CN116838955A - Two-stage oil and gas pipeline line interference identification method - Google Patents
Two-stage oil and gas pipeline line interference identification method Download PDFInfo
- Publication number
- CN116838955A CN116838955A CN202310787617.1A CN202310787617A CN116838955A CN 116838955 A CN116838955 A CN 116838955A CN 202310787617 A CN202310787617 A CN 202310787617A CN 116838955 A CN116838955 A CN 116838955A
- Authority
- CN
- China
- Prior art keywords
- signal
- event
- gas pipeline
- identification method
- classification
- 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 35
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 30
- 239000013307 optical fiber Substances 0.000 claims abstract description 18
- 238000013145 classification model Methods 0.000 claims abstract description 10
- 239000013598 vector Substances 0.000 claims abstract description 6
- 238000000354 decomposition reaction Methods 0.000 claims description 12
- 230000006870 function Effects 0.000 claims description 8
- 230000002159 abnormal effect Effects 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 5
- 230000009467 reduction Effects 0.000 claims description 5
- 238000012935 Averaging Methods 0.000 claims description 3
- 238000009412 basement excavation Methods 0.000 claims description 3
- 230000007123 defense Effects 0.000 claims description 3
- 238000001228 spectrum Methods 0.000 claims description 2
- 230000009471 action Effects 0.000 abstract description 9
- 238000013135 deep learning Methods 0.000 abstract description 5
- 239000007789 gas Substances 0.000 description 8
- VNWKTOKETHGBQD-UHFFFAOYSA-N methane Chemical compound C VNWKTOKETHGBQD-UHFFFAOYSA-N 0.000 description 6
- 238000010801 machine learning Methods 0.000 description 5
- 238000012544 monitoring process Methods 0.000 description 4
- 238000012549 training Methods 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 238000001514 detection method Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 239000003345 natural gas Substances 0.000 description 3
- 238000007637 random forest analysis Methods 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 238000012706 support-vector machine Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000013527 convolutional neural network Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000009977 dual effect Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 239000000835 fiber Substances 0.000 description 2
- 239000012634 fragment Substances 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000007635 classification algorithm Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Abstract
The invention belongs to the field of deep learning and safety, and particularly relates to a two-stage oil and gas pipeline line interference identification method, which comprises the following steps: collecting original signals of distributed optical fibers of an optical fiber pipeline; dividing the noise-reduced signal to obtain a large number of signal samples; inputting all signal samples into a pre-classification model for pre-classification, and taking the signal samples with the pre-classification result being the interference event as a data set of an event identification main algorithm; extracting feature vectors of all signals in a data set; and identifying the intrusion signals of the distributed optical fiber pipeline through a cascade forest model, wherein the output result is an event type. The invention adds a pre-selection link to effectively reduce the possibility that the interference-free state is mistakenly identified as an interference event, and the main algorithm part emphasizes the capability of distinguishing common interference from third party damage events threatening the pipeline safety, and both enable the system to more accurately identify which events are threats needing to be alarmed and which actions do not need to be alarmed.
Description
Technical Field
The invention belongs to the field of deep learning and safety, and particularly relates to a two-stage oil and gas pipeline line interference identification method.
Background
In recent years, distributed optical fiber vibration sensing (DOVS) technology has attracted wide attention in the fields of intelligent security and protection and the like due to the advantages of high sensitivity, electromagnetic interference resistance, low price and the like. The method is applied to the fields of perimeter safety, oil and gas pipeline safety pre-warning and structural health monitoring, and particularly plays a very favorable role in protecting long-distance pipelines.
However, the sensing fiber is susceptible to environmental influences such as wind and rain, pedestrian walking or animal activity, and thus these innocuous events can lead to unexpected false positives in the system. Furthermore, the complexity and similarity of the vibration signals may lead to errors in the identification of the vibration type. Time delay also becomes one of the problems puzzling the sensing technology, and intrusion recognition time exceeds 7 seconds, so that emergency response is difficult to deal with. Therefore, the reliable real-time mode recognition method is researched to recognize the harmless vibration events, the false alarm rate is reduced, the recognition precision is improved, and the method has very important practical significance and guiding effect on guaranteeing the safe operation of the natural gas pipeline, wherein the practical application capability of the distributed optical fiber sensing is improved.
