, and the conditions of product damage and even operation safety accidents are caused. Therefore, the method is very important for detecting the faults of the aviation hydraulic pump station.
In the prior art, equipment managers usually perform inspection and maintenance on equipment regularly after professional technical training, or shut down to perform troubleshooting and analysis on fault reasons after faults occur. Therefore, the faults of the aviation hydraulic pump station cannot be detected quickly, and the production efficiency of airplane assembly is influenced.
Disclosure of Invention
The application mainly aims to provide a real-time abnormity detection method, a real-time abnormity detection device, a real-time abnormity detection equipment and a real-time abnormity detection medium for an aviation hydraulic pump station, and aims to solve the technical problem that faults of the aviation hydraulic pump station cannot be quickly detected in the prior art, so that the assembly production efficiency of an airplane is influenced.
In order to achieve the above object, a first aspect of the present application provides a method for detecting real-time abnormality of an aviation hydraulic pump station, where the method includes:
acquiring historical time sequence data of a target aviation hydraulic pump station; the historical time sequence data is used for representing the performance of the target aviation hydraulic pump station;
performing segmentation fitting on the historical time sequence data to obtain segmentation points;
classifying the historical time sequence data after segmentation fitting based on the segmentation points to obtain a training set;
training the constructed initial neural network model based on the training set to obtain a data prediction model;
and detecting the real-time sequence data of the target aviation hydraulic pump station based on the data prediction model.
Optionally, the performing segment fitting on the historical time series data to obtain segment points includes:
setting a segmentation threshold; the segmentation threshold is a preset error sum of a certain segment of historical time sequence data;
constructing a piecewise fitting model;
and performing segment fitting on the historical time series data based on the segment threshold and the segment fitting model to obtain segment points.
Optionally, the performing segment fitting on the historical time series data based on the segment threshold and the segment fitting model to obtain segment points includes:
and when the total segmentation error in the segmentation fitting model is larger than the segmentation threshold value, obtaining a segmentation point.
Optionally, the building a piecewise-fitting model includes:
constructing a piecewise fitting model through the following relation:
wherein x is i An observed value of historical time series data representing the ith time, F j (t i ) And representing a fitted linear function, J represents a segment error sum in a segment fitting model, jf represents an ending moment of a segment J, js represents a starting moment of the segment J, and K represents a Kth segment.
Optionally, the classifying, based on the segmentation point, the historical time series data after segment fitting to obtain a training set includes:
segmenting the historical time series data based on the number of the segmentation points;
based on the working condition types of the historical time sequence data, dividing the segmented historical time sequence data into a plurality of types;
and obtaining a training set based on the historical time sequence data which is divided into a plurality of classes.
Optionally, training the constructed initial neural network model based on the training set to obtain a data prediction model, including:
obtaining a training time step of each type of historical time sequence data; wherein the training time step is the average density of each type of historical time sequence data in a plurality of times;
and training the constructed initial neural network model based on the BP algorithm, the training set and the training time step to obtain a data prediction model.
Optionally, the obtaining a training set based on the historical time series data classified into several classes includes:
constructing an initial neural network model by the following relation:
z t ＝σ(W z ·[h t-1 ,x t ])
r t ＝σ(W r ·[h t-1 ,x t ])
wherein x is t Input data representing time t, h t Indicating a hidden state at time t, h t-1 Representing the hidden state at time t-1, z t Indicating that the door is updated at time t, r t Indicating that the gate is reset at time t,
representing candidate hidden states at time t, W representing a weight coefficient matrix, W z Represents the weight corresponding to the updated gate at time t, W r Represents the weight corresponding to the reset gate at time t, σ () represents the activation function, and tanh () represents the hyperbolic tangent function.
Optionally, the detecting, based on the data prediction model, the real-time series data of the target aviation hydraulic pump station includes:
acquiring real-time sequence data of the target aviation hydraulic pump station;
preprocessing the real-time sequence data of the target aviation hydraulic pump station;
inputting the preprocessed real-time sequence data of the target aviation hydraulic pump station into the data prediction model to obtain a predicted value of the target aviation hydraulic pump station;
obtaining an actual value of the target aviation hydraulic pump station based on inverse normalization;
and detecting the target aviation hydraulic station in real time based on the predicted value and the actual value of the target aviation hydraulic pump station.
Optionally, the detecting, in real time, the target aviation hydraulic station based on the predicted value and the actual value of the target aviation hydraulic pump station includes:
obtaining the working state of the target aviation hydraulic pump station;
obtaining the standard deviation of each type of historical time sequence data;
obtaining a judgment threshold value based on the working state of the target aviation hydraulic pump station and the standard deviation of each type of historical time sequence data;
and if the absolute value of the difference between the predicted value and the actual value is greater than the judgment threshold, the target aviation hydraulic station has an abnormal condition.
Optionally, before the step of performing segment fitting on the historical time series data to obtain segment points, the method further includes:
cleaning historical time sequence data of the target aviation hydraulic pump station to remove irrelevant historical time sequence data;
carrying out normalization processing on the cleaned historical time sequence data of the target aviation hydraulic pump station;
the performing segment fitting on the historical time series data to obtain segment points includes:
and performing segmentation fitting on the normalized historical time sequence data to obtain segmentation points.
In a second aspect, the application provides a real-time anomaly detection device for an aviation hydraulic pump station, and the device comprises:
the acquisition module is used for acquiring historical time sequence data of the target aviation hydraulic pump station; the historical time sequence data is used for representing the performance of the target aviation hydraulic pump station;
the first obtaining module is used for performing segmentation fitting on the historical time sequence data to obtain segmentation points;
a second obtaining module, configured to classify the historical time series data after segment fitting based on the segment points to obtain a training set;
a third obtaining module, configured to train the constructed initial neural network model based on the training set to obtain a data prediction model;
and the detection module is used for detecting the real-time sequence data of the target aviation hydraulic pump station based on the data prediction model.
In a third aspect, the present application provides a computer device, which includes a memory and a processor, where the memory stores a computer program, and the processor executes the computer program to implement the method described in the embodiment.
In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, wherein a processor executes the computer program to implement the method described in the embodiments.
