CN115859437A - Jacket underwater stress detection system based on distributed optical fiber sensing system - Google Patents
Jacket underwater stress detection system based on distributed optical fiber sensing system Download PDFInfo
- Publication number
- CN115859437A CN115859437A CN202211592492.9A CN202211592492A CN115859437A CN 115859437 A CN115859437 A CN 115859437A CN 202211592492 A CN202211592492 A CN 202211592492A CN 115859437 A CN115859437 A CN 115859437A
- Authority
- CN
- China
- Prior art keywords
- feature
- strain
- jacket
- training
- 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
Images
Landscapes
- Length Measuring Devices By Optical Means (AREA)
Abstract
The application relates to the field of intelligent detection, and particularly discloses a jacket underwater stress detection system based on a distributed optical fiber sensing system, which detects the stress distribution of a jacket by adopting an artificial intelligence detection technology based on deep learning and intelligently analyzing a strain cloud chart provided by the distributed optical fiber sensing system. Specifically, stress distribution characteristic information focused on the spatial position of the jacket in the strain cloud image is extracted through an artificial intelligence algorithm, and overall global correlation characteristic distribution extraction is performed based on the stress characteristic distribution of each local position, so that the precision of detecting the stress distribution of the jacket is improved. Like this, can be intelligently whether normally carry out accurate the detection to the stress distribution of offshore wind power field jacket to guarantee the security of offshore wind power jacket, reduce enterprise later stage cost of maintenance.
Description
Technical Field
The present application relates to the field of intelligent detection, and more particularly, to a jacket underwater stress detection system based on a distributed optical fiber sensing system.
Background
Under the large background that China vigorously advances 'carbon peak reaching' and 'carbon neutralization', the wind power, photovoltaic and pumped storage industries are rapidly developed. The development speed of offshore wind power, which is the most important component of the wind power industry, is not variable, and with the high-speed development of offshore wind power, the development trend of large-scale offshore wind turbines is inevitable. The jacket foundation is one of the most stable and reliable foundation forms in the offshore wind turbine foundation, occupies an indispensable position in the development of offshore wind power, and plays a key role in bearing the above and below as a transition section structure for connecting an upper wind turbine tower and a lower main body.
At present, the area where the jacket platform is located is mostly a sea area with high temperature, high salt and complex environmental load, and the stress condition, the structural loss and the corrosion condition of the jacket of the offshore wind farm need to be monitored in real time so as to ensure the safety of the jacket platform. When the stress condition of an offshore wind farm jacket is monitored and analyzed by the conventional scheme, the stress condition of the jacket is mostly acquired by a sensor, and the jacket is manually monitored by analyzing related technical personnel based on the stress condition, so that the problems of low manual monitoring efficiency, poor real-time performance and the like are caused, and the artificial error is easy to occur, and the influence is brought to the stress condition detection result of the jacket. Moreover, when the traditional stress monitoring scheme is used for deploying the sensor, the deployment of the sensor is difficult due to the harsh underwater environment.
Therefore, an optimized jacket underwater stress detection system is expected, which can overcome many defects of the traditional manual monitoring means, can objectively reflect the actual wind power foundation condition, and can intelligently detect whether the stress distribution of the offshore wind power plant jacket is normal or not in real time, so that the safety of the offshore wind power jacket is guaranteed, and the later maintenance cost of enterprises is reduced.
Disclosure of Invention
The present application is proposed to solve the above-mentioned technical problems. The embodiment of the application provides a jacket underwater stress detection system based on a distributed optical fiber sensing system, which adopts an artificial intelligence detection technology based on deep learning to carry out stress distribution detection on a jacket through intelligent analysis of a strain cloud chart provided by the distributed optical fiber sensing system. Specifically, stress distribution characteristic information focused on the spatial position of the jacket in the strain cloud image is extracted through an artificial intelligence algorithm, and overall global correlation characteristic distribution extraction is performed based on the stress characteristic distribution of each local position, so that the precision of detecting the stress distribution of the jacket is improved. Like this, can be intelligently whether normally carry out accurate the detection to the stress distribution of offshore wind power field jacket to guarantee the security of offshore wind power jacket, reduce enterprise later stage cost of maintenance.
According to one aspect of the application, a jacket underwater stress detection system based on a distributed optical fiber sensing system is provided, and comprises:
the strain data acquisition unit is used for acquiring a strain cloud picture provided by a distributed optical fiber sensing system deployed on a jacket to be tested;
the stress distribution characteristic extraction unit is used for enabling the strain cloud picture to pass through a convolutional neural network model serving as a characteristic extractor to obtain a strain characteristic picture;
the spatial enhancement unit is used for passing the strain characteristic map through a spatial attention module to obtain a spatial enhancement strain characteristic map;
the local feature expansion unit is used for expanding each feature matrix of the space enhanced strain feature map along the channel dimension into feature vectors so as to obtain a plurality of local region strain feature vectors;
a context encoding unit, configured to pass the plurality of local region strain feature vectors through a converter-based context encoder to obtain a classification feature vector; and
and the stress detection result generation unit is used for enabling the classified characteristic vectors to pass through a classifier to obtain a classification result, and the classification result is used for indicating whether the stress distribution of the jacket to be detected is normal or not.
In the jacket underwater stress detection system based on the distributed optical fiber sensing system, the stress distribution characteristic extraction unit is further configured to: performing, in a layer forward pass, input data using the layers of the convolutional neural network model as a feature extractor: performing convolution processing on input data to obtain a convolution characteristic diagram; pooling the convolution feature map based on a local feature matrix to obtain a pooled feature map; and performing nonlinear activation on the pooled feature map to obtain an activated feature map; and the output of the last layer of the convolutional neural network as the feature extractor is the strain feature map, and the input of the first layer of the convolutional neural network as the feature extractor is the strain cloud map.
In the above jacket underwater stress detection system based on the distributed optical fiber sensing system, the spatial enhancement unit is further configured to: calculating a global mean value of each feature matrix of the strain feature map along the spatial dimension to obtain a spatial feature vector; inputting the spatial feature vector into a Softmax activation function to obtain a spatial attention weight feature vector; and respectively weighting each feature matrix of the strain feature map along the spatial dimension by taking the feature value of each position in the spatial attention weight feature vector as a weight so as to obtain the spatial enhancement strain feature map.
In the jacket underwater stress detection system based on the distributed optical fiber sensing system, the local feature expansion unit is further configured to expand each feature matrix of the spatial enhanced strain feature map along the channel dimension into a plurality of local region strain feature vectors according to the row vector dimension or the column vector dimension.
In the above jacket underwater stress detection system based on the distributed optical fiber sensing system, the context encoding unit includes: a context associated feature extraction subunit, configured to pass the plurality of local area strain feature vectors through the converter-based context encoder to obtain a plurality of context semantic area strain feature vectors; and the cascading subunit is used for cascading the plurality of context semantic region strain feature vectors to obtain the classification feature vector.
In the jacket underwater stress detection system based on the distributed optical fiber sensing system, the context correlation feature extraction subunit is further configured to: performing one-dimensional arrangement on the plurality of local region strain characteristic vectors to obtain a global region strain characteristic vector; calculating a product between the global region strain eigenvector and a transposed vector of each region strain eigenvector in the plurality of local region strain eigenvectors to obtain a plurality of self-attention correlation matrices; respectively carrying out standardization processing on each self-attention correlation matrix in the self-attention correlation matrixes to obtain a plurality of standardized self-attention correlation matrixes; obtaining a plurality of probability values by passing each normalized self-attention correlation matrix in the plurality of normalized self-attention correlation matrices through a Softmax classification function; and respectively weighting each region strain characteristic vector in the local region strain characteristic vectors by taking each probability value in the probability values as a weight so as to obtain the context semantic region strain characteristic vectors.
In the above jacket underwater stress detection system based on the distributed optical fiber sensing system, the stress detection result generating unit includes: a full-concatenation coding subunit, configured to perform full-concatenation coding on the classification feature vector using a plurality of full-concatenation layers of the classifier to obtain a coded classification feature vector; and the classification result generation subunit is used for enabling the coded classification feature vector to pass through a Softmax classification function of the classifier so as to obtain the classification result.
