CN115993268A - Sampling system and method for anhydrous hydrogen fluoride - Google Patents
Sampling system and method for anhydrous hydrogen fluoride Download PDFInfo
- Publication number
- CN115993268A CN115993268A CN202310280557.4A CN202310280557A CN115993268A CN 115993268 A CN115993268 A CN 115993268A CN 202310280557 A CN202310280557 A CN 202310280557A CN 115993268 A CN115993268 A CN 115993268A
- Authority
- CN
- China
- Prior art keywords
- matrix
- classification
- feature
- pipeline
- internal pressure
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E60/00—Enabling technologies; Technologies with a potential or indirect contribution to GHG emissions mitigation
- Y02E60/30—Hydrogen technology
- Y02E60/32—Hydrogen storage
Landscapes
- Pipeline Systems (AREA)
- Examining Or Testing Airtightness (AREA)
Abstract
The application relates to the technical field of anhydrous hydrogen fluoride sampling, and particularly discloses a anhydrous hydrogen fluoride sampling system and a method thereof. Through the mode, whether leakage occurs in the sampling process of anhydrous hydrogen fluoride can be accurately pre-warned, so that the sampling safety of the anhydrous hydrogen fluoride is ensured.
Description
Technical Field
The present application relates to the field of anhydrous hydrogen fluoride sampling technology, and more particularly, to an anhydrous hydrogen fluoride sampling system and method.
Background
Anhydrous Hydrogen Fluoride (HF) is used as basic raw material of various fluorine refrigerants, fluoride salts and organic fluorine, and the application range of the anhydrous Hydrogen Fluoride (HF) covers the industries of air conditioning refrigeration, aerospace, automobiles, chemical industry, medicines and the like. Anhydrous hydrogen fluoride is very important as a raw material in the electrolytic production of nitrogen trifluoride gas. In order to strictly and effectively detect the quality of the incoming material of the anhydrous hydrogen fluoride, a sampling operation of the anhydrous hydrogen fluoride is indispensable. However, anhydrous hydrogen fluoride has a high risk as a chemical having high chemical activity and extremely strong corrosive stimulus, and it is necessary to design a device for sampling anhydrous hydrogen fluoride to eliminate the risk during the sampling process in order to safely perform the sampling operation without contamination.
In view of the above problems, chinese patent CN113624567a discloses a method and apparatus for sampling anhydrous hydrogen fluoride, which is to flow anhydrous hydrogen fluoride from a quantitative bottle into a sampling bottle, adjust a valve of a pipeline to connect the sampling pipeline to an acid mist absorption tower, and purge residual hydrogen fluoride in the sampling pipeline into the sampling bottle by using nitrogen in the process, thereby realizing sampling of anhydrous hydrogen fluoride. However, in the above-described sampling process, since the potential safety hazard is leakage, it is necessary to detect leakage of the sampling system and to early warn in time when sampling anhydrous hydrofluoric acid by the sampling system.
Accordingly, an optimized anhydrous hydrogen fluoride sampling system is desired.
Disclosure of Invention
The application provides a sampling system and a sampling method for anhydrous hydrogen fluoride, which can accurately early warn whether the anhydrous hydrogen fluoride leaks in the sampling process, so as to ensure the sampling safety of the anhydrous hydrogen fluoride.
In a first aspect, there is provided a sampling system for anhydrous hydrogen fluoride, the system comprising: a pressure data acquisition module for acquiring internal pressure values of the pipeline at a plurality of predetermined time points within a predetermined time period acquired by the pressure sensor; the internal pressure change module is used for arranging the internal pressure values of the pipelines at a plurality of preset time points into pipeline internal pressure input vectors according to the time dimension, and then calculating the difference value between the internal pressure values of every two adjacent positions in the pipeline internal pressure input vectors to obtain pipeline internal pressure change input vectors; the pressure time sequence association module is used for enabling the pressure change input vector in the pipeline to pass through a one-dimensional convolutional neural network model to obtain a pressure time sequence feature vector in the pipeline; the Gaussian strengthening module is used for strengthening the characteristic expression of the pressure time sequence characteristic vector in the pipeline based on a Gaussian density diagram to obtain a classification characteristic matrix; the optimizing module is used for optimizing the characteristic distribution of the classification characteristic matrix to obtain an optimized classification characteristic matrix; and the leakage early warning module is used for enabling the optimized classification feature matrix to pass through a classifier to obtain a classification result, wherein the classification result is used for indicating whether leakage early warning prompt is generated or not.
With reference to the first aspect, in an implementation manner of the first aspect, the pressure timing association module is configured to: each layer of the one-dimensional convolutional neural network model is used for respectively carrying out input data in forward transmission of the layer; performing convolution processing based on a one-dimensional convolution kernel on the input data to obtain a convolution feature map; pooling processing is carried out on the convolution feature images based on feature matrixes to obtain pooled feature images; non-linear activation is carried out on the pooled feature map so as to obtain an activated feature map; the input of the first layer of the one-dimensional convolutional neural network model is the pressure change input vector in the pipeline, and the output of the last layer of the one-dimensional convolutional neural network model is the pressure time sequence characteristic vector in the pipeline.
With reference to the first aspect, in an implementation manner of the first aspect, the gaussian enhancement module includes: a gaussian density map construction unit for constructing a gaussian density map of the pipeline internal pressure timing feature vector as follows Gao Sigong; wherein, the Gaussian formula is:
wherein->
Is the pressure time sequence characteristic vector in the pipeline, and +.>
Is the variance between the eigenvalues of the corresponding two locations in the pipeline internal pressure timing eigenvector,/ >
Variable representing the collaborative Gaussian density map, < >>
Representing a gaussian probability density function; and the Gaussian discrete unit is used for carrying out Gaussian discretization on the Gaussian distribution of each position in the Gaussian density map so as to obtain the classification characteristic matrix.
With reference to the first aspect, in an implementation manner of the first aspect, the optimizing module is configured to: carrying out eigenvoization bitwise displacement association matching optimization on the classification characteristic matrix by using the following optimization formula to obtain an optimized classification characteristic matrix; wherein, the optimization formula is:
wherein->
Is the classification feature matrix,/a>
And->
Is obtained by eigenvoicing the classification characteristic matrix>
Intrinsic value->
For said->
Edge of each eigenvalueThe obtained eigenvoization matrix is arranged diagonally, and +.>
And->
Are all diagonal matrix>
For the distance between the eigen-unitized matrix and the classification feature matrix,/i>
Representing matrix multiplication +.>
Representing matrix addition, ++>
Representing multiplication by location +.>
Classifying the feature matrix for the optimization.
With reference to the first aspect, in an implementation manner of the first aspect, the leakage early-warning module is configured to: processing the optimized classification feature matrix by using the classifier according to the following classification formula to obtain the classification result; wherein, the classification formula is: 
Wherein->
To->
As a matrix of weights, the weight matrix,
to->
For the bias vector +.>
Classifying the feature matrix for the optimization, +.>
Representing the projection of the optimized classification feature matrix as a vector,/->
Representing a normalized exponential function.
