An electric bridge plug intelligent blocking early warning method and device

Abstract

The present invention discloses an electric bridge plug intelligent plugging warning method and device, which relates to the field of data processing technology. The method includes: collecting real-time status data of the target electric bridge plug, combining with multivariate data processing channels, respectively performing data feature extraction and trend analysis, obtaining a multivariate state feature set and a multivariate state trend set; performing multi-source data fusion to obtain a multi-source data fusion result; training and constructing a plugging state prediction network, combining with the multi-source data fusion result, performing plugging state prediction and comprehensive risk assessment, obtaining multiple risk level coefficients, and performing warning processing according to multiple risk level coefficients. The present invention solves the technical problem that the implementation of real-time warning is difficult due to the limitations of the underground environment and the complexity of signal transmission in the prior art, and achieves the technical effect of performing fault prediction through synchronous fusion analysis of multi-source data to improve the real-time and effectiveness of the electric bridge plug intelligent plugging warning.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to an intelligent blocking and early warning method and device for an electric bridge plug. Background Art

[0002] An electric bridge plug is a tool used in oil and gas wells, mainly used to isolate specific areas in the well for testing, workover or other operations. Monitoring the plugging status of the electric bridge plug is a key link to ensure the plugging effect and safety. Through real-time monitoring and early warning systems, abnormal conditions that may occur during the plugging process, such as abnormal pressure, fluid leakage, etc., can be discovered in time, thereby preventing accidents such as blowouts and well leakage. Therefore, the early warning system needs to be able to issue early warning signals in time before or at the early stage of a fault so that operators can quickly take measures to deal with it. However, due to the limitations of the downhole environment and the complexity of signal transmission, it is difficult to implement real-time early warnings. Summary of the invention

[0003] The present application provides an electric bridge plug intelligent plugging warning method and device, which is used to solve the technical problem in the prior art that it is difficult to achieve real-time warning due to the limitations of the downhole environment and the complexity of signal transmission.

[0004] The first aspect of the present application provides an intelligent plugging warning method for an electric bridge plug, the method comprising: collecting real-time status data of a target electric bridge plug, including collecting downhole environmental data, fluid status data, electrical status data, and chemical status data; establishing a multivariate data processing channel, and based on the real-time status data, performing data feature extraction and trend analysis on multiple channels respectively to obtain a multivariate state feature set and a multivariate state trend set; performing multi-source data fusion on the multivariate state feature set and the multivariate state trend set respectively to obtain a multi-source data fusion result; obtaining historical fault records of the target electric bridge plug, combining fault record data of electric bridge plugs of the same specification, performing data fusion analysis and learning training, and constructing a plugging status prediction network; performing plugging status prediction based on the multi-source data fusion result and the plugging status prediction network to obtain a plugging status prediction result; performing a comprehensive evaluation based on the plugging status prediction result to obtain multiple risk level coefficients, and performing warning processing based on the multiple risk level coefficients.

[0005] The second aspect of the present application provides an intelligent plugging warning device for an electric bridge plug, the device comprising: a real-time status data acquisition module, the real-time status data acquisition module is used to collect real-time status data of a target electric bridge plug, including collecting downhole environment data, fluid status data, electrical status data, and chemical status data; a multi-channel data processing module, the multi-channel data processing module is used to establish a multivariate data processing channel, based on the real-time status data, the multi-channels respectively perform data feature extraction and trend analysis to obtain a multivariate state feature set and a multivariate state trend set; a multi-source data fusion module, the multi-source data fusion module is used to perform multi-source data fusion on the multivariate state feature set and the multivariate state trend set, respectively. Combine to obtain the multi-source data fusion result; a blocking state prediction network construction module, the blocking state prediction network construction module is used to obtain the historical fault records of the target electric bridge plug, combined with the fault record data of the electric bridge plug of the same specification, to perform data fusion analysis and learning training, and to construct a blocking state prediction network; a blocking state prediction module, the blocking state prediction module is used to perform blocking state prediction based on the multi-source data fusion result and in combination with the blocking state prediction network, to obtain a blocking state prediction result; a comprehensive risk assessment and early warning module, the comprehensive risk assessment and early warning module is used to perform a comprehensive assessment based on the blocking state prediction result, obtain multiple risk level coefficients, and perform early warning processing according to the multiple risk level coefficients.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0007] The present application provides an electric bridge plug intelligent plugging early warning method and device, which relate to the field of data processing technology. By collecting real-time status data of a target electric bridge plug, combined with a multivariate data processing channel, data feature extraction and trend analysis are performed respectively to obtain a multivariate status feature set and a multivariate status trend set, and multi-source data fusion is performed to obtain a multi-source data fusion result. Through a plugging status prediction network, plugging status prediction and comprehensive risk assessment are performed, multiple risk level coefficients are obtained and early warning processing is performed, which solves the technical problem in the prior art that real-time early warning is difficult to achieve due to the limitations of the downhole environment and the complexity of signal transmission, and realizes fault prediction through synchronous fusion analysis of multi-source data, thereby improving the real-time and effectiveness of the electric bridge plug intelligent plugging early warning. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. 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 creative work.