Through above-mentioned technical scheme, this application has following beneficial effect at least:
the method, the device, the equipment and the medium for detecting the real-time abnormity of the aviation hydraulic pump station are provided by the embodiment of the application, and the method comprises the steps of firstly acquiring historical time sequence data of a target aviation hydraulic pump station; the historical time sequence data is used for representing the performance of the target aviation hydraulic pump station; then processing the historical time sequence data of the target aviation hydraulic pump station; then, performing segmentation fitting on the historical time series data to obtain segmentation points; then based on the segmentation points, classifying the historical time sequence data after segmentation fitting to obtain a training set; training the constructed initial neural network model based on the training set to obtain a data prediction model; and finally, detecting the real-time sequence data of the target aviation hydraulic pump station based on the data prediction model. When the target aviation hydraulic pump is required to be judged whether to be abnormal or not, historical time sequence data of the target aviation hydraulic pump are obtained firstly, then the historical time sequence data are processed and then are subjected to segment fitting to obtain the number of segment points, then the historical time sequence data are classified, and then an initial neural network model is constructed on the basis of the classified historical time sequence data. And finally, inputting the real-time data of the target aviation hydraulic pump station into the data prediction model, and detecting the real-time data of the target aviation hydraulic pump station. Namely, the data prediction model capable of predicting or detecting the real-time sequence data of the target aviation pump station is constructed on the basis of the historical time sequence data of the target aviation hydraulic pump station. When the data prediction model is constructed, the historical time sequence data are subjected to segment fitting and classification, so that the accuracy of the constructed data prediction model is higher, when the target aviation hydraulic pump station is detected, only the real-time sequence data of the target flight hydraulic pump station are input into the data prediction model, and whether the real-time sequence data of the target aviation hydraulic pump station is abnormal or not can be known after certain comparison, and if the real-time data are detected to be abnormal, the target aviation hydraulic pump station is indicated to be abnormal. Therefore, on the premise of ensuring the prediction accuracy of the data prediction model, whether the aviation hydraulic pump station has a fault can be detected more quickly, the fault position of the target aviation hydraulic pump station can be known accurately due to the classification of the historical time sequence data, and related faults can be eliminated more efficiently by workers due to the fact that the fault position of the target aviation pump station can be detected quickly and accurately, so that the production efficiency of airplane assembly can be improved.
Drawings
FIG. 1 is a schematic diagram of a computer device in a hardware operating environment according to an embodiment of the present application;
FIG. 2 is a flowchart of a real-time anomaly detection method for an aviation hydraulic pump station according to an embodiment of the application;
FIG. 3 is a flowchart illustrating a specific implementation of step S11 of the present application;
FIG. 4 is a flowchart illustrating a specific implementation of step S12 of the present application;
FIG. 5 is a schematic structural diagram of an initial neural network model provided in an embodiment of the present application;
FIG. 6 is a flowchart illustrating a specific implementation of step S14 of the present application;
fig. 7 is a schematic diagram of a real-time abnormality detection device for an aviation hydraulic pump station according to an embodiment of the application.
The implementation, functional features and advantages of the objectives of the present application will be further explained with reference to the accompanying drawings.
Detailed Description
It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
The hydraulic pump station is widely applied to the industrial manufacturing industry, and the common hydraulic pump station mainly comprises a hydraulic oil pump, a control valve, an actuating cylinder, a hydraulic motor, an auxiliary device and the like. In the field of aircraft manufacturing, a hydraulic pump station is an important special device in the aircraft final assembly link, and is mainly used for providing hydraulic energy for testing actions such as undercarriage, speed reduction plate retraction and release in the aircraft final assembly experiment process. The final assembly link is a key stage formed by physical and electrical characteristics of the airplane, the final performance of the airplane is directly influenced by the final assembly quality, and in order to ensure the stability, reliability, accuracy and the like of the airplane after delivery, a hydraulic pump station is required to continuously and stably output high pressure in a working state.
Because aviation hydraulic power unit system structure is complicated and the part is many, and is higher to operational environment requirement simultaneously, therefore aviation hydraulic power unit breaks down the frequency higher in the use, and most fault conditions have proruption and contingency moreover, and the fault conditions are difficult to predict. Once a hydraulic pump station breaks down in the using process, the conditions of task suspension, plan delay and the like are caused, and the conditions of product damage and even operation safety accidents are caused. For the problem, usually, the equipment manager regularly checks and maintains the equipment after training by professional techniques, or stops the equipment after a fault occurs to check and analyze the fault reason. Obviously, this way seriously affects the efficiency of the aircraft assembly production. Therefore, according to the technical means of real-time data of equipment operation, improvement of the accuracy of real-time detection of abnormal operation of the equipment and the like, the enhancement of the abnormal real-time detection capability of the aviation hydraulic pump station is a key method for guaranteeing stable operation of the equipment and improving the use efficiency of the equipment.
The real-time anomaly detection technology for the time series data of the industrial equipment mainly comprises the following steps: state temporal data acquisition, data preprocessing, data prediction, abnormality judgment and the like. The data prediction method is the core of the equipment abnormity real-time detection technology, and the accurate analysis and prediction of the time sequence data is the key for improving the equipment abnormity detection accuracy. Commonly used real-time data prediction methods are mainly classified into two categories: the first is algorithms such as supervised convolutional neural networks, decision trees, support vector machines and the like. And secondly, unsupervised algorithms such as auto-regressive moving average (ARMA) and long-short term memory neural network (LSTM). In a time series data real-time abnormity detection model, a supervised time series data abnormity method needs to label all or part of data, and judgment is carried out by learning the difference between normal data and abnormal data. However, in practical applications, it is difficult and expensive to label time series data in most cases, especially to obtain abnormal data in an aviation hydraulic pump station system with a complicated structure.
A real-time anomaly detection algorithm based on autoregressive moving average establishes a prediction model by learning short-term correlation among time sequence data, judges anomaly according to a threshold value on a prediction result, and can extract short-term time sequence dependency relationship in data. The autoregressive moving average algorithm is not suitable for unstable time series data, and meanwhile, the algorithm cannot capture the nonlinear relation in the time series data and learn the long-term dependence relation in the time series data. The algorithm can learn the long-term dependence relationship in the time series data, but the algorithm has longer detection time, and the difference between the reconstruction error intervals of the normal data and the abnormal data is not obvious enough in the abnormal detection stage, so that the detection accuracy and the real-time performance in industrial application can not meet the requirements.