In a second aspect, there is provided a method of sampling anhydrous hydrogen fluoride, the method comprising: acquiring internal pressure values of the pipeline at a plurality of predetermined time points within a predetermined time period acquired by the pressure sensor; after arranging the internal pressure values of the pipelines at a plurality of preset time points into pipeline internal pressure input vectors according to the time dimension, calculating the difference value between the internal pressure values of every two adjacent positions in the pipeline internal pressure input vectors to obtain pipeline internal pressure change input vectors; the pressure change input vector in the pipeline is subjected to a one-dimensional convolutional neural network model to obtain a pressure time sequence characteristic vector in the pipeline; performing feature expression reinforcement based on a Gaussian density diagram on the pressure time sequence feature vector in the pipeline to obtain a classification feature matrix; performing feature distribution optimization on the classification feature matrix to obtain an optimized classification feature matrix; and the optimized classification feature matrix passes through a classifier to obtain a classification result, wherein the classification result is used for indicating whether leakage early warning prompt is generated or not.
With reference to the second aspect, in an implementation manner of the second aspect, the step of passing the pipeline internal pressure variation input vector through a one-dimensional convolutional neural network model to obtain a pipeline internal pressure time sequence feature vector includes: each layer of the one-dimensional convolutional neural network model is used for respectively carrying out input data in forward transmission of the layer; performing convolution processing based on a one-dimensional convolution kernel on the input data to obtain a convolution feature map; pooling processing is carried out on the convolution feature images based on feature matrixes to obtain pooled feature images; non-linear activation is carried out on the pooled feature map so as to obtain an activated feature map; the input of the first layer of the one-dimensional convolutional neural network model is the pressure change input vector in the pipeline, and the output of the last layer of the one-dimensional convolutional neural network model is the pressure time sequence characteristic vector in the pipeline.
With reference to the second aspect, in an implementation manner of the second aspect, performing feature expression enhancement on the pressure time sequence feature vector in the pipeline based on a gaussian density map to obtain a classification feature matrix, includes: constructing a gaussian density map of the pipeline internal pressure timing feature vector with a gaussian formula; wherein, the Gaussian formula is: 
Wherein
Is the pressure time sequence characteristic vector in the pipeline, and +.>
Is the variance between the eigenvalues of the corresponding two locations in the pipeline internal pressure timing eigenvector,/>
Variable representing the collaborative Gaussian density map, < >>
Representing a gaussian probability density function; and the Gaussian discrete unit is used for carrying out Gaussian discretization on the Gaussian distribution of each position in the Gaussian density map so as to obtain the classification characteristic matrix.
With reference to the second aspect, in an implementation manner of the second aspect, performing feature distribution optimization on the classification feature matrix to obtain an optimized classification feature matrix includes: carrying out eigenvoization bitwise displacement association matching optimization on the classification characteristic matrix by using the following optimization formula to obtain an optimized classification characteristic matrix; wherein, the optimization formula is:
wherein->
Is a matrix of the classification characteristic,
and->
Is obtained by eigenvoicing the classification characteristic matrix>
Intrinsic value->
For said->
The eigenvalues are arranged diagonally to obtain an eigenvoization matrix, and +.>
And->
Are all diagonal matrix>
For the distance between the eigen-unitized matrix and the classification feature matrix,/i >
Representing matrix multiplication +.>
Representing matrix addition, ++>
Representing multiplication by location +.>
Classifying the feature matrix for the optimization.
With reference to the second aspect, in an implementation manner of the second aspect, the step of passing the optimized classification feature matrix through a classifier to obtain a classification result includes: processing the optimized classification feature matrix by using the classifier according to the following classification formula to obtain the classification result; wherein, the classification formula is:
wherein->
To->
Is a weight matrix>
To->
For the bias vector +.>
For the optimized classification feature matrix,
representing the projection of the optimized classification feature matrix as a vector,/->
Representing a normalized exponential function.
The anhydrous hydrogen fluoride sampling system and the anhydrous hydrogen fluoride sampling method can accurately early warn whether leakage occurs in the sampling process of the anhydrous hydrogen fluoride, so that the sampling safety of the anhydrous hydrogen fluoride is guaranteed.
Drawings
FIG. 1 is a diagram of an application scenario of an anhydrous hydrogen fluoride sampling system and method according to an embodiment of the present application.
Fig. 2 is a schematic block diagram of a sampling system for anhydrous hydrogen fluoride in an embodiment of the present application.
Fig. 3 is a schematic flow chart of a sampling method of anhydrous hydrogen fluoride according to an embodiment of the present application.
Fig. 4 is a schematic diagram of a model architecture of a sampling method of anhydrous hydrogen fluoride according to an embodiment of the present application.
Detailed Description
The technical solutions in the present application will be described below with reference to the accompanying drawings.
Because of the deep learning-based deep neural network model, related terms and concepts of the deep neural network model that may be related to embodiments of the present application are described below.
1. Deep neural network model in the deep neural network model, the hidden layer may be a convolution layer and a pooling layer. The set of weight values corresponding to the convolutional layer is referred to as a filter, also referred to as a convolutional kernel. The filter and the input eigenvalue are both represented as a multi-dimensional matrix, correspondingly, the filter represented as a multi-dimensional matrix is also called a filter matrix, the input eigenvalue represented as a multi-dimensional matrix is also called an input eigenvalue, of course, besides the input eigenvalue, the eigenvector can also be input, and the input eigenvector is only exemplified by the input eigenvector. The operation of the convolution layer is called a convolution operation, which is to perform an inner product operation on a part of eigenvalues of the input eigenvalue matrix and weight values of the filter matrix.
The operation process of each convolution layer in the deep neural network model can be programmed into software, and then the output result of each layer of network, namely the output characteristic matrix, is obtained by running the software in an operation device. For example, the software performs inner product operation by taking the upper left corner of the input feature matrix of each layer of network as a starting point and taking the size of the filter as a window in a sliding window mode, and extracting data of one window from the feature value matrix each time. After the inner product operation is completed between the data of the right lower corner window of the input feature matrix and the filter, a two-dimensional output feature matrix of each layer of network can be obtained. The software repeats the above process until the entire output feature matrix for each layer of network is generated.
The convolution layer operation process is to slide a window with a filter size across the whole input image (i.e. the input feature matrix), and at each moment, to perform inner product operation on the input feature value covered in the window and the filter, wherein the step length of window sliding is 1. Specifically, the upper left corner of the input feature matrix is used as a starting point, the size of the filter is used as a window, the sliding step length of the window is 1, the input feature value of one window is extracted from the feature value matrix each time and the filter performs inner product operation, and when the data of the lower right corner of the input feature matrix and the filter complete inner product operation, a two-dimensional output feature matrix of the input feature matrix can be obtained.