[0009] Figure 1 A schematic diagram of a flow chart of an electric bridge plug intelligent plugging warning method provided in an embodiment of the present application;

[0010] Figure 2 A schematic diagram of a process for obtaining multi-source data fusion results in an electric bridge plug intelligent plugging warning method provided in an embodiment of the present application;

[0011] Figure 3 A schematic structural diagram of an electric bridge plug intelligent blocking and warning device provided in an embodiment of the present application.

[0012] Explanation of the accompanying drawings: real-time status data acquisition module 11, multi-channel data processing module 12, multi-source data fusion module 13, blocking status prediction network construction module 14, blocking status prediction module 15, comprehensive risk assessment and early warning module 16. DETAILED DESCRIPTION

[0013] The present application provides an electric bridge plug intelligent plugging warning method and device, which is used to solve the technical problem in the prior art that it is difficult to achieve real-time warning due to the limitations of the downhole environment and the complexity of signal transmission.

[0014] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0015] It should be noted that the terms "first", "second", etc. in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products, or devices.

[0016] Embodiment 1

[0017] like Figure 1 As shown, the present application provides an electric bridge plug intelligent plugging warning method, the method comprising:

[0018] P10: Collect real-time status data of the target electric bridge plug, including downhole environment data, fluid status data, electrical status data, and chemical status data.

[0019] Furthermore, step P10 of the embodiment of the present application further includes:

[0020] P11: According to the working characteristics of the target electric bridge plug, key state factors are obtained, and the key state factors include but are not limited to environmental factors, fluid state factors, electrical state factors and chemical state factors; P12: The basic structure data of the target electric bridge plug is obtained, and key monitoring point extraction is performed in combination with the key state factors and historical maintenance records to obtain multiple key monitoring points; P13: Based on the multiple key monitoring points, a monitoring sensor module is deployed, and real-time status monitoring is performed through the monitoring sensor module to obtain the real-time status data.

[0021] It should be understood that the comprehensive collection of real-time status data of the target electric bridge plug includes multiple dimensions such as downhole environmental data, fluid state data, electrical state data, and chemical state data. First, a series of key state factors are screened out based on the unique working characteristics of the target electric bridge plug. These factors are core indicators for evaluating the operating status of electric bridge plugs, including but not limited to environmental factors (such as temperature and pressure changes), fluid state factors (such as flow, flow velocity, fluid composition and flow stability), electrical state factors (such as voltage fluctuations, current stability, power consumption) and chemical state factors (such as corrosion rate, chemical reaction product concentration). The selection of these factors is based on long-term experience accumulation and professional knowledge, ensuring the pertinence and effectiveness of subsequent analysis.

[0022] Next, the basic structural data of the target electric bridge plug is obtained, which is the basis for understanding its physical characteristics and potential failure points. Combined with key status factors and historical maintenance records, advanced data analysis techniques such as failure mode and effect analysis (FMEA) and fault tree analysis (FTA) are used to extract multiple key monitoring points. These points are the locations where problems are most likely to occur during the operation of the electric bridge plug or the locations that best reflect the changes in its operating status, which are of great significance for early warning and fault diagnosis.