In the jacket underwater stress detection system based on the distributed optical fiber sensing system, the system further comprises a training module for training the convolutional neural network model serving as the feature extractor, the space attention module, the context encoder based on the converter and the classifier; wherein the training module comprises: the training data acquisition unit is used for acquiring a training strain cloud picture and a true value of whether the stress distribution of the jacket to be tested is normal or not; the training stress distribution characteristic extraction unit is used for enabling the training strain cloud picture to pass through the convolutional neural network model serving as the characteristic extractor so as to obtain a training strain characteristic picture; the training space enhancement unit is used for enabling the training strain characteristic map to pass through the space attention module to obtain a training space enhancement strain characteristic map; the training local feature expansion unit is used for expanding each feature matrix of the training space enhanced strain feature map along the channel dimension into feature vectors to obtain a plurality of training local area strain feature vectors; a training context encoding unit, configured to pass the training local region strain feature vectors through the converter-based context encoder to obtain training classification feature vectors; the classification loss unit is used for enabling the training classification feature vector to pass through the classifier to obtain a classification loss function value; a training unit for training the convolutional neural network model as a feature extractor, the spatial attention module, the converter-based context encoder, and the classifier by gradient descent direction propagation based on the classification loss function values, wherein, in each iteration of the training, a cross-classifier soft similarity-based free label optimization factor of the training classification feature vector is calculated as a weighted weight to perform weighted iteration on the training classification feature vector.
In the jacket underwater stress detection system based on the distributed optical fiber sensing system, in each iteration of the training, a free label optimization factor based on cross-classifier soft similarity of the training classification feature vector is calculated according to the following formula to serve as a label value of the classifier;
wherein the formula is:
v is the classification feature vector, M is a weight matrix of the classifier on the classification feature vector,
and &>
Respectively representing tensor multiplication and tensor addition, d (·,) represents the distance between vectors, | | · includes 2 Denotes a two-norm vector, and α and β are weight superparameters, exp (-) denotes an exponential operation of a vector, which denotes calculation of a natural exponential function value raised to the eigenvalue of each position in the vector.
According to another aspect of the application, a jacket underwater stress detection method based on a distributed optical fiber sensing system is provided, and comprises the following steps:
strain data acquisition: acquiring a strain cloud chart provided by a distributed optical fiber sensing system deployed on a jacket to be tested;
a stress distribution characteristic extraction step: passing the strain cloud picture through a convolutional neural network model serving as a feature extractor to obtain a strain feature picture;
a spatial enhancement step: passing the strain signature through a spatial attention module to obtain a spatially enhanced strain signature;
local feature unfolding step: expanding each feature matrix of the spatial enhancement strain feature map along the channel dimension into feature vectors to obtain a plurality of local area strain feature vectors;
context coding step: passing the plurality of local region strain feature vectors through a transducer-based context encoder to obtain a classification feature vector; and
a stress detection result generation step: and passing the classified characteristic vectors through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the stress distribution of the jacket to be tested is normal or not.
According to still another aspect of the present application, there is provided an electronic apparatus including: a processor; and a memory having stored therein computer program instructions which, when executed by the processor, cause the processor to perform the method of jacket subsea stress detection based on a distributed optical fiber sensing system as described above.
According to yet another aspect of the present application, there is provided a computer readable medium having stored thereon computer program instructions, which, when executed by a processor, cause the processor to execute the jacket underwater stress detection method based on the distributed optical fiber sensing system as described above.
Compared with the prior art, the jacket underwater stress detection system based on the distributed optical fiber sensing system provided by the application detects the stress distribution of the jacket by adopting the artificial intelligence detection technology based on deep learning and intelligently analyzing a strain cloud chart provided by the distributed optical fiber sensing system. Specifically, stress distribution characteristic information focused on the spatial position of the jacket in the strain cloud image is extracted through an artificial intelligence algorithm, and overall global correlation characteristic distribution extraction is performed based on the stress characteristic distribution of each local position, so that the precision of detecting the stress distribution of the jacket is improved. Like this, can be intelligently whether normally carry out accurate the detection to the stress distribution of offshore wind power field jacket to guarantee the security of offshore wind power jacket, reduce enterprise later stage cost of maintenance.
Drawings
The above and other objects, features and advantages of the present application will become more apparent by describing in more detail embodiments of the present application with reference to the attached drawings. The accompanying drawings are included to provide a further understanding of the embodiments of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and together with the description serve to explain the principles of the application. In the drawings, like reference numbers generally indicate like parts or steps.
Fig. 1 is an application scenario diagram of a jacket underwater stress detection system based on a distributed optical fiber sensing system according to an embodiment of the application;
FIG. 2 is a block diagram of a jacket underwater stress detection system based on a distributed optical fiber sensing system according to an embodiment of the application;
FIG. 3 is a block diagram of a jacket subsea stress detection system based on a distributed fiber optic sensing system according to an embodiment of the present application;
FIG. 4 is a system architecture diagram of a jacket subsea stress detection system based on a distributed fiber optic sensing system in accordance with an embodiment of the present application;
FIG. 5 is a flow chart of convolutional neural network coding in a jacket underwater stress detection system based on a distributed optical fiber sensing system according to an embodiment of the present application;
FIG. 6 is a flow chart of a spatial enhancement process in a jacket underwater stress detection system based on a distributed optical fiber sensing system according to an embodiment of the present application;
FIG. 7 is a flow chart of a context encoding process in a jacket underwater stress detection system based on a distributed optical fiber sensing system according to an embodiment of the present application;
FIG. 8 is a system architecture diagram of a training module in a jacket underwater stress detection system based on a distributed optical fiber sensing system according to an embodiment of the application;
FIG. 9 is a flow chart of a jacket underwater stress detection method based on a distributed optical fiber sensing system according to an embodiment of the present application;
fig. 10 is a block diagram of an electronic device according to an embodiment of the application.
Detailed Description
Hereinafter, example embodiments according to the present application will be described in detail with reference to the accompanying drawings. It should be understood that the described embodiments are only some embodiments of the present application and not all embodiments of the present application, and that the present application is not limited by the example embodiments described herein.
Overview of a scene
As for the background technology, most of the traditional stress monitoring schemes for the jacket of the offshore wind farm acquire the stress condition of the jacket through a sensor, and perform manual monitoring through analysis of related technical personnel based on the stress condition, which causes the problems of low manual monitoring efficiency, poor real-time performance and the like, and is easy to cause human errors, thereby affecting the stress condition detection result of the jacket. Moreover, when the traditional stress monitoring scheme is used for deploying the sensor, the deployment of the sensor is difficult due to the harsh underwater environment. Accordingly, an optimized jacket subsea stress detection system is desired.
At present, deep learning and neural networks have been widely applied in the fields of computer vision, natural language processing, speech signal processing, and the like. In addition, deep learning and neural networks also exhibit a level close to or even exceeding that of humans in the fields of image classification, object detection, semantic segmentation, text translation, and the like.
In recent years, deep learning and development of neural networks provide new solutions and schemes for underwater stress detection of jackets.
Correspondingly, the traditional scheme for detecting the underwater stress of the jacket of the offshore wind farm has high requirements on the operating skills of personnel, low cost effectiveness ratio and low intelligence level, and is difficult to deploy a sensor, so that the detection accuracy of the stress distribution of the jacket is reduced. At the same time, it is also considered that distributed detection does not rely on manufactured sensors, but rather utilizes optical fibers, unlike conventional sensors that rely on discrete sensors measured at predetermined points. The optical fiber itself is the sensing element without any additional sensors in the optical path. Since the fiber is a sensor, it is also a cost effective method that can be easily deployed even in the harshest and most unusual environments.
Based on this, in the technical scheme of this application, adopt artificial intelligence detection technique based on deep learning to carry out the stress distribution detection of jacket through the intelligent analysis of the strain cloud picture that distributed optical fiber sensing system provided. Specifically, stress distribution characteristic information focused on the spatial position of the jacket in the strain cloud image is extracted through an artificial intelligence algorithm, and overall global correlation characteristic distribution extraction is performed based on the stress characteristic distribution of each local position, so that the precision of detecting the stress distribution of the jacket is improved. Like this, can be intelligently whether normally carry out accurate the detection to the stress distribution of offshore wind power field jacket to guarantee the security of offshore wind power jacket, reduce enterprise later stage cost of maintenance.