In summary, at present, equipment managers regularly check and maintain the equipment after training with professional skills, or shut down the equipment to check and analyze fault reasons after faults occur. Therefore, the faults of the aviation hydraulic pump station cannot be detected quickly, and the production efficiency of airplane assembly is influenced.
In order to solve the technical problems, the application provides a real-time anomaly detection method, a real-time anomaly detection device, an aviation hydraulic pump station real-time anomaly detection equipment and a medium.
Referring to fig. 1, fig. 1 is a schematic structural diagram of a computer device in a hardware operating environment according to an embodiment of the present application.
As shown in fig. 1, the computer apparatus may include: a processor 1001, such as a Central Processing Unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. Wherein a communication bus 1002 is used to enable connective communication between these components. The user interface 1003 may include a Display screen (Display), an input unit such as a Keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface, a wireless interface. The network interface 1004 may optionally include a standard wired interface, a WIreless interface (e.g., a WIreless-FIdelity (WI-FI) interface). The Memory 1005 may be a Random Access Memory (RAM) Memory, or may be a Non-Volatile Memory (NVM), such as a disk Memory. The memory 1005 may alternatively be a storage device separate from the processor 1001.
Those skilled in the art will appreciate that the configuration shown in FIG. 1 does not constitute a limitation of a computer device and may include more or fewer components than those shown, or some components may be combined, or a different arrangement of components.
As shown in fig. 1, a memory 1005, which is a storage medium, may include therein an operating system, a data storage module, a network communication module, a user interface module, and an electronic program.
In the computer device shown in fig. 1, the network interface 1004 is mainly used for data communication with a network server; the user interface 1003 is mainly used for data interaction with a user; the processor 1001 and the memory 1005 in the computer device of the present invention may be disposed in the computer device, and the computer device invokes the real-time anomaly detection apparatus of the aviation hydraulic pump station stored in the memory 1005 through the processor 1001, and executes the real-time anomaly detection method of the aviation hydraulic pump station provided in the embodiment of the present invention.
Referring to fig. 2, based on the hardware environment of the foregoing embodiment, an embodiment of the present application provides a real-time anomaly detection method for an aviation hydraulic pump station, where the method includes:
s10: acquiring historical time sequence data of a target aviation hydraulic pump station; and the historical time sequence data is used for representing the performance of the target aviation hydraulic pump station.
In the specific implementation process, the target aviation hydraulic pump station refers to an aviation hydraulic pump station which needs to detect whether an abnormal problem exists. The time sequence data refers to time sequence data, the time sequence data is a data sequence of the same unified index recorded according to the time sequence, the historical time sequence data refers to time sequence data before the detection time of the target aviation hydraulic pump station, the time sequence data can directly reflect whether the aviation hydraulic pump station has an abnormal problem or not, the historical time sequence data of the target aviation hydraulic pump station comprises real-time data such as the temperature, the pressure and the flow of the hydraulic pump station, the performance of the target aviation hydraulic pump station can be exactly presented by the real-time data such as the temperature, the pressure and the flow of the hydraulic pump station, and the real-time data such as the temperature, the pressure and the flow of the hydraulic pump station are easier to obtain, so that the abnormal situation of the target aviation hydraulic pump station is easier to detect.
S11: and performing segmentation fitting on the historical time sequence data to obtain segmentation points.
In the specific implementation process, the obtained historical time sequence data is long and needs to be segmented, and when the historical time sequence data needs to be segmented, segmentation points need to be obtained, and then the historical time sequence data is segmented according to the segmentation points.
S12: and classifying the historical time sequence data after the segmentation fitting based on the segmentation points to obtain a training set.
In the specific implementation process, based on the number of the segmentation points obtained in the step S11, the historical time sequence data is segmented, then the segmented historical time sequence data is divided into a plurality of classes by adopting a K-means algorithm according to the working condition types of the aviation hydraulic pump station data, and then the historical time sequence data divided into the plurality of classes is divided into a training set and a test set, so that the accuracy of a data prediction model for subsequent training can be higher based on the training set containing the classification standard.
S13: and training the constructed initial neural network model based on the training set to obtain a data prediction model.
In a specific implementation process, a sliding window is used to create a data sample, historical time series data is used as the data sample, and the data sample (the classified historical time series data) is divided into a training set and a test set, for example, 70% of the historical time series data is used as the training set, and the remaining 30% of the historical time series data is used as the test set. Training the initial neural network model through the data of the training set, testing the trained initial neural network model through the data of the testing set, and obtaining a data prediction model capable of detecting real-time sequence data of the target aviation hydraulic pump station after the test is passed.
S14: and detecting the real-time sequence data of the target aviation hydraulic pump station based on the data prediction model.
In the specific implementation process, when the real-time sequence data of the target aviation hydraulic pump station are detected, the real-time sequence data of the target aviation hydraulic pump station are input into the data prediction model, the data prediction model automatically gives a predicted value, whether the real-time sequence data of the target aviation hydraulic pump station are abnormal or not can be known through the predicted value, and if the real-time sequence data of the target aviation hydraulic pump station are abnormal, the target aviation hydraulic pump station at the moment is considered to have abnormal conditions.
In summary, when it is necessary to determine whether the target aviation hydraulic pump is abnormal, historical time series data of the target aviation hydraulic pump is obtained first, then the historical time series data is processed and then is subjected to segment fitting to obtain the number of segment points, then the historical time series data is classified, and then an initial neural network model is constructed based on the classified historical time series data. And finally, inputting the real-time data of the target aviation hydraulic pump station into the data prediction model, and detecting the real-time data of the target aviation hydraulic pump station. Namely, the data prediction model capable of predicting or detecting the real-time sequence data of the target aviation pump station is constructed on the basis of the historical time sequence data of the target aviation hydraulic pump station. When the data prediction model is constructed, the historical time sequence data are subjected to segment fitting and classification, so that the accuracy of the constructed data prediction model is higher, when the target aviation hydraulic pump station is detected, only the real-time sequence data of the target flight hydraulic pump station are input into the data prediction model, and whether the real-time sequence data of the target aviation hydraulic pump station is abnormal or not can be known after certain comparison, and if the real-time data are detected to be abnormal, the target aviation hydraulic pump station is indicated to be abnormal. Therefore, on the premise of ensuring the prediction accuracy of the data prediction model, whether the aviation hydraulic pump station has a fault can be detected more quickly, the fault position of the target aviation hydraulic pump station can be known accurately due to the classification of the historical time sequence data, and related faults can be eliminated more efficiently by workers due to the fact that the fault position of the target aviation pump station can be detected quickly and accurately, so that the production efficiency of airplane assembly can be improved.