Since it is often necessary to reduce the number of training parameters, the convolutional layer often requires a periodic introduction of a pooling layer, the only purpose of which is to reduce the spatial size of the image during image processing. The pooling layer may include an average pooling operator and/or a maximum pooling operator for sampling the input image to obtain a smaller size image. The average pooling operator may calculate pixel values in the image over a particular range to produce an average as a result of the average pooling. The max pooling operator may take the pixel with the largest value in a particular range as the result of max pooling. In addition, just as the size of the weighting matrix used in the convolutional layer should be related to the image size, the operators in the pooling layer should also be related to the image size. The size of the image output after the processing by the pooling layer can be smaller than the size of the image input to the pooling layer, and each pixel point in the image output by the pooling layer represents the average value or the maximum value of the corresponding sub-region of the image input to the pooling layer.
Since the functions actually required to be simulated in the deep neural network are nonlinear, but the previous rolling and pooling can only simulate linear functions, in order to introduce nonlinear factors in the deep neural network model to increase the characterization capability of the whole network, an activation layer is further arranged after the pooling layer, an activation function is arranged in the activation layer, and the commonly used activation functions include sigmoid, tanh, reLU functions and the like.
2. Softmax sort function is also known as soft max function, normalized exponential function. One K-dimensional vector containing arbitrary real numbers can be "compressed" into another K-dimensional real vector such that each element ranges between (0, 1) and the sum of all elements is 1. The Softmax classification function is commonly used to classify problems.
Having described the relevant terms and concepts of the deep neural network model that may be involved in embodiments of the present application, the following description will be made of the basic principles of the present application for the convenience of understanding by those skilled in the art.
As described above, anhydrous hydrogen fluoride is highly dangerous as a chemical having high chemical activity and extremely strong corrosive stimulus, and it is necessary to design a set of equipment for sampling anhydrous hydrogen fluoride to eliminate the danger during the sampling process in order to safely perform the sampling operation without contamination. In the anhydrous hydrogen fluoride sampling method and device disclosed in chinese patent CN113624567a, the potential safety hazard is leakage, so when the anhydrous hydrofluoric acid is sampled by the sampling system, the leakage detection and early warning of the sampling system are required. Accordingly, an optimized anhydrous hydrogen fluoride sampling system is desired.
Accordingly, in order to ensure the safety of sampling, it is considered that in the sampling process of actually performing anhydrous hydrogen fluoride, the sampling system needs to be subjected to leakage detection and timely leakage early warning. In performing leak detection, it is considered that it is difficult to directly detect whether anhydrous hydrogen fluoride leaks or not due to a long line, resulting in difficulty in detection. In particular, in the process of sampling anhydrous hydrogen fluoride without leakage, the pressure value inside the pipeline has a specific time sequence change characteristic, so in the technical scheme of the application, whether the anhydrous hydrogen fluoride leaks or not is accurately detected by utilizing the time sequence change condition of the pressure value inside the pipeline, and the safety of sampling the anhydrous hydrogen fluoride is ensured. In the process, the difficulty is how to accurately dig out the dynamic change characteristic information of the pressure value in the pipeline in the time dimension so as to accurately perform the leakage early warning of the anhydrous hydrogen fluoride.
In recent years, deep learning and neural networks have been widely used in the fields of computer vision, natural language processing, text signal processing, and the like.
The development of deep learning and neural networks provides new solutions and schemes for mining dynamic change characteristic information of pressure values inside the pipeline in the time dimension.
Specifically, in the technical solution of the present application, first, internal pressure values of a pipeline at a plurality of predetermined time points within a predetermined period of time are acquired by a pressure sensor disposed inside the pipeline. Next, considering that the internal pressure value of the pipeline has a dynamic change rule in the time dimension, in order to accurately extract the change characteristic information of the internal pressure value of the pipeline in the time dimension, in the technical scheme of the application, the internal pressure values of the pipelines at a plurality of preset time points are arranged into pipeline internal pressure input vectors according to the time dimension, so that the distribution information of the internal pressure values of the pipelines in time sequence is integrated.
Then, considering that the time-series change information of the internal pressure value of the pipeline is weak, the weak change feature is small-scale change feature information relative to the internal pressure value of the pipeline, if the time-series dynamic change feature extraction of the flow velocity value is performed by using absolute change information, not only the calculated amount is large, but also the small-scale weak change feature of the internal pressure value of the pipeline in the time dimension is difficult to be perceived, and the accuracy of the subsequent classification is further affected. Therefore, in the technical scheme of the application, the difference value between the internal pressure values of every two adjacent positions in the pipeline internal pressure input vector is calculated to obtain the pipeline internal pressure change input vector, so that the characteristic extraction of the time sequence dynamic change of the pipeline internal pressure value is performed by replacing absolute change information with the relative time sequence change information of the pipeline internal pressure.
Further, since the pressure value inside the pipeline has a dynamic change rule in the time dimension, that is, the time sequence relative change characteristics of the pressure value inside the pipeline at every two adjacent preset time points are dynamically associated in the time dimension, in order to fully extract the time sequence dynamic change characteristic information of the pressure value inside the pipeline, so as to perform leakage detection, in the technical scheme of the application, the time sequence dynamic associated characteristic distribution information of the pressure inside the pipeline in the time dimension is extracted by performing characteristic mining on the pressure change input vector inside the pipeline in a one-dimensional convolutional neural network model, so as to obtain the time sequence dynamic associated characteristic distribution information of the pressure inside the pipeline in the time dimension.
Then, in order to improve the accuracy of time-varying feature extraction of the internal pressure values of the pipeline, data enhancement is required for dynamic implicit correlation features of the internal pressure values of the pipeline in a high-dimensional feature space, taking into account that the internal pressure values of the pipeline have volatility and uncertainty in the time dimension. It should be understood that, as a learning target of the neural network model, the gaussian density map may represent a joint distribution in the case where a plurality of feature values constitute an overall distribution due to probability density thereof, that is, a feature distribution is taken as a priori distribution, and probability density under the influence of correlation with other a priori distribution positions at each a priori distribution position is obtained as a posterior distribution, thereby describing the feature distribution more accurately in a higher dimension. Therefore, in the technical solution of the present application, the data enhancement can be performed on the time-series dynamic correlation characteristics of the internal pressure of the pipeline through the prior distribution of the pressure values, namely, gaussian distribution. Specifically, a gaussian density map of the pressure time series feature vector in the pipeline is constructed, and in particular, herein, the mean value vector of the gaussian density map is the pressure time series feature vector in the pipeline, and the feature value of each position in the covariance matrix of the gaussian density map is the variance between the feature values of the corresponding two positions in the pressure time series feature vector in the pipeline.