Specifically, in the technical scheme of the application, firstly, a strain cloud chart is provided through a distributed optical fiber sensing system deployed on a jacket to be tested. It should be understood that the strain cloud pictures are cloud pictures of stress magnitude and distribution of each part of a mechanism generated in the finite element software under the action of given external force, and can well reflect the stress distribution condition of the jacket to be measured. And then, processing the strain cloud picture in a convolutional neural network model which has excellent performance in the aspect of image implicit feature extraction and serves as a feature extractor to extract implicit feature distribution information of the jacket to be detected in the strain cloud picture, so that a strain feature picture is obtained.
Then, considering that the stress distribution situation at the spatial position of the jacket to be tested should be focused more when detecting the stress distribution of the jacket to be tested, the implicit feature extraction of the jacket to be tested should be performed focusing on the stress distribution feature information of the jacket to be tested at the spatial position. Specifically, in the technical solution of the present application, the strain characteristic map is passed through a spatial attention module to obtain a spatial enhanced strain characteristic map. It should be understood that the image features extracted by the spatial attention reflect the feature difference weights in the spatial dimension of the stress distribution information about the jacket in the strain cloud, so as to suppress or strengthen the features of different spatial positions.
Further, when the stress distribution of the jacket to be tested is determined to be normal, the stresses in the local regions on the jacket to be tested have mutual influence, that is, the stress features in the local regions on the jacket to be tested have associated feature distribution information in a high-dimensional space, so that in order to be able to sufficiently extract the stress distribution feature information on the jacket to accurately determine whether the stress distribution is normal, it is necessary to further extract the globally associated feature information of the stress distribution in the space. Specifically, in the technical solution of the present application, each feature matrix along the channel dimension of the spatial enhanced strain feature map is further expanded into feature vectors in rows or columns, so as to obtain a plurality of local area strain feature vectors having a plurality of local area spatial implicit features of the jacket to be measured.
Then, the local region strain feature vectors are encoded in a context encoder based on a converter, so that stress distribution implicit features of each local region of the catheter rack in the spatial position are extracted based on global relevance feature distribution information, and classification feature vectors are obtained. That is, based on the transform idea, with the converter being able to capture the long-distance context-dependent characteristic, each of the local region strain feature vectors is subjected to global context-based semantic encoding to obtain a context-semantic association feature representation, i.e., the classification feature vector, with the global semantic association of the local region strain feature vectors as context. It should be understood that in the technical solution of the present application, the context semantic association feature representation of the implicit feature of stress distribution of each local area of the jacket in spatial position with respect to the implicit feature of stress distribution of the jacket as a whole can be captured by the converter-based encoder.
And then, further enabling the classified characteristic vectors to pass through a classifier to obtain a classification result for indicating whether the stress distribution of the jacket to be tested is normal or not. Like this, can be intelligently whether normally carry out accurate the detection to the stress distribution of offshore wind power field jacket to guarantee the security of offshore wind power jacket, reduce enterprise later stage cost of maintenance.
In particular, in the technical solution of the present application, when the plurality of local area strain feature vectors are obtained by the context encoder based on the converter, the plurality of local area strain feature vectors are directly concatenated by the context encoder based on the converter to obtain the classification feature vector, so even if the context encoder based on the converter can extract the correlation between the plurality of local area strain feature vectors to improve the correlation of the feature semantic distribution between the plurality of context local area strain feature vectors, in the case of directly concatenating the feature vectors, since the explicit correlation between the plurality of context local area strain feature vectors is low, the overall distribution of the classification feature vectors still has strong discreteness, so that the training of the classifier is difficult, especially the convergence of the label value of the classifier is difficult.
Therefore, soft label learning is preferably used instead of the commonly used hard label learning, in particular, at each iteration, a free label optimization factor based on cross-classifier soft similarity of the classification feature vectors is calculated as the label value of the classifier;
wherein the formula is:
v is the classification feature vector, M is a weighting matrix of the classifier on the classification feature vector,
and &>
Respectively representing tensor multiplication and tensor addition, d (·,) represents the distance between vectors, | | · includes 2 Representing the two norms of the vector, and α and β are weight hyperparameters, exp (·) represents an exponential operation of the vector, which represents the computation of a natural exponential function value raised by eigenvalues of various locations in the vector.
Here, the cross-classifier soft similarity-based free label optimization factor simulates pseudo classes based on classifier weight matrices through the cross-classifier soft similarity of the classification feature vectors and the weight matrices of the classifier by performing bidirectional clustering on the classification feature vectors and the classifier before calculating the classification probability of the feature vectors with hard label values, so that the classification quantization loss caused by hard label learning is avoided through soft similarity learning, the free label optimization focusing more on the intrinsic weight structure of the classifier is realized, the training of the label values of the classifier is optimized, and the training speed of the classifier is improved. Like this, can be intelligently whether normally carry out accurate the detection to the stress distribution of offshore wind power field jacket to guarantee the security of offshore wind power jacket, reduce enterprise later stage cost of maintenance.
Based on this, this application has proposed a jacket underwater stress detecting system based on distributed optical fiber sensing system, and it includes: the strain data acquisition unit is used for acquiring a strain cloud picture provided by a distributed optical fiber sensing system deployed on a jacket to be tested; the stress distribution characteristic extraction unit is used for enabling the strain cloud picture to pass through a convolutional neural network model serving as a characteristic extractor to obtain a strain characteristic picture; the spatial enhancement unit is used for passing the strain characteristic map through a spatial attention module to obtain a spatial enhancement strain characteristic map; the local feature expansion unit is used for expanding each feature matrix of the space enhanced strain feature map along the channel dimension into feature vectors so as to obtain a plurality of local region strain feature vectors; a context coding unit, configured to pass the plurality of local area strain feature vectors through a converter-based context coder to obtain a classification feature vector; and the stress detection result generation unit is used for enabling the classified characteristic vectors to pass through a classifier to obtain a classification result, and the classification result is used for indicating whether the stress distribution of the jacket to be detected is normal or not.
Fig. 1 is a view of an application scenario of a jacket underwater stress detection system based on a distributed optical fiber sensing system according to an embodiment of the application. As shown in fig. 1, in this application scenario, a strain cloud is obtained by a distributed optical fiber sensing system (e.g., S1 as illustrated in fig. 1) deployed on a jacket under test. Then, the images are input into a server (for example, S2 in fig. 1) deployed with a jacket underwater stress detection algorithm for the distributed optical fiber sensing system, wherein the server can process the images by the jacket underwater stress detection algorithm for the distributed optical fiber sensing system to generate a classification result for indicating whether the stress distribution of the jacket to be tested is normal or not.
Having described the general principles of the present application, various non-limiting embodiments of the present application will now be described with reference to the accompanying drawings.
Exemplary System
Fig. 2 is a block diagram of a jacket underwater stress detection system based on a distributed optical fiber sensing system according to an embodiment of the application. As shown in fig. 2, a jacket underwater stress detection system 300 based on a distributed optical fiber sensing system according to an embodiment of the present application includes an inference module, wherein the inference module includes: a strain data acquisition unit 310; a stress distribution feature extraction unit 320; a spatial enhancement unit 330; a local feature deployment unit 340; a context encoding unit 350; and a stress detection result generation unit 360.
The strain data acquisition unit 310 is configured to acquire a strain cloud provided by a distributed optical fiber sensing system deployed in a jacket to be tested; the stress distribution feature extraction unit 320 is configured to pass the strain cloud graph through a convolutional neural network model serving as a feature extractor to obtain a strain feature graph; the spatial enhancement unit 330 is configured to pass the strain feature map through a spatial attention module to obtain a spatially enhanced strain feature map; the local feature expansion unit 340 is configured to expand each feature matrix of the spatial enhanced strain feature map along a channel dimension into a feature vector to obtain a plurality of local region strain feature vectors; the context encoding unit 350 is configured to pass the plurality of local area strain feature vectors through a context encoder based on a converter to obtain a classification feature vector; and the stress detection result generating unit 360 is configured to pass the classification feature vectors through a classifier to obtain a classification result, where the classification result is used to indicate whether the stress distribution of the jacket to be tested is normal.