In order to better segment the historical time series data, in some embodiments, as shown in fig. 3, an optional technical solution is provided, that is, the step of performing segment fitting on the historical time series data to obtain segment points includes:
s111: setting a segmentation threshold; the segmentation threshold is the sum of errors preset by a certain segment of historical time sequence data.
In the specific implementation process, the segmentation threshold is preset and can reflect the value of the sum of errors of a certain section of historical time sequence data.
S112: and constructing a piecewise fitting model.
In a specific implementation process, a sliding window method is adopted to perform segmented fitting on historical time series data to obtain segmented points of the data, the segmented points can be obtained more quickly and accurately through a segmented fitting model, and specifically, the segmented fitting model is constructed through the following relational expression:
wherein x is i Showing the observed value of the historical time series data at the ith time, F j (t i ) And a linear function of the fitting is represented, J represents the sum of segment errors in the segment fitting model, jf represents the ending moment of the segment J, js represents the starting moment of the segment J, and K represents the Kth segment.
S113: and performing segment fitting on the historical time sequence data based on the segment threshold and the segment fitting model to obtain segment points.
In a specific implementation process, when the sum of the segmentation errors in the segmentation fitting model is greater than a segmentation threshold value, a segmentation point is obtained, and a new segmentation is started.
To build more accurate, in some embodiments, as shown in fig. 4, the step of classifying the historical time series data after segment fitting based on the segmentation points to obtain a training set includes:
s121: segmenting the historical timing data based on the number of segmentation points.
In a specific implementation process, for example, W segmentation points are obtained, the historical time series data may be divided into W +1 segments.
S122: and based on the working condition types of the historical time sequence data, dividing the segmented historical time sequence data into a plurality of types.
In the specific implementation process, according to the working condition type of the target aviation hydraulic pump station data, the historical time sequence data are divided into a plurality of types by adopting a K-means algorithm.
S123: and obtaining a training set based on the historical time sequence data which is divided into a plurality of classes.
In the specific implementation process, the historical time sequence data is segmented, and then the segmented historical time sequence data is classified. Therefore, more accurate data information can be provided for the training of the subsequent initial neural network model, the initial neural network model can be conveniently fitted with the actual situation better, and the real-time data of the target aviation hydraulic pump station detected by the initial neural network model can be more accurate.
Specifically, an initial neural network model is constructed through the following relational expression:
z t ＝σ(W z ·[h t-1 ,x t ])
r t ＝σ(W r ·[h t-1 ,x t ])
wherein x is t Input data representing time t, h t Indicating a hidden state at time t, h t-1 Representing the hidden state at time t-1, z t Indicating that the door is updated at time t, r t Indicating that the gate is reset at time t,
representing candidate hidden states at time t, W representing a weight coefficient matrix, W z Indicates the weight corresponding to the updated gate at time t, W r Represents the weight corresponding to the reset gate at time t, σ () represents the activation function, and tanh () represents the hyperbolic tangent function.
Fig. 5 shows the constructed initial neural network model, where fig. 5 is a schematic diagram of the initial neural network model provided in the embodiment of the present application, and in fig. 5, σ represents an activation function, tanh represents a hyperbolic tangent function, and x t Input data representing time t, h t Indicating a hidden state at time t, h t-1 Representing the hidden state at time t-1, z t Indicating that the door is updated at time t, r t Indicating that the gate is reset at time t,
representing the candidate hidden state at time t. The initial neural network model preferably selects a GRU neural network, an input layer, a hidden layer and an output layer of the GRU neural network are firstly constructed, and then the number of neurons in each layer and corresponding activation functions are determined.
In order to train the initial neural network model better, in some embodiments, the following preferred technical solution is implemented, that is, the step of training the constructed initial neural network model based on the training set to obtain the data prediction model includes: firstly, obtaining the training time step of each type of historical time sequence data; wherein the training time step is the average density of each type of historical time sequence data in a plurality of times; dividing the historical time sequence data into a training set and a testing set; and finally, training the constructed initial neural network model based on the BP algorithm, the training set and the training time step to obtain a data prediction model.
In this embodiment, the average density of each type of historical time series data in the time t is calculated
And taking the average density of each type of historical time sequence data as a time step when the type of historical time sequence data is trained. The data samples are created using a sliding window, preferably 70% of the data samples are used as a training set, and the remaining 30% are used as a testing set. And then training the GRU initial neural network model by adopting a BP algorithm, iteratively calculating network errors and updating weights after the input of each data segment is finished, terminating the training of the GRU initial neural network model when the accuracy on the test set is not improved any more, and storing the trained GRU data prediction model to obtain the data prediction model which can be directly used for real-time sequence data of the aviation hydraulic pump station.
When the trained data prediction model is used for predicting real-time sequence data of an aviation pump station, as shown in fig. 6, a preferred embodiment is shown, that is, the step of detecting the real-time sequence data of the target aviation hydraulic pump station based on the data prediction model includes:
s141: and acquiring real-time sequence data of the target aviation hydraulic pump station.
In a specific implementation process, the real-time sequence data refers to current real-time sequence data of the target aviation hydraulic pump station, the real-time sequence data includes, but is not limited to, real-time temperature, real-time pressure, real-time flow and the like of the target aviation hydraulic pump station, and the real-time sequence data can be obtained through a conventional means.
S142: and preprocessing the real-time sequence data of the target aviation hydraulic pump station.
In a specific implementation process, the preprocessing of the real-time sequence data comprises formatting the real-time sequence data detected by the hydraulic pump station and normalizing the real-time sequence data detected by the hydraulic pump station.