And then, further carrying out Gaussian discretization on the Gaussian density map based on joint probability density distribution along the rows or the columns, thereby obtaining a classification characteristic matrix. Here, the gaussian discretization processing mode is utilized, so that information loss is not generated when the characteristics of the data are amplified, and the accuracy of subsequent classification is further improved. It should be understood that, because each row or each column of the gaussian density map corresponds to the joint distribution of the eigenvalues of a certain position of the original vector with respect to the eigenvalues of other positions, the discretization of the gaussian density map also obtains one vector of each row or each column based on the joint distribution of each row or each column, and then two-dimensionally arranges the one vector of each row or each column to obtain the discretized matrix.
Further, the classification feature matrix is subjected to classification processing through a classifier to obtain a classification result used for indicating whether leakage early warning prompt is generated or not. That is, classification is performed with time-series dynamic change characteristics concerning the pressure value inside the pipeline in the classification characteristic matrix, so that detection and capturing are performed on abnormal pressure change information inside the pipeline, thereby giving an early warning as to whether or not anhydrous hydrogen fluoride leakage is generated.
In particular, in the technical scheme of the application, when the characteristic expression strengthening based on the Gaussian density map is performed on the pressure time sequence characteristic vector in the pipeline to obtain the classification characteristic matrix, abnormal characteristic values deviating from the overall characteristic distribution are introduced into the classification characteristic matrix in consideration of randomness of Gaussian discretization, so that the training effect of the model is affected.
In the solution of the present application, it is therefore preferable to first control the sampling frequency of the gaussian discretization so that the classification feature matrix is a diagonal matrix, for example, denoted as
And then +/for the classification feature matrix>
Performing eigenvoization bitwise displacement associated matching optimization, expressed as: />
And->
Is the classification feature matrix->
Intrinsic decomposition of the obtained->
Intrinsic value->
For said->
The eigenvalue matrix obtained by arranging the eigenvalues along a diagonal is also a diagonal matrix,/-j->
For the eigenvoization matrix->
And the classification characteristic matrix->
Distance between them.
That is, by being based on the classification feature matrix
Is obtained by eigenvoization of the matrix 
To the classification characteristic matrix->
Performing bit-by-bit displacement correlation, and using the classification feature matrix +.>
The matching of the characteristic association relation is carried out relative to the projection distance in the eigenvoization space, so that the problem of mismatching of the optimization direction caused by weak association distribution of the characteristics in the back propagation of model parameters is solved, and the classification characteristic matrix is avoided>
The feature values at the edges of the class object domain are mismatching constrained in the opposite optimization direction, resulting in poor training results. Therefore, the early warning can be accurately carried out on whether the anhydrous hydrogen fluoride leaks in the sampling process, so that the sampling safety of the anhydrous hydrogen fluoride is ensured.
FIG. 1 is a diagram of an application scenario of an anhydrous hydrogen fluoride sampling system and method according to an embodiment of the present application. As shown in fig. 1, in this application scenario, the internal pressure values (e.g., P illustrated in fig. 1) of the pipeline (e.g., L illustrated in fig. 1) at a plurality of predetermined time points within a predetermined time period acquired by the pressure sensor (e.g., C illustrated in fig. 1). The collected internal pressure values of the pipeline at the plurality of predetermined time points are then input into a server (e.g., S illustrated in fig. 1) deployed with a sampling algorithm of anhydrous hydrogen fluoride, wherein the server is capable of processing the internal pressure values of the pipeline at the plurality of predetermined time points using the sampling algorithm of anhydrous hydrogen fluoride to generate a classification result indicative of whether a leak warning prompt is generated.
Having described the basic principles of the present application, various non-limiting embodiments of the present application are described in detail below with reference to the accompanying drawings.
Fig. 2 is a schematic block diagram of a sampling system for anhydrous hydrogen fluoride in an embodiment of the present application. As shown in fig. 2, the anhydrous hydrogen fluoride sampling system 100 comprises: a pressure data acquisition module 110 for acquiring internal pressure values of the pipeline at a plurality of predetermined time points within a predetermined time period acquired by the pressure sensor. It should be understood that, in order to ensure the safety of sampling, it is necessary to perform leak detection on the sampling system and timely perform leak early warning in the sampling process of anhydrous hydrogen fluoride. In performing leak detection, it is considered that it is difficult to directly detect whether anhydrous hydrogen fluoride leaks or not due to a long line, resulting in difficulty in detection. In particular, in the process of sampling anhydrous hydrogen fluoride without leakage, the pressure value inside the pipeline has a specific time sequence change characteristic, so in the technical scheme of the application, whether the anhydrous hydrogen fluoride leaks or not is accurately detected by utilizing the time sequence change condition of the pressure value inside the pipeline, and the safety of sampling the anhydrous hydrogen fluoride is ensured. In the process, the difficulty is how to accurately dig out the dynamic change characteristic information of the pressure value in the pipeline in the time dimension so as to accurately perform the leakage early warning of the anhydrous hydrogen fluoride.
The internal pressure change module 120 is configured to calculate a difference between internal pressure values of each two adjacent positions in the pipeline internal pressure input vector after arranging the internal pressure values of the pipeline at the plurality of predetermined time points into the pipeline internal pressure input vector according to a time dimension, so as to obtain a pipeline internal pressure change input vector. It should be understood that, considering that the internal pressure value of the pipeline has a dynamic change rule in the time dimension, in order to accurately extract the change feature information of the internal pressure value of the pipeline in the time dimension, in the technical solution of the present application, the internal pressure values of the pipelines at the plurality of predetermined time points are arranged as pipeline internal pressure input vectors according to the time dimension, so as to integrate the distribution information of the internal pressure values of the pipelines in time sequence. Then, considering that the time-series change information of the internal pressure value of the pipeline is weak, the weak change feature is small-scale change feature information relative to the internal pressure value of the pipeline, if the time-series dynamic change feature extraction of the flow velocity value is performed by using absolute change information, not only the calculated amount is large, but also the small-scale weak change feature of the internal pressure value of the pipeline in the time dimension is difficult to be perceived, and the accuracy of the subsequent classification is further affected. Therefore, in the technical scheme of the application, the difference value between the internal pressure values of every two adjacent positions in the pipeline internal pressure input vector is calculated to obtain the pipeline internal pressure change input vector, so that the characteristic extraction of the time sequence dynamic change of the pipeline internal pressure value is performed by replacing absolute change information with the relative time sequence change information of the pipeline internal pressure.
The pressure time sequence correlation module 130 is configured to pass the pipeline internal pressure variation input vector through a one-dimensional convolutional neural network model to obtain a pipeline internal pressure time sequence feature vector. It should be understood that, since the pressure value inside the pipeline has a dynamic change rule in the time dimension, that is, the time sequence relative change characteristics of the pressure value inside the pipeline at every two adjacent predetermined time points are dynamically associated in the time dimension, in order to fully extract the time sequence dynamic change characteristic information of the pressure value inside the pipeline, so as to perform leakage detection, in the technical scheme of the application, the time sequence dynamic associated characteristic distribution information of the pressure inside the pipeline in the time dimension is extracted by performing characteristic mining on the pressure change input vector inside the pipeline in a one-dimensional convolutional neural network model, so as to obtain the time sequence dynamic associated characteristic distribution information of the pressure inside the pipeline in the time dimension.
