Intelligent control system and method for inspection and maintenance of natural gas purification equipment

Abstract

The present application discloses an intelligent management and control system and method for the inspection and maintenance of a natural gas purification device. It uses sensors and monitoring equipment to collect the equipment operation data of the natural gas purification device in real time, and then transmits these equipment operation data to the cloud computing platform for analysis, so as to use data processing and analysis algorithms based on artificial intelligence and deep learning technology to learn and capture the time series collaborative association and interactive feature information between the operation data of each device of the natural gas purification device, so as to identify the equipment operation status and determine whether to perform equipment pre-maintenance. In this way, it is possible to timely discover potential signs of equipment abnormality or performance degradation, and to warn of possible failure risks in advance, so as to perform equipment pre-maintenance, so as to realize intelligent management and control of the inspection and maintenance of the natural gas purification device.

Description

Technical Field

[0001] The present application relates to the field of intelligent inspection and maintenance, and specifically to an intelligent management and control system and method for inspection and maintenance operations of a natural gas purification device. Background Art

[0002] Natural gas purification equipment plays a vital role in industrial production, and is used to remove impurities, hydrogen sulfide and other harmful substances in natural gas to ensure the quality and safety of natural gas. However, as the equipment runs longer, it may wear, age or fail, resulting in performance degradation or even cessation of operation.

[0003] However, traditional inspection and maintenance methods for natural gas purification equipment are usually based on fixed time or cycle, and regular maintenance usually does not take into account the real-time operating status of the equipment, which may lead to failure to timely discover and deal with problems before the equipment performance deteriorates or fails. Moreover, since maintenance is based on a fixed schedule, even if the equipment is in good condition, unnecessary inspections and replacement of parts will be carried out, resulting in a waste of manpower, material and financial resources. In addition, if the natural gas purification equipment is not maintained and inspected in time, when the equipment fails, it may take a long time to diagnose the problem and arrange maintenance. During this period, the equipment cannot work normally, affecting production efficiency and safety, and may also cause environmental pollution problems.

[0004] Therefore, an intelligent management and control system for inspection and maintenance operations of natural gas purification equipment is desired. Summary of the invention

[0005] This application is made in consideration of the above problems. One purpose of this application is to provide an intelligent management and control system and method for inspection and maintenance operations of a natural gas purification device.

[0006] An embodiment of the present application provides an intelligent management and control method for inspection and maintenance operations of a natural gas purification device, which includes: collecting a time series of equipment operation data of the natural gas purification device through a sensor network, wherein the equipment operation data includes temperature values, pressure values, vibration amplitude values, current values ​​and voltage values; embedding and encoding each equipment operation data in the time series of the equipment operation data to obtain a time series of equipment operation data embedded coding vectors; passing the time series of the equipment operation data embedded coding vectors through a context-associated encoder based on data semantic association and data time span to obtain a sequence of context-associated equipment operation semantic feature vectors; performing node message propagation fusion on the sequence of context-associated equipment operation semantic feature vectors based on the importance of fused nodes to obtain equipment operation status feature time series propagation aggregation representation information; based on the equipment operation status feature time series propagation aggregation representation information, determining whether to perform equipment pre-maintenance.

[0007] For example, according to the intelligent management and control method of inspection and maintenance operations of a natural gas purification device according to an embodiment of the present application, each device operation data in the time series of the device operation data is embedded and encoded respectively to obtain a time series of the device operation data embedded coding vector, including: each device operation data in the time series of the device operation data is respectively passed through a device data embedding coding module based on a fully connected layer to obtain a time series of the device operation data embedded coding vector.

[0008] For example, according to the intelligent management and control method for inspection and maintenance operations of a natural gas purification device according to an embodiment of the present application, the time series of the equipment operation data embedded coding vector is passed through a context-associated encoder based on data semantic association and data time span to obtain a sequence of context-associated equipment operation semantic feature vectors, including: extracting a predetermined equipment operation data embedded coding vector from the time series of the equipment operation data embedded coding vector; calculating the semantic association between the predetermined equipment operation data embedded coding vector and all other equipment operation data embedded coding vectors to obtain a sequence of equipment operation data embedded coding semantic temporal associations; calculating the global mean of the sequence of equipment operation data embedded coding semantic temporal associations to obtain the global average association of equipment operation data embedded coding semantic temporal; using the global average association of equipment operation data embedded coding semantic temporal as a weight, weightedly strengthening the predetermined equipment operation data embedded coding vector to obtain the context-associated equipment operation semantic feature vector corresponding to the predetermined equipment operation data embedded coding vector.

[0009] For example, according to the intelligent management and control method for inspection and maintenance operations of a natural gas purification device according to an embodiment of the present application, the semantic association between the predetermined equipment operation data embedded coding vector and all other equipment operation data embedded coding vectors is calculated to obtain a sequence of equipment operation data embedded coding semantic temporal associations, including: calculating the semantic similarity between the predetermined equipment operation data embedded coding vector and all other equipment operation data embedded coding vectors to obtain a sequence of equipment operation data embedded coding semantic temporal similarities; calculating the number of feature vectors between the predetermined equipment operation data embedded coding vector and all other equipment operation data embedded coding vectors to obtain a sequence of equipment operation data embedded coding semantic temporal distance counter numbers; dividing the sequence of equipment operation data embedded coding semantic temporal similarities by the sequence of equipment operation data embedded coding semantic temporal distance counter numbers by position to obtain a sequence of equipment operation data embedded coding semantic temporal associations.

[0010] For example, according to the intelligent management and control method of inspection and maintenance operations of a natural gas purification device according to an embodiment of the present application, the sequence of the context-associated equipment operation semantic feature vectors is subjected to node message propagation fusion based on the importance of the fusion node to obtain the equipment operation status feature time series propagation aggregation representation information, including: inputting the sequence of the context-associated equipment operation semantic feature vectors into the node message propagation fusion network of the fusion node importance to obtain the equipment operation status feature time series propagation aggregation representation vector as the equipment operation status feature time series propagation aggregation representation information.

[0011] For example, according to the intelligent management and control method for inspection and maintenance operations of a natural gas purification device according to an embodiment of the present application, the sequence of the context-associated device operation semantic feature vectors is input into a node message propagation fusion network of the fusion node importance to obtain the device operation status feature time series propagation aggregation representation vector as the device operation status feature time series propagation aggregation representation information, including: using a scoring function to measure the significance factors of the first k-1 context-associated device operation semantic feature vectors in the sequence of the context-associated device operation semantic feature vectors to obtain k-1 context-associated device operation semantic time series significance factors; using the k-1 context-associated device operation semantic time series association significance factors as weights, calculating the position-weighted sum of the first k-1 context-associated device operation semantic feature vectors in the sequence of the context-associated device operation semantic feature vectors to aggregate the context-associated device operation semantic feature vectors. The first k-1 context-associated device operation semantic feature vectors in the sequence of the context-associated device operation semantic feature vectors are used to obtain the context-associated device operation semantic time-series associated aggregate representation vector of the rear node set; the kth context-associated device operation semantic feature vector in the sequence of the context-associated device operation semantic feature vectors is linearly modulated using a preset hyperparameter to obtain the modulated kth context-associated device operation semantic feature vector; the position-by-position sum of the modulated kth context-associated device operation semantic feature vector and the rear node set context-associated device operation semantic time-series associated aggregate representation vector is calculated to obtain the context-associated device operation semantic time-series associated full-node aggregate representation vector; the context-associated device operation semantic time-series associated full-node aggregate representation vector is input into a multi-layer perceptron model to obtain the device operation status feature time-series propagation aggregate representation vector.