The realization of vibration signal type recognition depends on a classifier with good performance, and is responsible for inputting the characteristics of various intrusion signals into a classification model for training, and then, the classification model is utilized for carrying out rapid type recognition on the vibration signals. This requires a strong sample learning ability of the model to achieve high accuracy classification. At present, the distributed optical fiber field mostly uses one-dimensional time sequence data as an original sample to extract the characteristics of signals, and a classifier mainly selects model classification algorithms based on machine learning and deep learning. Among the machine learning based models include Support Vector Machines (SVMs), correlation vector machines (RVMs), linear Discriminant Analysis (LDA), gaussian Mixture Models (GMMs), random Forests (RF), and the like. Machine learning methods such as support vector machines, while achieving better classification on a small sample basis, have significant shortcomings for multi-classification tasks, and some methods such as GMM, the classification depends on the choice of initial values. Algorithm models based on deep learning, such as Convolutional Neural Network (CNN), long-short-term memory network (LSTM), echo State Network (ESN), etc., are applied. Deep learning, which performs feature extraction in an automated manner, can learn more useful features by constructing a model with multiple hidden layers and a large amount of training data, thereby improving classification accuracy. High-precision methods based on a combination of Empirical Mode Decomposition (EMD) and Radial Basis Function (RBF) neural networks have been proposed, which use the energy ratio of the Intrinsic Mode Function (IMF) obtained from the intrusion signal by EMD decomposition as the basis for the classification task. However, the EMD method causes a problem of mode mixing of discontinuous signals, severely affecting accuracy.
In summary, the prior art has the following problems:
1) The real-time performance is not high. Most of the existing researches need to collect a large amount of historical data for training, obvious delay exists in practical application, and the real-time capability required by field operation is difficult to ensure.
2) No locality was bound. The surrounding environment of the long-distance oil and gas pipeline in the cross region is highly complex, and the characteristics of mountain regions, farmlands, rivers, highways and the like are various. The sensing signals are greatly different according to the geological characteristics, and most of the current researches do not pay attention to the fact.
3) The applicability is not strong. Most research remains in the laboratory and is not deployed in the actual long haul pipeline network and therefore cannot accommodate the actual demand.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention provides a two-stage oil and gas pipeline line interference identification method. The method is used for accurately identifying which actions need to be alarmed in actions occurring along the pipeline in the long-distance pipeline safety monitoring process, and which actions do not need to be alarmed.
The invention aims at realizing the following technical scheme:
a two-stage oil and gas pipeline line interference identification method comprises the following steps:
step 1, acquiring original signals of distributed optical fibers of an optical fiber pipeline according to a time sequence and performing noise reduction treatment;
step (a)2, dividing the noise-reduced signal to obtain a large number of signal samples, if the window length is w l Step length s l A signal of length l may be partitioned into n samples;
step 3, inputting all signal samples into a pre-classification model for pre-classification, wherein the pre-classification type is divided into interference events and normal events, and taking the signal samples with the pre-classification result being the interference events as a data set of an event identification main algorithm;
step 4, extracting the feature vector of the signal screened in the step 3;
and 5, the event identification main algorithm adopts a cascading forest model, the distributed optical fiber pipeline intrusion signals are identified through the cascading forest model, and the output result is the event type.
Further, the step 1 uses a variation mode decomposition to reduce noise of the signal, through which a given signal can be decomposed into K modes, through which the original signal is decomposed into several natural mode functions, and then the high frequency mode is reconstructed to obtain a clean signal.
Further, the variation mode used in the step 1 is decomposed into:
where f represents the original signal to be processed, t represents time, x is the convolution operator, j is the complex symbol,representing the differential, delta (t) is a dirac function, mu k Represents the decomposed kth modal component (IMF), ω k Representing the corresponding center frequency of each mode, the spectrum of each mode being modulated by Hilbert transform to a corresponding baseband represented as
Further, a dual sliding window algorithm is used in the step 3 as a pre-classification model.
Further, in the step 3, the classification method of the interference event is as follows: judging whether a disturbance event occurs or not by adopting the fluctuation amplitude, setting X to represent the size of a first sliding window, recording the size as a small window, setting Y to represent the size of a second sliding window, recording the size as a large window, and marking the small window as an abnormal window if the amplitude of more than N points in the small window exceeds a set threshold value; if more than M abnormal windows are found in the large window, the signal segment is marked as a disturbance event.