Optionally, in an embodiment of the present application, the pressure timing correlation module 130 is configured to: each layer of the one-dimensional convolutional neural network model is used for respectively carrying out input data in forward transmission of the layer; performing convolution processing based on a one-dimensional convolution kernel on the input data to obtain a convolution feature map; pooling processing is carried out on the convolution feature images based on feature matrixes to obtain pooled feature images; non-linear activation is carried out on the pooled feature map so as to obtain an activated feature map; the input of the first layer of the one-dimensional convolutional neural network model is the pressure change input vector in the pipeline, and the output of the last layer of the one-dimensional convolutional neural network model is the pressure time sequence characteristic vector in the pipeline.
And the Gaussian strengthening module 140 is used for strengthening the characteristic expression of the pressure time sequence characteristic vector in the pipeline based on the Gaussian density diagram to obtain a classification characteristic matrix. It will be appreciated that the gaussian density map is further subjected to gaussian discretization based on a joint probability density distribution along a row or column, resulting in a classification feature matrix. Here, the gaussian discretization processing mode is utilized, so that information loss is not generated when the characteristics of the data are amplified, and the accuracy of subsequent classification is further improved. It should be understood that, because each row or each column of the gaussian density map corresponds to the joint distribution of the eigenvalues of a certain position of the original vector with respect to the eigenvalues of other positions, the discretization of the gaussian density map also obtains one vector of each row or each column based on the joint distribution of each row or each column, and then two-dimensionally arranges the one vector of each row or each column to obtain the discretized matrix.
Optionally, in an embodiment of the present application, the gaussian enhancement module 140 includes: a gaussian density map construction unit for constructing a gaussian density map of the pipeline internal pressure timing feature vector as follows Gao Sigong; wherein, the Gaussian formula is: 
Wherein->
Is the pressure time sequence characteristic vector in the pipeline, and +.>
Is the variance between the eigenvalues of the corresponding two locations in the pipeline internal pressure timing eigenvector,/>
Variable representing the collaborative Gaussian density map, < >>
Representing a gaussian probability density function; and the Gaussian discrete unit is used for carrying out Gaussian discretization on the Gaussian distribution of each position in the Gaussian density map so as to obtain the classification characteristic matrix.
An optimizing module 150, configured to optimize the feature distribution of the classification feature matrix to obtain an optimized scoreAnd (5) a class feature matrix. It should be understood that, in the technical solution of the present application, when the characteristic expression strengthening based on the gaussian density map is performed on the pressure time sequence characteristic vector in the pipeline to obtain the classification characteristic matrix, an abnormal characteristic value deviating from the overall characteristic distribution is introduced into the classification characteristic matrix in consideration of randomness of gaussian discretization, so as to affect the training effect of the model. Therefore, in the technical solution of the present application, it is preferable that, first, when performing feature expression enhancement based on a gaussian density map on the pressure timing feature vector in the pipeline, the sampling frequency of the gaussian discretization is controlled so that the classification feature matrix is a diagonal matrix, and then the classification feature matrix is applied to the sampling frequency 
An eigenvoice bitwise displacement correlation matching optimization is performed, optionally, in an embodiment of the present application, the optimization module 150 is configured to: carrying out eigenvoization bitwise displacement association matching optimization on the classification characteristic matrix by using the following optimization formula to obtain an optimized classification characteristic matrix; wherein, the optimization formula is:
wherein->
Is a matrix of the classification characteristic,
and->
Is obtained by eigenvoicing the classification characteristic matrix>
Intrinsic value->
For said->
The eigenvalues are arranged diagonally to obtain an eigenvoization matrix, and +.>
And->
Are all diagonal matrix>
For the distance between the eigen-unitized matrix and the classification feature matrix,/i>
Representing matrix multiplication +.>
Representing matrix addition, ++>
Representing multiplication by location +.>
Classifying the feature matrix for the optimization.
That is, by being based on the classification feature matrix
Is obtained by eigenvoization of the matrix
To the classification characteristic matrix->
Performing bit-by-bit displacement correlation, and using the classification feature matrix +.>
Matching of characteristic association relation is carried out relative to projection distance in eigenvoization space, so that the problem that model parameters are caused by weak association distribution of characteristics in back propagation is solved Is avoided by avoiding the mismatch problem of the optimized direction of the classification feature matrix>
The feature values at the edges of the class object domain are mismatching constrained in the opposite optimization direction, resulting in poor training results. Therefore, the early warning can be accurately carried out on whether the anhydrous hydrogen fluoride leaks in the sampling process, so that the sampling safety of the anhydrous hydrogen fluoride is ensured. />
And the leakage early warning module 160 is configured to pass the optimized classification feature matrix through a classifier to obtain a classification result, where the classification result is used to indicate whether the leakage early warning prompt is generated. It is to be understood that classification is performed with time-series dynamic change characteristics concerning the pressure value inside the pipeline in the classification characteristic matrix, so that detection and capturing are performed on abnormal pressure change information inside the pipeline, thereby giving an early warning as to whether or not anhydrous hydrogen fluoride leakage is generated.
Optionally, in an embodiment of the present application, the leakage pre-warning module 160 is configured to: processing the optimized classification feature matrix by using the classifier according to the following classification formula to obtain the classification result; wherein, the classification formula is:
wherein->
To->
Is a weight matrix >
To->
For the bias vector +.>
Classifying the feature matrix for the optimization, +.>
Representing the projection of the optimized classification feature matrix as a vector,/->
Representing a normalized exponential function.
That is, in the technical solution of the present application, the labeling of the classifier includes generating a leakage early warning prompt (a first label) and not generating a leakage early warning prompt (a second label), specifically, the optimized classification feature matrix is projected as a vector and then fully connected and encoded to obtain a classification feature vector, and then the classification feature vector is determined to which classification label the optimized classification feature matrix belongs through a normalized exponential function. After the classification result is obtained, whether the leakage early warning prompt is generated or not can be controlled based on the classification result, and an alarm lamp can be arranged, and when the classification result is that the leakage early warning prompt is generated, the alarm lamp is controlled to alarm.
In summary, the anhydrous hydrogen fluoride sampling system and the anhydrous hydrogen fluoride sampling method provided by the application can accurately early warn whether leakage occurs in the sampling process of the anhydrous hydrogen fluoride, so that the sampling safety of the anhydrous hydrogen fluoride is ensured.
