Intelligent early warning method and system for offshore fracturing sand blockage

Abstract

The invention provides an intelligent early warning method and system for offshore fracturing sand blockage, and relates to the technical field of offshore oil and gas exploitation, and the method comprises the steps: collecting real-time data of offshore fracturing construction equipment; performing standardization processing on the data, and training PGNN and PINN models based on the processed data; storing the processed data in a regional database, and merging the processed data into a data lake; and the PGNN and PINN models are adopted to predict the construction pressure change in real time, and sand blocking early warning is carried out when the pressure prediction value is abnormal. The system comprises a data acquisition module, a data processing and model training module, a database establishment module and a sand blockage intelligent early warning module. According to the method, misjudgment and missed judgment caused by human factors are reduced, the scientificity of sand blockage judgment is improved, sand blockage early warning can be carried out in advance, a fracturing site can be helped to take measures in time, the site sand blockage risk and the operation cost are effectively reduced, and the construction safety is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of offshore oil and gas production, and in particular to an intelligent early warning method and system for offshore fracturing sand plugging. Background Art

[0002] The extraction of offshore oil and gas resources is an important component of global energy supply. In the process of offshore oil and gas extraction, fracturing is an important production-increasing technology. It increases the flow efficiency of oil and gas by artificially creating cracks in the oil and gas layer.

[0003] However, during offshore fracturing construction, sand plugging may occur, that is, the accumulation of proppant in formation cracks, perforations or wellbores blocks the fracturing channel, causing an increase in construction pressure. Sand plugging is an important factor affecting construction quality, effect and fracturing efficiency. It not only increases operating costs, but in severe cases may lead to fracturing failure.

[0004] Traditional offshore fracturing monitoring methods rely on technicians to predict sand plugging based on pressure trends. This approach has the following problems:

[0005] 1. Low monitoring efficiency, making it difficult to capture early signals of sand plugging in real time;

[0006] 2. Lack of accuracy and susceptibility to human factors, leading to misjudgment or omission;

[0007] 3. Insufficient data utilization, and failure to fully utilize the large amount of data collected on-site for in-depth analysis.

[0008] Therefore, there is an urgent need for an intelligent early warning method and system for offshore fracturing sand plugging to improve the safety and efficiency of offshore fracturing operations. Summary of the Invention

[0009] The purpose of the present invention is to provide an intelligent early warning method and system for offshore fracturing sand plugging, which can automatically collect and transmit construction data, analyze the construction data using advanced data processing and machine learning technologies, thereby accurately predicting and promptly warning of the risk of sand plugging during the fracturing process, and improving the safety and efficiency of offshore fracturing operations.

[0010] The various technical effects that can be produced by the preferred technical solutions among the various technical solutions provided by the present invention are described in detail below.

[0011] To achieve the above objectives, the present invention provides the following technical solutions:

[0012] The present invention provides an intelligent early warning method for offshore fracturing sand plugging, comprising:

[0013] Collect real-time data from offshore fracturing equipment;

[0014] Standardizing the data and training PGNN and PINN models based on the processed data;

[0015] Store the processed data in regional databases and merge them into the data lake;

[0016] The PGNN and PINN models are used to predict construction pressure changes in real time, and sand blockage warnings are issued when the predicted pressure values ​​are abnormal.

[0017] Preferably, the data is normalized, and the PGNN and PINN models are trained based on the normalized data, including:

[0018] Data standardization: remove samples with missing values ​​and standardize the data set;

[0019] Training set and test set division, randomly divide the standardized data set into training set and test set;

[0020] PGNN model training, training machine learning models based on physical constraint loss functions;

[0021] PINN model training, the model is trained by the principle of structural risk minimization.

[0022] Preferably, in the division of the training set and the test set, the entire standardized data set is randomly divided into the training set and the test set according to the ratio of 75% and 25%.

[0023] Preferably, the data set is standardized and the calculation formula is:

[0024]

[0025]

[0026] Where: z i The standardized value of the i-th sample; x i The value of the i-th sample before normalization; μ is the mean of all samples; σ is the standard deviation of all samples; n is the number of samples.

[0027] Preferably, the PGNN and PINN models are used to predict construction pressure changes in real time, and to issue sand blockage warnings when the predicted pressure values ​​are abnormal, including:

[0028] When the pressure prediction value is less than the sum of the data mean and 3 times the standard deviation, it is considered normal;

[0029] When the predicted pressure value rises sharply and the slope of the pressure curve changes, it is considered an abnormal situation.