Further, a dynamic threshold setting method is used in the dual sliding window algorithm, the threshold is adjusted between different defense sectors according to the statistical attribute of the signal, and the threshold setting comprises two steps:
firstly, setting an average value of waveforms under normal events as an initial threshold value of each defending sector;
the threshold is then adjusted in conjunction with a double sliding window algorithm.
Further, the time domain features of the waveform extracted in the step 4 include a plurality of maximum value, minimum value, peak-to-peak value, average value, absolute average value, root mean square, variance, standard deviation, energy, peak factor, skewness factor, gap factor, waveform factor, pulse factor, and margin factor.
Further, in the step 5, the cascade structure is used by the cascade forest model to sequentially refine the prediction of the model, and in each cascade level, a set of weak classifiers is trained on the data subset; each classifier will generate an estimate of class distribution by calculating the percentage of different classes on the leaf node where the relevant instance is located, and then averaging all trees in the same forest; each cascaded level receives the characteristic information processed by the previous level and outputs the processing result to the next level; the weak classifiers combine to form a strong classifier, which is then used to classify the remaining data.
Further, the event types in the step 5 include various of highway disturbance, railway disturbance, mountain forest disturbance and excavation damage 4.
Compared with the prior art, the invention has the beneficial effects that:
1) And (3) reducing the false alarm rate: the pre-selection link is added, so that the possibility that the interference-free state is mistakenly identified as an interference event is effectively reduced, the main algorithm part emphasizes the capability of distinguishing common interference from third party damage events threatening the pipeline safety, and the main algorithm part emphasizes the capability of distinguishing the common interference from third party damage events threatening the pipeline safety, so that the system can accurately identify which events are threats needing to be alarmed and which actions do not need to be alarmed;
2) The instantaneity is improved: a pre-selection step is carried out before the main algorithm to reduce the number of samples to be classified by the main algorithm, so that the calculation efficiency can be remarkably improved.
Drawings
FIG. 1 is a flow chart of the present invention;
FIG. 2 is a schematic diagram of a pre-classification model structure according to the present invention;
FIG. 3 is a schematic diagram of the structure of the event recognition main algorithm of the present invention.
Detailed Description
The invention is further described below with reference to the accompanying drawings.
The geological environment along which the oil and gas pipeline passes, along which the facilities are very complex, such as the pipeline may pass through different environments such as densely populated towns, villages, farmlands, rivers, highways and railways. The amount of vibration along different geographic characteristics varies greatly, such as along roads and railways with more disturbing behavior, and in relatively uncommon farms and mountainous regions, the intensity of waveforms defining the failure event varies greatly. How to accurately identify which actions need to be alarmed in actions occurring along a pipeline in the long-distance pipeline safety monitoring process and which actions do not need to be alarmed, has become a major task of a natural gas pipeline optical cable early warning leakage detection system.
The present invention aims to provide an effective two-stage strategy to identify third party damage to the entire route, while identifying different environments such as mountains, forests, roads, railways, etc. In the long-distance pipeline safety monitoring process, the system can accurately identify which events are threats needing to be alarmed and which actions do not need to be alarmed. The first stage aims to identify and delete the normal operating state without interference from the input dataset, ensuring that only interference events can be input to the second stage main algorithm. This is a pre-selection step to reduce the number of samples that the main algorithm needs to sort, which can significantly improve the computational efficiency; in addition, the method can filter out the non-interference fragments to help improve the accuracy of interference identification and reduce the false alarm rate. And in the second stage, the machine learning algorithm is utilized to learn the basic characteristics of images corresponding to different intrusion signals, so that the identification of different types of intrusion events is realized.
Referring to fig. 1-3, a two-stage oil and gas pipeline line interference identification method includes the following steps:
step 1, acquiring original signals of distributed optical fibers of an optical fiber pipeline according to a time sequence and performing noise reduction treatment.
The invention first performs noise reduction processing on all signals, and a variation mode decomposition (variational mode decomposition, VMD) is used. The variational modal decomposition is a variational model which adaptively determines the relevant frequency band and simultaneously estimates the corresponding modes, and can solve the noise problem existing in the input signal. The model has strong interpretability and is supported by strong mathematical theory, and the most important advantage of the method is that modal aliasing is avoided. By this method, a given signal can be decomposed into K modes. The invention decomposes the original signal into 8 eigenmode functions (IMFs) by VMD and then reconstructs the high frequency modes to obtain a clean signal. This approach may improve noise reduction performance without causing signal distortion.