[0023] Finally, based on the extracted key monitoring points, monitoring sensor modules are deployed. These sensor modules integrate a variety of high-precision sensors, which can sense and collect various status data of key monitoring points in real time. Through wireless or wired means, the monitoring sensor modules quickly transmit the collected real-time status data to the data processing center, providing a solid foundation for subsequent data analysis and status prediction.

[0024] P20: Establish a multivariate data processing channel. Based on the real-time status data, multiple channels perform data feature extraction and trend analysis respectively to obtain a multivariate state feature set and a multivariate state trend set.

[0025] Furthermore, step P20 of the embodiment of the present application further includes:

[0026] P21: Based on the key state factors, sample feature data sets are collected for training, and a multivariate data processing channel is constructed, wherein the multivariate data processing channel includes an environmental data processing channel, a fluid data processing channel, an electrical data processing channel, and a chemical data processing channel;

[0027] P22: Using the multivariate data processing channel, perform multi-channel synchronous feature extraction on the real-time state data to obtain a multivariate state feature set, wherein each state feature set corresponds to a key state factor; P23: Based on the multivariate state feature set, perform time series analysis and smoothing to obtain the multivariate state trend set.

[0028] Optionally, implement a multivariate data processing process to extract valuable feature information and trend changes from complex real-time status data. This process achieves refined processing and analysis of different types of data by building a dedicated multivariate data processing channel.

[0029] Specifically, first, a large number of sample feature data sets were collected based on the previously determined key state factors (including environmental factors, fluid state factors, electrical state factors, and chemical state factors). These data sets are the basis for building multivariate data processing channels, covering the various performances of electric bridge plugs under different working conditions and states. Through the training of machine learning or deep learning algorithms, corresponding data processing channels are established for each key state factor, namely environmental data processing channels, fluid data processing channels, electrical data processing channels, and chemical data processing channels. These channels have built-in specialized data processing algorithms and models that can efficiently process data in their respective fields.

[0030] Next, use the constructed multivariate data processing channel to perform multi-channel synchronous feature extraction on the real-time state data. In this process, each channel extracts representative feature information from the real-time data for its corresponding key state factor. These feature information may be numerical (such as temperature value, current value) or structured (such as waveform characteristics, spectrum distribution). For example, environmental characteristics such as temperature, pressure, and humidity are extracted through the environmental data processing channel. The fluid data processing channel extracts fluid characteristics such as flow rate, flow rate, and viscosity. The electrical data processing channel extracts electrical characteristics such as current and voltage. The chemical data processing channel extracts chemical characteristics such as chemical composition and pH. A multivariate state feature set is obtained, in which each subset corresponds to a key state factor, which comprehensively reflects the real-time status of the electric bridge plug in all aspects.

[0031] Finally, time series analysis is performed based on the multivariate state feature set, and the analysis results are smoothed to obtain a multivariate state trend set. Time series analysis is a method to study the law of data change over time, which can reveal the trend and periodicity behind the data. By performing time series analysis on the characteristic data set of each key state factor, the respective trend change curve can be obtained. However, since actual data often contain noise and outliers, these trend curves need to be smoothed to eliminate the interference of noise and make the trend changes clearer and more accurate. The final multivariate state trend set not only reflects the current state change trend of the electric bridge plug, but also provides an important reference for its future state prediction.

[0032] P30: Perform multi-source data fusion on the multi-state feature set and the multi-state trend set respectively to obtain a multi-source data fusion result.

[0033] Further, such as Figure 2 As shown, step P30 of the embodiment of the present application also includes:

[0034] P31: After standardizing the multivariate state feature set and the multivariate state trend set, perform correlation analysis between different features and trends to obtain multiple correlation coefficients; P32: Based on the multiple correlation coefficients, identify and obtain multiple groups of strongly correlated features; P33: Based on the multiple groups of strongly correlated features, perform data fusion processing to obtain multi-source data fusion results.