[0030] Preferably, the processed data is stored in a regional database and merged into the data lake, including:

[0031] Collection and transmission: The processed data is transmitted to the DMZ network through the data acquisition room communication network or 4G / VPN;

[0032] Data storage, transferring data in the DMZ network to the regional database;

[0033] Data merging: Merge data from regional databases into the data lake.

[0034] Preferably, the collecting of real-time data of offshore fracturing equipment includes:

[0035] The real-time data of the offshore fracturing construction equipment is transmitted to the data acquisition instrument through the serial port or network port, wherein the network port transmission adopts the WITS protocol or TCP / UDP protocol.

[0036] Preferably, the real-time data collected from offshore fracturing construction equipment includes: construction well number, construction time, oil pressure, casing pressure, displacement, liquid addition displacement, stage liquid volume, cumulative liquid volume, sand concentration, and implementation data of sand volume.

[0037] An intelligent early warning system for offshore fracturing sand plugging, the system comprising:

[0038] Data acquisition module, used to collect real-time data of offshore fracturing construction equipment;

[0039] A data processing and model training module is used to standardize the data and train PGNN and PINN models based on the processed data;

[0040] Database creation module, used to store processed data in regional databases and merge them into the data lake;

[0041] The sand plugging intelligent early warning module is used to predict construction pressure changes in real time using the PGNN and PINN models, and to issue a sand plugging warning when the predicted pressure value is abnormal.

[0042] Preferably, the system further comprises a user interface for displaying real-time data and warning information.

[0043] The present invention provides an intelligent early warning method for offshore fracturing sand plugging, comprising collecting real-time data of offshore fracturing construction equipment; standardizing the data, and training PGNN and PINN models based on the processed data; storing the processed data in a regional database and merging it into a data lake; using PGNN and PINN models to predict construction pressure changes in real time, and issuing a sand plugging warning when the pressure prediction value is abnormal. By analyzing the characteristics of historical sand plugging curves of offshore fracturing, based on machine learning and physical models, and utilizing real-time construction data and predicted construction pressure, an intelligent early warning method for offshore fracturing sand plugging is proposed, which reduces misjudgments and missed judgments caused by human factors and improves the scientific nature of sand plugging judgments. The method can provide early warning of sand plugging, effectively reduce the risk of sand plugging and operating costs on site, and improve construction safety. Through early warning of sand plugging, it can help the fracturing site take timely measures to avoid or reduce the impact of sand plugging problems on construction quality and efficiency, thereby improving fracturing efficiency and reducing operating costs.

[0044] The offshore fracturing sand plugging intelligent early warning system provided by the present invention includes a data acquisition module, a data processing and model training module, a database establishment module and a sand plugging intelligent early warning module. Among them, the data acquisition module can collect fracturing on-site construction data in real time, and transmit the data to the data processing center through efficient data transmission technology, monitor and analyze the pressure changes during the construction process in real time, and promptly discover potential sand plugging risks. The collected data is standardized through the data processing and model training module to predict future construction pressure change trends with high prediction accuracy. By setting up a database establishment module, not only real-time monitoring and early warning can be achieved, but the collected data can also be stored in the database, providing a rich data foundation for subsequent data analysis, machine learning model training and technical iterative updates. In addition, this system design takes into account compatibility with existing offshore oil and gas production equipment and is easy to integrate and deploy. At the same time, the modular design of the system is also convenient for future technology upgrades and functional expansions. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0046] Figure 1 This is the process of the intelligent early warning method for offshore fracturing sand plugging of the present invention Figure 1 ;

[0047] Figure 2 This is the process of the intelligent early warning method for offshore fracturing sand plugging of the present invention Figure 2 ;

[0048] Figure 3 It is a schematic diagram of the intelligent early warning system for offshore fracturing sand plugging according to the present invention. DETAILED DESCRIPTION

[0049] To make the objectives, technical solutions, and advantages of the present invention more apparent, the technical solutions of the present invention will be described in detail below. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other implementations obtained by those of ordinary skill in the art without inventive effort are within the scope of protection of the present invention.