Wherein the variational modal decomposition essence is to solve a constrained variational problem:
wherein mu is k Represents the kth IMF, omega obtained by decomposition k Representing the corresponding center frequency of each mode, the spectral modulation of each mode to the corresponding baseband is represented as
And 2, dividing the noise-reduced signal to obtain a large number of signal samples with the time span of 30 s.
Wherein if the window length is w l Step length s l A signal of length l may be partitioned into n samples;
and step 3, inputting all the signal samples into a pre-classification model to perform pre-classification, and taking the signal samples with the pre-classification result being the interference event as a data set of a main algorithm.
And step 4, extracting the feature vector of the signal screened in the step three.
And 5, identifying the distributed optical fiber pipeline intrusion signals by using a cascade forest model, wherein the output result is event type and is divided into 4 categories of expressway disturbance, railway disturbance, mountain forest disturbance and excavation damage.
In the natural gas pipeline optical cable early warning leakage detection system, mode identification is one of core technologies, and extraction of detection signal feature vectors is one of the most critical links in a mode identification module. The eigenvector of the optical fiber disturbance signal needs to satisfy the following three conditions:
the characteristics need to be unique, and disturbance signals of different types need to have unique characteristic attributes and overlap with the characteristics of other types of disturbance;
the characteristics need to have stability, and the characteristics cannot be changed along with the change of factors such as time, the number of samples, the external environment and the like, namely the inherent characteristics of the signals always exist and are stable and unchanged;
features need to be generic, i.e. different classes of events can characterize the differences between events under the same features. In particular, in a distributed optical fiber early warning system, characteristic data can be quantized, and quantitative processing is convenient.
The time domain characteristics of the waveform are the change condition of the distributed optical fiber waveform along with time, part of the time domain characteristics can be intuitively perceived by naked eyes, such as the maximum value, the minimum value and the like of the waveform, certain characteristics are needed to be obtained through certain operation, such as the average value, the variance, the short-time energy and the like of the waveform in a period of time, and waveform vibration caused by different reasons often has certain difference in the related time domain characteristics. As shown in table 1, the time domain features of the waveform extracted by the present invention specifically include the following 16: maximum, minimum, peak-to-peak, average, absolute average, root mean square, variance, standard deviation, energy, peak factor, skewness factor, gap factor, waveform factor, pulse factor, margin factor.
Table 1 time domain features extractable for distributed fiber waveforms
In practical applications, about 95% of the real-time signal represents that the device or system is operating properly, no disturbance event occurs, and only 5% represents that a disturbance event is present. If all signals are put into the machine learning classification model, resources will be occupied, resulting in low timeliness. In order to improve accuracy and computational efficiency, the invention uses a double sliding window (TDSW) algorithm as a pre-classification in the first stage to distinguish between intrusion events and normal operating conditions before identifying a particular class.
The waveform of the normal working state is smoother, but the waveform of the disturbance event fluctuates greatly, so that the fluctuation amplitude can be used for judging whether the disturbance event has occurred. Let X denote the size of the first layer sliding window, record as a small window, and let Y (as long as the sample made by the present invention) denote the size of the second layer sliding window, record as a large window. If the amplitudes of more than N points in the small window exceed the set threshold value, marking the small window as an abnormal window; if more than M abnormal windows are found in the large window, the signal segment is marked as a disturbance event.
The magnitude threshold among the algorithms uses a dynamic thresholding method, i.e. the threshold is adjusted between different defensive sectors according to the statistical properties of the signal, which allows for a more adaptive and robust screening. The threshold setting requires 2 steps: firstly, setting an average value of waveforms in a normal working state as an initial threshold value of each defense area; the threshold is then adjusted slightly in combination with the double sliding window.
For the second stage of event recognition main algorithm, the present invention uses a cascading forest based model to perform signal recognition and classification. The cascading forest model can be viewed as a set of depths of random forest algorithms that use a cascading structure to refine predictions of the model in order. At each level of the cascade, a set of weak classifiers is trained on a subset of the data. Each classifier will generate an estimate of the class distribution by calculating the percentage of the different types on the leaf node where the relevant instance is located and then averaging all trees in the same forest. Each level of the cascade receives the characteristic information processed by the previous level and outputs the processing result to the next level. The weak classifiers combine to form a strong classifier, which is then used to classify the remaining data.