[0012] For example, according to the intelligent management and control method for inspection and maintenance operations of a natural gas purification device according to an embodiment of the present application, a scoring function is used to measure the significance factors of the first k-1 context-associated device operation semantic feature vectors in the sequence of the context-associated device operation semantic feature vectors to obtain k-1 context-associated device operation semantic timing significance factors, including: respectively calculating the matrix product between each weight coefficient matrix and the first k-1 context-associated device operation semantic feature vectors to obtain the first k-1 context-associated device operation semantic timing associated weighted feature vectors; adding the corresponding bias vector to each context-associated device operation semantic timing associated weighted feature vector in the first k-1 context-associated device operation semantic timing associated weighted feature vectors to obtain the first k-1 context-associated device operation semantic timing associated biased feature vectors; using The function activates the first k-1 context-associated device operation semantic timing association bias feature vectors to obtain the first k-1 context-associated device operation semantic timing association activation feature vectors; calculates the product between each context-associated device operation semantic timing association activation feature vector in the first k-1 context-associated device operation semantic timing association activation feature vector and the weight coefficient vector to obtain the first k-1 context-associated device operation semantic timing association degree; inputs the first k-1 context-associated device operation semantic timing association degree into The function is used to obtain the k-1 context-related device operation semantic timing significance factors.

[0013] For example, according to the intelligent management and control method of the inspection and maintenance operation of the natural gas purification device according to the embodiment of the present application, whether to perform equipment pre-maintenance is determined based on the time-series propagation aggregation representation information of the equipment operation status characteristics, including: passing the equipment operation status characteristic time-series propagation aggregation representation vector through a classifier-based management and control result generator to obtain a management and control result, and the management and control result is used to indicate whether to perform equipment pre-maintenance.

[0014] For example, according to the intelligent management and control method for inspection and maintenance operations of a natural gas purification device according to an embodiment of the present application, it also includes a training step: for training the equipment data embedding coding module based on the fully connected layer, the context association encoder based on data semantic association and data time span, the node message propagation fusion network of the fusion node importance, and the classifier-based management and control result generator; wherein the training step includes: obtaining training data, the training data includes a time series of training equipment operation data of the natural gas purification device collected through a sensor network and actual management and control results, wherein the training equipment operation data includes training temperature values, training pressure values, training vibration amplitude values, training current values, and training voltage values; each training equipment operation data in the time series of the training equipment operation data is respectively passed through the equipment data embedding coding module based on the fully connected layer to obtain a time series of training equipment operation data embedding coding vectors; the training equipment The time series of the equipment operation data embedded coding vector is passed through the context-associated encoder based on data semantic association and data time span to obtain a sequence of training context-associated equipment operation semantic feature vectors; the sequence of training context-associated equipment operation semantic feature vectors is input into the node message propagation fusion network of the fusion node importance to obtain the training equipment operation status feature time series propagation aggregation representation vector; the training equipment operation status feature time series propagation aggregation representation vector is passed through the classifier-based control result generator to obtain the training control result; the cross entropy loss function value between the training control result and the real control result is calculated to obtain the classification loss function value; based on the classification loss function value, the equipment data embedded coding module based on the fully connected layer, the context-associated encoder based on data semantic association and data time span, the node message propagation fusion network of the fusion node importance and the classifier-based control result generator are trained.

[0015] An embodiment of the present application also provides an intelligent management and control system for inspection and maintenance operations of a natural gas purification device, which includes: an operation data acquisition module, which is used to collect a time series of equipment operation data of the natural gas purification device through a sensor network, wherein the equipment operation data includes temperature values, pressure values, vibration amplitude values, current values ​​and voltage values; an embedded coding module, which is used to embed and encode each equipment operation data in the time series of the equipment operation data to obtain a time series of equipment operation data embedded coding vectors; a context-associated coding module, which is used to embed the time series of the equipment operation data into coding vectors through a context-associated encoder based on data semantic association and data time span to obtain a sequence of context-associated equipment operation semantic feature vectors; a fusion module, which is used to perform node message propagation fusion on the sequence of context-associated equipment operation semantic feature vectors based on the importance of fused nodes to obtain equipment operation status feature time series propagation aggregation representation information; a maintenance analysis module, which is used to determine whether to perform equipment pre-maintenance based on the equipment operation status feature time series propagation aggregation representation information.

[0016] According to the intelligent management and control system and method for the inspection and maintenance of natural gas purification equipment in the embodiment of the present application, it uses sensors and monitoring equipment to collect the equipment operation data of the natural gas purification equipment in real time, and then transmits these equipment operation data to the cloud computing platform for analysis, so as to use the data processing and analysis algorithm based on artificial intelligence and deep learning technology to learn and capture the time series collaborative association and interactive feature information between the operation data of each device of the natural gas purification device, so as to identify the equipment operation status, and thus determine whether to perform equipment pre-maintenance. In this way, it is possible to timely discover potential signs of equipment abnormality or performance degradation, and to warn of possible failure risks in advance, so as to perform equipment pre-maintenance, so as to realize intelligent management and control of the inspection and maintenance of natural gas purification equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings of the embodiments of the present application are briefly introduced below. Obviously, the drawings described below only relate to some embodiments of the present application, and are not intended to limit the present application.

[0018] Figure 1 A schematic diagram of the application architecture of the intelligent management and control method for inspection and maintenance operations of a natural gas purification device in an embodiment of the present application is shown.

[0019] Figure 2 A flow chart of an intelligent management and control method for inspection and maintenance operations of a natural gas purification device in an embodiment of the present application is shown.

[0020] Figure 3 A flowchart of sub-step S530 of the intelligent management and control method for inspection and maintenance operations of a natural gas purification device in an embodiment of the present application is shown.

[0021] Figure 4 A structural schematic diagram of an intelligent management and control system for inspection and maintenance operations of a natural gas purification device in an embodiment of the present application is shown.

[0022] Figure 5 An application scenario diagram of the intelligent management and control method for inspection and maintenance operations of a natural gas purification device in an embodiment of the present application is shown.

[0023] Figure 6 A schematic diagram of a storage medium according to an embodiment of the present application is shown. DETAILED DESCRIPTION

[0024] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work also fall within the scope of protection of the present application.

[0025] The terms used in this specification are those common terms currently widely used in the art in consideration of the functions of the present application, but these terms may vary according to the intention of a person of ordinary skill in the art, precedents, or new technologies in the art. In addition, specific terms may be selected by the applicant, and in this case, their detailed meanings will be described in the detailed description of the present application. Therefore, the terms used in the specification should not be understood as simple names, but rather as a general description based on the meaning of the terms and the present application.

[0026] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules can be used and run on the user terminal and / or server. The modules are only illustrative, and different aspects of the system and method can use different modules.

[0027] Flowcharts are used in the present application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps may be processed in reverse order or simultaneously as required. At the same time, other operations may also be added to these processes, or a certain step or several steps of operations may be removed from these processes.

[0028] Figure 1 A schematic diagram of the application architecture of the intelligent management and control method for inspection and maintenance operations of a natural gas purification device in an embodiment of the present application is shown, including a server 100 and a terminal device 200.