[0050] Figure 1 FIG. 1 is a flow chart of the intelligent early warning method for offshore fracturing sand plugging according to this embodiment. Figure 1 As shown, this embodiment provides an intelligent early warning method for offshore fracturing sand plugging, aiming to improve the safety and efficiency of offshore fracturing operations. The method mainly includes:

[0051] S1: Collect real-time data of offshore fracturing equipment;

[0052] S2: normalizing the data and training the PGNN and PINN models based on the processed data;

[0053] S3: Stores the processed data in regional databases and merges them into the data lake;

[0054] S4: Use PGNN (physics-driven neural network) and PINN (physical information neural network) models to predict construction pressure changes in real time and issue sand blockage warnings when the pressure prediction value is abnormal.

[0055] Through data collection, data processing and model training, database establishment and sand plugging early warning, it is used to achieve real-time monitoring and early warning of sand plugging during offshore fracturing construction, thereby improving the quality of offshore fracturing construction and the effect of oil and gas reservoir development.

[0056] like Figure 2 As shown, in this embodiment, real-time data of offshore fracturing equipment is collected, which specifically includes the following steps:

[0057] S1: Data collection, including:

[0058] S11: Determine key data.

[0059] In this embodiment, the key data types determined for construction collection include at least: construction well number, construction time, oil pressure MPa, casing pressure MPa, displacement m 3 min -1 , Liquid addition displacement m 3 min -1, stage liquid volume m 3 , cumulative liquid volume m 3 , sand concentration kg / m 3 , sand quantity t.

[0060] S12: Monitor and record data.

[0061] In this embodiment, pressure data is monitored and recorded by a pressure sensor; liquid volume, displacement and sand volume are monitored and recorded by a flow meter; and sand concentration data is monitored and recorded by a mixing wheel on a sand mixer. Combined with the above, an intelligent early warning method and system for offshore fracturing sand blockage is implemented to collect data.

[0062] S2: Data transmission and storage, including real-time data transmission and storage from the on-site fracturing equipment during the fracturing operation. In this embodiment, the on-site fracturing equipment transmits data to the data acquisition instrument via a serial port or network port. Furthermore, the network port transmission adopts the WITS protocol or the TCP / UDP protocol.

[0063] Serial transmission is a computer hardware interface used to achieve data communication between devices. Its simplicity and cost-effectiveness make it very important in the field of petroleum engineering.

[0064] Ethernet transmission is a method of data transmission via a network connection, which makes data exchange between devices fast, reliable and efficient.

[0065] The WITS protocol is a communication protocol for well integrity and tubing services. It is commonly used in oilfield automation and monitoring systems to ensure safe and efficient well operations. The protocol allows data exchange and communication between different devices and systems, enabling remote monitoring, control, and diagnostics.

[0066] TCP (Transmission Control Protocol) is a connection-oriented, reliable, byte-stream-based transport layer communication protocol, and UDP (User Datagram Protocol) is a connectionless, unreliable transport layer protocol. The choice of these two commonly used network transport layer protocols depends on the application's needs. If the application requires reliable, ordered data transmission, choose TCP; if the application requires fast, real-time data transmission and can tolerate a certain degree of data loss, choose UDP.

[0067] The real-time data collected by the fracturing equipment is transmitted to the data acquisition instrument in the above manner.

[0068] In terms of data acquisition and transmission, this embodiment achieves efficient transmission of real-time data from fracturing equipment through serial and Ethernet transmission technologies. The Ethernet transmission uses the WITS protocol or TCP / UDP protocol to ensure fast and reliable data transmission.

[0069] S2: Standardize the data and train the PGNN and PINN models based on the processed data, including:

[0070] S21: Standardize the data acquired by the data acquisition device, delete samples with missing values ​​and standardize the data set;

[0071] In this embodiment, in order to ensure the representativeness of the samples, samples with missing values ​​are deleted.

[0072] In addition, in order to eliminate the influence of unit differences between features so that each feature is treated equally, the data of the feature set is standardized so that the mean value of each dimension feature is 0 and the standard deviation is 1. The calculation formula is:

[0073]

[0074] Where: zi is the value of the i-th sample after standardization; xi is the value of the i-th sample before standardization; μ is the mean of all samples; σ is the standard deviation of all samples; n is the number of samples.

[0075] S22: training set and test set division, randomly divide the standardized data set into training set and test set;

[0076] For the dataset, the entire dataset is randomly divided into a training set and a test set according to the ratio of 75% and 25%.

[0077] The training set is used to train and optimize the models of each algorithm, while the test set is used to evaluate the accuracy of the model.

[0078] To ensure the fairness of the evaluation, the test set does not participate in the training of any model throughout the process and is only used to evaluate model performance.

