An intelligent linkage system and device for an oilfield gathering station

Abstract

The application discloses an intelligent linkage system and device for an oilfield gathering and transferring station, belongs to the technical field of abnormal treatment of oilfield gathering and transferring stations, and comprises an oilfield data acquisition module, an edge data fusion module, a gathering and transferring scene simulation and decision selection module and a decision execution and linkage control module.The oilfield data acquisition module comprises a multi-modal sensing unit, equipment data in an oilfield gathering and transferring process is acquired based on a unified timestamp, the sensitivity of equipment in the multi-modal sensing unit is calibrated in combination with the external environment of different oilfield gathering and transferring stations, and the acquired data is transmitted into the edge data fusion module.The edge data fusion module.The application discloses an intelligent linkage system and device for an oilfield gathering and transferring station, an intelligent system is integrated with multi-modal sensing, edge data fusion, scene simulation decision and linkage control, so that the accuracy of abnormal identification, the scientificity of decision and the stability of system operation are improved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of abnormal treatment of oilfield gathering and transportation station, and particularly relates to an intelligent linkage system and device for an oilfield gathering and transportation station. BACKGROUND

[0002] As a key node of crude oil exploitation and transportation, the oilfield gathering and transportation station undertakes important tasks such as oil and gas separation, metering, and transportation, and its operation stability directly affects the production efficiency and safety of the oilfield. However, the gathering and transportation station has a large number of devices and complex processes, involving pump groups, pipelines, separators and other devices, and is affected by factors such as geological conditions, climate environment, and equipment aging, and is prone to single or composite abnormalities such as pressure abnormalities, leaks, and equipment failures. The traditional operation and management mode faces many challenges.

[0003] Traditional management of the gathering and transportation station relies on manual inspection and decentralized monitoring, the data acquisition method is single, and each sensor operates independently, lacking unified time calibration and environmental sensitivity calibration, making it difficult to achieve real-time and accurate capture of the whole process chain situation, leading to delayed abnormal identification or misjudgment. In addition, abnormal treatment relies on manual experience, and it is difficult to quickly generate a scientific adjustment strategy when facing composite abnormalities, and the decision-making and device control lack closed-loop linkage, which may expand the impact of faults due to improper adjustment, and there are problems of low practicality and functionality. SUMMARY

[0004] The present application aims to solve at least one of the technical problems existing in the prior art. To this end, the present application proposes an intelligent linkage system and device for an oilfield gathering and transportation station, which improves the detection method and processing method to solve the above technical problems.

[0005] In order to achieve the above-mentioned purpose, the present application adopts the following technical solutions:

[0006] An intelligent linkage system for an oilfield gathering and transportation station, comprising an oilfield data acquisition module, an edge data fusion module, a gathering and transportation scenario simulation and decision selection module, and a decision execution and linkage control module.

[0007] The oilfield data acquisition module comprises a multi-modal sensing unit, and acquires device data during the oilfield gathering and transportation process based on a unified timestamp, and calibrates the device sensitivity in the multi-modal sensing unit in combination with the external environment of different oilfield gathering and transportation stations, and transmits the acquired data into the edge data fusion module;

[0008] The edge data fusion module normalizes the received acquired data, outputs the acquired data into a uniform format time sequence structure to obtain a health vector, sets an edge decision model for different types of acquired data in the multi-modal sensing unit, performs single abnormality determination and composite abnormality identification on the acquired data, and transmits the identification result into the subsequent module.

[0009] The gathering scene simulation and decision selection module generates a station simulation model through three-dimensional point cloud data based on the environment and equipment data of different oilfield gathering stations, runs a simulation equation set in the station simulation model, combines the service life and failure frequency of the equipment, calibrates the state of different equipment according to historical data, and assigns a corresponding compensation value in the station simulation model, inputs an abnormal health vector into the station simulation model to generate an output decision, simulates the output decision, selects an optimal adjustment strategy, and outputs the optimal adjustment strategy to a subsequent module;

[0010] The decision execution and linkage control module converts the received optimal adjustment strategy into a structured control instruction, binds the specific equipment action logic of different oilfield gathering stations to execute the optimal adjustment strategy, and collects the adjusted equipment data through a multi-modal sensing unit and performs abnormal verification to realize closed-loop adjustment control.

[0011] Further, the oilfield data acquisition module includes a multi-modal sensing unit, which acquires equipment data during the oilfield gathering process based on a unified timestamp, calibrates the sensitivity of the equipment in the multi-modal sensing unit in combination with the external environment of different oilfield gathering stations, and transmits the collected data into an edge data fusion module. The specific steps are as follows:

[0012] The multi-modal sensing unit is deployed in different oilfield gathering stations to collect equipment data during the oilfield gathering process. The multi-modal sensing unit integrates temperature sensors, pressure sensors, flow sensors, liquid level sensors, vibration sensors, infrared thermal imaging acquisition probes, visible light camera probes, and gas detection probes.

[0013] The data collected by the multi-modal sensing unit is compensated for clock offset and a timestamp is attached to each collected data , wherein represents the clock offset compensation value, represents the master clock sending time, the slave clock receiving time, the slave clock reply time, and the master clock receiving time, respectively, represents the timestamp time set for the data, represents the time when the data is collected.

[0014] In combination with the external environment data of different oilfield gathering stations, the sensitivity of the sensors in the multi-modal sensing unit is adjusted , wherein represents the sensor sensitivity adjustment value, represents the current sensor sensitivity reference value, represents the environmental attenuation coefficient and the interference intensity, respectively.