S143: inputting the preprocessed real-time sequence data of the target aviation hydraulic pump station into the data prediction model to obtain a predicted value of the target aviation hydraulic pump station.
In the specific implementation process, the processed real-time sequence data of the target aviation hydraulic pump station is input into the data prediction model in a conventional mode, and the predicted value of the target aviation hydraulic pump station can be obtained after the data prediction model is processed.
S144: and obtaining the actual value of the target aviation hydraulic pump station based on inverse normalization.
In a specific implementation process, the inverse normalization is a conventional means, and after the inverse normalization is performed, an actual value of the target aviation hydraulic pump station can be obtained, wherein the actual value of the target aviation hydraulic pump station refers to an actual value of the real-time sequence data of the target aviation hydraulic pump station.
S145: and detecting the target aviation hydraulic station in real time based on the predicted value and the actual value of the target aviation hydraulic pump station.
In the specific implementation process, whether the real-time sequence data of the target aviation hydraulic pump station is abnormal or not can be detected by comparing a predicted value obtained by the data prediction model with an actual value obtained by inverse normalization, and the specific detection mode is as follows:
firstly, obtaining the working state of the target aviation hydraulic pump station; then obtaining the standard deviation of each type of historical time sequence data; obtaining a judgment threshold value based on the working state of the target aviation hydraulic pump station and the standard deviation of each type of historical time sequence data; and finally, if the absolute value of the difference between the predicted value and the actual value is greater than the judgment threshold, the target aviation hydraulic station has an abnormal condition.
In the embodiment, the current working state of the target hydraulic pump station is determined according to the actual running condition of the current target hydraulic pump station, then a judgment threshold value is determined and selected according to the current working state, the judgment threshold value refers to a preset range between a predicted value and an actual value, the absolute value of the difference between the predicted value and the actual value is compared with the judgment threshold value according to the current working state, if the absolute value of the difference between the predicted value and the actual value is greater than the judgment threshold value, the real-time sequence data of the target aviation hydraulic pump station is judged to be abnormal, and if the real-time sequence data is abnormal, the abnormal situation of the target aviation hydraulic pump station is indicated; and if the absolute value of the difference between the predicted value and the actual value is less than or equal to a judgment threshold value, judging that the real-time sequence data is normal.
In order to better establish a data prediction model for historical time series data, in some embodiments, an optional technical solution is provided, that is, before the step of performing segment fitting on the historical time series data to obtain segment points, the method further includes:
cleaning historical time sequence data of the target aviation hydraulic pump station to remove irrelevant historical time sequence data; and then, carrying out normalization processing on the cleaned historical time sequence data of the target aviation hydraulic pump station.
The step of segment fitting the historical time series data to obtain segment points comprises: and performing segmentation fitting on the normalized historical time series data to obtain segmentation points.
In this embodiment, a local mean interpolation method is used to perform missing value processing on historical time series data, specifically, the missing value processing may be performed through the following relation:
wherein the content of the first and second substances,
representing an estimated value at time t, n representing a look-back at t times, y i Indicating the observed value at time i. Then, the date format data such as time and the like is converted into the long integer format type data for subsequent calculation. And removing meaningless repeated data of the equipment in the standby state, counting the positions and the number of the standby state data subsequences in the data, only keeping C data when the continuous number exceeds a threshold value C, removing noise from historical time sequence data, and removing obvious noise data. And finally, performing normalization processing on the historical time sequence data by adopting a Z-score method to ensure that the historical time sequence data conforms to standard normal distribution, so that the convergence speed of the historical time sequence data is accelerated when the historical time sequence data is trained. After the historical time series data are processed, some noise data can be removed, so that the efficiency of subsequently constructing a data prediction model can be improved, and the accuracy of the data prediction model can be improved.
The scheme of the invention adopts a dynamic time step-based time sequence data real-time anomaly detection algorithm of GRU combined working condition types. An example of real-time anomaly detection of pressure time sequence data of an aviation hydraulic pump station by applying the method is listed. Specific examples are:
1. and (5) preprocessing the time sequence data of the hydraulic pump station.
(1) Firstly, acquiring key real-time sequence data such as pressure, time, equipment number and the like of a hydraulic pump station from a redis time sequence database by using Python, and storing the key real-time sequence data in a local disk file system in a CSV format;
(2) reading CSV data into a two-dimensional array, wherein the row of the array represents the state of the pressure of a hydraulic pump station at a certain moment, and the column of the array represents the observed value of one pressure at all moments;
(3) traversing each element in the array, and filling missing elements with the missing values by adopting the average value of the previous 15 data at the current moment;
(4) sequencing the behavior units of the array from small to large according to time, and converting the time into timestamp digits;
(5) counting the number and positions of continuous zero values in all pressure data, and ensuring that no more than 180 continuous zero-value sequences exist according to actual use conditions;
(6) denoising the time sequence data of the hydraulic pump station, judging that the process file is noisy if the numerical value is greater than a given threshold value of the process file (if the pressure exceeds 54 Mpa) according to the process file used by the hydraulic pump station, and deleting the data in the original data set, wherein the data smaller than the threshold value are not processed, and finally 4000 target data are obtained;
(7) and normalizing the pressure time sequence data of the hydraulic pump station by adopting a z-score method.