[0029] The terminal device 200 and the server 100 can be connected via the Internet to achieve mutual communication. Optionally, the above-mentioned Internet uses standard communication technology and / or protocol. The Internet is usually the Internet, but it can also be any network, including but not limited to any combination of a local area network (LAN), a metropolitan area network (MAN), a wide area network (WAN), a mobile, wired or wireless network, a private network or a virtual private network. In some embodiments, the data exchanged through the network can be represented by technologies and / or formats including Hyper Text Markup Language (HTML), Extensible Markup Language (XML), etc. In addition, conventional encryption technologies such as Secure Socket Layer (SSL), Transport Layer Security (TLS), Virtual Private Network (VPN), Internet Protocol Security (IPsec), etc. can be used to encrypt all or some links. In other embodiments, customized and / or dedicated data communication technologies can also be used to replace or supplement the above-mentioned data communication technologies.

[0030] The server 100 can provide various network services for the terminal device 200, wherein the server 100 can be a single server, a server cluster consisting of several servers, or a cloud computing center. Specifically, the server 100 may include a processor 110 (Center Processing Unit, CPU), a memory 120, an input device 130, and an output device 140, etc. The input device 130 may include a keyboard, a mouse, a touch screen, etc., and the output device 140 may include a display device, such as a liquid crystal display (Liquid Crystal Display, LCD), a cathode ray tube (Cathode Ray Tube, CRT), etc.

[0031] The memory 120 may include a read-only memory (ROM) and a random access memory (RAM), and provides the processor 110 with program instructions and data stored in the memory 120. In the embodiment of the present application, the memory 120 may be used to store the program of the intelligent management and control method for the inspection and maintenance operation of the natural gas purification device in the embodiment of the present application.

[0032] The processor 110 calls the program instructions stored in the memory 120, and the processor 110 is used to execute the steps of any one of the intelligent management and control methods for inspection and maintenance operations of a natural gas purification device in the embodiments of the present application according to the obtained program instructions.

[0033] In addition, the application architecture diagram in the embodiment of the present application is intended to more clearly illustrate the technical solution in the embodiment of the present application, and does not constitute a limitation on the technical solution provided in the embodiment of the present application. Of course, for other application architectures and business applications, the technical solution provided in the embodiment of the present application is also applicable to similar problems.

[0034] The following is a non-restrictive description of the intelligent management and control method for inspection and maintenance operations of a natural gas purification device provided according to at least one embodiment of the present application through several examples or embodiments. As described below, different features in these specific examples or embodiments can be combined with each other without conflicting with each other to obtain new examples or embodiments, and these new examples or embodiments also fall within the scope of protection of the present application.

[0035] In response to the above technical problems, in the technical solution of this application, a natural gas purification device inspection and maintenance operation intelligent management and control system is proposed, which can use sensors and monitoring equipment to collect the equipment operation data of the natural gas purification device in real time, and then transmit these equipment operation data to the cloud computing platform for analysis, so as to use the data processing and analysis algorithm based on artificial intelligence and deep learning technology to learn and capture the time series collaborative association and interactive feature information between the operation data of each device of the natural gas purification device, so as to identify the equipment operation status and determine whether to perform equipment pre-maintenance. In this way, potential signs of equipment abnormality or performance degradation can be discovered in time, and possible failure risks can be warned in advance, so as to perform equipment pre-maintenance and realize intelligent management and control of the inspection and maintenance operations of the natural gas purification device.

[0036] Figure 2 The flowchart of the intelligent control method for the inspection and maintenance operation of the natural gas purification device in the embodiment of the present application is shown. For example, the intelligent control method for the inspection and maintenance operation of the natural gas purification device can be executed by a server, which can be Figure 1 The server 100 shown in FIG. Figure 2As shown, according to the embodiment of the present application, the intelligent management and control method for the inspection and maintenance operation of the natural gas purification device includes the following steps: S510, collecting the time series of equipment operation data of the natural gas purification device through a sensor network, wherein the equipment operation data includes temperature value, pressure value, vibration amplitude value, current value and voltage value; S520, respectively embedding and encoding each equipment operation data in the time series of the equipment operation data to obtain a time series of equipment operation data embedded coding vectors; S530, passing the time series of the equipment operation data embedded coding vectors through a context association encoder based on data semantic association and data time span to obtain a sequence of context-associated equipment operation semantic feature vectors; S540, performing node message propagation fusion based on the importance of the fusion node on the sequence of context-associated equipment operation semantic feature vectors to obtain equipment operation status feature time series propagation aggregation representation information; S550, determining whether to perform equipment pre-maintenance based on the equipment operation status feature time series propagation aggregation representation information.

[0037] Specifically, in the technical solution of the present application, first, the time series of the equipment operation data of the natural gas purification device is collected through the sensor network, wherein the equipment operation data includes temperature value, pressure value, vibration amplitude value, current value and voltage value. It should be understood that, considering that the equipment operation data of the natural gas purification device contains temperature value, pressure value, vibration amplitude value, current value and voltage value, these operation data have correlation relationship and mutual influence in the time dimension, for example, the changes in temperature and pressure may interact and influence each other, and this correlation relationship can reflect the hidden equipment operation state of the natural gas purification device. However, since these operation data have different ranges and units, they are not comparable in the time dimension, and are not convenient for subsequent processing and comprehensive characterization of equipment operation state characteristics. Based on this, in the technical solution of the present application, each equipment operation data in the time series of the equipment operation data is respectively embedded in the equipment data encoding module based on the fully connected layer to obtain the time series of the equipment operation data embedded encoding vector. Through the processing of the device data embedding coding module based on the fully connected layer, different device operation status data can be embedded and mapped into a common space, so as to facilitate the subsequent extraction of the complementary characteristics of the time series operation data of each device and the analysis of the correlation. In particular, the fully connected layer can learn the nonlinear relationship between the input device operation data by introducing a nonlinear activation function, which is crucial for capturing complex patterns and abnormal behaviors in device operation.

[0038] Correspondingly, in step S520, each device operation data in the time series of the device operation data is embedded and encoded respectively to obtain a time series of device operation data embedded coding vectors, including: each device operation data in the time series of the device operation data is passed through a device data embedding coding module based on a fully connected layer to obtain a time series of device operation data embedded coding vectors.

[0039] Then, considering that each device operation data embedded coding vector in the time series of the device operation data embedded coding vector respectively contains the fully connected coding features between each device operation data of the natural gas purification device at a specific time point, and each device operation data of the natural gas purification device will continue to change over time. Therefore, there is a temporal correlation relationship and synergy between the fully connected coding features of the device operation data at each time point, and this correlation provides basic information for subsequent device operation status detection and equipment pre-maintenance judgment. Based on this, in the technical solution of the present application, the time series of the device operation data embedded coding vector is further passed through a context-related encoder based on data semantic association and data time span to obtain a sequence of context-related device operation semantic feature vectors. Through the processing of the context-related encoder based on data semantic association and data time span, the temporal correlation features and context dependency between the fully connected coding features of the device operation data at each time point can be learned and captured. In this way, the similarity semantics and temporal span information between the fully connected coding features of the device operation data at each time point can be used to optimize the expression of the embedded coding vectors of each device operation data, thereby obtaining a richer and more comprehensive feature representation. It is worth mentioning that here, by analyzing the similarity of the equipment status at different time points, it helps to identify similar failure modes or performance issues that may occur in the equipment in different time periods. By considering the time span, the encoder can better understand long-term trends and short-term fluctuations, providing a more comprehensive perspective for fault prediction.