[0015] Further, the edge data fusion module normalizes the received collected data, and outputs a health vector in a uniform time sequence structure, including the following steps:

[0016] By receiving various equipment data collected by the multi-modal sensing unit in the oilfield data collection module , wherein represents the type of sensor, and the collected data is integrated into a unified data set, and the data in the data set is normalized by Z-score standardization:

[0017] ;

[0018] , wherein represents the th sensor data obtained after normalization, represents the sliding mean of the historical data of the th sensor, represents the standard deviation of the th sensor, and the normalized data is time-aligned by dynamic time warping, a data statistics library is established, and the data under the same time sequence is coded into the same data set to generate a health vector :

[0019] ;

[0020] , wherein represents the time sequence, represents the normalized data of the th sensor under the time sequence, respectively represent the collected image feature vector and gas concentration.

[0021] Further, the edge data fusion module normalizes the received collected data, and outputs a health vector in a uniform time sequence structure, including the following steps:

[0022] For the data collected by the multi-modal sensing unit, the data types are divided, including image data, time sequence data and numerical data, and single abnormality determination and composite abnormality determination are performed respectively:

[0023] In the single abnormality determination process, a convolutional neural network model is constructed for image data to identify abnormal targets in the image, a long short-term memory network model is constructed for time sequence data to identify abnormalities, and a preset threshold is used for abnormality identification of numerical data;

[0024] Based on the single abnormality determination result, the types of each single abnormality under the same time sequence are integrated, and the current abnormality under the same time sequence is confirmed combined with the composite abnormality determination rule. The specific steps are as follows:

[0025] Under the same time sequence, all single abnormality determination results are collected, and the collected determination results are integrated into a determination vector , wherein represents the determination result of different data categories, represents the determination result of data under the time sequence , and is substituted into the output rule to determine the composite abnormality, and the composite abnormality determination result is output , wherein , is the preset composite abnormality determination rule.

[0026] Further, the gathering scene simulation and decision selection module, based on the environment and equipment data of different oilfield gathering stations, generates a station simulation model through three-dimensional point cloud data, and runs the simulation equation set in the station simulation model, and at the same time, according to the service life and failure frequency of the equipment, the state of different equipment is calibrated according to the historical data, and the corresponding compensation value is given in the station simulation model, the abnormal health vector is input into the station simulation model to generate the output decision, the output decision is simulated and the optimal adjustment strategy is selected and output to the subsequent module, including the following steps:

[0027] Statistical analysis of the service life and failure frequency of different equipment in the current oilfield gathering station, combined with historical data to divide the equipment state under different service life and failure frequency, and to determine the service life and failure frequency state of the running equipment in the current gathering station. The determination result is selected to obtain the equipment state;

[0028] Three-dimensional point cloud data of the current oilfield gathering station is obtained by three-dimensional laser scanning to construct a three-dimensional simulation model of the station, and according to the physical principles and process flow of the gathering station, a simulation equation set is established to describe the running of the station. According to the equipment state of different equipment, the corresponding parameter compensation value is given to the equipment in the station simulation model, the abnormal health vector is input to generate the output decision, and the optimal decision is obtained by simulation evaluation and screening.

[0029] Further, the statistical analysis of the service life and failure frequency of different equipment in the current oilfield gathering station, combined with historical data to divide the equipment state under different service life and failure frequency, and to determine the service life and failure frequency state of the running equipment in the current gathering station. The determination result is selected to obtain the equipment state, and the specific steps are as follows:

[0030] Collecting the time of putting into use and the record of failure of different equipment in the current oilfield gathering and transferring station, and calculating the service life of different equipment , wherein respectively represent the current time and the time of putting into use of the equipment, and the failure frequency of the equipment in unit time is calculated , wherein represents the number of failures of the equipment in unit time

[0031] By comprehensively considering the service life and the failure frequency of the same equipment, the service life state of the equipment, including good, normal and general, is counted, and according to the high threshold value and the low threshold value of the failure frequency of the equipment in unit time, the failure frequency state of different equipment is determined separately, and the current equipment state is obtained by taking the worse one of the service life state and the failure frequency state of the equipment. The specific steps are as follows:

[0032] For each equipment, the service life is determined, when ≤ , the service life state of the equipment is good, when < ≤ , the service life state of the equipment is normal, and when > , the service life state of the equipment is general, wherein , represent the low threshold value and the high threshold value of the current service life of the equipment

[0033] For each equipment, the failure frequency is determined, when ≤ , the failure frequency state of the equipment is good, when < ≤ , the failure frequency state of the equipment is normal, and when > , the failure frequency state of the equipment is general, wherein , represent the low threshold value and the high threshold value of the current failure frequency of the equipment

[0034] The service life state and the failure frequency state of the equipment are taken to obtain the current equipment state.