Fig. 4 is a system architecture diagram of a jacket underwater stress detection system based on a distributed optical fiber sensing system according to an embodiment of the application. As shown in fig. 4, in the system architecture of the jacket underwater stress detection system 300 based on the distributed optical fiber sensing system, in the inference process, a strain cloud provided by the distributed optical fiber sensing system deployed in the jacket to be detected is first obtained through the strain data acquisition unit 310; then, the stress distribution feature extraction unit 320 obtains a strain feature map by passing the strain cloud map obtained by the strain data acquisition unit 310 through a convolutional neural network model as a feature extractor; the spatial enhancement unit 330 passes the strain feature map obtained by the stress distribution feature extraction unit 320 through a spatial attention module to obtain a spatially enhanced strain feature map; then, the local feature expansion unit 340 expands each feature matrix along the channel dimension of the spatial enhanced strain feature map obtained by the spatial enhancement unit 330 into a feature vector to obtain a plurality of local region strain feature vectors; the context encoding unit 350 passes the plurality of local region strain feature vectors obtained by the local feature expansion unit 340 through a context encoder based on a converter to obtain a classification feature vector; further, the stress detection result generating unit 360 passes the classified feature vectors through a classifier to obtain a classification result, where the classification result is used to indicate whether the stress distribution of the jacket to be tested is normal.
Specifically, in the operation process of the jacket underwater stress detection system 300 based on the distributed optical fiber sensing system, the strain data acquisition unit 310 is configured to acquire a strain cloud provided by the distributed optical fiber sensing system deployed on the jacket to be tested. The conventional scheme for detecting the underwater stress of the jacket of the offshore wind farm has the advantages of high requirement on personnel operation skills, low efficiency-cost ratio and low intelligence level, and is difficult to deploy a sensor, so that the detection accuracy of the stress distribution of the jacket is reduced. At the same time, it is also considered that distributed detection does not rely on manufactured sensors, but rather utilizes optical fibers, unlike conventional sensors that rely on discrete sensors measured at predetermined points. The optical fiber itself is the sensing element without any additional sensors in the optical path. Since the optical fiber is a sensor, it is also a cost effective and efficient method to deploy easily even in the most hostile and unusual environments, and therefore in the solution of the present application, the stress distribution detection of the jacket is performed by intelligent analysis of the strain cloud provided by the distributed optical fiber sensing system, which in one specific example of the present application can be provided by the distributed optical fiber sensing system deployed on the jacket under test.
Specifically, during the operation of the jacket underwater stress detection system 300 based on the distributed optical fiber sensing system, the stress distribution feature extraction unit 320 is configured to pass the strain cloud through a convolutional neural network model as a feature extractor to obtain a strain feature map. It should be understood that the strain cloud pictures are cloud pictures of stress magnitude and distribution of each part of a mechanism generated in the finite element software under the action of given external force, and can well reflect the stress distribution condition of the jacket to be measured. And then, processing the strain cloud picture in a convolutional neural network model which has excellent performance in the aspect of image implicit feature extraction and serves as a feature extractor to extract implicit feature distribution information of the jacket to be detected in the strain cloud picture, so that a strain feature picture is obtained. In one particular example, the first convolutional neural network comprises a plurality of neural network layers cascaded with one another, wherein each neural network layer comprises a convolutional layer, a pooling layer, and an activation layer. In the encoding process of the convolutional neural network, each layer of the convolutional neural network performs convolution processing based on a convolution kernel on input data by using the convolutional layer in the forward transmission process of the layer, performs pooling processing on a convolutional feature map output by the convolutional layer by using the pooling layer, and performs activation processing on the pooled feature map output by the pooling layer by using the activation layer, wherein the input data of the first layer of the convolutional neural network is the strain cloud map, and the output data of the last layer of the convolutional neural network is the strain feature map.
Fig. 5 is a flowchart of convolutional neural network coding in a jacket underwater stress detection system based on a distributed optical fiber sensing system according to an embodiment of the present application. As shown in fig. 5, in the convolutional neural network coding process, the convolutional neural network coding method includes: performing, in a layer forward pass, input data using the layers of the convolutional neural network model as a feature extractor: s210, performing convolution processing on input data to obtain a convolution characteristic diagram; s220, pooling the convolution characteristic diagram based on a local characteristic matrix to obtain a pooled characteristic diagram; s230, carrying out nonlinear activation on the pooling feature map to obtain an activation feature map; the output of the last layer of the convolutional neural network as the feature extractor is the strain feature map, and the input of the first layer of the convolutional neural network as the feature extractor is the strain cloud map.
Specifically, during the operation of the jacket underwater stress detection system 300 based on the distributed optical fiber sensing system, the spatial enhancement unit 330 is configured to pass the strain characteristic map through a spatial attention module to obtain a spatially enhanced strain characteristic map. Considering that the stress distribution situation of the jacket to be tested on the spatial position should be focused more when detecting the stress distribution of the jacket to be tested, therefore, the implicit feature extraction of the jacket to be tested should be performed focusing on the stress distribution feature information of the jacket to be tested on the space. Specifically, in the technical solution of the present application, the strain characteristic map is passed through a spatial attention module to obtain a spatial enhanced strain characteristic map. It should be understood that the image features extracted by the spatial attention reflect the feature difference weights in the spatial dimension of the stress distribution information about the jacket in the strain cloud, so as to suppress or strengthen the features of different spatial positions. More specifically, the passing the strain signature through a spatial attention module to obtain a spatially enhanced strain signature includes: calculating a global mean value of each feature matrix of the strain feature map along the spatial dimension to obtain a spatial feature vector, and inputting the spatial feature vector into a Softmax activation function to obtain a spatial attention weight feature vector; and then, weighting each feature matrix of the strain feature map along the spatial dimension by taking the feature value of each position in the spatial attention weight feature vector as a weight to obtain the spatial enhancement strain feature map.
Fig. 6 is a flow chart illustrating a spatial enhancement process in a jacket underwater stress detection system based on a distributed optical fiber sensing system according to an embodiment of the application. As shown in fig. 6, the spatial enhancement process includes: s310, calculating a global mean value of each feature matrix of the strain feature map along the space dimension to obtain a space feature vector; s320, inputting the space feature vector into a Softmax activation function to obtain a space attention weight feature vector; and S330, respectively weighting each feature matrix of the strain feature map along the spatial dimension by taking the feature value of each position in the spatial attention weight feature vector as a weight so as to obtain the spatial enhancement strain feature map.
Specifically, during the operation of the jacket underwater stress detection system 300 based on the distributed optical fiber sensing system, the local feature expansion unit 340 is configured to expand each feature matrix of the spatial enhanced strain feature map along the channel dimension into feature vectors to obtain a plurality of local region strain feature vectors. It should be understood that, when the detection and determination are performed on whether the stress distribution of the jacket to be tested is normal, the stresses in the local regions on the jacket to be tested have mutual influence, that is, the stress features in the local regions on the jacket to be tested have associated feature distribution information in a high-dimensional space, so that in order to be able to sufficiently extract the stress distribution feature information on the jacket to perform accurate detection on whether the stress distribution is normal, it is necessary to further extract global associated feature information of the stress distribution in the space. Specifically, in the technical solution of the present application, each feature matrix along the channel dimension of the spatial enhanced strain feature map is further expanded into feature vectors in rows or columns, so as to obtain a plurality of local area strain feature vectors having a plurality of local area spatial implicit features of the jacket to be measured. In one specific example of the present application, each feature matrix along the channel dimension of the spatially enhanced strain signature may be expanded into a plurality of local region strain signature vectors in either the row vector or column vector dimensions.