[0040] Accordingly, in step S530, if Figure 3As shown, the time series of the device operation data embedded coding vector is passed through a context-associated encoder based on data semantic association and data time span to obtain a sequence of context-associated device operation semantic feature vectors, including: S531, extracting a predetermined device operation data embedded coding vector from the time series of the device operation data embedded coding vector; S532, calculating the semantic association between the predetermined device operation data embedded coding vector and all other device operation data embedded coding vectors to obtain a sequence of device operation data embedded coding semantic temporal associations; S533, calculating the global mean of the sequence of device operation data embedded coding semantic temporal associations to obtain the global average temporal association of device operation data embedded coding semantics; S534, using the global average temporal association of the device operation data embedded coding semantics as a weight, weightedly strengthening the predetermined device operation data embedded coding vector to obtain the context-associated device operation semantic feature vector corresponding to the predetermined device operation data embedded coding vector.

[0041] Among them, in step S532, the semantic association between the predetermined device operation data embedded coding vector and all other device operation data embedded coding vectors is calculated to obtain a sequence of device operation data embedded coding semantic temporal associations, including: calculating the semantic similarity between the predetermined device operation data embedded coding vector and all other device operation data embedded coding vectors to obtain a sequence of device operation data embedded coding semantic temporal similarities; calculating the number of feature vectors between the predetermined device operation data embedded coding vector and all other device operation data embedded coding vectors to obtain a sequence of device operation data embedded coding semantic temporal distance counter numbers; dividing the sequence of device operation data embedded coding semantic temporal similarities by the sequence of device operation data embedded coding semantic temporal distance counter numbers by position to obtain a sequence of device operation data embedded coding semantic temporal associations.

[0042] In a specific example, the time series of the device operation data embedded in the coding vector is passed through a context-associated encoder based on data semantic association and data time span to obtain a sequence of context-associated device operation semantic feature vectors, including: the time series of the device operation data embedded in the coding vector is processed through the context-associated encoder based on data semantic association and data time span with the following context-associated formula to obtain the sequence of context-associated device operation semantic feature vectors; wherein the context-associated formula is: ;in, and respectively represent the first and The device operation data is embedded in the coding vector, express and The number of eigenvectors between express and The similarity between The total number of feature vectors in the time series representing the embedding of the encoding vector of the device operation data minus one, and The values ​​of are equal, The first one in the sequence of the semantic feature vectors of the context-associated device operation A context-associated device operation semantic feature vector.

[0043] Furthermore, since each context-associated device operation semantic feature vector in the sequence of the context-associated device operation semantic feature vectors respectively contains semantic features about the device operation status at each time point after being expressed through the context temporal semantic association enhancement, and there is a correlation and importance relationship between the semantics of the device operation status at these different time points. In other words, the operation status of the equipment may continue to change in the time dimension, and the equipment operation status at different time points will have different contributions to the subsequent equipment status prediction and pre-maintenance management and control tasks. Based on this, in order to be able to comprehensively focus on the semantic feature information of the device operation status at different time points, thereby providing a more comprehensive representation of the device operation status features, in the technical solution of the present application, the sequence of the context-associated device operation semantic feature vectors is further input into the node message propagation fusion network of the fusion node importance to obtain the device operation status feature temporal propagation aggregation representation vector. The node message propagation fusion network of the fusion node importance can further deeply learn the implicit correlation and interaction between the equipment operation status time series features after feature enhancement expression at each time node, so as to identify the key time node semantics related to subsequent tasks, so as to weight different node semantics in the process of fusing the context-related equipment operation semantic features at these different time points, so as to highlight those more critical features related to equipment status prediction and pre-maintenance tasks, so as to carry out the operation status time series information propagation and aggregation representation of the natural gas purification device equipment, and provide an important basis for equipment pre-maintenance.

[0044] Correspondingly, in step S540, the sequence of context-associated device operation semantic feature vectors is subjected to node message propagation fusion based on fusion node importance to obtain device operation status feature time series propagation aggregate representation information, including: inputting the sequence of context-associated device operation semantic feature vectors into a node message propagation fusion network of fusion node importance to obtain a device operation status feature time series propagation aggregate representation vector as the device operation status feature time series propagation aggregate representation information.

[0045] The sequence of the context-associated device operation semantic feature vectors is input into a node message propagation fusion network of the fusion node importance to obtain a device operation state feature time series propagation aggregate representation vector as the device operation state feature time series propagation aggregate representation information, including: using a scoring function to measure the significance factors of the first k-1 context-associated device operation semantic feature vectors in the sequence of the context-associated device operation semantic feature vectors to obtain k-1 context-associated device operation semantic time series significance factors; using the k-1 context-associated device operation semantic time series significance factors as weights, calculating the position-weighted sum of the first k-1 context-associated device operation semantic feature vectors in the sequence of the context-associated device operation semantic feature vectors to aggregate the first k-1 context-associated device operation semantic feature vectors in the sequence of the context-associated device operation semantic feature vectors. k-1 context-associated device operation semantic feature vectors are obtained to obtain a context-associated device operation semantic time-series associated aggregate representation vector of a rear node set; a k-th context-associated device operation semantic feature vector in a sequence of the context-associated device operation semantic feature vectors is linearly modulated using a preset hyperparameter to obtain a modulated k-th context-associated device operation semantic feature vector; a position-wise sum of the modulated k-th context-associated device operation semantic feature vector and the context-associated device operation semantic time-series associated aggregate representation vector of the rear node set is calculated to obtain a context-associated device operation semantic time-series associated full-node aggregate representation vector; and the context-associated device operation semantic time-series associated full-node aggregate representation vector is input into a multi-layer perceptron model to obtain a device operation status feature time-series propagation aggregate representation vector.

[0046] Specifically, a scoring function is used to measure the significance factors of the first k-1 context-associated device operation semantic feature vectors in the sequence of the context-associated device operation semantic feature vectors to obtain k-1 context-associated device operation semantic temporal significance factors, including: respectively calculating the matrix product between each weight coefficient matrix and the first k-1 context-associated device operation semantic feature vectors to obtain the first k-1 context-associated device operation semantic temporal association weighted feature vectors; adding the corresponding bias vector to each context-associated device operation semantic temporal association weighted feature vector in the first k-1 context-associated device operation semantic temporal association weighted feature vectors to obtain the first k-1 context-associated device operation semantic temporal association biased feature vectors; using The function activates the first k-1 context-associated device operation semantic timing association bias feature vectors to obtain the first k-1 context-associated device operation semantic timing association activation feature vectors; calculates the product between each context-associated device operation semantic timing association activation feature vector in the first k-1 context-associated device operation semantic timing association activation feature vector and the weight coefficient vector to obtain the first k-1 context-associated device operation semantic timing association degree; inputs the first k-1 context-associated device operation semantic timing association degree into The function is used to obtain the k-1 context-related device operation semantic timing significance factors.