[0035] ​​Further, the three-dimensional point cloud data of the current oilfield gathering and transferring station is acquired by three-dimensional laser scanning to construct a three-dimensional simulation model of the station, and a simulation equation set describing the operation of the station is established according to the physical principle and process flow of the gathering and transferring station, corresponding parameter compensation values are given to the equipment in the station simulation model according to the equipment states of different equipment, an output decision is generated by inputting the abnormal health vector, and the optimal decision is obtained by simulation evaluation and screening, including the following steps:

[0036] The three-dimensional point cloud data of the current oilfield gathering and transferring station is acquired by three-dimensional laser scanning, the point cloud data is filtered, the geometric information and topological structure of the station are extracted to construct a three-dimensional simulation model of the station, and a simulation equation set describing the operation of the station is established according to the physical principle and process flow of the oilfield gathering and transferring station, and the equation set is embedded into the station simulation model;

[0037] Different compensation coefficients are given to the equipment according to different equipment states The set value of the current equipment parameter after compensation is calculated , wherein represents the set value of the current equipment based on the equipment state parameter compensation, represents the original parameter of the equipment;

[0038] The abnormal health vector obtained by the edge data fusion module is input into the station simulation model, the running state of the station is updated in real time by solving the simulation equation set, the equipment adjustment strategy preset adjustment rule library under different abnormal parameters is collected, the running abnormal parameters in the current station simulation model are combined, the cosine similarity is calculated to match the rules in the adjustment rule library, K rounds of rule matching are performed to generate a response strategy, different effect adjustment rules in the adjustment rule library are combined, and one adjustment strategy is selected under each adjustment rule through M times of simple and non-repetitive random sampling, M adjustment response paths are generated by combination, and each adjustment response path is simulated in the station simulation model. The adjustment response path with the shortest response time is selected as the optimal adjustment strategy and output to the subsequent module.

[0039] Further, the decision execution and linkage control module converts the received optimal adjustment strategy into a structured control instruction, binds the specific equipment action logic of different oilfield gathering and transferring stations to execute the optimal adjustment strategy, and collects the adjusted equipment data through the multi-modal sensing unit and performs abnormal verification to realize closed-loop adjustment control, including the following steps:

[0040] According to the layout and equipment configuration information of different oilfield gathering and transferring stations, the mapping relationship between the control instruction and the specific equipment is established, the corresponding equipment action logic is bound for each control instruction, the received optimal adjustment strategy is converted into a structured control instruction to control the equipment, and the structured control instruction is sent to the equipment control system of the corresponding oilfield gathering and transferring station through an industrial protocol adaptation module for instruction execution.

[0041] The data of the oilfield gathering and transferring station after the execution of the instruction is collected through the multi-modal sensing unit in the oilfield data acquisition module, and the abnormality is identified in the edge data fusion module. When the abnormality still exists, the optimal adjustment strategy is generated again through the gathering and transferring scene simulation and decision selection module for abnormal adjustment. The above steps are repeated until the abnormality disappears or the maximum number of cycles is reached, and the abnormality report is transmitted to the on-duty personnel.

[0042] The intelligent linkage device for the oilfield gathering and transferring station is applied to the intelligent linkage system of the oilfield gathering and transferring station, and includes a memory, a processor, and a program stored in the memory and executable on the processor. The processor executes the program to implement the steps of the intelligent linkage system of the oilfield gathering and transferring station as described in the application.

[0043] Compared with the prior art, the application has the following beneficial effects:

[0044] 1. In the application, a plurality of sensors are integrated through the multi-modal sensing unit, data is collected based on a unified timestamp, and sensitivity is calibrated in combination with an external environment, so that real-time collection of the running situation of the whole process chain of the oilfield gathering and transferring station is realized. The edge data fusion module performs data normalization processing and abnormality determination, can accurately identify single and composite abnormalities, and provides accurate and comprehensive basic data for subsequent decision-making of the system, thereby improving the accuracy and timeliness of abnormality identification.

[0045] 2. In the application, the gathering and transferring scene simulation and decision selection module is provided, a station simulation model is generated using three-dimensional point cloud data, a compensation value is given in combination with the service life and failure frequency of the equipment by embedding a simulation equation set, an optimal adjustment strategy is generated and selected by inputting an abnormal health vector, and the output decision is more suitable for the actual situation of different oilfield gathering and transferring stations, thereby enhancing the practicality.

[0046] 3. In the application, the decision execution and linkage control module converts the optimal adjustment strategy into a structured control instruction and binds the equipment action logic, collects the adjusted data through the multi-modal sensing unit for abnormality verification to realize closed-loop adjustment control, and repeats the adjustment until the abnormality disappears or notifies manual intervention when the abnormality is not eliminated, thereby ensuring stable operation of the system and reducing the loss caused by the fault. BRIEF DESCRIPTION OF DRAWINGS

[0047] Figure 1This is a block diagram of an intelligent linkage system for oilfield gathering and transportation stations according to the present invention. Detailed Implementation

[0048] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0049] Example 1:

[0050] like Figure 1 As shown, an intelligent linkage system for oilfield gathering and transportation stations includes an oilfield data acquisition module, an edge data fusion module, a gathering and transportation scenario simulation and decision selection module, and a decision execution and linkage control module.

[0051] The oilfield data acquisition module includes a multimodal sensing unit that collects equipment data during the oilfield gathering and transportation process based on a unified timestamp. Simultaneously, it calibrates the sensitivity of the equipment in the multimodal sensing unit by considering the external environment of different oilfield gathering and transportation stations. The collected data is then transmitted to the edge data fusion module. The specific steps are as follows:

[0052] Multimodal sensing units are deployed in different oilfield gathering and transportation stations to collect equipment data during the oilfield gathering and transportation process. The multimodal sensing units integrate temperature sensors, pressure sensors, flow sensors, liquid level sensors, vibration sensors, infrared thermal imaging acquisition probes, visible light camera probes, and gas detection probes.