Fig. 3 is a schematic flow chart of a sampling method of anhydrous hydrogen fluoride according to an embodiment of the present application. As shown in fig. 3, the method for sampling anhydrous hydrogen fluoride includes: s110, acquiring internal pressure values of pipelines at a plurality of preset time points in a preset time period acquired by a pressure sensor; s120, after arranging the internal pressure values of the pipelines at a plurality of preset time points into pipeline internal pressure input vectors according to a time dimension, calculating the difference value between the internal pressure values of every two adjacent positions in the pipeline internal pressure input vectors to obtain pipeline internal pressure change input vectors; s130, the pipeline internal pressure change input vector is passed through a one-dimensional convolutional neural network model to obtain a pipeline internal pressure time sequence feature vector; s140, performing feature expression reinforcement based on a Gaussian density diagram on the pressure time sequence feature vector in the pipeline to obtain a classification feature matrix; s150, performing feature distribution optimization on the classification feature matrix to obtain an optimized classification feature matrix; and S160, enabling the optimized classification feature matrix to pass through a classifier to obtain a classification result, wherein the classification result is used for indicating whether leakage early warning prompt is generated or not.
Fig. 4 is a schematic diagram of a model architecture of a sampling method of anhydrous hydrogen fluoride according to an embodiment of the present application. As shown in fig. 4, the input of the model architecture of the anhydrous hydrogen fluoride sampling method is the internal pressure values of the pipeline at a plurality of predetermined time points. Firstly, arranging the internal pressure values of the pipelines at a plurality of preset time points into pipeline internal pressure input vectors according to a time dimension, and then calculating the difference value between the internal pressure values of every two adjacent positions in the pipeline internal pressure input vectors to obtain pipeline internal pressure change input vectors. And then, the pipeline internal pressure change input vector is passed through a one-dimensional convolutional neural network model to obtain a pipeline internal pressure time sequence characteristic vector. And performing feature expression reinforcement based on a Gaussian density diagram on the pressure time sequence feature vector in the pipeline to obtain a classification feature matrix. And then, carrying out feature distribution optimization on the classification feature matrix to obtain an optimized classification feature matrix. And finally, the optimized classification feature matrix passes through a classifier to obtain a classification result, wherein the classification result is used for indicating whether leakage early warning prompt is generated or not.
Optionally, in an embodiment of the present application, the step of passing the pipeline internal pressure variation input vector through a one-dimensional convolutional neural network model to obtain a pipeline internal pressure time sequence feature vector includes: each layer of the one-dimensional convolutional neural network model is used for respectively carrying out input data in forward transmission of the layer; performing convolution processing based on a one-dimensional convolution kernel on the input data to obtain a convolution feature map; pooling processing is carried out on the convolution feature images based on feature matrixes to obtain pooled feature images; non-linear activation is carried out on the pooled feature map so as to obtain an activated feature map; the input of the first layer of the one-dimensional convolutional neural network model is the pressure change input vector in the pipeline, and the output of the last layer of the one-dimensional convolutional neural network model is the pressure time sequence characteristic vector in the pipeline.
Optionally, in an embodiment of the present application, performing feature expression enhancement on the pressure time sequence feature vector in the pipeline based on the gaussian density map to obtain a classification feature matrix, including: constructing a gaussian density map of the pipeline internal pressure timing feature vector with a gaussian formula; wherein, the Gaussian formula is:
wherein->
Is the pressure time sequence characteristic vector in the pipeline, and +.>
Is the variance between the eigenvalues of the corresponding two locations in the pipeline internal pressure timing eigenvector,/>
Variable representing the collaborative Gaussian density map, < >>
Representing a gaussian probability density function; and the Gaussian discrete unit is used for carrying out Gaussian discretization on the Gaussian distribution of each position in the Gaussian density map so as to obtain the classification characteristic matrix.
Optionally, in an embodiment of the present application, performing feature distribution optimization on the classification feature matrix to obtain an optimized classification feature matrix includes: carrying out eigenvoization bitwise displacement association matching optimization on the classification characteristic matrix by using the following optimization formula to obtain an optimized classification characteristic matrix; wherein, the optimization formula is:
wherein->
Is a matrix of the classification characteristic, 
And->
Is obtained by eigenvoicing the classification characteristic matrix>
Intrinsic value->
For said->
The eigenvalues are arranged diagonally to obtain an eigenvoization matrix, and +.>
And->
Are all diagonal matrix>
For the distance between the eigen-unitized matrix and the classification feature matrix,/i>
Representing matrix multiplication +.>
Representing matrix addition, ++>
Representing multiplication by location +.>
Classifying the feature matrix for the optimization.
Optionally, in an embodiment of the present application, the step of passing the optimized classification feature matrix through a classifier to obtain a classification result includes: processing the optimized classification feature matrix using the classifier in the following classification formula to obtain the classifierThe classification result; wherein, the classification formula is:
wherein->
To->
Is a weight matrix>
To->
For the bias vector +.>
Classifying the feature matrix for the optimization, +.>
Representing the projection of the optimized classification feature matrix as a vector,/->
Representing a normalized exponential function.
Here, it will be understood by those skilled in the art that the specific operation of each step in the above-described anhydrous hydrogen fluoride sampling method has been described in detail in the above description of the anhydrous hydrogen fluoride sampling system with reference to fig. 2, and thus, a repetitive description thereof will be omitted.
It should be understood that the specific examples herein are intended only to facilitate a better understanding of the embodiments of the present application by those skilled in the art and are not intended to limit the scope of the embodiments of the present application.
It should also be understood that, in various embodiments of the present application, the size of the sequence number of each process does not mean that the execution sequence of each process should be determined by its functions and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.
It should also be understood that the various embodiments described in this specification may be implemented alone or in combination, and that the examples herein are not limited in this regard.
Unless defined otherwise, all technical and scientific terms used in the examples of this application have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in the present application is for the purpose of describing particular embodiments only and is not intended to limit the scope of the present application. The term "and/or" as used herein includes any and all combinations of one or more of the associated listed items. As used in the examples and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.
In the several embodiments provided in this application, it should be understood that the disclosed systems, devices, and methods may be implemented in other manners. For example, the apparatus embodiments described above are merely illustrative, e.g., the division of the units is merely a logical function division, and there may be additional divisions when actually implemented, e.g., multiple units or components may be combined or integrated into another system, or some features may be omitted or not performed.
The foregoing is merely specific embodiments of the present application, but the scope of the present application is not limited thereto, and any person skilled in the art can easily think about changes or substitutions within the technical scope of the present application, and the changes and substitutions are intended to be covered by the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.