2. And (5) an off-line training stage.
(1) Using the formula
For target data D = { x 1 ,x 2 ,...,x n Piecewise linear fitting is carried out to obtain a fitted piecewise point D '= { x' 1 ,x′ 2 ,...,x′ k A total of 39 segmentation points;
(2) dividing 4000 data into 40 sections by using 39 section points, and dividing the section data into 5 categories by adopting a K-means algorithm according to 5 working conditions (including preparation, pressure rising, pressure stabilizing, pressure lowering and stopping) of the hydraulic pump station;
(3) calculating the mean value u of the 5 data in the category i And standard deviation σ i ；
(4) Calculating the average density of the data within 30S seconds for each type of data to obtain the average density of the five types of data within 30S
Taking T as the time step of each type of data training;
(5) a sliding window of size 3 is used to create data samples within each segment, and a single training sample is shaped as: x t ＝(x t-3 ,x t-2 ,x t-1 )，Y t ＝x t Wherein X is t As a characteristic attribute, Y t For the sample label, 3732 sample data are obtained in total, and the number of samples S = { S } in each segment 1 ,s 2 ,..,s 40 }；
(6) Taking 70% of 3732 sample data as a training set, and taking the rest 30% as a test set;
(7) constructing a GRU initial neural network model with an input layer size of 3, a hidden layer size of 48 and an output layer size of 1, wherein a Sigmoid function is used as an activation function by a forgetting gate, an input gate and an output gate; when generating the candidate memory, using a hyperbolic tangent function tanh as an activation function;
(8) will be provided with
The samples are output in 40 times in the GRU initial neural network model in a form, and the time step of each time
Determining according to the category of the data;
(9) adopting a back propagation training GRU initial neural network model, after current data is input in a segmented mode, iteratively calculating network errors and updating weights until all samples are trained, testing the accuracy on a test set, repeatedly iterating the training model, and finally terminating network training when the accuracy on the test set is not improved any more;
the trained GRU model is finally saved in r.
3. And (5) an online detection stage.
(1) For real-time data y t Carrying out pretreatment;
(2) load-trained GRU initial neural network model prediction current value
(3) Inverse normalization is carried out on the predicted value to obtain an actual value
(4) Selecting a judgment threshold value sigma according to the working condition of the current target hydraulic pump station 2 ；
(5) Judgment of
If the inequality is not established, judging the inequality to be normal, and if the inequality is not established, judging the inequality to be abnormal;
(6) and outputting the data value detected as abnormal and the corresponding time.
In conclusion, before the real-time sequence data of the target hydraulic pump station is predicted, the priori knowledge of the application process is fused according to the use characteristics of the aviation hydraulic pump station, the historical time sequence data of the target aviation hydraulic pump station is analyzed and subjected to segment fitting, then classification is carried out, more accurate data information is provided for the subsequent training of the initial neural network model, and the initial neural network model can be conveniently used for better fitting the actual situation. According to the segmentation of the time sequence data, the neural network is trained by adopting the dynamic time step length, so that the network model is more consistent with the real situation of the change of the time sequence data, and the accuracy of network prediction is further improved. The density, the data mean value and the variation trend of time sequence data generated by the target aviation hydraulic pump station in different working states are different, and the same fixed threshold value is adopted, so that the abnormal conditions under different working conditions can not be accurately judged obviously, and even misjudgment can be caused. Therefore, according to the working state of the hydraulic pump station, the judgment threshold is selected by combining the standard differential state of the data type, and the abnormal condition of the complex aviation hydraulic pump station under the complex condition can be judged more accurately.
In another embodiment, as shown in fig. 7, based on the same inventive concept as the foregoing embodiment, an embodiment of the present application further provides a real-time abnormality detection apparatus for an aviation hydraulic pump station, where the apparatus includes:
the acquisition module is used for acquiring historical time sequence data of the target aviation hydraulic pump station; the historical time sequence data is used for representing the performance of the target aviation hydraulic pump station;
the first obtaining module is used for performing segmentation fitting on the historical time sequence data to obtain segmentation points;
a second obtaining module, configured to classify the historical time series data after segment fitting based on the segment points to obtain a training set;
a third obtaining module, configured to train the constructed initial neural network model based on the training set to obtain a data prediction model;
and the detection module is used for detecting the real-time sequence data of the target aviation hydraulic pump station based on the data prediction model.
It should be noted that, in this embodiment, each module in the real-time abnormality detection apparatus for an aviation hydraulic pump station corresponds to each step in the real-time abnormality detection method for an aviation hydraulic pump station in the foregoing embodiment one to one, and therefore, the specific implementation and the achieved technical effect of this embodiment may refer to the implementation of the real-time abnormality detection method for an aviation hydraulic pump station, which is not described herein again.
Furthermore, in an embodiment, the present application also provides a computer device comprising a processor, a memory and a computer program stored in the memory, which when executed by the processor implements the method in the preceding embodiment.
Furthermore, in an embodiment, the present application further provides a computer storage medium having a computer program stored thereon, where the computer program is executed by a processor to implement the method in the foregoing embodiment.
In some embodiments, the computer-readable storage medium may be memory such as FRAM, ROM, PROM, EPROM, EEPROM, flash, magnetic surface memory, optical disk, or CD-ROM; or may be various devices including one or any combination of the above memories. The computer may be a variety of computing devices including intelligent terminals and servers.
In some embodiments, the executable instructions may be in the form of a program, software module, script, or code written in any form of programming language, including compiled or interpreted languages, or declarative or procedural languages, and it may be deployed in any form, including as a stand-alone program or as a module, component, subroutine, or other unit suitable for use in a computing environment.
By way of example, executable instructions may correspond, but do not necessarily have to correspond, to files in a file system, and may be stored in a portion of a file that holds other programs or data, such as in one or more scripts in a hypertext Markup Language (HTML) document, in a single file dedicated to the program in question, or in multiple coordinated files (e.g., files that store one or more modules, sub-programs, or portions of code).
By way of example, executable instructions may be deployed to be executed on one computing device or on multiple computing devices at one site or distributed across multiple sites and interconnected by a communication network.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or system. Without further limitation, an element defined by the phrases "comprising one of 8230; \8230;" 8230; "does not exclude the presence of additional like elements in a process, method, article, or system that comprises the element.
The above-mentioned serial numbers of the embodiments of the present application are merely for description and do not represent the merits of the embodiments.
Through the above description of the embodiments, those skilled in the art will clearly understand that the method of the above embodiments can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware, but in many cases, the former is a better implementation manner. Based on such understanding, the technical solutions of the present application or portions thereof contributing to the prior art may be embodied in the form of a software product, which is stored in a storage medium (such as a rom/ram, a magnetic disk, and an optical disk), and includes instructions for enabling a multimedia terminal device (which may be a mobile phone, a computer, a television receiver, or a network device) to execute the method according to the embodiments of the present application.
The above description is only a preferred embodiment of the present application, and not intended to limit the scope of the present application, and all modifications of equivalent structures and equivalent processes, which are made by the contents of the specification and the drawings of the present application, or which are directly or indirectly applied to other related technical fields, are included in the scope of the present application.