[0053] Data acquired by the multimodal sensing unit is processed via clock skew. To compensate for errors, a timestamp is added to each collected data point. ,in Represents the clock offset compensation value. These represent the master clock transmission time, slave clock reception time, slave clock reply time, and master clock reception time, respectively. This represents the timestamp time set for the data. This represents the time when the data was collected;

[0054] It should be noted that the multimodal sensing units need to be deployed at key equipment locations such as pump sets, burners, gas gathering pipelines, water mixing nodes, and pressure regulating equipment in different oilfield gathering and transportation stations to achieve real-time acquisition of the operational status of the entire process chain.

[0055] Combining external environmental data from different oilfield gathering and transportation stations, through The sensitivity of the sensors in the multimodal sensing unit is adjusted in a targeted manner, among which... represent the sensor sensitivity adjustment value, represent the current sensor sensitivity reference value, represent the environmental attenuation coefficient and the interference intensity, respectively.

[0056] It should be noted that, among them represent the current sensor sensitivity reference value, i.e. the calibration value of the sensor in the standard laboratory environment, which is usually the factory setting value of the sensor, is the environmental attenuation coefficient, which needs to be obtained by fitting the environmental attenuation curve through testing the sensor in different simulation cabins for different sensors, represent the environmental interference intensity, which is estimated by establishing a mathematical model according to meteorological data and geographic information. The environmental interference intensity is estimated by using meteorological parameters such as wind speed, wind direction, humidity, temperature in weather forecast, combining with the topographic information of the current oilfield gathering and transportation station, such as whether it is close to the desert, mountains and other factors that may affect the distribution of sand or snow, through historical data and machine learning algorithm to train the environmental interference prediction intensity model to estimate the interference intensity of the current environment;

[0057] Embodiment 2:

[0058] The edge data fusion module normalizes the received collected data, outputs the collected data to a uniform format time sequence structure to obtain a health vector, sets edge determination models for different types of collected data in the multi-modal sensor unit, and performs single abnormality determination and composite abnormality recognition on the collected data. The recognition result is transmitted into the subsequent module, including the following steps:

[0059] Through receiving various equipment data collected by the multi-modal sensor unit in the oilfield data collection module Among them represent the sensor type, integrate the collected data into a unified data set, and normalize the data in the data set by Z-score standardization:

[0060] ;

[0061] Among them, represent the data obtained after normalization of the first sensor, represent the historical data sliding mean of the first sensor, represent the standard deviation of the first sensor, align the normalized data in time sequence by dynamic time warping, establish a data statistics library, and encode the data under the same time sequence into the same data set to generate a health vector :

[0062] ;

[0063] wherein, represents a time sequence, represents the first standardized data of the sensor in the time sequence, respectively represent the image feature vector and the gas concentration collected.

[0064] It should be noted that the health vector is a structured feature vector formed by fusing multi-source data, which is used to represent the health state of the object, the image feature vector is a visual feature extracted from the data collected by the image acquisition probe through the convolutional neural network, and the dynamic time warping algorithm DTW is used to match the time sequence to determine the health vector in the same time sequence when the time sequence is aligned.

[0065] For the data collected by the multi-modal sensing unit, the data types are divided, including image data, time sequence data and numerical data, and single abnormality judgment and composite abnormality judgment are performed respectively:

[0066] In the single abnormality judgment process, a convolutional neural network model is constructed for image data to identify abnormal targets in the image, a long short-term memory network model is constructed for time sequence data to identify abnormalities, and a preset threshold is used for abnormality identification based on numerical data;

[0067] It should be noted that in the image data identification process, image data sets containing normal and abnormal targets need to be collected in advance, the image is preprocessed to adjust the image size, ResNet is selected as the CNN architecture, the model is constructed and the parameters and structures of the input layer, convolutional layer, pooling layer, fully connected layer, etc. are determined, the preprocessed image data set is divided into training set and validation set, cross-entropy loss function is used as the loss function, stochastic gradient descent SGD is used as the optimizer, the model is trained using the training set, the loss is calculated by calculating the predicted result and the true label through forward propagation, the model parameters are updated through back propagation, the performance of the model is evaluated using the validation set to calculate the accuracy, recall rate and F1 value, and the model parameters are adjusted according to the evaluation result to finally obtain the determination model.

[0068] When constructing a long short-term memory network model for time sequence data, time sequence data such as temperature, pressure and other data that changes with time is collected, the time sequence data is divided into input sequence and corresponding target sequence for training the model, and for numerical data, the upper and lower threshold values of normal data are determined through the 3sigma principle to identify abnormalities according to the production process indicators and safety specification requirements, combined with business knowledge and historical data statistics.

[0069] ​Based on the single abnormality determination result, the types of each single abnormality under the same time sequence are integrated, and the current abnormality under the same time sequence is confirmed combined with the composite abnormality determination rule. The specific steps are as follows:

[0070] Under the same time sequence, all single abnormality determination results are collected, and the collected determination results are integrated into a determination vector , wherein represents the determination result of different data categories, represents the determination result of data under the time sequence , and is substituted into the output rule to determine the composite abnormality, and the composite abnormality determination result is output , wherein , is the preset composite abnormality determination rule.

[0071] It should be noted that the composite abnormality determination rule needs to be expressed by a logical expression. For example, assuming that the rule is "when abnormality type 1 and abnormality type 2 occur at the same time, determine as composite abnormality type A", then the rule can be expressed as , represents a logical AND operator, and the composite abnormality determination rule can be a combination of multiple logical operators. Different device types need to be combined with historical abnormality data categories to determine the composite abnormality determination rule of different devices, so as to determine whether the device under the current time sequence is abnormal and which type of abnormality occurs.