[0035] It should be understood that the multi-state feature set and the multi-state trend set are deeply integrated to form a more comprehensive and accurate data representation. This process realizes the effective integration of information from different data sources and dimensions through multi-source data fusion technology.

[0036] Specifically, first, the multivariate state feature set and the multivariate state trend set are standardized. Standardization is an important step in data preprocessing, which aims to eliminate the dimensional differences between different data items so that the data can be on the same comparative scale in subsequent processing. After completing the standardization, the correlation analysis between different features and trends is further performed. Correlation analysis is a statistical method used to quantify the degree of association between two or more variables. By calculating the correlation coefficient (such as the Pearson correlation coefficient, the Spearman rank correlation coefficient, etc.), it is identified which features and trends have strong correlations. These strong correlation features are of great significance for subsequent data fusion and state prediction.

[0037] Next, according to the size of the correlation coefficient, multiple groups of strongly correlated features are screened out. These features have similar change trends or interdependencies in the operating state of the electric bridge plug, and their combination can more accurately reflect the overall state of the electric bridge plug.

[0038] Finally, data fusion processing is performed based on multiple groups of strongly correlated features. Data fusion is a technology that merges, associates and integrates information from multiple data sources, aiming to improve the overall value and availability of data. In this step, advanced data fusion algorithms (such as weighted fusion, Bayesian fusion, neural network fusion, etc.) can be used to organically fuse multiple groups of strongly correlated features to form a multi-source data fusion result. This result not only contains rich information from the original data, but also reveals the deeper state characteristics and trend changes of the electric bridge plug through the association and integration between features. It will provide more comprehensive and accurate data support for subsequent plugging status prediction and comprehensive evaluation.

[0039] Furthermore, step P33 of the embodiment of the present application also includes:

[0040] P33-1: Use deep learning methods to construct a data fusion model, and the data fusion model includes an attention mechanism layer; P33-2: Through the data fusion model, perform data fusion processing on the multiple groups of strongly correlated features to obtain the multi-source data fusion result, and the multi-source data fusion result includes key prediction features.

[0041] Specifically, in the process of obtaining multi-source data fusion results, deep learning methods are introduced, especially data fusion models combined with attention mechanisms, to further improve the accuracy and efficiency of data fusion. This not only enhances the model's ability to process complex data, but also makes the extraction of key prediction features more accurate.

[0042] Specifically, we design and build a deep learning model specifically for data fusion. The core of this model lies in its built-in attention mechanism layer, which is a technology that simulates the human attention mechanism and can automatically learn and evaluate the importance of each part of the input data and adjust its processing weight accordingly. In the data fusion task, the attention mechanism layer can identify the key information in multiple groups of strongly correlated features, thereby giving higher weights in the fusion process and ensuring that this information is fully reflected in the final multi-source data fusion result.

[0043] Subsequently, the constructed data fusion model is used to perform in-depth data fusion processing on multiple groups of strongly correlated features. In this process, the model automatically captures the intrinsic connections and potential laws between features through its complex network structure and learning algorithm, and effectively integrates them together to obtain multi-source data fusion results. This result not only contains rich information of the original features, but also highlights the importance of key prediction features through the guidance of the attention mechanism. These key prediction features play a vital role in the subsequent plugging status prediction and comprehensive evaluation, and can directly reflect the operating status and potential risks of the electric bridge plug.

[0044] P40: Obtain the historical fault records of the target electric bridge plug, combine the fault record data of electric bridge plugs of the same specification, perform data fusion analysis and learning training, and build a plugging status prediction network.

[0045] Furthermore, step P40 of the embodiment of the present application also includes:

[0046] P41: Collect the historical fault records of the target electric bridge plug and the fault record data of the electric bridge plugs with the same specifications, and obtain basic fault samples after data preprocessing; P42: Perform data fusion based on the basic fault samples, and extract multiple groups of fault event data, wherein the multiple groups of fault event data include multivariate feature data before the fault of different fault events, and corresponding fault feature data; P43: For the multiple groups of fault event data, perform data clustering with the fault type as the clustering target, and obtain multiple sample fault data sets; P44: Perform supervised training based on the multiple sample fault data sets to construct the blocking state prediction network.