[0079] S23: Train the PGNN and PINN models based on the processed data, wherein the PGNN model training trains the machine learning model based on the loss function of physical constraints; the PINN model training trains the model through the principle of structural risk minimization.

[0080] Specifically, if the PGNN constructs an accurate physical driving model that is very close to the true value, then the HPD modeling method can enable the PINN model to learn the correspondence between data and physical knowledge; if the physical model is inaccurate and there is a deviation from the true value, the HPD modeling method can also extract features from the driving data to supplement the learning.

[0081] It also includes extracting sufficient historical fracturing data from the regional data set and the data lake, and standardizing the historical fracturing data according to step S21.

[0082] The processed historical fracturing data is input into the PGNN model for training. In essence, it is a semi-supervised deep learning method that combines machine learning with physical methods. A loss function based on physical constraints is used to train the machine learning model to ensure the consistency of the learning process with physical knowledge.

[0083] In this example, data processing and model training are first standardized, including removing missing value samples and normalizing features to ensure data representativeness and consistency. Subsequently, the data is divided into training and test sets for training and evaluating the PGNN and PINN models. The PGNN model combines physics-driven and data-driven approaches, while the PINN model addresses overfitting and improves model generalization through the principle of structural risk minimization.

[0084] S3: Stores processed data in regional databases and merges it into the data lake, including:

[0085] S31: Collect and transmit the real-time data from the fracturing site to the DMZ network via the data acquisition room communication network or 4G / VPN;

[0086] The data acquisition room communication network in this embodiment refers to the network system used for data transmission and communication in the data acquisition room. In the field of fracturing oil and gas development involved in this embodiment, the data acquisition room communication network is a key infrastructure responsible for collecting, processing, and transmitting data from various sensors and equipment.

[0087] The combination of 4G and VPN technology can ensure that fracturing data is not eavesdropped or tampered with during transmission. 4G technology, short for fourth-generation mobile communications, is a wireless communication standard primarily used to enhance the data transmission capabilities of mobile communication networks. VPN is a technology commonly used to improve network security and privacy, establishing an encrypted, secure connection over public networks.

[0088] Data is transferred to the regional database via the DMZ network. In the field of computer network security, DMZ refers to a specific network area that is located between the internal trusted network and the external untrusted network, protecting the internal network where the fracturing data is transmitted from unauthorized access.

[0089] S32: Data storage, transferring data in the DMZ network to the regional database;

[0090] In this example, a regional database refers to a database system serving a specific geographic region or organization. This database is designed to store, manage, and retrieve data related to a specific region to support various business and operational activities within that region. Regional databases can be deployed locally or in the cloud.

[0091] S33: Data merging: Merging data from regional databases into a data lake. A data lake is a large, diverse data repository that stores raw data from various sources. It supports big data analysis, machine learning, data mining, and other applications. This approach implements an intelligent early warning method and system for offshore fracturing sand plugging.

[0092] In this implementation, real-time data from the fracturing site is transmitted to the DMZ network via the data acquisition room communication network or 4G / VPN. This data is then transferred to the regional database and ultimately merged into the data lake. This process not only ensures data security but also supports subsequent data analysis and construction decision-making.

[0093] S4: Use the PGNN and PINN models to predict construction pressure changes in real time and issue sand plugging warnings when the predicted pressure value is abnormal, including:

[0094] When the pressure prediction value is less than the sum of the data mean and three times the standard deviation, the model-predicted pressure is considered to meet the physical laws and is considered normal;

[0095] When the predicted pressure value rises sharply and the slope of the pressure curve changes, it is considered to violate the basic physical laws and is regarded as an abnormal situation.

[0096]

[0097] f(P)-f(μ+3σ)<0

[0098]

[0099] Where, P i is the fracturing operation data; μ is the data mean; n is the data size; f(P) is the pressure prediction value; f(μ+3σ) is the data mean and 3 times the standard deviation; is the slope of the construction curve; K is the slope of the prediction curve.

[0100] This embodiment uses the PINN model for sand plugging early warning, which is characterized by converting the machine learning problem into an optimization problem, namely, the empirical risk minimization criterion.