The use of cascading forests requires only limited training time costs and has good pattern matching performance. These advantages are fully consistent with the original purpose of the present invention of designing a quick, efficient, accurate strategy. The key idea behind cascading forest algorithms is to use cascading structures to increase model complexity, which is why it can process data with a large number of features or variables and capture their non-linear relationships. Furthermore, it provides a simple mechanism by which the number of cascade levels can be adaptively determined so that the model complexity can be automatically set. Meanwhile, the method has stronger robustness and can process noisy data or peripheral data points.
In summary, the two-stage interference identification strategy provided by the invention effectively reduces the possibility that the interference-free state is mistakenly identified as the interference event while realizing the basic function of identifying various interference events, and enhances the capability of distinguishing the interference event which does not threaten the pipeline safety and the third party damage event which can cause damage. The invention provides a double sliding window algorithm, which aims at carrying out a pre-selection step before a main algorithm so as to reduce the number of samples to be classified of the main algorithm and remarkably improve the calculation efficiency; in addition, the method can filter out the non-interference fragments to help improve the accuracy of interference identification and reduce the false alarm rate.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the same; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some or all of the technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit of the invention.
Claims (9)
1. The two-stage oil and gas pipeline line interference identification method is characterized by comprising the following steps of:
step 1, acquiring original signals of distributed optical fibers of an optical fiber pipeline according to a time sequence and performing noise reduction treatment;
step 2, dividing the noise-reduced signal to obtain a large number of signal samples, if the window length is w l Step length s l A signal of length l may be partitioned into n samples;
step 3, inputting all signal samples into a pre-classification model for pre-classification, wherein the pre-classification type is divided into interference events and normal events, and taking the signal samples with the pre-classification result being the interference events as a data set of an event identification main algorithm;
step 4, extracting the feature vector of the signal screened in the step 3;
and 5, the event identification main algorithm adopts a cascading forest model, the distributed optical fiber pipeline intrusion signals are identified through the cascading forest model, and the output result is the event type.
2. The two-stage oil and gas pipeline line interference identification method according to claim 1, wherein the step 1 uses a variation mode decomposition to reduce noise of signals, a given signal can be decomposed into K modes by the variation mode decomposition, an original signal is decomposed into a plurality of inherent mode functions by the variation mode decomposition, and then a high frequency mode is reconstructed to obtain a clean signal.
3. The two-stage oil and gas pipeline line interference identification method according to claim 1, wherein the variational modal decomposition core concept used in the step 1 is to construct and solve a variational problem:
where f represents the original signal to be processed, t represents time, x is the convolution operator, j is the complex symbol,representing the differential, delta (t) is a dirac function, mu k Represents the decomposed kth modal component (IMF), ω k Representing the corresponding center frequency of each mode, the spectrum of each mode being modulated by Hilbert transform to a corresponding baseband represented as
4. The two-stage oil and gas pipeline line interference identification method according to claim 1, wherein the step 3 uses a double sliding window algorithm as a pre-classification model.
5. The two-stage oil and gas pipeline line interference identification method according to claim 4, wherein in the step 3, the classification method of the interference event is as follows: judging whether a disturbance event occurs or not by adopting the fluctuation amplitude, setting X to represent the size of a first sliding window, recording the size as a small window, setting Y to represent the size of a second sliding window, recording the size as a large window, and marking the small window as an abnormal window if the amplitude of more than N points in the small window exceeds a set threshold value; if more than M abnormal windows are found in the large window, the signal segment is marked as a disturbance event.
6. The two-stage oil and gas pipeline line interference identification method according to claim 5, wherein a dynamic threshold setting method is used in the double sliding window algorithm, the threshold is adjusted between different defense sectors according to statistical properties of signals, and the threshold setting comprises the following steps:
firstly, setting an average value of waveforms under normal events as an initial threshold value of each defending sector;
the threshold is then adjusted in conjunction with a double sliding window algorithm.
7. The two-stage oil and gas pipeline line interference identification method according to claim 1, wherein the time domain features of the waveform extracted in the step 4 comprise a plurality of maximum values, minimum values, peak-to-peak values, average values, absolute average values, root mean square values, variances, standard deviations, energies, peak factors, skewness factors, gap factors, waveform factors, pulse factors, and margin factors.