[0047] It should be understood that by comprehensively analyzing the historical fault records of the target electric bridge plug and its equipment of the same specifications, a neural network model that can accurately predict the plugging status is constructed. This process integrates multiple technical means such as data preprocessing, data fusion, data clustering, and supervised learning to ensure the accuracy and generalization ability of the prediction network.

[0048] Specifically, first, the historical fault records of the target electric bridge plug are collected, and the fault record data of the electric bridge plug of the same specification is obtained at the same time. These data cover key information such as various fault types, environmental conditions before and after the fault occurs, operating parameters, and maintenance records. Subsequently, these raw data are preprocessed, including data cleaning (removing noise, outliers, etc.), data formatting (unifying data format, units, etc.), and data transformation (such as normalization, standardization, etc.) to obtain basic fault samples that can be used for subsequent analysis.

[0049] Next, data fusion is performed based on the basic fault samples to extract multiple groups of fault event data. In this process, special attention is paid to the multivariate feature data before the fault occurs and the corresponding fault feature data. Through data fusion technology, data from different sources and dimensions are integrated together to form a more comprehensive and rich description of fault events. These fault event data not only include the direct causes and manifestations of the fault, but also imply potential signs and change patterns before the fault occurs.

[0050] Then, for multiple groups of fault event data, data clustering is performed with fault type as the clustering target. Through cluster analysis, events with similar fault characteristics are classified into one category, thus obtaining multiple sample fault data sets. These data sets not only reflect the unique characteristics of different fault types, but also contain rich fault samples, providing a solid foundation for subsequent model training.

[0051] Finally, supervised training is performed based on multiple sample fault data sets to construct a blocking state prediction network. Supervised learning is a method of training a model when the input-output pairs are known. It can optimize the model parameters through continuous iteration so that the model can accurately predict the output results of unknown data. In this step, advanced technologies such as deep learning and machine learning are used to build a prediction network, and sample fault data sets are used for training. By continuously adjusting the network structure and parameter settings, a neural network model that can accurately predict the blocking state of the electric bridge plug is finally obtained. This model not only has high prediction accuracy and generalization ability, but also can provide strong support for the maintenance and management of electric bridge plugs.

[0052] P50: According to the multi-source data fusion result, combined with the blocking state prediction network, a blocking state prediction is performed to obtain a blocking state prediction result.

[0053] Optionally, the multi-source data fusion results obtained in the previous steps are integrated with the plugging state prediction network to perform accurate prediction of the plugging state of the electric bridge plug. This process not only makes full use of the comprehensiveness and accuracy brought by multi-source data fusion, but also gives play to the advantages of the prediction network in pattern recognition and trend prediction.

[0054] Specifically, first, we review the results of multi-source data fusion. This result is obtained by deeply integrating the multivariate state feature set with the multivariate state trend set, which contains the state information of the electric bridge plug in different dimensions and at different time points. After a series of processing such as standardization, correlation analysis, and data fusion, this information has formed a high-value data set that can reflect the overall operating status of the electric bridge plug.

[0055] Next, this multi-source data fusion result is used as input to the already constructed blocking state prediction network. The blocking state prediction network is a deep learning or machine learning model trained with a large amount of sample data, which has the ability to extract key features from the input data, identify potential patterns, and predict future states. In this scenario, the prediction network will focus on analyzing the key prediction features in the input data, which are determined by data fusion and correlation analysis in the previous steps and are highly correlated with the blocking state.

[0056] During the prediction process, the blocking state prediction network uses its internal complex network structure and algorithm logic to deeply process the input data. It extracts the spatiotemporal features of the data through multiple network layers such as convolutional layers, loop layers, and attention layers, and comprehensively evaluates the impact of these features on the blocking state through the fully connected layer, and finally outputs a blocking state prediction result. This prediction result may be a probability value (indicating the possibility of the blocking state occurring) or a specific classification label (such as "normal", "mild blocking", "severe blocking", etc.).