[0047] In a specific example, the sequence of the context-associated device operation semantic feature vectors is input into the node message propagation fusion network of the fusion node importance to obtain the device operation status feature time series propagation aggregation representation vector as the device operation status feature time series propagation aggregation representation information, including: inputting the sequence of the context-associated device operation semantic feature vectors into the node message propagation fusion network of the fusion node importance to process with the following message propagation fusion formula to obtain the device operation status feature time series propagation aggregation representation vector; wherein the message propagation fusion formula is: ;in, The first character in the sequence of the semantic feature vector of the context-related device operation Context-associated device operation semantic feature vector, represents the weight coefficient matrix, is the bias vector, represents the weight coefficient vector, express function, express function, The context-related device operation semantics temporal correlation significant factors at each position in the sequence representing the context-related device operation semantics temporal correlation significant factors, is the number of feature vectors in the sequence of semantic feature vectors run by the context association device, The first character in the sequence of the semantic feature vector of the context-related device operation context-associated device operation semantic feature vector, is the preset hyperparameter, represents a multi-layer perceptron, A representation vector is aggregated for the time series propagation of the characteristics of the equipment operation status.

[0048] Then, the equipment operation status feature time series propagation aggregation representation vector is passed through the classifier-based control result generator to obtain the control result, which is used to indicate whether to perform equipment pre-maintenance. In other words, the operation status of the natural gas purification device equipment is classified and processed by using the time series information propagation aggregation representation feature to identify the equipment operation status, thereby determining whether to perform equipment pre-maintenance. In this way, potential signs of equipment abnormality or performance degradation can be discovered in time, and possible failure risks can be warned in advance, so that equipment pre-maintenance can be performed to realize intelligent control of the inspection and maintenance operations of the natural gas purification device.

[0049] Accordingly, in step S550, based on the time series propagation aggregation representation information of the equipment operating status characteristics, it is determined whether to perform equipment pre-maintenance, including: passing the equipment operating status characteristic time series propagation aggregation representation vector through a classifier-based control result generator to obtain a control result, and the control result is used to indicate whether to perform equipment pre-maintenance.

[0050] In a preferred embodiment, the sequence of the context-associated device operation semantic feature vector expresses the associated coding features of the device operation data under the local timing nodes. When the node message propagation fusion network of the input fusion node importance performs feature fusion under the global timing, due to the non-correspondence of the associated coding feature distributions of each local timing node, the device operation status feature timing propagation aggregation representation vector will have a global timing associated coding feature distribution knowledge offset based on the difference in the local timing node importance propagation weights, thereby causing the global timing associated coding feature distribution to lack class regression interpretability relative to the classifier, affecting the accuracy of the classification results.

[0051] Based on this, when the device operation status feature time series propagation aggregation representation vector is passed through a classifier-based control result generator to obtain a control result, the device operation status feature time series propagation aggregation representation vector is optimized, including the steps of: calculating the autocorrelation matrix of the device operation status feature time series propagation aggregation representation vector, that is: ;in, represents the time series propagation aggregation representation vector of the operating status characteristics of the device, represents the transpose of a vector, represents matrix multiplication, Denotes the autocorrelation matrix.

[0052] The device operating status feature time series propagation aggregation representation vector is query-multiplied with the autocorrelation matrix to obtain an intermediate feature vector, that is: ;in, represents the time series propagation aggregation representation vector of the operating status characteristics of the device, represents the autocorrelation matrix, represents matrix multiplication, Denotes the intermediate eigenvector.

[0053] Calculate the first Row vector and Inner product sum of row vectors distance, to obtain the first weight matrix and the second weight matrix , expressed as: ;in, and The autocorrelation matrix is ​​represented by Row vector and row vector, represents the transpose of a vector, represents matrix multiplication, represents the autocorrelation matrix, represents the first weight matrix, represents the first weight matrix The eigenvalues ​​of the positions, represents the second weight matrix, represents the second weight matrix The eigenvalues ​​of the positions, Represents the calculation vector distance.

[0054] The first weight matrix and the second weight matrix After matrix multiplication, further matrix multiplication is performed with the intermediate feature vector to obtain an optimized device operation status feature timing propagation aggregation representation vector, that is: ;in, represents the first weight matrix, represents the second weight matrix, represents the intermediate eigenvector, represents matrix multiplication, The optimized device operation status feature time series propagation aggregation representation vector is represented. Finally, the optimized device operation status feature time series propagation aggregation representation vector is passed through a classifier-based control result generator to obtain a control result.

[0055] That is, for the bit-by-bit scattering association response space of the eigenvalues ​​of the device operating status feature time series propagation aggregation representation vector, a deep embedding structure in the high-order association space is obtained through the joint embedding space of the vector inner product embedding space and the vector independent embedding space, and the decomposable association response sub-dimensions of the bit-by-bit scattering association response space of the eigenvalues ​​of the device operating status feature time series propagation aggregation representation vector are represented as a set of spatial sub-dimensions to further perform dependency space mapping of the decomposable association response sub-dimensions of the deep embedding structure of the device operating status feature time series propagation aggregation representation vector, so as to improve the interpretability of the solution space of the class regression solution of the device operating status feature time series propagation aggregation representation vector based on the association distribution through the complex dimensional representation of the association response, thereby improving the accuracy of the control result obtained by the device operating status feature time series propagation aggregation representation vector through the classifier-based control result generator.

[0056] Furthermore, in the technical solution of the present application, the intelligent management and control method for inspection and maintenance operations of the natural gas purification device also includes a training step: for training the equipment data embedding coding module based on the fully connected layer, the context association encoder based on data semantic association and data time span, the node message propagation fusion network of the fusion node importance, and the classifier-based management and control result generator.

[0057] Wherein, the training step includes: obtaining training data, wherein the training data includes a time series of training equipment operation data of a natural gas purification device collected through a sensor network and actual control results, wherein the training equipment operation data includes training temperature values, training pressure values, training vibration amplitude values, training current values ​​and training voltage values; each training equipment operation data in the time series of the training equipment operation data is respectively passed through the equipment data embedding coding module based on the fully connected layer to obtain a time series of training equipment operation data embedded coding vectors; the time series of the training equipment operation data embedded coding vectors is passed through the context association encoder based on data semantic association and data time span to obtain a sequence of training context association equipment operation semantic feature vectors. ; Input the sequence of the training context-associated device operation semantic feature vectors into the node message propagation fusion network of the fused node importance to obtain the training device operation status feature time series propagation aggregation representation vector; Pass the training device operation status feature time series propagation aggregation representation vector through the classifier-based control result generator to obtain the training control result; Calculate the cross entropy loss function value between the training control result and the real control result to obtain the classification loss function value; Based on the classification loss function value, train the fully connected layer-based device data embedding coding module, the context-associated encoder based on data semantic association and data time span, the node message propagation fusion network of the fused node importance and the classifier-based control result generator.

[0058] In a preferred embodiment, the present application optimizes the training equipment operating status feature time series propagation aggregation representation vector when passing the training equipment operating status feature time series propagation aggregation representation vector through the classifier-based control result generator to obtain the training control result, including the steps of: determining the pre-maintenance class probability value indicating equipment pre-maintenance, and the non-pre-maintenance class probability value corresponding to pre-maintenance without equipment pre-maintenance, obtained by inputting the training equipment operating status feature time series propagation aggregation representation vector into the classifier-based control result generator; respectively calculating the power function of each feature value of the training equipment operating status feature time series propagation aggregation representation vector with the pre-maintenance class probability value as an exponent and the power function with the non-pre-maintenance class probability value as an exponent to obtain the training equipment operating status feature time series propagation aggregation pre-maintenance class vector and the training Training equipment operating status feature timing propagation aggregation without pre-maintenance class vector; multiplying the pre-maintenance class probability value with the training equipment operating status feature timing propagation aggregation without pre-maintenance class vector to obtain the first equipment operating status feature timing propagation aggregation cross class vector, and multiplying the non-pre-maintenance class probability value with the training equipment operating status feature timing propagation aggregation pre-maintenance class vector to obtain the second equipment operating status feature timing propagation aggregation cross class vector; after multiplying the first equipment operating status feature timing propagation aggregation cross class vector with the second equipment operating status feature timing propagation aggregation cross class vector, further perform dot addition with the dot product result of the pre-maintenance class probability value and the training equipment operating status feature timing propagation aggregation pre-maintenance class vector to obtain the optimized training equipment operating status feature timing propagation aggregation representation vector.

[0059] Through the above-mentioned optimization steps, when the applicant of the present application classifies the training equipment operating status feature time series propagation aggregation representation vector through the classifier-based control result generator, in order to achieve unsupervised domain adaptation of the feature distribution domain of the training equipment operating status feature time series propagation aggregation representation vector to the probability distribution domain of the classification probability, the pre-maintenance class probability value and the non-pre-maintenance class probability value obtained by the training equipment operating status feature time series propagation aggregation representation vector through the classifier-based control result generator are used as domain agents, and the moving average of the probability distribution of the training equipment operating status feature time series propagation aggregation representation vector is performed through the interaction of the power function class distribution as an exponential, and the knowledge transfer from unlabeled classification features to labeled probability distribution is achieved by superimposing the feature domain probability distribution knowledge of the training equipment operating status feature time series propagation aggregation representation vector itself, thereby promoting the classification operation of the training equipment operating status feature time series propagation aggregation representation vector through the classifier-based control result generator, that is, improving the classification training speed of the classifier-based control result generator and the accuracy of the classification results. In this way, the equipment operating status of the natural gas purification device can be identified more accurately, which helps to promptly discover potential equipment abnormalities or signs of performance degradation, and provide early warning of possible failure risks, so as to carry out equipment pre-maintenance and realize intelligent management and control of the inspection and maintenance operations of the natural gas purification device.

[0060] Based on the above embodiments, see Figure 4 As shown, it is a structural schematic diagram of an intelligent management and control system 800 for inspection and maintenance of a natural gas purification device in an embodiment of the present application. The intelligent management and control system 800 for inspection and maintenance of a natural gas purification device includes: an operation data acquisition module 810, which is used to collect the time series of equipment operation data of the natural gas purification device through a sensor network, wherein the equipment operation data includes temperature value, pressure value, vibration amplitude value, current value and voltage value; an embedded coding module 820, which is used to embed and code each equipment operation data in the time series of the equipment operation data to obtain a time series of equipment operation data embedded coding vectors; a context-related coding module 830, which is used to pass the time series of the equipment operation data embedded coding vectors through a context-related encoder based on data semantic association and data time span to obtain a sequence of context-related equipment operation semantic feature vectors; a fusion module 840, which is used to perform node message propagation fusion based on the importance of fusion nodes on the sequence of context-related equipment operation semantic feature vectors to obtain equipment operation status feature time series propagation aggregation representation information; a maintenance analysis module 850, which is used to determine whether to perform equipment pre-maintenance based on the equipment operation status feature time series propagation aggregation representation information.

[0061] Here, those skilled in the art can understand that the specific functions and operations of each module in the above-mentioned natural gas purification device inspection and maintenance operation intelligent management and control system 800 have been referred to above. Figure 2 to Figure 3 The intelligent management and control method for inspection and maintenance operations of a natural gas purification device has been introduced in detail, and therefore, its repeated description will be omitted.

[0062] Figure 5 FIG. 1 is an application scenario diagram of the intelligent management and control method for the inspection and maintenance of a natural gas purification device according to an embodiment of the present application. Figure 5 As shown, in this application scenario, first, the time series of equipment operation data of the natural gas purification device is collected through the sensor network (for example, Figure 5 D), wherein the equipment operation data includes temperature value, pressure value, vibration amplitude value, current value and voltage value, and then the time series of the equipment operation data is input into a server (for example, Figure 5 In S) shown in , the server can use the intelligent management and control algorithm for inspection and maintenance operations of the natural gas purification device to process the time series of the equipment operation data to obtain a management and control result for indicating whether to perform equipment pre-maintenance.

[0063] Based on the above embodiments, another exemplary embodiment of an electronic device is also provided in the embodiments of the present application. In some possible implementations, the electronic device in the embodiments of the present application may include a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor can implement the steps of the intelligent management and control method for inspection and maintenance operations of a natural gas purification device in the above embodiments when executing the program.

[0064] For example, in the case of electronic equipment Figure 1 Taking the server 100 in the example as an example, the processor in the electronic device is the processor 110 in the server 100, and the memory in the electronic device is the memory 120 in the server 100.

[0065] An embodiment of the present application also provides a computer-readable storage medium. Figure 6 1 shows a schematic diagram of a computer-readable storage medium 1000 according to an embodiment of the present application. Figure 6As shown, the computer-readable storage medium 1000 stores computer-executable instructions 1001. When the computer-executable instructions 1001 are executed by the processor, the intelligent management and control method for the inspection and maintenance operation of the natural gas purification device according to the embodiment of the present application described with reference to the above figures can be executed. The computer-readable storage medium includes, but is not limited to, for example, volatile memory and / or non-volatile memory. The volatile memory may include, for example, a random access memory (RAM) and / or a cache memory (cache), etc. The non-volatile memory may include, for example, a read-only memory (ROM), a hard disk, a flash memory, etc.

[0066] The embodiment of the present application also provides a computer program product or a computer program, which includes computer executable instructions, and the computer executable instructions are stored in a computer readable storage medium. The processor of the computer device reads the computer executable instructions from the computer readable storage medium, and the processor executes the computer executable instructions, so that the computer device executes the intelligent management and control method for the inspection and maintenance operation of the natural gas purification device according to the embodiment of the present application.

[0067] Those skilled in the art will appreciate that the contents disclosed in this application may be subject to various variations and improvements. For example, the various devices or components described above may be implemented by hardware, or by software, firmware, or a combination of some or all of the three.

[0068] In addition, although the present application makes various references to certain units in the system according to embodiments of the present application, any number of different units can be used and run on the client and / or server. The units are only illustrative, and different aspects of the system and method can use different units.

[0069] Those skilled in the art will appreciate that all or part of the steps in the above method can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium, such as a read-only memory, a disk or an optical disk. Optionally, all or part of the steps in the above embodiment can also be implemented using one or more integrated circuits. Accordingly, each module / unit in the above embodiment can be implemented in the form of hardware or in the form of software function modules. The present application is not limited to any particular form of combination of hardware and software.

[0070] Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by those of ordinary skill in the art to which this application belongs. It should also be understood that terms such as those defined in common dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an idealized or extremely formal sense, unless explicitly defined as such herein.

[0071] The above is an explanation of the present application and should not be considered as a limitation thereof. Although several exemplary embodiments of the present application have been described, it will be readily understood by those skilled in the art that many modifications may be made to the exemplary embodiments without departing from the novel teachings and advantages of the present application. Therefore, all of these modifications are intended to be included within the scope of the present application as defined by the claims. It should be understood that the above is an explanation of the present application and should not be considered to be limited to the specific embodiments disclosed, and modifications to the disclosed embodiments and other embodiments are intended to be included within the scope of the appended claims. The present application is defined by the claims and their equivalents.

Claims

1. An intelligent management and control method for inspection and maintenance of a natural gas purification device, characterized in that: include: Collecting a time series of equipment operation data of the natural gas purification device through a sensor network, wherein the equipment operation data includes temperature value, pressure value, vibration amplitude value, current value and voltage value; respectively embedding and encoding each device operation data in the time series of the device operation data to obtain a time series of device operation data embedded coding vectors; The device operation data is embedded into the time series of the coding vector through a context-associated encoder based on data semantic association and data time span to obtain a sequence of context-associated device operation semantic feature vectors; Performing node message propagation fusion based on the importance of fusion nodes on the sequence of the context-related device operation semantic feature vectors to obtain device operation status feature time series propagation aggregation representation information; Determining whether to perform equipment pre-maintenance based on the time series propagation aggregation representation information of the equipment operation status characteristics; The sequence of the context-associated device operation semantic feature vectors is subjected to node message propagation fusion based on the importance of the fusion node to obtain the device operation status feature time series propagation aggregate representation information, including: inputting the sequence of the context-associated device operation semantic feature vectors into a node message propagation fusion network of the importance of the fusion node to obtain the device operation status feature time series propagation aggregate representation vector as the device operation status feature time series propagation aggregate representation information; The sequence of the context-related device operation semantic feature vectors is input into the node message propagation fusion network of the fusion node importance to obtain the device operation status feature time series propagation aggregation representation vector as the device operation status feature time series propagation aggregation representation information, including: Using a scoring function to measure the significance factors of the first k-1 context-associated device operation semantic feature vectors in the sequence of the context-associated device operation semantic feature vectors to obtain k-1 context-associated device operation semantic temporal significance factors; Taking the k-1 context-associated device operation semantics temporal association significant factors as weights, calculating the weighted sum of the first k-1 context-associated device operation semantics feature vectors in the sequence of the context-associated device operation semantics feature vectors by position, and aggregating the first k-1 context-associated device operation semantics feature vectors in the sequence of the context-associated device operation semantics feature vectors to obtain a rear node set context-associated device operation semantics temporal association aggregate representation vector; Linearly modulating the kth context-associated device operation semantic feature vector in the sequence of the context-associated device operation semantic feature vectors using a preset hyperparameter to obtain a modulated kth context-associated device operation semantic feature vector; Calculate the positional sum of the kth context-associated device operation semantic feature vector after the modulation and the rear node set context-associated device operation semantic time-series associated aggregate representation vector to obtain a context-associated device operation semantic time-series associated full-node aggregate representation vector; Inputting the context-associated device operation semantics time series associated full-node aggregate representation vector into a multi-layer perceptron model to obtain the device operation status feature time series propagation aggregate representation vector; Wherein, determining whether to perform equipment pre-maintenance based on the equipment operation status feature time series propagation aggregation representation information includes: passing the equipment operation status feature time series propagation aggregation representation vector through a classifier-based control result generator to obtain a control result, wherein the control result is used to indicate whether to perform equipment pre-maintenance; Wherein, when the equipment operation status feature time series propagation aggregation representation vector is passed through the classifier-based control result generator to obtain the control result, the equipment operation status feature time series propagation aggregation representation vector is optimized, including the steps of: Calculate the autocorrelation matrix of the device operation status feature time series propagation aggregation representation vector, that is: ;in, represents the time series propagation aggregation representation vector of the operating status characteristics of the device, represents the transpose of a vector, represents matrix multiplication, represents the autocorrelation matrix; The device operating status feature time series propagation aggregation representation vector is query-multiplied with the autocorrelation matrix to obtain an intermediate feature vector, that is: ;in, represents the time series propagation aggregation representation vector of the operating status characteristics of the device, represents the autocorrelation matrix, represents matrix multiplication, represents the intermediate eigenvector; Calculate the autocorrelation matrix Row vector and Inner product sum of row vectors distance, to obtain the first weight matrix and the second weight matrix , expressed as: ;in, and The autocorrelation matrix is ​​represented by Row vector and row vector, represents the transpose of a vector, represents matrix multiplication, represents the autocorrelation matrix, represents the first weight matrix, represents the first weight matrix The eigenvalues ​​of the positions, represents the second weight matrix, represents the second weight matrix The eigenvalues ​​of the positions, Represents the calculation vector distance; The first weight matrix and the second weight matrix After matrix multiplication, further matrix multiplication is performed with the intermediate feature vector to obtain an optimized device operation status feature timing propagation aggregation representation vector, that is: ;in, represents the first weight matrix, represents the second weight matrix, represents the intermediate eigenvector, represents matrix multiplication, The optimized device operation status feature time series propagation aggregation representation vector is represented; finally, the optimized device operation status feature time series propagation aggregation representation vector is passed through a classifier-based control result generator to obtain a control result.

2. The intelligent management and control method for inspection and maintenance of a natural gas purification device according to claim 1 is characterized in that: Each device operation data in the time series of the device operation data is embedded and encoded respectively to obtain a time series of device operation data embedded coding vectors, including: each device operation data in the time series of the device operation data is passed through a device data embedding coding module based on a fully connected layer to obtain a time series of device operation data embedded coding vectors.

3. The intelligent management and control method for inspection and maintenance of a natural gas purification device according to claim 2 is characterized in that: The time series of the device operation data embedded in the coding vector is passed through a context-associated encoder based on data semantic association and data time span to obtain a sequence of context-associated device operation semantic feature vectors, including: extracting a predetermined device operation data embedding code vector from the time series of the device operation data embedding code vector; Calculating the semantic association between the predetermined device operation data embedding coding vector and all other device operation data embedding coding vectors to obtain a sequence of device operation data embedding coding semantic temporal associations; Calculating the global mean of the sequence of the device operation data embedded coding semantic time series correlation to obtain the device operation data embedded coding semantic time series global average correlation; The global average correlation degree of the device operation data embedded coding semantic time series is used as a weight to weight-enhance the predetermined device operation data embedded coding vector to obtain a context-associated device operation semantic feature vector corresponding to the predetermined device operation data embedded coding vector.

4. The intelligent management and control method for inspection and maintenance of a natural gas purification device according to claim 3 is characterized in that: Calculating the semantic association between the predetermined device operation data embedding coding vector and all other device operation data embedding coding vectors to obtain a sequence of device operation data embedding coding semantic temporal associations, including: Calculating the semantic similarity between the predetermined device operation data embedding coding vector and all other device operation data embedding coding vectors to obtain a sequence of device operation data embedding coding semantic temporal similarity; Calculate the number of feature vectors between the predetermined device operation data embedding coding vector and all other device operation data embedding coding vectors to obtain a sequence of device operation data embedding coding semantic temporal distance counter numbers; The sequence of the device operation data embedded coding semantic temporal similarity and the sequence of the device operation data embedded coding semantic temporal distance counts are divided by position to obtain the sequence of the device operation data embedded coding semantic temporal association.