[0101] When the sample size is large enough, empirical risk minimization can achieve excellent learning results. However, in most practical situations, it is not possible to obtain an unlimited number of training samples, and the training samples are often only a small subset of the real data or contain a certain amount of noise, which does not fully reflect the true distribution of the entire data. In this case, using the principle of empirical risk minimization to train the model can easily lead to a very low error rate on the training set, but a high error rate on unknown data, a phenomenon known as overfitting. To address this overfitting problem, parameter regularization is generally introduced on top of empirical risk minimization to limit the model's ability to minimize the empirical risk excessively. This learning principle is called structural risk minimization.

[0102] In actual operation, in addition to inputting the standardized fracturing construction site data set into the PINN model, the results of the PGNN model must also be used as an input to the PINN model.

[0103] In this embodiment, the PINN model uses the ReLU function as the activation function to perform nonlinear transformation on the input value to increase the expressive power of the model. Two modifications are made to the function:

[0104] ReLU(x)=max(0,x)

[0105] Where ReLU is the activation function; if the input value x is positive, the output value is x; if x is negative, the output value is 0.

[0106] The physical inconsistency of all pairs of points needs to be considered, not just a single pair, so the values ​​of all pairs can be averaged.

[0107] Minimizing the weight of the physical inconsistency penalty term is also critical, which can be done by multiplying the average physical inconsistency term by the hyperparameter. The physical loss function is:

[0108]

[0109] Where PHY·Loss is the physical loss function; n t Field dataset size; n l The size of the historical dataset.

[0110] The PINN model data is returned to the fracturing site to predict changes in operating pressure approximately 60 seconds into the future. If the slope of the curve changes dramatically and the pressure rises suddenly, it is determined that sand plugging is about to occur.

[0111] The fracturing field data is transmitted to the data acquisition instrument, and the data is standardized according to S21. The processed field real-time data is sent to the PINN model for training to obtain an intelligent early warning model for offshore fracturing sand plugging based on the PINN neural network.

[0112] The PINN sand plugging intelligent early warning model predicts construction pressure approximately 60 seconds into the future. If the predicted pressure is less than the sum of the data mean and three times the standard deviation, it is considered normal. If the predicted pressure rises sharply and the slope of the pressure curve changes, it is considered abnormal, indicating a warning of sand plugging.

[0113] During operation, real-time data and early warning information can also be displayed through the user interface of the monitoring system, so that monitoring personnel can understand the construction status in a timely manner and take corresponding measures.

[0114] This intelligent early warning method for offshore fracturing sand plugging can predict in real time the construction pressure changes in the next 60 seconds or so, and by monitoring the changes in the slope of the pressure curve, timely warn of the occurrence of sand plugging, thereby providing an effective real-time monitoring and early warning method for offshore fracturing construction, which can improve the quality of offshore fracturing construction and the effectiveness of oil and gas reservoir development.

[0115] Figure 3 It is a structural diagram of this embodiment, as shown in FIG. Figure 3 As shown, this embodiment also provides an intelligent early warning system for offshore fracturing sand plugging, which includes a data acquisition module, a data processing and model training module, a database establishment module and a sand plugging intelligent early warning module.

[0116] The data acquisition module is used to collect real-time data from offshore fracturing equipment, including:

[0117] Serial transmission interface uses computer hardware interface, such as RS-232 or USB, to achieve data communication between devices.

[0118] Network transmission interface: Through the Ethernet interface, using the WITS protocol or 4TCP / UDP protocol, it is used to achieve fast and reliable data transmission through the network.

[0119] The database creation module is used to store processed data in the regional database and merge it into the data lake, including:

[0120] Data transmission unit, where data is securely transmitted to the DMZ network via the data acquisition room communication network or 4G / VPN;

[0121] The data merging unit is used to merge data in the DMZ network into the regional database and store the data uniformly in the data lake.

[0122] The data processing and model training module is used to standardize the data and train the PGNN and PINN models based on the processed data, including:

[0123] Data standardization processing unit, used to delete samples with missing values ​​and standardize feature set data;

[0124] The training set and test set division unit is used to randomly divide the data set into a training set and a test set;

[0125] PGNN model training unit, used to train machine learning models based on physical constraint loss functions;

[0126] PINN model training unit, used to train the model based on the structural risk minimization principle.

[0127] Specifically, in this embodiment, the collected data is transmitted to the data processing center. In the data processing and model training module, the data is first preprocessed, including removing samples with missing values ​​and normalizing the feature set data. Next, the dataset is randomly divided into a training set and a test set with a ratio of 75% and 25%.

[0128] The PGNN and PINN models are trained using the training set. The PGNN model combines physical knowledge and data-driven methods, while the PINN model avoids overfitting by minimizing structural risk.