Claims (10)
1. A sampling system for anhydrous hydrogen fluoride, comprising: a pressure data acquisition module for acquiring internal pressure values of the pipeline at a plurality of predetermined time points within a predetermined time period acquired by the pressure sensor; the internal pressure change module is used for arranging the internal pressure values of the pipelines at a plurality of preset time points into pipeline internal pressure input vectors according to the time dimension, and then calculating the difference value between the internal pressure values of every two adjacent positions in the pipeline internal pressure input vectors to obtain pipeline internal pressure change input vectors; the pressure time sequence association module is used for enabling the pressure change input vector in the pipeline to pass through a one-dimensional convolutional neural network model to obtain a pressure time sequence feature vector in the pipeline; the Gaussian strengthening module is used for strengthening the characteristic expression of the pressure time sequence characteristic vector in the pipeline based on a Gaussian density diagram to obtain a classification characteristic matrix; the optimizing module is used for optimizing the characteristic distribution of the classification characteristic matrix to obtain an optimized classification characteristic matrix; and the leakage early warning module is used for enabling the optimized classification feature matrix to pass through a classifier to obtain a classification result, wherein the classification result is used for indicating whether leakage early warning prompt is generated or not.
2. The anhydrous hydrogen fluoride sampling system of claim 1, wherein the pressure timing correlation module is configured to: each layer of the one-dimensional convolutional neural network model is used for respectively carrying out input data in forward transmission of the layer; performing convolution processing based on a one-dimensional convolution kernel on the input data to obtain a convolution feature map; pooling processing is carried out on the convolution feature images based on feature matrixes to obtain pooled feature images; non-linear activation is carried out on the pooled feature map so as to obtain an activated feature map; the input of the first layer of the one-dimensional convolutional neural network model is the pressure change input vector in the pipeline, and the output of the last layer of the one-dimensional convolutional neural network model is the pressure time sequence characteristic vector in the pipeline.
3. The anhydrous hydrogen fluoride sampling system of claim 2, wherein the gaussian enhancement module comprises: a gaussian density map construction unit for constructing a gaussian density map of the pipeline internal pressure timing feature vector as follows Gao Sigong; wherein, the Gaussian formula is:
wherein->
Is the pressure time sequence characteristic vector in the pipeline, and +. >
Is the variance between the eigenvalues of the corresponding two locations in the pipeline internal pressure timing eigenvector,/>
Variable representing the collaborative Gaussian density map, < >>
Representing a gaussian probability density function; and the Gaussian discrete unit is used for carrying out Gaussian discretization on the Gaussian distribution of each position in the Gaussian density map so as to obtain the classification characteristic matrix.
4. The anhydrous hydrogen fluoride sampling system of claim 3, wherein the optimization module is configured to: carrying out eigenvoization bitwise displacement association matching optimization on the classification characteristic matrix by using the following optimization formula to obtain an optimized classification characteristic matrix; wherein, the optimization formula is:
wherein->
Is the classification feature matrix,/a>
And->
Is obtained by eigenvoicing the classification characteristic matrix>
Intrinsic value->
For said->
The eigenvalues are arranged diagonally to obtain an eigenvoization matrix, and +.>
And->
Are all diagonal matrix>
For the distance between the eigen-unitized matrix and the classification feature matrix,/i>
Representing matrix multiplication +.>
Representing matrix addition, ++>
Representing multiplication by location +.>
Classifying the feature matrix for the optimization.
5. The anhydrous hydrogen fluoride sampling system of claim 4, wherein the leakage pre-warning module is configured to: processing the optimized classification feature matrix by using the classifier according to the following classification formula to obtain the classification result; wherein, the classification formula is: 
Wherein->
To->
Is a weight matrix>
To->
For the bias vector +.>
Classifying the feature matrix for the optimization, +.>
Representing the projection of the optimized classification feature matrix as a vector,/->
Representing a normalized exponential function.
6. A method for sampling anhydrous hydrogen fluoride, comprising: acquiring internal pressure values of the pipeline at a plurality of predetermined time points within a predetermined time period acquired by the pressure sensor; after arranging the internal pressure values of the pipelines at a plurality of preset time points into pipeline internal pressure input vectors according to the time dimension, calculating the difference value between the internal pressure values of every two adjacent positions in the pipeline internal pressure input vectors to obtain pipeline internal pressure change input vectors; the pressure change input vector in the pipeline is subjected to a one-dimensional convolutional neural network model to obtain a pressure time sequence characteristic vector in the pipeline; performing feature expression reinforcement based on a Gaussian density diagram on the pressure time sequence feature vector in the pipeline to obtain a classification feature matrix; performing feature distribution optimization on the classification feature matrix to obtain an optimized classification feature matrix; and the optimized classification feature matrix passes through a classifier to obtain a classification result, wherein the classification result is used for indicating whether leakage early warning prompt is generated or not.
7. The anhydrous hydrogen fluoride sampling method of claim 6, wherein passing the pipeline internal pressure variation input vector through a one-dimensional convolutional neural network model to obtain a pipeline internal pressure timing feature vector comprises: each layer of the one-dimensional convolutional neural network model is used for respectively carrying out input data in forward transmission of the layer; performing convolution processing based on a one-dimensional convolution kernel on the input data to obtain a convolution feature map; pooling processing is carried out on the convolution feature images based on feature matrixes to obtain pooled feature images; non-linear activation is carried out on the pooled feature map so as to obtain an activated feature map; the input of the first layer of the one-dimensional convolutional neural network model is the pressure change input vector in the pipeline, and the output of the last layer of the one-dimensional convolutional neural network model is the pressure time sequence characteristic vector in the pipeline.
8. The method for sampling anhydrous hydrogen fluoride according to claim 7, wherein the characteristic expression enhancement based on a gaussian density map is performed on the pipeline internal pressure timing characteristic vector to obtain a classification characteristic matrix, comprising: constructing a gaussian density map of the pipeline internal pressure timing feature vector with a gaussian formula; wherein, the Gaussian formula is: 
Wherein->
Is the pressure time sequence characteristic vector in the pipeline, and +.>
Is the variance between the eigenvalues of the corresponding two locations in the pipeline internal pressure timing eigenvector,/>
Variable representing the collaborative Gaussian density map, < >>
Representing a gaussian probability density function; and performing Gaussian discretization processing on the Gaussian distribution of each position in the Gaussian density map to obtain the classification feature matrix.
9. The method of claim 8, wherein optimizing the classification feature matrix for feature distribution to obtain an optimized classification feature matrix comprises: carrying out eigenvoization bitwise displacement association matching optimization on the classification characteristic matrix by using the following optimization formula to obtain an optimized classification characteristic matrix; wherein, the optimization formula is:
wherein->
Is a matrix of the classification characteristic,
and->
Is obtained by eigenvoicing the classification characteristic matrix>
Intrinsic value->
For said->
The eigenvalues are arranged diagonally to obtain eigenvaluesMatrix formation, and->
And->
Are all diagonal matrix>
For the distance between the eigen-unitized matrix and the classification feature matrix,/i>
Representing matrix multiplication +. >
Representing matrix addition, ++>
Representing multiplication by location +.>
Classifying the feature matrix for the optimization.