Specifically, during the operation of the jacket underwater stress detection system 300 based on the distributed optical fiber sensing system, the context encoding unit 350 is configured to pass the plurality of local area strain feature vectors through a context encoder based on a converter to obtain a classification feature vector. Namely, the local region strain feature vectors are encoded by a context encoder based on a converter, so as to extract the stress distribution implicit feature of each local region of the catheter rack on the spatial position based on global relevance feature distribution information, and thus, a classification feature vector is obtained. That is, based on the transform idea, with the converter being able to capture the long-distance context-dependent characteristic, each of the local region strain feature vectors is subjected to global context-based semantic encoding to obtain a context-semantic association feature representation, i.e., the classification feature vector, with the global semantic association of the local region strain feature vectors as context. It should be understood that in the technical solution of the present application, the context semantic association feature representation of the implicit feature of stress distribution of each local area of the jacket in spatial position with respect to the implicit feature of stress distribution of the jacket as a whole can be captured by the converter-based encoder. More specifically, passing the plurality of local region strain feature vectors through the transducer-based context encoder to obtain a plurality of context semantic region strain feature vectors; and cascading the plurality of context semantic region strain feature vectors to obtain the classification feature vector. In one example, the plurality of context semantic region strain feature vectors are fused to obtain a classification feature vector in the following formula; wherein the formula is: v c ＝Concat[V 1 ，V 2 ，...V n ]In which V is 1 ，V 2 ，...V n A strain feature vector, concat [, representing the plurality of context semantic regions]Representing a cascade function, V c Representing the classification feature vector.
Fig. 7 is a flowchart of a context encoding process in a jacket underwater stress detection system based on a distributed optical fiber sensing system according to an embodiment of the application. As shown in fig. 7, in the context coding process, the following steps are included: s410, performing one-dimensional arrangement on the local region strain characteristic vectors to obtain a global region strain characteristic vector; s420, calculating the product of the global region strain characteristic vector and the transposed vector of each region strain characteristic vector in the plurality of local region strain characteristic vectors to obtain a plurality of self-attention correlation matrixes; s430, respectively carrying out standardization processing on each self-attention correlation matrix in the self-attention correlation matrices to obtain a plurality of standardized self-attention correlation matrices; s440, obtaining a plurality of probability values by passing each normalized self-attention correlation matrix in the plurality of normalized self-attention correlation matrices through a Softmax classification function; s450, weighting each region strain characteristic vector in the local region strain characteristic vectors by taking each probability value in the probability values as a weight to obtain the context semantic region strain characteristic vectors.
Specifically, in the operation process of the jacket underwater stress detection system 300 based on the distributed optical fiber sensing system, the stress detection result generation unit 360 is configured to pass the classification feature vector through a classifier to obtain a classification result, where the classification result is used to indicate whether the stress distribution of the jacket to be detected is normal. That is, the classified feature vectors are passed through a classifier to obtain a classification result indicating whether the stress distribution of the jacket to be tested is normal or not. Like this, can be intelligently whether normally go on accurately detecting to the stress distribution of offshore wind power plant jacket to guarantee offshore wind power plant jacket's security, reduce enterprise later stage cost of maintenance. In particular, the classifier includes a plurality of fully-connected layers and a Softmax layer cascaded with a last fully-connected layer of the plurality of fully-connected layers. In a specific example, the classification feature vector is subjected to multiple full-join encoding using multiple full-join layers of the classifier to obtain an encoded classification feature vector; and then inputting the coding classification feature vector into a Softmax layer of the classifier, namely classifying the coding classification feature vector by using the Softmax classification function to obtain a classification result for indicating whether the stress distribution of the jacket to be tested is normal or not. More specifically, the classification feature vector is full-connected encoded using a plurality of full-connection layers of the classifier to obtain an encoded classification feature vector; and passing the encoding classification feature vector through a Softmax classification function of the classifier to obtain the classification result. More specifically, the passing the classified feature vector through a classifier to obtain a classification result includes: processing the classification feature vector using the classifier to obtain a classification result with the following formula:
O＝softmax{(W n ,B n ):…:(W 1 ,B 1 ) I X }, wherein W 1 To W n As a weight matrix, B 1 To B n Is the bias vector and X is the classification feature vector.
It will be appreciated that the convolutional neural network model as a feature extractor, the spatial attention module, the transformer-based context encoder, and the classifier need to be trained prior to making inferences using the neural network model described above. That is to say, in the jacket underwater stress detection system based on the distributed optical fiber sensing system of the present application, a training module is further included, which is used for training the convolutional neural network model as the feature extractor, the spatial attention module, the context encoder based on the converter, and the classifier.
Fig. 3 is a block diagram of a jacket underwater stress detection system based on a distributed optical fiber sensing system according to an embodiment of the application. As shown in fig. 3, the jacket underwater stress detection system 300 based on the distributed optical fiber sensing system according to the embodiment of the present application further includes a training module 400, which includes: a training data acquisition unit 410; training stress distribution feature extraction unit 420; a training space enhancing unit 430; training a local feature expansion unit 440; a training context encoding unit 450; a classification loss unit 460; and a training unit 470.
The training data obtaining unit 410 is configured to obtain a training strain cloud chart and a true value of whether the stress distribution of the jacket to be tested is normal; the training stress distribution feature extraction unit 420 is configured to pass the training strain cloud graph through the convolutional neural network model serving as a feature extractor to obtain a training strain feature graph; the training spatial enhancement unit 430 is configured to pass the training strain feature map through the spatial attention module to obtain a training spatial enhancement strain feature map; the training local feature expanding unit 440 is configured to expand each feature matrix of the training space enhanced strain feature map along a channel dimension into a feature vector to obtain a plurality of training local region strain feature vectors; the training context encoding unit 450 is configured to pass the plurality of training local area strain feature vectors through the converter-based context encoder to obtain training classification feature vectors; the classification loss unit 460 is configured to pass the training classification feature vector through the classifier to obtain a classification loss function value; the training unit 470 is configured to train the convolutional neural network model as the feature extractor, the spatial attention module, the converter-based context encoder, and the classifier through gradient descent direction propagation based on the classification loss function values, wherein in each iteration of the training, a free label optimization factor based on cross-classifier soft similarity of the training classification feature vector is calculated as a weighting weight to perform weighted iteration on the training classification feature vector.
Fig. 8 is a system architecture diagram of a training module in a jacket underwater stress detection system based on a distributed optical fiber sensing system according to an embodiment of the present application. As shown in fig. 8, in the system architecture of the jacket underwater stress detection system 300 based on the distributed optical fiber sensing system, in the training process, a training strain cloud chart and a true value of whether the stress distribution of the jacket to be tested is normal or not are first obtained through the training data obtaining unit 410; the training stress distribution feature extraction unit 420 passes the training strain cloud acquired by the training data acquisition unit 410 through the convolutional neural network model as a feature extractor to obtain a training strain feature map; then, the training space enhancing unit 430 passes the training strain feature map extracted by the training stress distribution feature extracting unit 420 through the space attention module to obtain a training space enhancing strain feature map; the training local feature expanding unit 440 expands each feature matrix along the channel dimension of the training space enhanced strain feature map obtained by the training space enhancing unit 430 into feature vectors to obtain a plurality of training local region strain feature vectors; then, the training context encoding unit 450 passes the plurality of training local region strain feature vectors obtained by the training local feature expanding unit 440 through the converter-based context encoder to obtain training classification feature vectors; the classification loss unit 460 passes the training classification feature vector through the classifier to obtain a classification loss function value; further, the training unit 470 trains the convolutional neural network model as a feature extractor, the spatial attention module, the converter-based context encoder, and the classifier through gradient descent direction propagation based on the classification loss function values, wherein, in each iteration of the training, a free label optimization factor based on cross-classifier soft similarity of the training classification feature vector is calculated as a weighting weight to perform weighted iteration on the training classification feature vector.
In particular, in the technical solution of the present application, when the plurality of local area strain feature vectors are obtained by the context encoder based on the converter, the plurality of local area strain feature vectors are directly concatenated by the context encoder based on the converter to obtain the classification feature vector, so even if the context encoder based on the converter can extract the correlation between the plurality of local area strain feature vectors to improve the correlation of the feature semantic distribution between the plurality of context local area strain feature vectors, in the case of directly concatenating the feature vectors, since the explicit correlation between the plurality of context local area strain feature vectors is low, the overall distribution of the classification feature vectors still has strong discreteness, so that the training of the classifier is difficult, especially the convergence of the label value of the classifier is difficult. Therefore, soft label learning is preferably used instead of the commonly used hard label learning, in particular, at each iteration, the free label optimization factors of the classification feature vectors based on cross-classifier soft similarity are calculated for the label values of the classifiers;
wherein the formula is:
v is the classification feature vector, M is a weighting matrix of the classifier on the classification feature vector,
and &>
Respectively representing tensor multiplication and tensor addition, d (·,) represents the distance between vectors, | | · includes 2 Denotes a two-norm vector, and α and β are weight superparameters, exp (-) denotes an exponential operation of a vector, which denotes calculation of a natural exponential function value raised to the eigenvalue of each position in the vector.