[0129] The sand plugging intelligent early warning module is used to predict construction pressure changes in real time using the PGNN and PINN models, and to issue a sand plugging warning when the predicted pressure value is abnormal.

[0130] Specifically, during offshore fracturing operations, data acquisition instruments and network transmission signals transmit construction data in real time to a real-time database server at the base. The PINN sand plugging intelligent early warning model then predicts pressure changes over the next 60 seconds. If the predicted pressure is less than the sum of the data mean and three standard deviations, it's considered normal. If the predicted pressure rises sharply and the slope of the pressure curve changes, it's considered abnormal, signaling a warning of sand plugging.

[0131] Optionally, the offshore fracturing sand plugging intelligent early warning system in this embodiment further includes a user interface for displaying real-time data and early warning information, so that construction personnel can understand the construction status in a timely manner and take corresponding measures.

[0132] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. An intelligent early warning method for offshore fracturing sand plugging, characterized in that: include: Collect real-time data from offshore fracturing equipment; Standardizing the data and training PGNN and PINN models based on the processed data; Store the processed data in regional databases and merge them into the data lake; The PGNN and PINN models are used to predict construction pressure changes in real time, and sand blockage warnings are issued when the predicted pressure values ​​are abnormal.

2. The intelligent early warning method for offshore fracturing sand plugging according to claim 1 is characterized in that: The data is standardized and the PGNN and PINN models are trained based on the processed data, including: Data standardization: remove samples with missing values ​​and standardize the data set; Training set and test set division, randomly divide the standardized data set into training set and test set; PGNN model training, training machine learning models based on physical constraint loss functions; PINN model training, the model is trained by the principle of structural risk minimization.

3. The intelligent early warning method for offshore fracturing sand plugging according to claim 2, characterized in that: In the division of the training set and the test set, the entire standardized data set is randomly divided into the training set and the test set according to the ratio of 75% and 25%.

4. The intelligent early warning method for offshore fracturing sand plugging according to any one of claims 1 to 3, characterized in that: The data set is standardized and the calculation formula is: Where: z i The standardized value of the i-th sample; x i The value of the i-th sample before normalization; μ is the mean of all samples; σ is the standard deviation of all samples; n is the number of samples.

5. The intelligent early warning method for offshore fracturing sand plugging according to claim 4 is characterized by: Use the PGNN and PINN models to predict construction pressure changes in real time and issue sand blockage warnings when the predicted pressure value is abnormal, including: When the pressure prediction value is less than the sum of the data mean and 3 times the standard deviation, it is considered normal; When the predicted pressure value rises sharply and the slope of the pressure curve changes, it is considered an abnormal situation.

6. The intelligent early warning method for offshore fracturing sand plugging according to any one of claims 1 to 3, characterized in that: Store the processed data in regional databases and merge them into the data lake, including: Collection and transmission: The processed data is transmitted to the DMZ network through the data acquisition room communication network or 4G / VPN; Data storage, transferring data in the DMZ network to the regional database; Data merging: Merge data from regional databases into the data lake.

7. The intelligent early warning method for offshore fracturing sand plugging according to any one of claims 1 to 3, characterized in that: The real-time data collection of offshore fracturing equipment includes: The real-time data of the offshore fracturing construction equipment is transmitted to the data acquisition instrument through the serial port or network port, wherein the network port transmission adopts the WITS protocol or TCP / UDP protocol.

8. The intelligent early warning method for offshore fracturing sand plugging according to claim 7, characterized in that: The real-time data collected from offshore fracturing construction equipment includes: construction well number, construction time, oil pressure, casing pressure, displacement, liquid addition displacement, stage liquid volume, cumulative liquid volume, sand concentration, and implementation data of sand volume.

9. An intelligent early warning system for offshore fracturing sand plugging, characterized by: The system comprises: Data acquisition module, used to collect real-time data of offshore fracturing construction equipment; A data processing and model training module is used to standardize the data and train PGNN and PINN models based on the processed data; Database creation module, used to store processed data in regional databases and merge them into the data lake; The sand plugging intelligent early warning module is used to predict construction pressure changes in real time using the PGNN and PINN models, and to issue a sand plugging warning when the predicted pressure value is abnormal.

10. The intelligent early warning system for offshore fracturing sand plugging according to claim 9, characterized in that: The system also includes a user interface for displaying real-time data and early warning information.

Images