[0072] Embodiment 3:

[0073] The gathering and transportation scene simulation and decision selection module is based on the environment and device data of different oilfield gathering and transportation stations. The station simulation model is generated through three-dimensional point cloud data, and the simulation equation set is run in the station simulation model. At the same time, according to the historical data, the state of different devices is calibrated and the corresponding compensation value is given in the station simulation model. The abnormal health vector is input into the station simulation model to generate the output decision. The output decision is simulated and the optimal adjustment strategy is selected and output to the subsequent module, including the following steps:

[0074] The service life and failure frequency of different devices in the current oilfield gathering and transportation station are counted, and the device state under different service life and failure frequency is divided according to the historical data. The service life and failure frequency state of the running device in the current gathering and transportation station is determined, and the worst selection is made to obtain the device state. The specific steps are as follows:

[0075] Collect the time of putting into use and the record of failure of different equipment in the current oilfield gathering and transferring station in the historical data, and calculate the service life of different equipment , wherein respectively represent the current time and the time of putting into use of the equipment, and the failure frequency of the equipment in unit time is calculated , wherein represents the number of failures of the equipment in unit time

[0076] The service life and the failure frequency of the same equipment are comprehensively considered to count the service life state of the equipment, including good, normal and general, and the failure frequency state of different equipment is determined according to the high threshold and the low threshold of the failure frequency of the equipment in unit time, and the current equipment state is obtained by selecting the worse one of the service life state and the failure frequency state, and the specific steps are as follows:

[0077] For each equipment, the service life is determined, when ≤ , the service life state of the equipment is good, when < ≤ , the service life state of the equipment is normal, and when > , the service life state of the equipment is general, wherein , represent the low threshold and the high threshold of the service life of the current equipment

[0078] For each equipment, the failure frequency is determined, when ≤ , the failure frequency state of the equipment is good, when < ≤ , the failure frequency state of the equipment is normal, and when > , the failure frequency state of the equipment is general, wherein , represent the low threshold and the high threshold of the failure frequency of the current equipment

[0079] It should be noted that when the failure frequency in unit time is calculated, the unit time ​​Generally set as 3 months, can also be adjusted according to actual use, the low threshold of the service life and the high threshold of the service life need to be calculated by the 3 sigma principle according to the historical failure frequency of the same device at different service life, and the low threshold of the failure frequency and the high threshold of the failure frequency need to be set according to the type of the device, historical data and industry standards combined with experience method.

[0080] The device service life state and the device failure frequency state are selected to obtain the current device state.

[0081] It should be noted that by selecting the device service life state and the device failure frequency state to obtain the current device state, the inaccurate service life state caused by the special failure frequency of the device due to its own problems can be avoided.

[0082] The three-dimensional point cloud data of the current oilfield gathering and transferring station is obtained by three-dimensional laser scanning to construct a three-dimensional simulation model of the station, and according to the physical principle and process flow of the gathering and transferring station, a simulation equation set describing the operation of the station is established, the corresponding parameter compensation value is given to the equipment in the station simulation model according to the equipment state of different equipment, the output decision is generated by inputting the abnormal health vector, the optimal decision is obtained by simulation evaluation and screening, including the following steps:

[0083] The three-dimensional point cloud data of the current oilfield gathering and transferring station is obtained by three-dimensional laser scanning, the point cloud filtering is performed on the three-dimensional point cloud data, the geometric information and topological structure of the station are extracted to construct a three-dimensional simulation model of the station, and according to the physical principle and process flow of the oilfield gathering and transferring station, a simulation equation set describing the operation of the station is established, and the equation set is embedded into the station simulation model;

[0084] It should be noted that when performing point cloud filtering, Gaussian filtering is selected to perform weighted average on the entire image, so that the value of each pixel point is obtained by weighted average of its own and other pixel values in the neighborhood to eliminate noise, wherein the equation set includes oil-water-gas three-phase flow, heat exchange, mixing water balance, pipe network pressure difference, etc. The equation set is solved by numerical calculation method to simulate the running state of the station. Taking the heat conduction equation as an example:

[0085] ;

[0086] Where, T represents temperature, t represents time, x represents spatial coordinate, and β represents thermal diffusion coefficient. The finite difference method is used for numerical solution. The equation is converted into a discrete form that can be iteratively solved by computer through numerical calculation method, and the running state of the station is simulated by point-by-point and time-step-by-time-step calculation.

[0087] Different compensation coefficients are given to different equipment states , the current device parameter compensated set value is calculated , wherein represents the current device set value compensated based on the device state parameter, represents the original parameter of the device;

[0088] It should be noted that the determination of the compensation coefficient needs to be set in combination with different devices, historical operation data of the device under different states is collected, the data is grouped according to the device state, the change of the device parameter in each group of data is analyzed, the change rate of the device parameter under different states relative to the parameter in the normal state, i.e. the good state, is calculated, and the calculated parameter change rate is taken as the compensation coefficient of the current device under different device states.

[0089] The abnormal health vector obtained by the edge data fusion module is input into the station simulation model, the running state of the station is updated in real time by solving the simulation equation set, the preset adjustment rule library of the device adjustment strategy under different abnormal parameters is collected, the running abnormal parameters in the current station simulation model are combined, the cosine similarity is calculated to match the rules in the adjustment rule library, K rounds of rule matching are generated to generate coping strategies, different effect adjustment rules in the adjustment rule library are combined, one adjustment strategy is selected under each adjustment rule through M times of simple and non-repeated random sampling, M kinds of adjustment response paths are generated by combination, and each adjustment response path is simulated in the station simulation model. The adjustment response path with the shortest response time is selected as the optimal adjustment strategy and output to the subsequent module.