Claims (13)
1. A real-time anomaly detection method for an aviation hydraulic pump station is characterized by comprising the following steps:
acquiring historical time sequence data of a target aviation hydraulic pump station; the historical time sequence data is used for representing the performance of the target aviation hydraulic pump station;
performing segmentation fitting on the historical time series data to obtain segmentation points;
classifying the historical time sequence data after segmentation fitting based on the segmentation points to obtain a training set;
training the constructed initial neural network model based on the training set to obtain a data prediction model;
and detecting real-time sequence data of the target aviation hydraulic pump station based on the data prediction model.
2. The real-time anomaly detection method for the aviation hydraulic pump station according to claim 1, wherein the step of performing segment fitting on the historical time series data to obtain segment points comprises the following steps:
setting a segmentation threshold; the segmentation threshold is a preset error sum of a certain segment of historical time series data;
constructing a piecewise fitting model;
and performing segment fitting on the historical time sequence data based on the segment threshold and the segment fitting model to obtain segment points.
3. The real-time anomaly detection method for the aviation hydraulic pump station according to claim 2, wherein the step of performing segment fitting on the historical time series data based on the segment threshold and the segment fitting model to obtain segment points comprises the following steps:
and when the total segmentation error in the segmentation fitting model is larger than the segmentation threshold value, obtaining a segmentation point.
4. The real-time anomaly detection method for the aviation hydraulic pump station according to claim 3,
the building of the piecewise fitting model comprises the following steps:
constructing a piecewise fitting model by the following relation:
wherein x is i Showing the observed value of the historical time series data at the ith time, F j (t i ) And J represents the sum of the segment errors in the segment fitting model, jf represents the ending time of the segment J, and js represents the starting time of the segment J.
5. The real-time anomaly detection method for the aviation hydraulic pump station according to claim 1, wherein the classifying the historical time series data after segment fitting based on the segment points to obtain a training set comprises:
segmenting the historical time series data based on the number of the segmentation points;
based on the working condition types of the historical time sequence data, dividing the segmented historical time sequence data into a plurality of types;
and obtaining a training set based on the historical time sequence data which is divided into a plurality of classes.
6. The real-time anomaly detection method for the aviation hydraulic pump station according to claim 1, wherein the training of the constructed initial neural network model based on the training set to obtain the data prediction model comprises:
obtaining a training time step of each type of historical time sequence data; wherein the training time step is the average density of each type of historical time sequence data in a plurality of times;
and training the constructed initial neural network model based on the BP algorithm, the training set and the training time step to obtain a data prediction model.
7. The real-time anomaly detection method for the aviation hydraulic pump station according to claim 6, wherein the training of the constructed initial neural network model based on the BP algorithm, the training set and the training time step to obtain the data prediction model comprises:
constructing an initial neural network model by the following relation:
z t ＝σ(W z ·[h t-1 ,x t ])
r t ＝σ(W r ·[h t-1 ,x t ])
wherein x is t Input data representing time t, h t Indicating a hidden state at time t, h t-1 Representing the hidden state at time t-1, z t Indicates the update of the gate at time t, r t Indicating that the gate is reset at time t,
representing candidate hidden states at time t, W representing a weight coefficient matrix, W z Indicates the weight corresponding to the updated gate at time t, W r Represents the weight corresponding to the reset gate at time t, σ () represents the activation function, and tanh () represents the hyperbolic tangent function.
8. The real-time anomaly detection method for the aviation hydraulic pump station according to claim 1, wherein the detecting the real-time sequence data of the target aviation hydraulic pump station based on the data prediction model comprises:
acquiring real-time sequence data of the target aviation hydraulic pump station;
preprocessing the real-time sequence data of the target aviation hydraulic pump station;
inputting the preprocessed real-time sequence data of the target aviation hydraulic pump station into the data prediction model to obtain a predicted value of the target aviation hydraulic pump station;
obtaining an actual value of the target aviation hydraulic pump station based on inverse normalization;
and detecting the target aviation hydraulic station in real time based on the predicted value and the actual value of the target aviation hydraulic pump station.
9. The real-time anomaly detection method for the aviation hydraulic pump station according to claim 8, wherein the real-time detection of the target aviation hydraulic pump station based on the predicted value and the actual value of the target aviation hydraulic pump station comprises:
obtaining the working state of the target aviation hydraulic pump station;
obtaining a standard deviation of each type of the historical time sequence data;
obtaining a judgment threshold value based on the working state of the target aviation hydraulic pump station and the standard deviation of each type of historical time sequence data;
and if the absolute value of the difference between the predicted value and the actual value is greater than the judgment threshold, the target aviation hydraulic station has an abnormal condition.
10. The method for detecting the real-time abnormality of the aviation hydraulic pump station according to claim 1, wherein before the step of performing segment fitting on the historical time series data to obtain segment points, the method further comprises:
cleaning historical time sequence data of the target aviation hydraulic pump station to remove irrelevant historical time sequence data;
carrying out normalization processing on the cleaned historical time sequence data of the target aviation hydraulic pump station;
the performing segment fitting on the historical time series data to obtain segment points includes:
and performing segmentation fitting on the normalized historical time sequence data to obtain segmentation points.
11. The utility model provides an aviation hydraulic power unit real-time anomaly detection device which characterized in that, the device includes:
the acquisition module is used for acquiring historical time sequence data of the target aviation hydraulic pump station; the historical time sequence data is used for representing the performance of the target aviation hydraulic pump station;
the first obtaining module is used for performing segmentation fitting on the historical time sequence data to obtain segmentation points;
a second obtaining module, configured to classify the historical time series data after segment fitting based on the segment points to obtain a training set;
a third obtaining module, configured to train the constructed initial neural network model based on the training set to obtain a data prediction model;
and the detection module is used for detecting the real-time sequence data of the target aviation hydraulic pump station based on the data prediction model.
12. A computer arrangement, characterized in that the computer arrangement comprises a memory in which a computer program is stored and a processor which executes the computer program for implementing the method as claimed in any one of claims 1-10.