8. The two-stage oil and gas pipeline line interference identification method according to claim 1, wherein the cascade structure is used by the cascade forest model in step 5 to refine the predictions of the models sequentially, and in each stage of the cascade, a set of weak classifiers is trained on a subset of data; each classifier will generate an estimate of class distribution by calculating the percentage of different classes on the leaf node where the relevant instance is located, and then averaging all trees in the same forest; each cascaded level receives the characteristic information processed by the previous level and outputs the processing result to the next level; the weak classifiers combine to form a strong classifier, which is then used to classify the remaining data.
9. The two-stage oil and gas pipeline line disturbance identification method according to claim 1, wherein the event types in the step 5 include a plurality of highway disturbances, railway disturbances, mountain forest disturbances and excavation damages 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310787617.1A CN116838955A (en) | 2023-06-30 | 2023-06-30 | Two-stage oil and gas pipeline line interference identification method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310787617.1A CN116838955A (en) | 2023-06-30 | 2023-06-30 | Two-stage oil and gas pipeline line interference identification method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116838955A true CN116838955A (en) | 2023-10-03 |
Family
ID=88172014
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310787617.1A Pending CN116838955A (en) | 2023-06-30 | 2023-06-30 | Two-stage oil and gas pipeline line interference identification method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116838955A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117386344A (en) * | 2023-12-13 | 2024-01-12 | 西南石油大学 | Drilling abnormal condition diagnosis method and system based on two-stage learning |
-
2023
- 2023-06-30 CN CN202310787617.1A patent/CN116838955A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117386344A (en) * | 2023-12-13 | 2024-01-12 | 西南石油大学 | Drilling abnormal condition diagnosis method and system based on two-stage learning |
| CN117386344B (en) * | 2023-12-13 | 2024-02-23 | 西南石油大学 | Drilling abnormal condition diagnosis method and system based on two-stage learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103617684B (en) | Interference-type optical fiber circumference vibrating intruding recognizer | |
| Scarpetta et al. | Automatic classification of seismic signals at Mt. Vesuvius volcano, Italy, using neural networks | |
| CN104240455B (en) | A kind of disturbance event recognition methods in distribution type fiber-optic pipeline safety early warning system | |
| EP3321859B1 (en) | Optical fiber perimeter intrusion signal identification method and device, and perimeter intrusion alarm system | |
| CN106874833B (en) | Vibration event pattern recognition method | |
| Shaw et al. | Multifractal modelling of radio transmitter transients for classification | |
| Priyadarshani et al. | Wavelet filters for automated recognition of birdsong in long‐time field recordings | |
| CN103968933A (en) | Fuzzy domain characteristics based optical fiber vibration signal identifying method | |
| CN116838955A (en) | Two-stage oil and gas pipeline line interference identification method | |
| Kim et al. | Machine-learning-based prediction of vortex-induced vibration in long-span bridges using limited information | |
| Carmichael et al. | Unfooling perturbation-based post hoc explainers | |
| CN114510960A (en) | Method for recognizing distributed optical fiber sensor system mode | |
| Das et al. | Crime analysis against women from online newspaper reports and an approach to apply it in dynamic environment | |
| CN111951505B (en) | Fence vibration intrusion positioning and mode identification method based on distributed optical fiber system | |
| Zhu et al. | A novel visual transformer for long-distance pipeline pattern recognition in complex environment | |
| Lyu et al. | Two-stage intrusion events recognition for vibration signals from distributed optical fiber sensors | |
| CN112232329A (en) | Multi-core SVM training and alarming method, device and system for intrusion signal recognition | |
| Pinto et al. | Nonlinear prediction of time series obtained from an experimental pendulum | |
| Draayer et al. | Pattern-of-life activity recognition in seismic data | |
| Al-Nafjan et al. | Intrusion detection using PCA based modular neural network | |
| Zhao et al. | Feature extraction and identification of pipeline intrusion based on phase-sensitive optical time domain reflectometer | |
| CN114093106B (en) | Intrusion signal alarm method and system based on EfficientNET classification network | |
| Huang et al. | A recognition method of less databases by using decoupling parallel CNN based on multiple feature with DAS | |
| CN112836591B (en) | Method for extracting optical fiber early warning signal characteristics of oil gas long-distance pipeline | |
| CN116738259B (en) | Multi-harmonic-based electromagnetic leakage radiation source fingerprint extraction and identification method and device |
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 |