Here, the cross-classifier soft similarity-based free label optimization factor simulates pseudo classes based on classifier weight matrices through the cross-classifier soft similarity of the classification feature vectors and the weight matrices of the classifier by performing bidirectional clustering on the classification feature vectors and the classifier before calculating the classification probability of the feature vectors with hard label values, so that the classification quantization loss caused by hard label learning is avoided through soft similarity learning, the free label optimization focusing more on the intrinsic weight structure of the classifier is realized, the training of the label values of the classifier is optimized, and the training speed of the classifier is improved. Like this, can be intelligently whether normally carry out accurate the detection to the stress distribution of offshore wind power field jacket to guarantee the security of offshore wind power jacket, reduce enterprise later stage cost of maintenance.
In summary, a jacket underwater stress detection system 300 based on a distributed optical fiber sensing system according to an embodiment of the present application is illustrated, which performs stress distribution detection of a jacket by employing an artificial intelligence detection technique based on deep learning to perform intelligent analysis of a strain cloud provided by the distributed optical fiber sensing system. Specifically, stress distribution characteristic information focused on the spatial position of the jacket in the strain cloud image is extracted through an artificial intelligence algorithm, and overall global correlation characteristic distribution extraction is performed based on the stress characteristic distribution of each local position, so that the precision of detecting the stress distribution of the jacket is improved. Like this, can be intelligently whether normally carry out accurate the detection to the stress distribution of offshore wind power field jacket to guarantee the security of offshore wind power jacket, reduce enterprise later stage cost of maintenance.
As described above, the jacket underwater stress detection system based on the distributed optical fiber sensing system according to the embodiment of the present application can be implemented in various terminal devices. In one example, the jacket underwater stress detection system 300 based on the distributed optical fiber sensing system according to the embodiment of the present application may be integrated into a terminal device as one software module and/or hardware module. For example, the jacket subsea stress detection system 300 based on a distributed fiber optic sensing system may be a software module in the operating system of the terminal device, or may be an application developed for the terminal device; of course, the jacket underwater stress detection system 300 based on the distributed optical fiber sensing system can also be one of numerous hardware modules of the terminal equipment.
Alternatively, in another example, the distributed optical fiber sensing system based jacket subsea stress detection system 300 and the terminal device may also be separate devices, and the distributed optical fiber sensing system based jacket subsea stress detection system 300 may be connected to the terminal device via a wired and/or wireless network and transmit the interactive information in an agreed data format.
Exemplary method
Fig. 9 is a flowchart of a jacket underwater stress detection method based on a distributed optical fiber sensing system according to an embodiment of the application. As shown in fig. 9, a jacket underwater stress detection method based on a distributed optical fiber sensing system according to an embodiment of the present application includes the steps of: s110, strain data acquisition: acquiring a strain cloud chart provided by a distributed optical fiber sensing system deployed on a jacket to be tested; s120, stress distribution characteristic extraction: passing the strain cloud picture through a convolutional neural network model serving as a feature extractor to obtain a strain feature picture; s130, a space enhancing step: passing the strain signature through a spatial attention module to obtain a spatially enhanced strain signature; s140, local feature unfolding: expanding each feature matrix of the spatial enhancement strain feature map along the channel dimension into feature vectors to obtain a plurality of local area strain feature vectors; s150, context coding: passing the plurality of local region strain feature vectors through a transducer-based context encoder to obtain a classification feature vector; and S160, a stress detection result generating step: and passing the classified characteristic vectors through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the stress distribution of the jacket to be tested is normal or not.
In an example, in the method for detecting underwater stress of a jacket based on a distributed optical fiber sensing system, the step S120 includes: respectively performing, in forward pass of layers, input data using layers of the convolutional neural network model as a feature extractor: performing convolution processing on input data to obtain a convolution characteristic diagram; pooling the convolution feature map based on a local feature matrix to obtain a pooled feature map; and performing nonlinear activation on the pooled feature map to obtain an activated feature map; the output of the last layer of the convolutional neural network as the feature extractor is the strain feature map, and the input of the first layer of the convolutional neural network as the feature extractor is the strain cloud map.
In an example, in the method for detecting underwater stress of a jacket based on a distributed optical fiber sensing system, the step S130 includes: calculating a global mean value of each feature matrix of the strain feature map along a spatial dimension to obtain a spatial feature vector; inputting the spatial feature vector into a Softmax activation function to obtain a spatial attention weight feature vector; and respectively weighting each feature matrix of the strain feature map along the spatial dimension by taking the feature value of each position in the spatial attention weight feature vector as a weight so as to obtain the spatial enhancement strain feature map.
In an example, in the method for detecting underwater stress of a jacket based on a distributed optical fiber sensing system, the step S140 includes: and expanding each characteristic matrix of the spatial enhancement strain characteristic diagram along the channel dimension according to the row vector or the column vector dimension to obtain a plurality of local area strain characteristic vectors.
In an example, in the jacket underwater stress detection method based on the distributed optical fiber sensing system, the step S150 includes: passing the plurality of local region strain feature vectors through the transducer-based context encoder to obtain a plurality of context semantic region strain feature vectors; and cascading the plurality of context semantic region strain feature vectors to obtain the classification feature vector. More specifically, the passing the plurality of local area strain feature vectors through the converter-based context encoder to obtain a plurality of context semantic area strain feature vectors includes: performing one-dimensional arrangement on the plurality of local region strain characteristic vectors to obtain a global region strain characteristic vector; calculating a product between the global region strain eigenvector and a transposed vector of each region strain eigenvector in the plurality of local region strain eigenvectors to obtain a plurality of self-attention correlation matrices; respectively carrying out standardization processing on each self-attention correlation matrix in the self-attention correlation matrixes to obtain a plurality of standardized self-attention correlation matrixes; obtaining a plurality of probability values by passing each normalized self-attention correlation matrix in the plurality of normalized self-attention correlation matrices through a Softmax classification function; and respectively weighting each region strain characteristic vector in the local region strain characteristic vectors by taking each probability value in the probability values as a weight so as to obtain the context semantic region strain characteristic vectors.
In an example, in the method for detecting underwater stress of a jacket based on a distributed optical fiber sensing system, the step S160 includes: performing full-join encoding on the classification feature vectors using a plurality of full-join layers of the classifier to obtain encoded classification feature vectors; and passing the encoding classification feature vector through a Softmax classification function of the classifier to obtain the classification result.
In summary, a jacket underwater stress detection method based on a distributed optical fiber sensing system according to an embodiment of the present application is illustrated, which performs stress distribution detection of a jacket by using an artificial intelligence detection technology based on deep learning to perform intelligent analysis of a strain cloud provided by the distributed optical fiber sensing system. Specifically, stress distribution characteristic information focused on the spatial position of the jacket in the strain cloud image is extracted through an artificial intelligence algorithm, and overall global correlation characteristic distribution extraction is performed based on the stress characteristic distribution of each local position, so that the precision of detecting the stress distribution of the jacket is improved. Like this, can be intelligently whether normally carry out accurate the detection to the stress distribution of offshore wind power field jacket to guarantee the security of offshore wind power jacket, reduce enterprise later stage cost of maintenance.
Exemplary electronic device
Next, an electronic apparatus according to an embodiment of the present application is described with reference to fig. 10.
FIG. 10 illustrates a block diagram of an electronic device in accordance with an embodiment of the present application.
As shown in fig. 10, the electronic device 10 includes one or more processors 11 and memory 12.
The processor 11 may be a Central Processing Unit (CPU) or other form of processing unit having data processing capabilities and/or instruction execution capabilities, and may control other components in the electronic device 10 to perform desired functions.
In one example, the electronic device 10 may further include: an input device 13 and an output device 14, which are interconnected by a bus system and/or other form of connection mechanism (not shown).