13. A computer-readable storage medium, having stored thereon a computer program, which, when executed by a processor, performs the method of any one of claims 1-10.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211112811.1A CN115526241A (en) | 2022-09-13 | 2022-09-13 | Real-time abnormity detection method, device, equipment and medium for aviation hydraulic pump station |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211112811.1A CN115526241A (en) | 2022-09-13 | 2022-09-13 | Real-time abnormity detection method, device, equipment and medium for aviation hydraulic pump station |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115526241A true CN115526241A (en) | 2022-12-27 |
Family
ID=84697010
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211112811.1A Pending CN115526241A (en) | 2022-09-13 | 2022-09-13 | Real-time abnormity detection method, device, equipment and medium for aviation hydraulic pump station |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115526241A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115879248A (en) * | 2023-03-03 | 2023-03-31 | 山东亿宁环保科技有限公司 | Full life cycle management method and system suitable for vacuum pump |
| CN116517921A (en) * | 2023-07-03 | 2023-08-01 | 成都飞机工业（集团）有限责任公司 | On-line detection method and system for aviation hydraulic oil vehicle state |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110879874A (en) * | 2019-11-15 | 2020-03-13 | 北京工业大学 | Astronomical big data optical variation curve abnormity detection method |
| CN111523707A (en) * | 2020-04-08 | 2020-08-11 | 江苏方天电力技术有限公司 | Method for modeling variable working condition characteristics of comprehensive energy system equipment based on segmented data fitting |
| WO2021212752A1 (en) * | 2020-04-23 | 2021-10-28 | 平安科技（深圳）有限公司 | Device index data-based anomaly detection method and apparatus, device, and storage medium |
| CN114579397A (en) * | 2022-02-22 | 2022-06-03 | 湖北美和易思教育科技有限公司 | Anomaly detection method and system based on data mining |
| CN114648227A (en) * | 2022-03-23 | 2022-06-21 | 成都飞机工业（集团）有限责任公司 | Aviation hydraulic pump station time sequence abnormity detection method, device, equipment and storage medium |
-
2022
- 2022-09-13 CN CN202211112811.1A patent/CN115526241A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110879874A (en) * | 2019-11-15 | 2020-03-13 | 北京工业大学 | Astronomical big data optical variation curve abnormity detection method |
| CN111523707A (en) * | 2020-04-08 | 2020-08-11 | 江苏方天电力技术有限公司 | Method for modeling variable working condition characteristics of comprehensive energy system equipment based on segmented data fitting |
| WO2021212752A1 (en) * | 2020-04-23 | 2021-10-28 | 平安科技（深圳）有限公司 | Device index data-based anomaly detection method and apparatus, device, and storage medium |
| CN114579397A (en) * | 2022-02-22 | 2022-06-03 | 湖北美和易思教育科技有限公司 | Anomaly detection method and system based on data mining |
| CN114648227A (en) * | 2022-03-23 | 2022-06-21 | 成都飞机工业（集团）有限责任公司 | Aviation hydraulic pump station time sequence abnormity detection method, device, equipment and storage medium |
Non-Patent Citations (1)
| Title |
|---|
| 车畅畅;王华伟;倪晓梅;付强;: "基于改进GRU的航空发动机剩余寿命预测", 航空计算技术, no. 01 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115879248A (en) * | 2023-03-03 | 2023-03-31 | 山东亿宁环保科技有限公司 | Full life cycle management method and system suitable for vacuum pump |
| CN116517921A (en) * | 2023-07-03 | 2023-08-01 | 成都飞机工业（集团）有限责任公司 | On-line detection method and system for aviation hydraulic oil vehicle state |
| CN116517921B (en) * | 2023-07-03 | 2023-11-10 | 成都飞机工业（集团）有限责任公司 | On-line detection method and system for aviation hydraulic oil vehicle state |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2023071217A1 (en) | Multi-working-condition process industrial fault detection and diagnosis method based on deep transfer learning | |
| Javed et al. | State of the art and taxonomy of prognostics approaches, trends of prognostics applications and open issues towards maturity at different technology readiness levels | |
| CN115526241A (en) | Real-time abnormity detection method, device, equipment and medium for aviation hydraulic pump station | |
| JP6740247B2 (en) | Anomaly detection system, anomaly detection method, anomaly detection program and learned model generation method | |
| CN113496262B (en) | Data-driven active power distribution network abnormal state sensing method and system | |
| CN111210024A (en) | Model training method and device, computer equipment and storage medium | |
| CN113312447B (en) | Semi-supervised log anomaly detection method based on probability label estimation | |
| CN113642754B (en) | Complex industrial process fault prediction method based on RF noise reduction self-coding information reconstruction and time convolution network | |
| CN109766583A (en) | Based on no label, unbalanced, initial value uncertain data aero-engine service life prediction technique | |
| CN108490923B (en) | System design method for detecting and positioning tiny faults of electric traction system | |
| CN111709447A (en) | Power grid abnormality detection method and device, computer equipment and storage medium | |
| CN112257755A (en) | Method and device for analyzing operating state of spacecraft | |
| CN110570544A (en) | method, device, equipment and storage medium for identifying faults of aircraft fuel system | |
| CN115760484A (en) | Method, device and system for improving potential danger identification capability of power distribution station area and storage medium | |
| JP6971514B1 (en) | Information processing equipment, information processing methods and programs | |
| CN113760670A (en) | Cable joint abnormity early warning method and device, electronic equipment and storage medium | |
| CN110757510A (en) | Method and system for predicting remaining life of robot | |
| Kefalas et al. | Automated machine learning for remaining useful life estimation of aircraft engines | |
| Tang et al. | Prediction of bearing performance degradation with bottleneck feature based on LSTM network | |
| US20230039740A1 (en) | Anomaly detection apparatus, anomaly detection method and program | |
| CN109324591B (en) | fault diagnosis device and method of stochastic hybrid system based on robust estimator | |
| CN114962390A (en) | Hydraulic system fault diagnosis method and system and working machine | |
| CN114500004A (en) | Anomaly detection method based on conditional diffusion probability generation model | |
| CN108459933B (en) | Big data computer system fault detection method based on deep recursion network | |
| CN116893924B (en) | Equipment fault processing method, device, electronic equipment and storage medium |
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 |