10. The method of claim 9, wherein passing the optimized classification feature matrix through a classifier to obtain classification results comprises: processing the optimized classification feature matrix by using the classifier according to the following classification formula to obtain the classification result;
wherein, the classification formula is:
wherein->
To->
Is a weight matrix>
To->
For the bias vector +.>
Classifying the feature matrix for the optimization, +.>
Representing the projection of the optimized classification feature matrix as a vector,/->
Representing a normalized exponential function. />
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310280557.4A CN115993268B (en) | 2023-03-22 | 2023-03-22 | Sampling system and method for anhydrous hydrogen fluoride |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310280557.4A CN115993268B (en) | 2023-03-22 | 2023-03-22 | Sampling system and method for anhydrous hydrogen fluoride |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115993268A true CN115993268A (en) | 2023-04-21 |
| CN115993268B CN115993268B (en) | 2023-06-02 |
Family
ID=85992330
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310280557.4A Active CN115993268B (en) | 2023-03-22 | 2023-03-22 | Sampling system and method for anhydrous hydrogen fluoride |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115993268B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116321620A (en) * | 2023-05-11 | 2023-06-23 | 杭州行至云起科技有限公司 | Intelligent lighting switch control system and method thereof |
| CN116597635A (en) * | 2023-05-18 | 2023-08-15 | 杭州新立仪表有限公司 | Wireless communication intelligent gas meter controller and control method thereof |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180291728A1 (en) * | 2015-10-02 | 2018-10-11 | Schlumberger Technology Corporation | Detection of influx and loss of circulation |
| CN109556797A (en) * | 2018-11-19 | 2019-04-02 | 浙江工业大学 | The pipeline leakage detection and location method with convolutional neural networks is decomposed based on spline local mean value |
| CN113723210A (en) * | 2021-08-05 | 2021-11-30 | 中冶南方城市建设工程技术有限公司 | Method and system for detecting leakage edge equipment of water supply pipe in intelligent park |
| CN115063584A (en) * | 2022-06-28 | 2022-09-16 | 上海威士顿信息技术股份有限公司 | Pipeline steam leakage detection method and detection system based on deep learning |
| CN115100456A (en) * | 2022-05-27 | 2022-09-23 | 福建天甫电子材料有限公司 | Intelligent toxic and harmful gas alarm system for preparing electronic-grade hydrogen peroxide |
| CN115719460A (en) * | 2022-07-20 | 2023-02-28 | 贵州大学 | Gas leakage detection method and system based on deep learning and terminal equipment |
-
2023
- 2023-03-22 CN CN202310280557.4A patent/CN115993268B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180291728A1 (en) * | 2015-10-02 | 2018-10-11 | Schlumberger Technology Corporation | Detection of influx and loss of circulation |
| CN109556797A (en) * | 2018-11-19 | 2019-04-02 | 浙江工业大学 | The pipeline leakage detection and location method with convolutional neural networks is decomposed based on spline local mean value |
| CN113723210A (en) * | 2021-08-05 | 2021-11-30 | 中冶南方城市建设工程技术有限公司 | Method and system for detecting leakage edge equipment of water supply pipe in intelligent park |
| CN115100456A (en) * | 2022-05-27 | 2022-09-23 | 福建天甫电子材料有限公司 | Intelligent toxic and harmful gas alarm system for preparing electronic-grade hydrogen peroxide |
| CN115063584A (en) * | 2022-06-28 | 2022-09-16 | 上海威士顿信息技术股份有限公司 | Pipeline steam leakage detection method and detection system based on deep learning |
| CN115719460A (en) * | 2022-07-20 | 2023-02-28 | 贵州大学 | Gas leakage detection method and system based on deep learning and terminal equipment |
Non-Patent Citations (2)
| Title |
|---|
| CONGCONG SUN 等: "Leak Localization in Water Distribution Networks Using Pressure and Data-Driven Classifier Approach", 《WATER》 * |
| 付磊: "一维卷积神经网络在天然气管道泄漏孔径识别中的应用研究", 《中国优秀硕士学位论文全文数据库 工程科技Ⅰ辑》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116321620A (en) * | 2023-05-11 | 2023-06-23 | 杭州行至云起科技有限公司 | Intelligent lighting switch control system and method thereof |
| CN116321620B (en) * | 2023-05-11 | 2023-08-11 | 杭州行至云起科技有限公司 | Intelligent lighting switch control system and method thereof |
| CN116597635A (en) * | 2023-05-18 | 2023-08-15 | 杭州新立仪表有限公司 | Wireless communication intelligent gas meter controller and control method thereof |
| CN116597635B (en) * | 2023-05-18 | 2024-02-23 | 杭州新立仪表有限公司 | Wireless communication intelligent gas meter controller and control method thereof |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115993268B (en) | 2023-06-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115993268B (en) | Sampling system and method for anhydrous hydrogen fluoride | |
| US11663489B2 (en) | Machine learning systems and methods for improved localization of image forgery | |
| CN108470332B (en) | Multi-target tracking method and device | |
| CN107563433B (en) | Infrared small target detection method based on convolutional neural network | |
| CN102722892B (en) | SAR (synthetic aperture radar) image change detection method based on low-rank matrix factorization | |
| CN105825200A (en) | High-spectrum abnormal object detection method based on background dictionary learning and structure sparse expression | |
| CN110348437B (en) | Target detection method based on weak supervised learning and occlusion perception | |
| CN112132012B (en) | High-resolution SAR ship image generation method based on generation countermeasure network | |
| CN108171119B (en) | SAR image change detection method based on residual error network | |
| CN114240892A (en) | Unsupervised industrial image anomaly detection method and system based on knowledge distillation | |
| CN111898639A (en) | Dimension reduction-based hierarchical time memory industrial anomaly detection method and device | |
| CN112560948B (en) | Fundus image classification method and imaging method under data deviation | |
| CN109298633A (en) | Chemical production process fault monitoring method based on adaptive piecemeal Non-negative Matrix Factorization | |
| CN112257741B (en) | Method for detecting generative anti-false picture based on complex neural network | |
| CN113139594A (en) | Airborne image unmanned aerial vehicle target self-adaptive detection method | |
| CN116311005A (en) | Apparatus, method and storage medium for moving image processing | |
| CN117115743A (en) | Mining safety production monitoring system and method thereof | |
| CN116310850A (en) | Remote sensing image target detection method based on improved RetinaNet | |
| Rahadian et al. | Image encoding selection based on Pearson correlation coefficient for time series anomaly detection | |
| Yu et al. | Intelligent corner synthesis via cycle-consistent generative adversarial networks for efficient validation of autonomous driving systems | |
| CN117134958A (en) | Information processing method and system for network technology service | |
| CN117152486A (en) | Image countermeasure sample detection method based on interpretability | |
| CN116704208A (en) | Local interpretable method based on characteristic relation | |
| Jafrasteh et al. | Generative adversarial networks as a novel approach for tectonic fault and fracture extraction in high resolution satellite and airborne optical images | |
| CN116258877A (en) | Land utilization scene similarity change detection method, device, medium and equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |