Analysis method and system for the impact of slope on oil and gas pipeline hazards

Abstract

The present application provides a method and system for analyzing the impact of slope hazards on oil and gas pipelines. The hazard attribute description feature of the specified slope event is extracted from the slope hazard data example. The hazard attribute description feature is unique and can represent the location information of the specified slope event. Furthermore, the slope risk factors associated with the hazard attribute description feature of the specified slope event can be analyzed from the slope hazard data based on the artificial intelligence analysis thread. In this way, a separate oil and gas pipeline hazard impact analysis result for the specified slope event can be generated based on the analyzed slope risk factors. The present application can accurately analyze the hazard points of the slope, so that the hazard points can be treated in advance, which can ensure the safety of the oil and gas pipeline and avoid subsequent safety hazards and economic losses caused by damage to the oil and gas pipeline due to geological disasters.

Description

Technical Field

[0001] The present application relates to the field of data analysis technology, and in particular to a method and system for analyzing the impact of slope hazards on oil and gas pipelines. Background Art

[0002] Some geological disasters are prone to occur on a slope. There are oil and gas pipelines on the slope. Therefore, there is a relatively large hazard to the oil and gas pipelines. Therefore, it is urgent to analyze and evaluate the hidden dangers on each slope and what scale of geological disasters will cause harm to the oil and gas pipelines. Therefore, a method for analyzing the impact of slope hazards on oil and gas pipelines is urgently needed to protect the oil and gas pipelines. Summary of the invention

[0003] In order to improve the technical problems existing in the related technologies, the present application provides a method and system for analyzing the impact of slope hazards on oil and gas pipelines.

[0004] In a first aspect, a method for analyzing the impact of slope hazards on oil and gas pipelines is provided, the method comprising:

[0005] Obtaining slope hazard data that needs to be processed, wherein the slope hazard data includes slope risk factors of no less than two slope events;

[0006] Obtaining a slope hazard data example of a designated slope event among the at least two slope events; the designated slope event refers to one of the at least two slope events; the slope hazard data example includes slope risk factors of the designated slope event;

[0007] Extracting the hazard attribute description features of the designated slope event from the slope hazard data examples, and analyzing the slope risk factors associated with the hazard attribute description features of the designated slope event from the slope hazard data based on the artificial intelligence analysis thread;

[0008] Based on the slope risk factors obtained through analysis, the oil and gas pipeline hazard impact analysis results of the designated slope event are generated.

[0009] It should be understood that the slope hazard data to be processed is obtained, and the slope hazard data includes the slope risk elements of no less than two slope matters; if there is a need to analyze the slope risk elements generated by the specified slope matters in the no less than two slope matters, then a section of slope hazard data example of the specified slope matter can be obtained; the specified slope matter can be one of no less than two slope matters. Then, the hazard attribute description feature of the specified slope matter is extracted from the slope hazard data example, and the hazard attribute description feature is unique and can represent the location information of the specified slope matter. Further, the slope risk elements associated with the hazard attribute description feature of the specified slope matter can be analyzed from the slope hazard data based on the artificial intelligence analysis thread, so that the slope risk elements obtained by the analysis can be used to generate a separate oil and gas pipeline hazard impact analysis result of the specified slope matter. The present application can accurately analyze the hazard points of the slope, so that the hazard points can be treated in advance, so that the safety of the oil and gas pipeline can be guaranteed, and the subsequent safety hazards and economic losses caused by the damage of the oil and gas pipeline due to geological disasters can be avoided.

[0010] In the present application, the hidden danger attribute description feature of the designated slope event is represented by a hidden danger attribute description vector; the artificial intelligence-based analysis thread analyzes the slope risk factors associated with the hidden danger attribute description feature of the designated slope event from the slope hidden danger data, including:

[0011] Performing feature extraction processing on the slope hidden danger data to obtain the occurrence characteristics of geological disaster hidden dangers corresponding to the slope hidden danger data;

[0012] Based on the artificial intelligence analysis thread, the correlation between the hidden danger attribute description vector and the geological disaster hidden danger occurrence feature is calculated to obtain the geological disaster hidden danger occurrence feature byte associated with the hidden danger attribute description feature;

[0013] The geological disaster hazard occurrence characteristic bytes are processed for risk factor identification to obtain slope risk factors associated with the hazard attribute description characteristics.

[0014] It should be understood that traditional technology uses manual analysis, which requires a large number of technicians and a lot of time for analysis when the amount of data is large. Therefore, the traditional technology processing efficiency is skewed and prone to errors. In view of the existing technology, I would like to adopt the artificial intelligence analysis thread. Based on the artificial intelligence analysis thread, the slope risk factors associated with the hazard attribute description characteristics of the specified slope event can be accurately and quickly analyzed from the slope hazard data.

[0015] In the present application, the correlation calculation is implemented by a feature analysis network; the feature analysis network includes a feature extraction local network and a local mining network, and the feature extraction local network and the local mining network are connected through a feature extraction unit;

[0016] The feature extraction local network and the local mining network are associated; the feature extraction local network includes x feature extraction units ranked by risk level, and the local mining network includes mining units corresponding to each feature extraction unit, where x is a positive integer; the feature extraction unit and the mining unit include a plurality of feature extraction networks connected in sequence;

[0017] Among them, an artificial intelligence analysis thread is spliced ​​in the global or local network unit in the feature analysis network, and the splicing positioning of the artificial intelligence analysis thread in several feature extraction networks corresponding to the network unit is not fixed; the network unit includes the feature extraction unit and the mining unit.

[0018] In the present application, each network unit in the feature analysis network is spliced ​​with an artificial intelligence analysis thread; the artificial intelligence analysis thread performs correlation calculation on the hidden danger attribute description vector and the geological disaster hidden danger occurrence feature to obtain the geological disaster hidden danger occurrence feature byte associated with the hidden danger attribute description feature, including:

[0019] Loading the hidden danger attribute description vector into each network unit in the feature analysis network;

[0020] According to the artificial intelligence analysis threads spliced ​​in the respective network units, the correlation between the hidden danger attribute description vector and the first characteristic relationship chain of the corresponding network unit is calculated to obtain the second characteristic relationship chain output by the corresponding network unit; the hidden danger attribute description feature represented by the second characteristic relationship chain output by the corresponding network unit is associated with the hidden danger attribute description feature represented by the hidden danger attribute description vector;

[0021] The second feature relationship chain output by the 2x+1th network unit in the feature analysis network is used as a geological disaster hazard occurrence feature byte associated with the hazard attribute description feature represented by the hazard attribute description vector.

[0022] It should be understood that when the correlation between the hidden danger attribute description vector and the geological disaster hidden danger occurrence characteristics is calculated based on the artificial intelligence analysis thread, the problem of inaccurate calculation of the network unit is improved, so that the geological disaster hidden danger occurrence characteristic bytes associated with the hidden danger attribute description characteristics can be obtained more accurately.

[0023] In the present application, any network unit in the feature analysis network that is spliced ​​with an artificial intelligence analysis thread is represented as a target network unit; the target network unit is a feature extraction unit in the feature analysis network; the splicing positioning of the artificial intelligence analysis thread in the several feature extraction networks corresponding to the target network unit is: the first feature extraction network in the several feature extraction networks connected in sequence, and the feature extraction network in the several feature extraction networks connected in sequence that is adjacent to the first feature extraction network and located after the first feature extraction network; the correlation calculation of the hidden danger attribute description vector and the first feature relationship chain of the corresponding network unit is performed according to the artificial intelligence analysis threads spliced ​​in each network unit to obtain the second feature relationship chain output by the corresponding network unit, including:

[0024] The first feature extraction network in the target network unit is used to perform feature extraction processing on the first feature relationship chain of the target network unit to obtain a third feature relationship chain of the target network unit; wherein, when the target network unit is the first feature extraction unit ranked by danger level in the feature extraction local network, the first feature relationship chain of the target network unit is the occurrence characteristics of the geological disaster hazard; when the target network unit is a feature extraction unit other than the first feature extraction unit in the feature analysis network, the first feature relationship chain of the target network unit is obtained by performing reduced convolution processing on the feature relationship chain output by the upper-level network unit adjacent to the target network unit;

[0025] According to the artificial intelligence analysis thread spliced ​​in the target network unit, the correlation between the hidden danger attribute description vector and the third characteristic relationship chain of the target network unit is calculated to obtain a fourth characteristic relationship chain of the target network unit; the characteristic layer of the third characteristic relationship chain is the same as the characteristic layer of the fourth characteristic relationship chain;

[0026] The fourth feature relationship chain is subjected to feature extraction processing by using other feature extraction networks in the target network unit except the first feature extraction network to obtain a second feature relationship chain output by the target network unit.

[0027] It should be understood that the splicing positioning of the artificial intelligence analysis thread in the several feature extraction networks corresponding to the target network unit is: the positioning between the first feature extraction network among the several feature extraction networks connected sequentially, and the feature extraction network adjacent to the first feature extraction network and located after the first feature extraction network among the several feature extraction networks connected sequentially; when the artificial intelligence analysis thread spliced ​​in each network unit calculates the correlation between the hidden danger attribute description vector and the first feature relationship chain of the corresponding network unit, the problem of inaccurate calculation is improved, so that the second feature relationship chain output by the corresponding network unit can be accurately obtained.

[0028] In the present application, any network unit in the feature analysis network that is spliced ​​with an artificial intelligence analysis thread is represented as a target network unit; the target network unit is a mining unit in the feature analysis network; the splicing positioning of the artificial intelligence analysis thread in the several feature extraction networks corresponding to the target network unit is: the positioning after the tail feature extraction network in the several feature extraction networks connected in sequence; the correlation calculation of the hidden danger attribute description vector and the first feature relationship chain of the corresponding network unit is performed according to the artificial intelligence analysis threads spliced ​​in each network unit to obtain the second feature relationship chain output by the corresponding network unit, including:

[0029] Using a plurality of feature extraction networks sequentially connected in the target network unit, a target feature relationship chain is subjected to feature extraction processing to obtain a first feature relationship chain of the target network unit; the target feature relationship chain is obtained by feature splicing a feature relationship chain output by a feature extraction unit corresponding to the target network unit and a feature relationship chain output by a network unit at an upper level of the target network unit;

[0030] The artificial intelligence analysis thread spliced ​​in the target network unit is used to calculate the correlation between the hidden danger attribute description vector and the first feature relationship chain of the target network unit to obtain a second feature relationship chain output by the target network unit; the feature layer of the second feature relationship chain is the same as the feature layer of the first feature relationship chain.

[0031] It should be understood that the splicing positioning of the artificial intelligence analysis thread in the several feature extraction networks corresponding to the target network unit is: the positioning after the tail feature extraction network in the several feature extraction networks connected sequentially; when the correlation between the hidden danger attribute description vector and the first feature relationship chain of the corresponding network unit is calculated according to the artificial intelligence analysis thread spliced ​​in each network unit, the omission of calculation and the resulting calculation error are avoided, so that the second feature relationship chain output by the corresponding network unit can be accurately obtained.

[0032] In the present application, before the correlation calculation is performed on the hidden danger attribute description vector and the first characteristic relationship chain of the corresponding network unit according to the artificial intelligence analysis thread spliced ​​in each network unit to obtain the second characteristic relationship chain output by the corresponding network unit, it also includes:

[0033] Performing derivative processing on the hidden danger attribute description vector to obtain the hidden danger attribute description vector after derivative processing;

[0034] Among them, the feature layer of the hidden danger attribute description vector after derivative processing is the same as the feature layer of the feature relationship chain of the artificial intelligence analysis thread to be loaded into the corresponding network unit for splicing.

[0035] It should be understood that by performing derivative processing on the hidden danger attribute description vector, the hidden danger attribute description vector after the derivative processing can be obtained in more detail, so that the second characteristic relationship chain can be determined more accurately.

[0036] In the present application, if the number of network units spliced ​​with artificial intelligence analysis threads in the feature analysis network is greater than the target number, the method further includes:

[0037] Compressing the feature analysis network to obtain a compressed feature analysis network;

[0038] The correlation calculation is implemented by the feature analysis network after compression processing.

[0039] It is important to understand that performing derivative processing on the feature analysis network can improve the performance of the feature analysis network.

[0040] In the present application, the hidden danger attribute description feature of the specified slope event is represented by a hidden danger attribute description vector; the step of extracting the hidden danger attribute description feature of the specified slope event from the slope hidden danger data example includes:

[0041] Performing element analysis processing on the slope hidden danger data example to obtain a plurality of hidden danger elements corresponding to the slope hidden danger data example;

[0042] Performing feature extraction processing on the slope hidden danger data example to obtain reference geological hazard hidden danger occurrence features corresponding to the slope hidden danger data example, and performing element parsing processing on the reference geological hazard hidden danger occurrence features to obtain reference geological hazard hidden danger occurrence feature bytes corresponding to each hidden danger element;

[0043] Based on each of the hidden danger elements and the corresponding reference geological disaster hidden danger occurrence feature bytes, each of the hidden danger elements is subjected to important content identification processing to obtain a hidden danger attribute important content feature vector corresponding to each hidden danger element; the hidden danger attribute important content feature vector is used to indicate the feature that the hidden danger element has a great influence on the occurrence of geological disasters;

[0044] The hidden danger attribute important content feature vector queue is processed to obtain the hidden danger attribute description vector of the specified slope event; the hidden danger attribute important content feature vector queue includes the hidden danger attribute important content feature vectors corresponding to each hidden danger element.

[0045] It should be understood that the hazard attribute description features of the specified slope event need to be extracted from the slope hazard data example according to the importance of the data, so as to obtain the importance of the hazard attribute description features of the specified slope event.

[0046] In the present application, any one of the several hidden danger elements is represented as a target hidden danger element; the important content identification processing is performed on each hidden danger element based on each hidden danger element and the corresponding reference geological disaster hidden danger occurrence feature byte, and the hidden danger attribute important content feature vector corresponding to each hidden danger element is obtained, including:

[0047] Performing feature extraction processing on the target hidden danger elements to obtain a main hidden danger feature relationship chain;

[0048] Performing feature extraction processing on the reference geological hazard hazard occurrence feature bytes corresponding to the target hazard element to obtain a secondary hazard feature relationship chain;

[0049] The main hidden danger feature relationship chain and the secondary hidden danger feature relationship chain are spliced ​​to generate an important content feature vector of hidden danger attributes corresponding to the target hidden danger element.

[0050] It should be understood that when important content identification processing is performed on each hazard element based on the each hazard element and the corresponding reference geological disaster hazard occurrence feature bytes, the problem of inaccurate identification is improved, so that the important content feature vector of the hazard attribute corresponding to each hazard element can be accurately obtained.

[0051] In the present application, the number of feature extraction processes is a times, where a is an integer greater than 1; any feature extraction process is represented as the bth feature extraction process; the main hidden danger feature relationship chain and the secondary hidden danger feature relationship chain are spliced ​​to generate the hidden danger attribute important content feature vector corresponding to the target hidden danger element, including:

[0052] When b = 1, splice the transitional main hidden danger feature relation chain and the transitional secondary hidden danger feature relation chain obtained by the first feature extraction process to generate the first transitional feature vector after the first feature extraction process;

[0053] When 1 < b ≤ a, splice the transitional main hidden danger feature relation chain and the transitional secondary hidden danger feature relation chain obtained by the b-th feature extraction process, and the (b - 1)-th transitional feature vector obtained by the (b - 1)-th feature extraction process to generate the b-th transitional feature vector after the b-th feature extraction process;

[0054] Based on the a-th transitional feature vector after the a-th feature extraction process when b = a, generate the hidden danger attribute important content feature vector corresponding to the target hidden danger element.

[0055] It should be understood that when splicing the main hidden danger feature relation chain and the secondary hidden danger feature relation chain, the problem of inaccurate splicing is improved, so that the hidden danger attribute important content feature vector corresponding to the target hidden danger element can be generated.

[0056] In a second aspect, a system for analyzing the harm and influence of slopes on oil and gas pipelines is provided, including a processor and a memory that communicate with each other. The processor is configured to read and execute a computer program from the memory to implement the above method.

[0057] The method and system for analyzing the harm and influence of slopes on oil and gas pipelines provided by the embodiments of the present application perform knowledge fragment extraction processing on exemplary collapse hidden danger data to obtain secondary knowledge fragments and important knowledge fragments. Secondary description labels are built according to the secondary knowledge fragments corresponding to several exemplary collapse hidden danger data respectively, and important description labels are built according to the important knowledge fragments corresponding to several exemplary collapse hidden danger data respectively. In this way, similar collapse hidden danger data analysis for the to-be-processed collapse hidden danger data is realized in two branches, namely secondary and important, and the knowledge fragments of the collapse hidden danger data in different aspects can be simultaneously concerned, thereby improving the analysis accuracy, accurately obtaining the collapse hidden danger, and giving early warnings in advance. In this way, the harm caused by the collapse can be reduced as much as possible. Description of the Drawings

[0058] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present application and should not be regarded as limiting the scope. For those of ordinary skill in the art, other related drawings can be obtained based on these drawings without creative efforts.

[0059] Figure 1A flow chart of a method for analyzing the impact of slope hazards on oil and gas pipelines provided in an embodiment of the present application. DETAILED DESCRIPTION

[0060] In order to better understand the above technical scheme, the technical scheme of the present application is described in detail below through the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present application and the specific features in the embodiments are detailed descriptions of the technical scheme of the present application, rather than limitations on the technical scheme of the present application. In the absence of conflict, the embodiments of the present application and the technical features in the embodiments can be combined with each other.

[0061] See also Figure 1 , shows a method for analyzing the impact of slope on oil and gas pipeline hazards, which may include the technical solutions described in the following steps 301-304.

[0062] S301: Obtain slope hazard data that needs to be processed.

[0063] For example, slope hazard data can be obtained through manual actual survey, or through drone photography or radar detection, etc.

[0064] In this application, drone technology can be an important breakthrough compared to existing technologies. Drones can conduct exploration in places that cannot be reached by relevant technicians or equipment.

[0065] S302: Obtain slope hazard data examples of no less than two designated slope events.

[0066] Exemplarily, the slope hazard data example can be understood as sample data in the database, wherein the slope event refers to each slope object.

[0067] In steps S301-S302, the slope hazard data to be processed include slope risk factors of no less than two slope events.

[0068] S303: Extract the hazard attribute description features of the specified slope event from the slope hazard data example, and based on the artificial intelligence analysis thread, analyze the slope risk factors associated with the hazard attribute description features of the specified slope event from the slope hazard data.

[0069] Exemplarily, the artificial intelligence analysis thread is a model with analysis capabilities, such as: AARRR model, DuPont analysis, and RFM analysis model, etc.

[0070] Among them, slope risk factors may include: factors affecting landslides, etc.

[0071] S304: Based on the analyzed slope risk factors, generate oil and gas pipeline hazard impact analysis results for designated slope matters.

[0072] Exemplarily, the oil and gas pipeline hazard impact analysis results are obtained through analysis. In this application,

[0073] In steps S303-S304, the hidden danger attribute can be used to represent the location information of the slope event. Therefore, after obtaining a relatively real slope hidden danger data example of the specified slope event, the embodiment of the present application extracts the hidden danger attribute description feature of the specified slope event from the slope hidden danger data example, so as to facilitate the subsequent separate extraction of slope risk elements based on the unique hidden danger attribute description feature.

[0074] By analyzing the slope hazard data examples through the hazard attribute vector extraction model, it is possible to extract the hazard attribute description features used to represent the location information of the specified slope event from the slope hazard data examples. In practical applications, the hazard attribute vector extraction model outputs the hazard attribute description vector of the specified slope event, that is, the hazard attribute description features of the specified slope event are represented by the hazard attribute description vector. Furthermore, after extracting the hazard attribute description vector of the specified slope event, the hazard attribute vector extraction model can innovatively use vector embedding as the transmission of the hazard attribute information representation, and input it into the feature analysis network to participate in the calculation of the artificial intelligence analysis thread, so as to analyze the slope risk factors associated with the hazard attribute description features of the specified slope event from the slope hazard data.

[0075] Furthermore, each feature extraction unit and mining unit includes a plurality of feature extraction networks connected in sequence; the feature extraction local network feature extraction unit and mining unit can each include three feature extraction networks with a convolution kernel of 3*3. Among them, the feature extraction network is also called a convolutional neural network (CNN).

[0076] By embedding the hidden danger attribute description vector of the specified slope event into each network unit, mainly by performing attention calculation with the feature relationship chain of each network unit, the entire model can deeply perceive and learn the extracted hidden danger attribute description vector, so that each level of calculation can move closer to the hidden danger attribute description vector, ensuring that the slope risk factor finally extracted is associated with the hidden danger attribute description feature represented by the hidden danger attribute description vector. It is worth noting that the splicing positioning of the artificial intelligence analysis thread in several feature extraction networks corresponding to the network unit is not fixed.

[0077] The following is an introduction to the network structure of the feature analysis network. Taking the example of each network unit in the feature analysis network being spliced ​​with an artificial intelligence analysis thread, the feature analysis network analyzes the slope risk factors associated with the hazard attribute description characteristics of the specified slope event from the slope hazard data based on the artificial intelligence analysis thread, and generates the oil and gas pipeline hazard impact analysis results of the specified slope event based on the slope risk factors. The process may include but is not limited to steps (1)-(4), wherein:

[0078] (1) Feature extraction is performed on slope hazard data to obtain the occurrence characteristics of geological disaster hazards corresponding to the slope hazard data.

[0079] (2) Based on the artificial intelligence analysis thread, the correlation between the hidden danger attribute description vector and the geological disaster hidden danger occurrence characteristics is calculated to obtain the geological disaster hidden danger occurrence characteristic bytes associated with the hidden danger attribute description characteristics.

[0080] After obtaining the hidden danger attribute description vector of the specified slope event extracted by the hidden danger attribute vector extraction model, the hidden danger attribute description vector will be loaded into each network unit in the feature analysis network, specifically the artificial intelligence analysis thread spliced ​​in each network unit. In this way, according to the artificial intelligence analysis thread spliced ​​in each network unit, the hidden danger attribute description vector and the first feature relationship chain of the corresponding network unit can be correlated to obtain the second feature relationship chain output by the corresponding network unit; it is worth noting that according to the different positioning of the network unit in the feature analysis network, the first feature relationship chain of the network unit is also different, which will be introduced in the subsequent embodiments. Then, the second feature relationship chain output by the 2x+1th network unit (i.e., the last mining unit in the local mining network) in the feature analysis network can be used as the geological disaster hidden danger occurrence feature byte associated with the hidden danger attribute description vector. It can be seen from this that by loading the hidden danger attribute description vector used to represent the hidden danger attribute description characteristics of the specified slope event into each network unit in the feature analysis network to participate in the feature extraction process, each network unit in the feature analysis network can deeply perceive the hidden danger attribute information of the specified slope event, so that the slope risk factors output by the final analysis are closer to the hidden danger attribute description characteristics of the specified slope event, ensuring that the extracted slope risk factors are more accurate.

[0081] It should be understood that, depending on the different splicing positioning of the artificial intelligence analysis thread in the feature analysis network, the network units in the feature extraction local network and the local mining network have different specific implementation processes for calculating the correlation between the hidden danger attribute description vector and the first feature relationship chain output by the upper level of the corresponding network unit according to the spliced ​​artificial intelligence analysis thread.

[0082] In one possible implementation, it is assumed that the target network unit is a feature extraction unit in a feature analysis network, and the splicing positioning of the artificial intelligence analysis thread in the several feature extraction networks corresponding to the target network unit is: the first feature extraction network in the several feature extraction networks connected in sequence, and the feature extraction network in the several feature extraction networks connected in sequence that is adjacent to the first feature extraction network and located after the first feature extraction network. In this implementation, the target network unit calculates the correlation between the hidden danger attribute description vector and the first feature relationship chain of the target network unit to obtain the specific implementation process of the second feature relationship chain output by the target network unit may include: first, using the first feature extraction network in the target network unit, performing feature extraction processing on the first feature relationship chain of the target network unit to obtain the third feature relationship chain of the target network unit. Then, according to the artificial intelligence analysis thread spliced ​​in the target network unit, the correlation between the hidden danger attribute description vector and the third feature relationship chain of the target network unit is calculated to obtain the fourth feature relationship chain of the target network unit; the feature layer of the third feature relationship chain of the target network unit is the same as the feature layer of the fourth feature relationship chain, that is, the feature layer before and after the artificial intelligence analysis thread calculation is the same. Finally, the feature extraction network other than the first feature extraction network in the target network unit is used to perform feature extraction processing on the fourth feature relationship chain to obtain the second feature relationship chain output by the target network unit.

[0083] In other implementations, it is assumed that the target network unit is a mining unit in a feature analysis network, and the splicing positioning of the artificial intelligence analysis thread in the several feature extraction networks corresponding to the target network unit is: the positioning after the tail feature extraction network in the several feature extraction networks connected in sequence. In this implementation, the target network unit performs a correlation calculation on the hidden danger attribute description vector and the first feature relationship chain of the target network unit to obtain the second feature relationship chain output by the target network unit. The specific implementation process may include: first, using several feature extraction networks connected in sequence in the target network unit to perform feature extraction processing on the target feature relationship chain to obtain the first feature relationship chain of the target network unit. Then, using the artificial intelligence analysis thread spliced ​​in the target network unit, the hidden danger attribute description vector and the first feature relationship chain of the target network unit are calculated to obtain the second feature relationship chain output by the target network unit; the feature layer of the second feature relationship chain is the same as the feature layer of the first feature relationship chain.

[0084] It can be understood that when the artificial intelligence analysis thread is spliced ​​to an exemplary splicing position in the feature extraction unit of the feature extraction module and the local mining network unit respectively, the target network unit performs an exemplary process of correlation calculation; when the artificial intelligence analysis thread is spliced ​​to different splicing positions in the target network unit, the specific implementation process of the target network unit performing the correlation calculation is different.

[0085] (3) The geological hazard occurrence characteristic bytes associated with the hazard attribute description characteristics are processed to obtain the slope risk factors associated with the hazard attribute description characteristics.

[0086] Based on the above steps, after extracting the geological hazard hazard occurrence feature bytes associated with the hazard attribute description characteristics of the specified slope event from the geological hazard hazard occurrence characteristics corresponding to the slope hazard data, it is also necessary to process the geological hazard hazard occurrence feature bytes to obtain slope risk factors that meet the data transmission format.

[0087] (4) Generate the oil and gas pipeline hazard impact analysis results of the specified slope matters based on the slope risk factors associated with the hazard attribute description characteristics.

[0088] Based on the slope risk factor extraction process described in the above steps (1)-(4), it can be seen that the embodiment of the present application supports the calculation of the hidden danger attribute description vector and each network unit in the feature analysis network; however, considering that the feature layer of each network unit in the feature analysis network is different, specifically, the feature layer of each feature extraction network in each network unit is different. Therefore, before loading the hidden danger attribute description vector of the specified slope event into each network unit in the feature analysis network, it is necessary to first use a layer of network units to derive the hidden danger attribute description vector to obtain the hidden danger attribute description vector after the derivative processing. Among them, the feature layer of the hidden danger attribute description vector after the derivative processing is the same as the feature layer of the feature relationship chain of the artificial intelligence analysis thread to be loaded into the corresponding network unit for splicing; the feature relationship chain of the artificial intelligence analysis thread to be loaded into the corresponding network unit for splicing here can refer to the third feature relationship chain described above. In addition, the artificial intelligence analysis thread can be inserted between any two feature extraction networks in a plurality of feature extraction networks connected sequentially in the network unit; the first feature relationship chain of the network unit can refer to the feature relationship chain output by the feature extraction network in the network unit that is adjacent to the artificial intelligence analysis thread and located before the artificial intelligence analysis thread.

[0089] An exemplary embodiment of the present application provides a flowchart of another method for analyzing the impact of slope on oil and gas pipeline hazards; the method for analyzing the impact of slope on oil and gas pipeline hazards can be executed by a computer device in the aforementioned system, such as a terminal and / or a server; the method for analyzing the impact of slope on oil and gas pipeline hazards can include but is not limited to steps S801-S806:

[0090] S801: Obtain slope hazard data that needs to be processed.

[0091] S802: Obtain slope hazard data examples of designated slope events of no less than two slope events.

[0092] S803: Perform important content identification processing on the slope hazard data example of the designated slope matter.

[0093] S804: Processing the slope hazard data example of the designated slope event to obtain a hazard attribute description vector of the designated slope event.

[0094] In steps S803-S804, in order to learn clearer hazard attribute description features of the specified slope event from the slope hazard data examples of the specified slope event, the embodiment of the present application supports analyzing the slope hazard data examples in combination with short-time correlation and long-time correlation to extract a hazard attribute description vector that can fully express the hazard attribute description features of the specified slope event.

[0095] S805: Based on the artificial intelligence analysis thread, slope risk factors associated with the hazard attribute description characteristics of the specified slope event are analyzed from the slope hazard data.

[0096] It can be understood that the specific implementation process shown in step S805 and the relevant description of the local specific implementation process in step S302 regarding analyzing the slope risk factors associated with the hazard attribute description characteristics of the specified slope event from the slope hazard data based on the artificial intelligence analysis thread are not repeated here.

[0097] S806: Based on the analyzed slope risk factors, generate oil and gas pipeline hazard impact analysis results for designated slope matters.

[0098] On the basis of the above, a device for analyzing the impact of slope on oil and gas pipeline hazards is provided, the device comprising:

[0099] A data acquisition module, used to obtain slope hazard data that needs to be processed, wherein the slope hazard data includes slope risk factors of no less than two slope events;

[0100] An example obtaining module, used for obtaining a slope hazard data example of a designated slope event among the at least two slope events; the designated slope event refers to one of the at least two slope events; the slope hazard data example covers the slope risk elements of the designated slope event;

[0101] An element determination module is used to extract the hazard attribute description characteristics of the specified slope event from the slope hazard data example, and analyze the slope risk elements associated with the hazard attribute description characteristics of the specified slope event from the slope hazard data based on the artificial intelligence analysis thread;

[0102] The result analysis module is used to generate the oil and gas pipeline hazard impact analysis results of the specified slope matter based on the slope risk factors obtained through analysis.

[0103] On the basis of the above, a system for analyzing the impact of slope hazards on oil and gas pipelines is shown, including a processor and a memory that communicate with each other, and the processor is used to read and execute a computer program from the memory to implement the above method.

[0104] Based on the above, a computer-readable storage medium is also provided, on which a computer program stored implements the above method when running.

[0105] In summary, based on the above scheme, the slope hazard data that needs to be processed is obtained, and the slope hazard data includes the slope risk elements of no less than two slope matters; if there is a need to analyze the slope risk elements generated by the specified slope matters in the no less than two slope matters, then a section of slope hazard data example of the specified slope matter can be obtained; the specified slope matter can be one of the no less than two slope matters. Then, the hazard attribute description feature of the specified slope matter is extracted from the slope hazard data example, and the hazard attribute description feature is unique and can represent the location information of the specified slope matter. Further, the slope risk elements associated with the hazard attribute description feature of the specified slope matter can be analyzed from the slope hazard data based on the artificial intelligence analysis thread, so that the slope risk elements obtained by the analysis can be used to generate a separate oil and gas pipeline hazard impact analysis result of the specified slope matter. The present application can accurately analyze the hazard points of the slope, so that the hazard points can be treated in advance, so that the safety of the oil and gas pipeline can be guaranteed, and the subsequent safety hazards and economic losses caused by the damage of the oil and gas pipeline due to geological disasters can be avoided.

[0106] It should be understood that the system and its modules shown above can be implemented in various ways. For example, in some embodiments, the system and its modules can be implemented by hardware, software, or a combination of software and hardware. Among them, the hardware part can be implemented using dedicated logic; the software part can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or a dedicated design hardware. Those skilled in the art will understand that the above methods and systems can be implemented using computer executable instructions and / or included in a processor control code, for example, such as a programmable memory such as a read-only memory (firmware) or a data carrier such as an optical or electronic signal carrier on a carrier medium such as a disk CD or DVD-ROM. Such code is provided on the carrier medium such as a disk CD or DVD-ROM. The system and its modules of the present application can not only be implemented by hardware circuits such as semiconductors such as logic chip transistors or programmable hardware devices such as field programmable gate array programmable logic devices, but can also be implemented by software such as executed by various types of processors, and can also be implemented by a combination of the above hardware circuits and software (for example, firmware).

[0107] It should be noted that different embodiments may produce different beneficial effects. In different embodiments, the beneficial effects that may be produced may be any one or a combination of the above, or any other beneficial effects that may be obtained.

Claims

1. The method for analyzing the impact of slope on oil and gas pipeline hazards is characterized by: The method comprises: Obtaining slope hazard data that needs to be processed, wherein the slope hazard data includes slope risk factors of no less than two slope events; Obtaining a slope hazard data example of a designated slope event among the at least two slope events; the designated slope event refers to one of the at least two slope events; the slope hazard data example includes slope risk factors of the designated slope event; Extracting the hazard attribute description features of the designated slope event from the slope hazard data examples, and analyzing the slope risk factors associated with the hazard attribute description features of the designated slope event from the slope hazard data based on the artificial intelligence analysis thread; Based on the slope risk factors obtained through analysis, an oil and gas pipeline hazard impact analysis result of the designated slope event is generated; The hidden danger attribute description feature of the designated slope event is represented by a hidden danger attribute description vector; the artificial intelligence analysis thread analyzes the slope risk factors associated with the hidden danger attribute description feature of the designated slope event from the slope hidden danger data, including: Performing feature extraction processing on the slope hidden danger data to obtain the occurrence characteristics of geological disaster hidden dangers corresponding to the slope hidden danger data; Based on the artificial intelligence analysis thread, the correlation between the hidden danger attribute description vector and the geological disaster hidden danger occurrence feature is calculated to obtain the geological disaster hidden danger occurrence feature byte associated with the hidden danger attribute description feature; Performing risk factor identification processing on the geological disaster hidden danger occurrence characteristic bytes to obtain slope risk factors associated with the hidden danger attribute description characteristics; Wherein, the correlation calculation is realized by a feature analysis network; the feature analysis network includes a feature extraction local network and a local mining network, and the feature extraction local network and the local mining network are connected through a feature extraction unit; The feature extraction local network and the local mining network are associated; the feature extraction local network includes x feature extraction units ranked by risk level, and the local mining network includes mining units corresponding to each feature extraction unit, where x is a positive integer; the feature extraction unit and the mining unit include a plurality of feature extraction networks connected in sequence; Wherein, an artificial intelligence analysis thread is spliced ​​in a global or local network unit in the feature analysis network, and the splicing positioning of the artificial intelligence analysis thread in a plurality of feature extraction networks corresponding to the network unit is not fixed; the network unit includes the feature extraction unit and the mining unit; Among them, each network unit in the feature analysis network is spliced ​​with an artificial intelligence analysis thread; the artificial intelligence analysis thread is used to calculate the correlation between the hidden danger attribute description vector and the geological disaster hidden danger occurrence feature, and obtain the geological disaster hidden danger occurrence feature byte associated with the hidden danger attribute description feature, including: Loading the hidden danger attribute description vector into each network unit in the feature analysis network; According to the artificial intelligence analysis threads spliced ​​in the respective network units, the correlation between the hidden danger attribute description vector and the first characteristic relationship chain of the corresponding network unit is calculated to obtain the second characteristic relationship chain output by the corresponding network unit; the hidden danger attribute description feature represented by the second characteristic relationship chain output by the corresponding network unit is associated with the hidden danger attribute description feature represented by the hidden danger attribute description vector; The second feature relationship chain output by the 2x+1th network unit in the feature analysis network is used as a geological disaster hazard occurrence feature byte associated with the hazard attribute description feature represented by the hazard attribute description vector.

2. The method according to claim 1, characterized in that Any network unit in the feature analysis network that is spliced ​​with an artificial intelligence analysis thread is represented as a target network unit; the target network unit is a feature extraction unit in the feature analysis network; the splicing location of the artificial intelligence analysis thread in the plurality of feature extraction networks corresponding to the target network unit is: the first feature extraction network in the plurality of feature extraction networks connected in sequence, and the feature extraction network in the plurality of feature extraction networks connected in sequence that is adjacent to the first feature extraction network and located after the first feature extraction network; the correlation calculation is performed on the hidden danger attribute description vector and the first feature relationship chain of the corresponding network unit according to the artificial intelligence analysis threads spliced ​​in each network unit to obtain the second feature relationship chain output by the corresponding network unit, including: The first feature extraction network in the target network unit is used to perform feature extraction processing on the first feature relationship chain of the target network unit to obtain a third feature relationship chain of the target network unit; wherein, when the target network unit is the first feature extraction unit ranked by danger level in the feature extraction local network, the first feature relationship chain of the target network unit is the occurrence characteristics of the geological disaster hazard; when the target network unit is a feature extraction unit other than the first feature extraction unit in the feature analysis network, the first feature relationship chain of the target network unit is obtained by performing reduced convolution processing on the feature relationship chain output by the upper-level network unit adjacent to the target network unit; According to the artificial intelligence analysis thread spliced ​​in the target network unit, the correlation between the hidden danger attribute description vector and the third characteristic relationship chain of the target network unit is calculated to obtain a fourth characteristic relationship chain of the target network unit; the characteristic layer of the third characteristic relationship chain is the same as the characteristic layer of the fourth characteristic relationship chain; The fourth feature relationship chain is subjected to feature extraction processing by using other feature extraction networks in the target network unit except the first feature extraction network to obtain a second feature relationship chain output by the target network unit.

3. The method according to claim 1, characterized in that Any network unit in the feature analysis network that is spliced ​​with an artificial intelligence analysis thread is represented as a target network unit; the target network unit is a mining unit in the feature analysis network; the splicing location of the artificial intelligence analysis thread in the plurality of feature extraction networks corresponding to the target network unit is: the location after the tail feature extraction network in the plurality of feature extraction networks connected in sequence; the correlation calculation is performed on the hidden danger attribute description vector and the first feature relationship chain of the corresponding network unit according to the artificial intelligence analysis threads spliced ​​in each network unit to obtain the second feature relationship chain output by the corresponding network unit, including: Using a plurality of feature extraction networks sequentially connected in the target network unit, a target feature relationship chain is subjected to feature extraction processing to obtain a first feature relationship chain of the target network unit; the target feature relationship chain is obtained by feature splicing a feature relationship chain output by a feature extraction unit corresponding to the target network unit and a feature relationship chain output by a network unit at an upper level of the target network unit; The artificial intelligence analysis thread spliced ​​in the target network unit is used to calculate the correlation between the hidden danger attribute description vector and the first feature relationship chain of the target network unit to obtain a second feature relationship chain output by the target network unit; the feature layer of the second feature relationship chain is the same as the feature layer of the first feature relationship chain.

4. The method according to claim 1, characterized in that Before calculating the correlation between the hidden danger attribute description vector and the first characteristic relationship chain of the corresponding network unit according to the artificial intelligence analysis threads spliced ​​in each network unit to obtain the second characteristic relationship chain output by the corresponding network unit, the method further includes: Performing derivative processing on the hidden danger attribute description vector to obtain the hidden danger attribute description vector after derivative processing; Among them, the feature layer of the hidden danger attribute description vector after derivative processing is the same as the feature layer of the feature relationship chain of the artificial intelligence analysis thread to be loaded into the corresponding network unit for splicing.

5. The method according to claim 1, characterized in that If the number of network units in the feature analysis network spliced ​​with artificial intelligence analysis threads is greater than the target number, the method further includes: Compressing the feature analysis network to obtain a compressed feature analysis network; The correlation calculation is implemented by the feature analysis network after compression processing.

6. The method according to claim 1, characterized in that The hidden danger attribute description feature of the designated slope event is represented by a hidden danger attribute description vector; the step of extracting the hidden danger attribute description feature of the designated slope event from the slope hidden danger data example includes: Performing element analysis processing on the slope hidden danger data example to obtain a plurality of hidden danger elements corresponding to the slope hidden danger data example; Performing feature extraction processing on the slope hidden danger data example to obtain reference geological hazard hidden danger occurrence features corresponding to the slope hidden danger data example, and performing element parsing processing on the reference geological hazard hidden danger occurrence features to obtain reference geological hazard hidden danger occurrence feature bytes corresponding to each hidden danger element; Based on each of the hidden danger elements and the corresponding reference geological disaster hidden danger occurrence characteristic bytes respectively, important content identification processing is performed on each of the hidden danger elements to obtain a hidden danger attribute important content feature vector corresponding to each of the hidden danger elements; the hidden danger attribute important content feature vector is used to represent the characteristics of the hidden danger element having a great impact on the occurrence of geological disasters. Processing the hidden danger attribute important content feature vector queue to obtain a hidden danger attribute description vector of the specified slope matter; the hidden danger attribute important content feature vector queue includes the hidden danger attribute important content feature vectors corresponding to each of the hidden danger elements. Wherein, any one of the several hidden danger elements is represented as a target hidden danger element; the respectively performing important content identification processing on each of the hidden danger elements based on each of the hidden danger elements and the corresponding reference geological disaster hidden danger occurrence characteristic bytes to obtain a hidden danger attribute important content feature vector corresponding to each of the hidden danger elements includes: Performing feature extraction processing on the target hidden danger element to obtain a main hidden danger feature relationship chain. Performing feature extraction processing on the reference geological disaster hidden danger occurrence characteristic bytes corresponding to the target hidden danger element to obtain a secondary hidden danger feature relationship chain. Performing splicing processing on the main hidden danger feature relationship chain and the secondary hidden danger feature relationship chain to generate a hidden danger attribute important content feature vector corresponding to the target hidden danger element. Wherein, the number of times of the feature extraction processing is a times, a is an integer greater than 1; any one time of the feature extraction processing is represented as the bth feature extraction processing; the performing splicing processing on the main hidden danger feature relationship chain and the secondary hidden danger feature relationship chain to generate a hidden danger attribute important content feature vector corresponding to the target hidden danger element includes: When b = 1, performing splicing processing on the transitional main hidden danger feature relationship chain and the transitional secondary hidden danger feature relationship chain obtained by the first feature extraction processing to generate a first transitional feature vector after the first feature extraction processing. When 1 < b ≤ a, performing splicing processing on the transitional main hidden danger feature relationship chain and the transitional secondary hidden danger feature relationship chain obtained by the bth feature extraction processing, and the (b - 1)th transitional feature vector obtained by the (b - 1)th feature extraction processing to generate the bth transitional feature vector after the bth feature extraction processing. Based on the ath transitional feature vector after the ath feature extraction processing when b = a, generating a hidden danger attribute important content feature vector corresponding to the target hidden danger element.

7. The slope-based oil and gas pipeline hazard analysis system is characterized by: Including a processor and a memory that communicate with each other, the processor is configured to read and execute a computer program from the memory to implement the method according to any one of claims 1 - 6.

Images