5. The intelligent management and control method for inspection and maintenance of a natural gas purification device according to claim 4 is characterized in that: Using a scoring function to measure the significance factors of the first k-1 context-associated device operation semantic feature vectors in the sequence of the context-associated device operation semantic feature vectors to obtain k-1 context-associated device operation semantic temporal significance factors, including: Calculating the matrix products of each weight coefficient matrix and the first k-1 context-associated device operation semantic feature vectors to obtain the first k-1 context-associated device operation semantic time-series associated weighted feature vectors; Adding a corresponding bias vector to each context-associated device operation semantics temporal association weighted feature vector in the first k-1 context-associated device operation semantics temporal association weighted feature vectors to obtain the first k-1 context-associated device operation semantics temporal association biased feature vectors; use The function activates the first k-1 context-associated device operation semantics timing-associated bias feature vectors to obtain the first k-1 context-associated device operation semantics timing-associated activation feature vectors; Calculate the product of each context-associated device operation semantics time sequence associated activation feature vector in the first k-1 context-associated device operation semantics time sequence associated activation feature vectors and the weight coefficient vector to obtain the first k-1 context-associated device operation semantics time sequence association degree; Input the semantic temporal correlation of the first k-1 context-related devices into The function is used to obtain the k-1 context-related device operation semantic timing significance factors.

6. The intelligent management and control method for inspection and maintenance of a natural gas purification device according to claim 5 is characterized in that: It also includes a training step: for training the device data embedding encoding module based on the fully connected layer, the context association encoder based on data semantic association and data time span, the node message propagation fusion network of the fusion node importance, and the classifier-based management and control result generator; Wherein, the training step includes: Acquire training data, wherein the training data includes a time series of training equipment operation data of the natural gas purification device collected through a sensor network and actual control results, wherein the training equipment operation data includes training temperature values, training pressure values, training vibration amplitude values, training current values, and training voltage values; Each piece of training device operation data in the time series of the training device operation data is respectively passed through the device data embedding coding module based on the fully connected layer to obtain a time series of training device operation data embedding coding vectors; Passing the training device operation data embedded in the time series of the coding vector through the context-associated encoder based on data semantic association and data time span to obtain a sequence of training context-associated device operation semantic feature vectors; Inputting the sequence of the training context-associated device operation semantic feature vectors into the node message propagation fusion network of the fusion node importance to obtain a training device operation state feature time series propagation aggregation representation vector; Passing the training equipment running state feature time series propagation aggregation representation vector through the classifier-based control result generator to obtain a training control result; Calculating a cross entropy loss function value between the training control result and the actual control result to obtain a classification loss function value; Based on the classification loss function value, the fully connected layer-based device data embedding encoding module, the context-related encoder based on data semantic association and data time span, the node message propagation fusion network of fused node importance, and the classifier-based management and control result generator are trained.

7. An intelligent management and control system for inspection and maintenance of natural gas purification equipment, characterized in that: include: An operation data acquisition module, used to acquire a time series of equipment operation data of the natural gas purification device through a sensor network, wherein the equipment operation data includes temperature value, pressure value, vibration amplitude value, current value and voltage value; An embedding coding module, used for embedding coding each device operation data in the time series of the device operation data to obtain a time series of device operation data embedded coding vectors; A context-associated encoding module, used for embedding the device operation data into a time sequence of a coding vector through a context-associated encoder based on data semantic association and data time span to obtain a sequence of context-associated device operation semantic feature vectors; A fusion module, used for performing node message propagation fusion on the sequence of the context-related device operation semantic feature vectors based on the importance of the fusion node to obtain device operation status feature time series propagation aggregation representation information; A maintenance analysis module, used to determine whether to perform equipment pre-maintenance based on the time series propagation aggregation representation information of the equipment operation status characteristics; The sequence of the context-associated device operation semantic feature vectors is subjected to node message propagation fusion based on the importance of the fusion node to obtain the device operation status feature time series propagation aggregate representation information, including: inputting the sequence of the context-associated device operation semantic feature vectors into a node message propagation fusion network of the importance of the fusion node to obtain the device operation status feature time series propagation aggregate representation vector as the device operation status feature time series propagation aggregate representation information; The sequence of the context-related device operation semantic feature vectors is input into the node message propagation fusion network of the fusion node importance to obtain the device operation state feature time series propagation aggregation representation vector as the device operation state feature time series propagation aggregation representation information, including: Using a scoring function to measure the significance factors of the first k-1 context-associated device operation semantic feature vectors in the sequence of the context-associated device operation semantic feature vectors to obtain k-1 context-associated device operation semantic temporal significance factors; Taking the k-1 context-associated device operation semantics temporal association significant factors as weights, calculating the weighted sum of the first k-1 context-associated device operation semantics feature vectors in the sequence of the context-associated device operation semantics feature vectors by position, and aggregating the first k-1 context-associated device operation semantics feature vectors in the sequence of the context-associated device operation semantics feature vectors to obtain a rear node set context-associated device operation semantics temporal association aggregate representation vector; Linearly modulating the kth context-associated device operation semantic feature vector in the sequence of the context-associated device operation semantic feature vectors using a preset hyperparameter to obtain a modulated kth context-associated device operation semantic feature vector; Calculate the positional sum of the kth context-associated device operation semantic feature vector after the modulation and the rear node set context-associated device operation semantic time-series associated aggregate representation vector to obtain a context-associated device operation semantic time-series associated full-node aggregate representation vector; Inputting the context-associated device operation semantics time series associated full-node aggregate representation vector into a multi-layer perceptron model to obtain the device operation status feature time series propagation aggregate representation vector; Wherein, determining whether to perform equipment pre-maintenance based on the equipment operation status feature time series propagation aggregation representation information includes: passing the equipment operation status feature time series propagation aggregation representation vector through a classifier-based control result generator to obtain a control result, wherein the control result is used to indicate whether to perform equipment pre-maintenance; Wherein, when the equipment operation status feature time series propagation aggregation representation vector is passed through the classifier-based control result generator to obtain the control result, the equipment operation status feature time series propagation aggregation representation vector is optimized, including the steps of: Calculate the autocorrelation matrix of the device operation status feature time series propagation aggregation representation vector, that is: ;in, represents the time series propagation aggregation representation vector of the operating status characteristics of the device, represents the transpose of a vector, represents matrix multiplication, represents the autocorrelation matrix; The device operating status feature time series propagation aggregation representation vector is query-multiplied with the autocorrelation matrix to obtain an intermediate feature vector, that is: ;in, represents the time series propagation aggregation representation vector of the operating status characteristics of the device, represents the autocorrelation matrix, represents matrix multiplication, represents the intermediate eigenvector; Calculate the first Row vector and Inner product sum of row vectors distance, to obtain the first weight matrix and the second weight matrix , expressed as: ;in, and The autocorrelation matrix is ​​represented by Row vector and row vector, represents the transpose of a vector, represents matrix multiplication, represents the autocorrelation matrix, represents the first weight matrix, represents the first weight matrix The eigenvalues ​​of the positions, represents the second weight matrix, represents the second weight matrix The eigenvalues ​​of the positions, Represents the calculation vector distance; The first weight matrix and the second weight matrix After matrix multiplication, further matrix multiplication is performed with the intermediate feature vector to obtain an optimized device operation status feature timing propagation aggregation representation vector, that is: ;in, represents the first weight matrix, represents the second weight matrix, represents the intermediate eigenvector, represents matrix multiplication, The optimized device operation status feature time series propagation aggregation representation vector is represented; finally, the optimized device operation status feature time series propagation aggregation representation vector is passed through a classifier-based control result generator to obtain a control result.

Images