[0090] It should be noted that when solving the simulation equation set to update the running state of the station in real time, the model dynamically simulates the transmission and conversion of matter and energy, and the interaction between devices to reflect the real changes of the station under abnormal conditions. K rounds of rule matching are used to cope with complex abnormal conditions, and the value of K is usually set to 5, which can also be adjusted according to the actual situation. During the establishment of the adjustment rule library, data of various abnormal events that have occurred in the past need to be collected from data sources such as the monitoring system of the oilfield gathering and transportation station, device maintenance records, fault reports, etc. These data should include parameter information such as temperature, pressure, flow, liquid level, device state, abnormal type such as device failure, process fluctuation, external interference, and adjustment strategy and final processing result, according to the nature, source and influence degree of the abnormality, the collected abnormal events are classified including device level abnormality such as pump failure, valve blockage, system level abnormality such as process interruption, pressure imbalance and external environment abnormality such as temperature mutation, power failure, etc. For each type of abnormality, the adjustment strategy taken in the historical event is analyzed, and general adjustment rules are summarized according to the abnormal parameter characteristics and the corresponding adjustment strategy;

[0091] At the same time, the effect of the adjustment rule under each adjustment strategy is evaluated, and based on the influence on the operation efficiency, energy consumption and safety of the station yard, the preset rule is reserved through quantitative indexes, the quantitative indexes include abnormal recovery time, energy consumption change rate and equipment damage degree, each quantitative index reserves 2 preset rules, one adjustment strategy is selected under each adjustment rule through M times of simple and non-repeated random sampling, and M adjustment strategies are generated by combination, wherein M is usually set to 7 times.

[0092] The decision execution and linkage control module converts the received optimal adjustment strategy into structured control instructions, and binds the specific device action logic of different oilfield gathering stations to execute the optimal adjustment strategy, and at the same time collects the adjusted device data through the multi-modal sensing unit and performs abnormal verification to realize closed-loop adjustment control, including the following steps:

[0093] According to the layout and equipment configuration information of different oilfield gathering stations, the mapping relationship between the control instructions and the specific devices is established, the corresponding device action logic is bound for each control instruction, the received optimal adjustment strategy is converted into structured control instructions to control the devices, and the structured control instructions are sent to the device control system of the corresponding oilfield gathering station through the industrial protocol adaptation module for instruction execution.

[0094] The multi-modal sensing unit in the oilfield data acquisition module collects the oilfield gathering station data after instruction execution, and performs abnormal identification in the edge data fusion module, and when the abnormality still exists, the optimal adjustment strategy is generated again through the gathering scene simulation and decision selection module for abnormal adjustment, and the above steps are repeated until the abnormality disappears or the maximum number of cycles is reached, and the abnormal report is transmitted to the on-duty personnel.

[0095] It should be noted that the maximum number of cycles is usually set to 3 times, and when the abnormality cannot be eliminated after 3 cycles, the abnormal report is transmitted to the on-duty personnel to inform the on-duty personnel to intervene manually, so as to avoid further expansion of the loss.

[0096] The application discloses an intelligent linkage device for an oilfield gathering station, which is applied to an intelligent linkage system of the oilfield gathering station and comprises a memory, a processor and a program stored in the memory and capable of running on the processor.

[0097] The application realizes real-time collection of the operation situation of the whole process chain of the oilfield gathering and transferring station by setting the multi-modal sensing unit to integrate various sensors, collecting data based on a unified timestamp and combining external environment to calibrate sensitivity, the edge data fusion module performs data normalization processing and abnormality determination, can accurately identify single and composite abnormalities, provides accurate and comprehensive basic data for subsequent decision-making of the system, improves the accuracy and timeliness of abnormality identification, sets the gathering and transferring scene simulation and decision selection module, generates a station field simulation model using three-dimensional point cloud data, embeds simulation equation sets and gives compensation values in combination with the service life and failure frequency of equipment, inputs abnormal health vectors to generate and screen optimal adjustment strategies, makes the output decision more suitable for the actual situation of different oilfield gathering and transferring stations, and enhances the practicality.

[0098] In the embodiments provided by the present application, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic, for example, the division of modules is merely a logical function division, and an alternative division manner can be used in actual implementation; the modules illustrated as separate components can or can not be physically separate, and the components illustrated as modules can or can not be physical units, that is, they can be located in one place or distributed on a plurality of network units. According to actual needs, some or all of the modules can be selected to achieve the purpose of the embodiments of the present application.

[0099] The above embodiments are only used to illustrate the technical method of the present application but not limit the present application. Although the present application is described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical method of the present application can be modified or replaced equivalently without departing from the spirit and scope of the technical method of the present application.