The input device 13 may include, for example, a keyboard, a mouse, and the like.
The output device 14 can output various information including the classification result to the outside. The output devices 14 may include, for example, a display, speakers, a printer, and a communication network and its connected remote output devices, among others.
Of course, for simplicity, only some of the components of the electronic device 10 relevant to the present application are shown in fig. 10, and components such as buses, input/output interfaces, and the like are omitted. In addition, the electronic device 10 may include any other suitable components depending on the particular application.
Exemplary computer program product and computer-readable storage Medium
In addition to the above-described methods and apparatus, embodiments of the present application may also be a computer program product comprising computer program instructions that, when executed by a processor, cause the processor to perform the steps in the functions of the distributed optical fiber sensing system based jacket underwater stress detection method according to various embodiments of the present application described in the "exemplary systems" section of this specification above.
The computer program product may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, C + + or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages, for carrying out operations according to embodiments of the present application. The program code may execute entirely on the user's computing device, partly on the user's device, as a stand-alone software package, partly on the user's computing device and partly on a remote computing device, or entirely on the remote computing device or server.
Furthermore, embodiments of the present application may also be a computer readable storage medium having stored thereon computer program instructions that, when executed by a processor, cause the processor to perform the steps in the functions of the distributed optical fiber sensing system based jacket underwater stress detection method according to various embodiments of the present application described in the "exemplary systems" section of this specification above.
The computer-readable storage medium may take any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may include, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or a combination of any of the foregoing. More specific examples (a non-exhaustive list) of the readable storage medium include: an electrical connection having one or more wires, a portable disk, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
The foregoing describes the general principles of the present application in conjunction with specific embodiments, however, it is noted that the advantages, effects, etc. mentioned in the present application are merely examples and are not limiting, and they should not be considered essential to the various embodiments of the present application. Furthermore, the foregoing disclosure of specific details is provided for purposes of illustration and understanding only, and is not intended to limit the application to the details which are set forth in order to provide a thorough understanding of the present application.
The block diagrams of devices, apparatuses, systems referred to in this application are only given as illustrative examples and are not intended to require or imply that the connections, arrangements, configurations, etc. must be made in the manner shown in the block diagrams. These devices, apparatuses, devices, systems may be connected, arranged, configured in any manner, as will be appreciated by those skilled in the art. Words such as "including," "comprising," "having," and the like are open-ended words that mean "including, but not limited to," and are used interchangeably therewith. As used herein, the words "or" and "refer to, and are used interchangeably with, the word" and/or, "unless the context clearly dictates otherwise. The word "such as" is used herein to mean, and is used interchangeably with, the phrase "such as but not limited to".
It should also be noted that in the devices, apparatuses, and methods of the present application, the components or steps may be decomposed and/or recombined. These decompositions and/or recombinations are to be considered as equivalents of the present application.
The previous description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other aspects without departing from the scope of the application. Thus, the present application is not intended to be limited to the aspects shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
The foregoing description has been presented for purposes of illustration and description. Furthermore, the description is not intended to limit embodiments of the application to the form disclosed herein. While a number of example aspects and embodiments have been discussed above, those of skill in the art will recognize certain variations, modifications, alterations, additions and sub-combinations thereof.
Claims (10)
1. A jacket underwater stress detection system based on a distributed optical fiber sensing system is characterized by comprising:
the strain data acquisition unit is used for acquiring a strain cloud picture provided by a distributed optical fiber sensing system deployed on a jacket to be tested;
the stress distribution characteristic extraction unit is used for enabling the strain cloud picture to pass through a convolutional neural network model serving as a characteristic extractor to obtain a strain characteristic picture;
the spatial enhancement unit is used for passing the strain characteristic map through a spatial attention module to obtain a spatial enhancement strain characteristic map;
the local feature expansion unit is used for expanding each feature matrix of the space enhanced strain feature map along the channel dimension into feature vectors so as to obtain a plurality of local region strain feature vectors;
a context coding unit, configured to pass the plurality of local area strain feature vectors through a converter-based context coder to obtain a classification feature vector; and
and the stress detection result generation unit is used for enabling the classification characteristic vectors to pass through a classifier to obtain a classification result, and the classification result is used for indicating whether the stress distribution of the jacket to be detected is normal or not.
2. The jacket underwater stress detection system based on the distributed optical fiber sensing system according to claim 1, wherein the stress distribution feature extraction unit is further configured to: performing, in a layer forward pass, input data using the layers of the convolutional neural network model as a feature extractor:
carrying out convolution processing on input data to obtain a convolution characteristic diagram;
pooling the convolution feature map based on a local feature matrix to obtain a pooled feature map; and
performing nonlinear activation on the pooled feature map to obtain an activated feature map;
the output of the last layer of the convolutional neural network as the feature extractor is the strain feature map, and the input of the first layer of the convolutional neural network as the feature extractor is the strain cloud map.
3. The jacket underwater stress detection system based on the distributed optical fiber sensing system according to claim 2, wherein the space enhancement unit is further configured to:
calculating a global mean value of each feature matrix of the strain feature map along the spatial dimension to obtain a spatial feature vector;
inputting the spatial feature vector into a Softmax activation function to obtain a spatial attention weight feature vector; and
and respectively weighting each feature matrix of the strain feature map along the spatial dimension by taking the feature value of each position in the spatial attention weight feature vector as a weight so as to obtain the spatial enhancement strain feature map.
4. The jacket underwater stress detection system based on the distributed optical fiber sensing system according to claim 3, wherein the local feature expansion unit is further configured to expand each feature matrix along a channel dimension of the spatial enhanced strain feature map into a plurality of local region strain feature vectors according to a row vector dimension or a column vector dimension.
5. The jacket underwater stress detection system based on the distributed optical fiber sensing system as claimed in claim 4, wherein the context encoding unit comprises:
a context associated feature extraction subunit, configured to pass the plurality of local area strain feature vectors through the converter-based context encoder to obtain a plurality of context semantic area strain feature vectors;
and the cascading subunit is used for cascading the plurality of context semantic region strain feature vectors to obtain the classification feature vector.
6. The jacket underwater stress detection system based on the distributed optical fiber sensing system according to claim 5, wherein the context associated feature extraction subunit is further configured to:
performing one-dimensional arrangement on the plurality of local region strain characteristic vectors to obtain a global region strain characteristic vector;
calculating the product of the global area strain characteristic vector and the transposed vector of each area strain characteristic vector in the plurality of local area strain characteristic vectors to obtain a plurality of self-attention correlation matrixes;
respectively carrying out standardization processing on each self-attention correlation matrix in the self-attention correlation matrixes to obtain a plurality of standardized self-attention correlation matrixes;
obtaining a plurality of probability values by passing each normalized self-attention correlation matrix in the plurality of normalized self-attention correlation matrices through a Softmax classification function;
and respectively weighting each region strain characteristic vector in the local region strain characteristic vectors by taking each probability value in the probability values as a weight so as to obtain the context semantic region strain characteristic vectors.
7. The jacket underwater stress detection system based on the distributed optical fiber sensing system as claimed in claim 6, wherein the stress detection result generation unit comprises:
a full-concatenation coding subunit, configured to perform full-concatenation coding on the classification feature vector using a plurality of full-concatenation layers of the classifier to obtain a coded classification feature vector; and
and the classification result generating subunit is used for enabling the coded classification feature vector to pass through a Softmax classification function of the classifier so as to obtain the classification result.
8. The jacket underwater stress detection system based on the distributed optical fiber sensing system as claimed in claim 7, further comprising a training module for training the convolutional neural network model as a feature extractor, the spatial attention module, the converter-based context encoder and the classifier;
wherein, the training module includes:
the training data acquisition unit is used for acquiring a training strain cloud picture and a true value of whether the stress distribution of the jacket to be tested is normal or not;
the training stress distribution characteristic extraction unit is used for enabling the training strain cloud picture to pass through the convolutional neural network model serving as the characteristic extractor to obtain a training strain characteristic picture;
the training space enhancement unit is used for enabling the training strain characteristic map to pass through the space attention module to obtain a training space enhancement strain characteristic map;
the training local feature expansion unit is used for expanding each feature matrix of the training space enhanced strain feature map along the channel dimension into feature vectors to obtain a plurality of training local region strain feature vectors;
a training context encoding unit, configured to pass the training local region strain feature vectors through the converter-based context encoder to obtain training classification feature vectors;
the classification loss unit is used for enabling the training classification feature vector to pass through the classifier to obtain a classification loss function value;
a training unit for training the convolutional neural network model as a feature extractor, the spatial attention module, the converter-based context encoder, and the classifier through gradient descent direction propagation based on the classification loss function values, wherein, in each iteration of the training, a free label optimization factor based on cross-classifier soft similarity of the training classification feature vector is calculated as a weighting weight to perform weighted iteration on the training classification feature vector.
9. The jacket underwater stress detection system based on the distributed optical fiber sensing system according to claim 8, wherein in each iteration of the training, a free label optimization factor based on cross-classifier soft similarity of the training classification feature vector is calculated as a label value of the classifier according to the following formula;
wherein the formula is:
wherein V is the classification feature vector, M is a weighting matrix of the classifier on the training classification feature vector,
and &>
Respectively representing tensor multiplication and tensor addition, d (·,) representing the distance between vectors, | ·| | survival 2 Representing the two-norm of the vector, and alpha and beta are weight hyperparameters.
10. A jacket underwater stress detection method based on a distributed optical fiber sensing system is characterized by comprising the following steps:
strain data acquisition: acquiring a strain cloud chart provided by a distributed optical fiber sensing system deployed on a jacket to be tested;
and (3) stress distribution characteristic extraction: passing the strain cloud picture through a convolutional neural network model serving as a feature extractor to obtain a strain feature picture;
a spatial enhancement step: passing the strain signature through a spatial attention module to obtain a spatially enhanced strain signature;
local feature unfolding step: expanding each feature matrix of the spatial enhancement strain feature map along the channel dimension into feature vectors to obtain a plurality of local area strain feature vectors;
context coding step: passing the plurality of local region strain feature vectors through a transducer-based context encoder to obtain a classification feature vector; and
a stress detection result generation step: and passing the classified characteristic vectors through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the stress distribution of the jacket to be tested is normal or not.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211592492.9A CN115859437A (en) | 2022-12-13 | 2022-12-13 | Jacket underwater stress detection system based on distributed optical fiber sensing system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211592492.9A CN115859437A (en) | 2022-12-13 | 2022-12-13 | Jacket underwater stress detection system based on distributed optical fiber sensing system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115859437A true CN115859437A (en) | 2023-03-28 |
Family
ID=85672253
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211592492.9A Pending CN115859437A (en) | 2022-12-13 | 2022-12-13 | Jacket underwater stress detection system based on distributed optical fiber sensing system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115859437A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116400426A (en) * | 2023-06-06 | 2023-07-07 | 山东省煤田地质局第三勘探队 | Electromagnetic method-based data survey system |
| CN116680620A (en) * | 2023-07-28 | 2023-09-01 | 克拉玛依市紫光技术有限公司 | Preparation method and system of anti-emulsifying agent for fracturing |
| CN116678506A (en) * | 2023-08-02 | 2023-09-01 | 国检测试控股集团南京国材检测有限公司 | Wireless transmission heat loss detection device |
| CN116734750A (en) * | 2023-08-15 | 2023-09-12 | 山西锦烁生物医药科技有限公司 | Intelligent detection method and system for thickness of ice layer of ice rink based on optical fiber sensor |
| CN117065876A (en) * | 2023-09-05 | 2023-11-17 | 浙江艾领创矿业科技有限公司 | Intelligent Sand Mill System and Method |
| CN117113859A (en) * | 2023-10-24 | 2023-11-24 | 清华大学 | Method and device for predicting stress and stability of ocean disposal jacket |
| CN118709341A (en) * | 2024-08-27 | 2024-09-27 | 中国电建集团华东勘测设计研究院有限公司 | BIM technology-based hotel installation engineering pipeline arrangement optimization system and method |
-
2022
- 2022-12-13 CN CN202211592492.9A patent/CN115859437A/en active Pending
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116400426A (en) * | 2023-06-06 | 2023-07-07 | 山东省煤田地质局第三勘探队 | Electromagnetic method-based data survey system |
| CN116400426B (en) * | 2023-06-06 | 2023-08-29 | 山东省煤田地质局第三勘探队 | Electromagnetic method-based data survey system |
| CN116680620A (en) * | 2023-07-28 | 2023-09-01 | 克拉玛依市紫光技术有限公司 | Preparation method and system of anti-emulsifying agent for fracturing |
| CN116680620B (en) * | 2023-07-28 | 2023-10-27 | 克拉玛依市紫光技术有限公司 | Preparation method and system of anti-emulsifying agent for fracturing |
| CN116678506A (en) * | 2023-08-02 | 2023-09-01 | 国检测试控股集团南京国材检测有限公司 | Wireless transmission heat loss detection device |
| CN116678506B (en) * | 2023-08-02 | 2023-10-10 | 国检测试控股集团南京国材检测有限公司 | Wireless transmission heat loss detection device |
| CN116734750A (en) * | 2023-08-15 | 2023-09-12 | 山西锦烁生物医药科技有限公司 | Intelligent detection method and system for thickness of ice layer of ice rink based on optical fiber sensor |
| CN117065876A (en) * | 2023-09-05 | 2023-11-17 | 浙江艾领创矿业科技有限公司 | Intelligent Sand Mill System and Method |
| CN117065876B (en) * | 2023-09-05 | 2024-03-22 | 浙江艾领创矿业科技有限公司 | Control method and control system of intelligent sand mill |
| CN117113859A (en) * | 2023-10-24 | 2023-11-24 | 清华大学 | Method and device for predicting stress and stability of ocean disposal jacket |
| CN117113859B (en) * | 2023-10-24 | 2024-02-20 | 清华大学 | Method and device for predicting stress and stability of ocean disposal jacket |
| CN118709341A (en) * | 2024-08-27 | 2024-09-27 | 中国电建集团华东勘测设计研究院有限公司 | BIM technology-based hotel installation engineering pipeline arrangement optimization system and method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115859437A (en) | Jacket underwater stress detection system based on distributed optical fiber sensing system | |
| CN116168352B (en) | Power grid obstacle recognition processing method and system based on image processing | |
| CN112632972A (en) | Method for rapidly extracting fault information in power grid equipment fault report | |
| CN116245513B (en) | Automatic operation and maintenance system and method based on rule base | |
| CN114973222B (en) | Scene text recognition method based on explicit supervision attention mechanism | |
| CN116247824B (en) | Control method and system for power equipment | |
| CN115471216B (en) | Data management method of intelligent laboratory management platform | |
| CN113849648B (en) | Classification model training method, device, computer equipment and storage medium | |
| CN115186774B (en) | Intelligent cable stripping equipment and method thereof | |
| CN116257406A (en) | Gateway data management method and system for smart city | |
| CN115951883B (en) | Service component management system of distributed micro-service architecture and method thereof | |
| CN116308754B (en) | Bank credit risk early warning system and method thereof | |
| CN117158923A (en) | Remote home-care monitoring method based on meta universe | |
| CN116486308A (en) | Teaching management system and method based on intelligent education | |
| CN117521641A (en) | Automatic text proofreading system and method based on natural language processing | |
| CN115324843A (en) | Wind generating set fault diagnosis system and method based on monitoring data | |
| CN116340506A (en) | Text classification method based on BERT and pooling-free convolutional neural network | |
| CN117744635B (en) | English text automatic correction system and method based on intelligent AI | |
| CN117787287A (en) | Accident handling method, system and storage medium based on large model | |
| CN117231590A (en) | Fault prediction system and method for hydraulic system | |
| CN116467485B (en) | Video image retrieval construction system and method thereof | |
| CN117113987A (en) | Keyword intelligent distinguishing method and system based on user behavior characteristics | |
| CN116124448A (en) | Fault diagnosis system and method for wind power gear box | |
| CN116468894A (en) | Distance self-adaptive mask generation method for supervised learning of lithium battery pole piece | |
| CN116385733A (en) | High-precision positioning system and method for bucket wheel machine |
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 |