Claims

1. An intelligent linkage system for an oilfield gathering and processing station, the system comprising: The oil field data acquisition module, the edge data fusion module, the gathering and transportation scene simulation and decision selection module, and the decision execution and linkage control module are included. The oil field data acquisition module includes a multi-modal sensing unit, which acquires equipment data during the oil field gathering and transportation process based on a unified timestamp, and calibrates the sensitivity of the equipment in the multi-modal sensing unit in combination with the external environment of different oil field gathering and transportation stations, and transmits the acquired data into the edge data fusion module. The edge data fusion module normalizes the received acquisition data, outputs the acquisition data into a unified format time sequence structure to obtain a health vector, sets edge decision models for different types of acquisition data in the multi-modal sensing unit, performs single abnormality determination and composite abnormality recognition on the acquisition data, and transmits the recognition results into subsequent modules. The edge data fusion module normalizes the received acquisition data, outputs the acquisition data into a unified format time sequence structure to obtain a health vector, sets edge decision models for different types of acquisition data in the multi-modal sensing unit, performs single abnormality determination and composite abnormality recognition on the acquisition data, and transmits the recognition results into subsequent modules. The data collected by the multi-modal sensing unit is divided into different types, including image data, time sequence data, and numerical data, and single abnormality determination and composite abnormality determination are performed on the data. In the single abnormality determination process, a convolutional neural network model is constructed for image data to recognize abnormal targets in the image, a long short-term memory network model is constructed for time sequence data to recognize abnormalities, and a preset threshold is used for numerical data to recognize abnormalities. Based on the single abnormality determination result, the types of each single abnormality under the same time sequence are integrated, and the current abnormality under the same time sequence is confirmed in combination with the composite abnormality determination rule. In the same time sequence, all single abnormality determination results are collected, and the collected determination results are integrated into a determination vector wherein the determination results represent different data categories, the determination results represent a collection of data determination results in a time sequence of , the collection of data determination results is substituted into an output rule to perform composite abnormality determination, and a composite abnormality determination result is output wherein , is a preset composite abnormality determination rule;​ The gathering and transportation scene simulation and decision selection module generates a station simulation model based on the environment and equipment data of different oil field gathering stations through three-dimensional point cloud data, runs simulation equation sets in the station simulation model, combines the service life and fault frequency of the equipment, calibrates the state of different equipment according to historical data, and assigns corresponding compensation values in the station simulation model, inputs the abnormal health vector into the station simulation model to generate an output decision, simulates the output decision, selects the optimal adjustment strategy, and outputs the optimal adjustment strategy to the subsequent module. The decision execution and linkage control module converts the received optimal adjustment strategy into a structured control instruction, binds the specific equipment action logic of different oil field gathering stations to execute the optimal adjustment strategy, and collects the adjusted equipment data through the multi-modal sensing unit and performs abnormality verification to realize closed-loop adjustment control.

2. The intelligent linkage system for an oilfield gathering and transferring station according to claim 1, characterized in that: The oil field data acquisition module includes a multi-modal sensing unit, which acquires equipment data during the oil field gathering and transportation process based on a unified timestamp, and calibrates the sensitivity of the equipment in the multi-modal sensing unit in combination with the external environment of different oil field gathering and transportation stations, and transmits the acquired data into the edge data fusion module. By deploying multi-modal sensing units in different oilfield gathering stations to collect equipment data during oilfield gathering, wherein the multi-modal sensing units integrate temperature sensors, pressure sensors, flow sensors, liquid level sensors, vibration sensors, infrared thermal imaging acquisition probes, visible light camera probes, and gas detection probes; Compensation of clock offset for data collected by multi-modal sensor units Compensation of errors and time stamping of each collected data wherein represents the clock offset compensation value, respectively represents the master clock sending time, the slave clock receiving time, the slave clock reply time, the master clock receiving time, represents the time stamp time set for the data, represents the time when the data is collected; In combination with external environment data of different oilfield gathering and transferring stations, the sensitivity of the sensor in the multi-modal sensing unit is adjusted sensitivity of the sensor in the multi-modal sensing unit is adjusted, representing a sensor sensitivity adjustment value, representing a current sensor sensitivity reference value, representing an environment attenuation coefficient and an interference intensity, respectively.

3. An intelligent linkage system for an oilfield gathering and processing station according to claim 2, characterized in that: The edge data fusion module normalizes the received collected data, outputs the collected data into a uniform format time sequence structure to obtain a health vector, including the following steps: Various equipment data collected by the multi-modal sensing unit in the oilfield data acquisition module is received wherein The collected data is integrated into a unified data set, and the data in the data set is normalized by Z-score standardization: ； wherein, representing the first class sensor historical data sliding mean, representing the first class sensor historical data sliding mean, representing the first class sensor standard deviation, the normalized data is time-aligned by dynamic time warping, establishing a data statistics library, data under the same time sequence is coded into the same data set to generate a health vector : ； wherein, represent a time sequence, represent a first sensor in standardized data under a time sequence, respectively represent an image feature vector and a gas concentration.

4. The intelligent linkage system for an oilfield gathering and transferring station of claim 2, characterized in that: The gathering scene simulation and decision selection module generates a station simulation model based on the environment and equipment data of different oilfield gathering stations through three-dimensional point cloud data, runs simulation equation sets in the station simulation model, combines the service life and failure frequency of the equipment, calibrates the state of different equipment according to historical data, and gives corresponding compensation values in the station simulation model, inputs the abnormal health vector into the station simulation model to generate an output decision, simulates and filters the optimal adjustment strategy and outputs it to the subsequent module, including the following steps: The service life and failure frequency of different equipment in the current oilfield gathering station are counted, the equipment state under different service life and failure frequency is divided combined with historical data, the service life and failure frequency state of the running equipment in the current gathering station is determined, and the worst selection of the determination result is obtained to obtain the equipment state. Three-dimensional point cloud data of the current oilfield gathering station is obtained by three-dimensional laser scanning to construct a three-dimensional simulation model of the station, and simulation equation sets describing the operation of the station are established according to the physical principles and process flow of the gathering station. According to the equipment state of different equipment, the corresponding parameter compensation value is given to the equipment in the station simulation model, the abnormal health vector is input to generate an output decision, and the optimal decision is obtained by simulation evaluation and screening.

5. An intelligent linkage system for an oilfield gathering and processing station according to claim 4, characterized in that: The service life and failure frequency of different equipment in the current oilfield gathering station are counted, the equipment state under different service life and failure frequency is divided combined with historical data, the service life and failure frequency state of the running equipment in the current gathering station is determined, and the worst selection of the determination result is obtained to obtain the equipment state. Collect the time of putting into use and the fault record of different equipment in the current oilfield gathering and transferring station in the historical data, and calculate the service life of different equipment , wherein respectively represent the current time and the time of putting into use of the equipment, and the frequency of the fault of the equipment in unit time is calculated , wherein represents the number of times of the fault of the equipment in unit time ​​ The service life and failure frequency of different equipment in the current oilfield gathering station are counted, the equipment state under different service life and failure frequency is divided combined with historical data, the service life and failure frequency state of the running equipment in the current gathering station is determined, and the worst selection of the determination result is obtained to obtain the equipment state. For each device, a decision is made based on its age when ≤ the age status of the device is good, when < ≤ the age status of the device is normal, and when > the age status of the device is average, where , represent the low and high age thresholds for the current device. For each device, a decision is made based on its failure occurrence frequency when ≤ then the device failure frequency status is good, when < ≤ then the device failure frequency status is normal, and when > then the device failure frequency status is average, where , represent the low and high thresholds for the failure occurrence frequency of the current device. The service life and failure frequency of different equipment in the current oilfield gathering station are counted, the equipment state under different service life and failure frequency is divided combined with historical data, the service life and failure frequency state of the running equipment in the current gathering station is determined, and the worst selection of the determination result is obtained to obtain the equipment state.

6. An intelligent linkage system for an oilfield gathering and processing station according to claim 5, characterized in that: The three-dimensional point cloud data of the current oilfield gathering and transferring station is obtained by three-dimensional laser scanning to construct a three-dimensional simulation model of the station, and according to the physical principle and process flow of the gathering and transferring station, a simulation equation set describing the operation of the station is established, the corresponding parameter compensation value is given to the equipment in the station simulation model according to the equipment state of different equipment, the output decision is generated by inputting the abnormal health vector, the optimal decision is obtained by simulation evaluation and screening, including the following steps: The three-dimensional point cloud data of the current oilfield gathering and transferring station is obtained by three-dimensional laser scanning, the point cloud filtering is performed on the three-dimensional point cloud data, the geometric information and topological structure of the station are extracted to construct a three-dimensional simulation model of the station, and according to the physical principle and process flow of the oilfield gathering and transferring station, a simulation equation set describing the operation of the station is established and embedded into the station simulation model; Different compensation coefficients are given to the device according to different device states The set value compensated by the current device parameter is calculated Wherein The set value compensated by the current device based on the device state parameter, The original parameter of the device; The abnormal health vector obtained by the edge data fusion module is input into the station simulation model, the running state of the station is updated in real time by solving the simulation equation set, the equipment adjustment strategy preset adjustment rule library under different abnormal parameters is collected, the rules in the adjustment rule library are matched by calculating the cosine similarity according to the running abnormal parameters in the current station simulation model, K rounds of rule matching are performed to generate response strategies, different adjustment rules in the adjustment rule library are combined, and one adjustment strategy is selected under each adjustment rule by M times of simple and non-repeated random sampling to generate M kinds of adjustment response paths, and each adjustment response path is simulated in the station simulation model, the adjustment response path with the shortest response time is selected as the optimal adjustment strategy and output to the subsequent module.

7. An intelligent linkage system for an oilfield gathering and processing station according to claim 6, characterized in that: The decision execution and linkage control module converts the received optimal adjustment strategy into structured control instructions, binds the specific equipment action logic of different oilfield gathering and transferring stations to execute the optimal adjustment strategy, and collects the adjusted equipment data through the multi-modal sensing unit and performs abnormal verification to realize closed-loop adjustment control, including the following steps: According to the layout and equipment configuration information of different oilfield gathering and transferring stations, the mapping relationship between the control instructions and the specific equipment is established, the corresponding equipment action logic is bound for each control instruction, the received optimal adjustment strategy is converted into structured control instructions to control the equipment, and the structured control instructions are sent to the equipment control system of the corresponding oilfield gathering and transferring station through the industrial protocol adaptation module for instruction execution; The oilfield gathering and transferring station data after instruction execution is collected by the multi-modal sensing unit in the oilfield data acquisition module, abnormal identification is performed in the edge data fusion module, and when the abnormality still exists, the optimal adjustment strategy is generated again by the gathering and transferring scene simulation and decision selection module for abnormal adjustment, the above steps are repeated until the abnormality disappears or the maximum number of cycles is reached, and the abnormal report is transmitted to the on-duty personnel.

8. An intelligent linkage for use in an oilfield gathering and processing station, the linkage comprising: The device is applied to an intelligent linkage system of an oilfield gathering and transferring station, and includes a memory and a processor: The memory is used for non-transitory storage of computer readable instructions; The processor is used for running the computer readable instructions; When the computer readable instructions are run by the processor, the steps in the system of any one of claims 1-7 are executed.

Images