[0057] The entire prediction process not only depends on the comprehensiveness and accuracy of the multi-source data fusion results, but also on the training quality and generalization ability of the prediction network. By continuously optimizing the network structure and training parameters, the performance of the prediction network can be improved, enabling it to more accurately predict the blocking state of the electric bridge plug. The final blocking state prediction results will provide an important reference for the operation and maintenance management of the electric bridge plug, helping the operation and maintenance personnel to promptly identify potential problems and take corresponding treatment measures.

[0058] P60: Perform a comprehensive assessment based on the blocking status prediction result, obtain multiple risk level coefficients, and perform early warning processing according to the multiple risk level coefficients.

[0059] Furthermore, step P60 of the embodiment of the present application also includes:

[0060] P61: Setting a comprehensive evaluation framework, the comprehensive evaluation framework includes multiple evaluation indicators and multiple indicator weights; P62: Based on the blocking status prediction result, obtaining the predicted fault type and the corresponding predicted fault trend; P63: According to the comprehensive evaluation framework, combining the predicted fault type and the predicted fault trend, performing risk level assessment, obtaining multiple risk level coefficients, the multiple risk level coefficients corresponding to different fault types; P64: Performing risk warning processing based on the multiple risk level coefficients.

[0061] In a possible embodiment of the present application, a comprehensive comprehensive evaluation is conducted based on the plugging state prediction results to determine the current risk level faced by the electric bridge plug and to conduct early warning processing accordingly. This process not only involves an in-depth analysis of the prediction results, but also combines a preset comprehensive evaluation framework and risk assessment model to ensure the accuracy and reliability of the evaluation results.

[0062] Specifically, first, a comprehensive evaluation framework is set up, which is the basis and guide of the evaluation process. It contains multiple evaluation indicators, which cover various aspects of the operation status of the electric bridge plug, such as temperature, pressure, vibration, current, etc. At the same time, a corresponding weight is assigned to each evaluation indicator, which reflects the importance and influence of different indicators in the evaluation process. Through reasonable indicator selection and weight allocation, the comprehensiveness and objectivity of the evaluation results can be ensured.

[0063] Next, based on the plugging status prediction results, the predicted fault types and corresponding predicted fault trends were further obtained. These prediction information can provide specific clues and directions about possible problems with the electric bridge plug in the future. Through careful analysis of the prediction results, potential fault types and fault development trends can be identified, providing strong support for subsequent risk assessment.

[0064] Then, risk level assessment is performed based on the comprehensive assessment framework and predicted fault types and predicted fault trends. In this process, each fault type is quantitatively assessed based on the comprehensive assessment framework, and the corresponding risk level coefficient is calculated. These risk level coefficients not only reflect the severity of different fault types, but also consider the impact of fault development trends on risk levels. By comparing the risk level coefficients of different fault types, the main risks and potential threats currently faced by electric bridge plugs can be clearly understood.

[0065] Finally, risk warning processing is carried out based on multiple risk level coefficients. According to different risk levels, corresponding warning strategies and measures are formulated. For example, for high-risk level fault types, emergency plans are activated and personnel are organized to carry out emergency repairs or replace parts; for medium and low risk level fault types, measures such as strengthening monitoring and optimizing operating parameters are taken to reduce the risk level. Through timely and effective early warning processing, the impact of faults on the operation of the electric bridge plug is minimized to ensure its safe and stable operation.

[0066] In summary, the embodiments of the present application have at least the following technical effects:

[0067] The present application collects real-time status data of the target electric bridge plug, combines it with a multivariate data processing channel, performs data feature extraction and trend analysis respectively, obtains a multivariate state feature set and a multivariate state trend set, and performs multi-source data fusion to obtain a multi-source data fusion result. Through the plugging state prediction network, plugging state prediction and comprehensive risk assessment are performed, multiple risk level coefficients are obtained, and early warning processing is performed.

[0068] The technical effect of predicting faults through synchronous fusion analysis of multi-source data and improving the real-time and effectiveness of the intelligent blocking warning of the electric bridge plug has been achieved.

[0069] Embodiment 2

[0070] Based on the same inventive concept as the electric bridge plug intelligent blocking warning method in the aforementioned embodiment, Figure 3 As shown, the present application provides an electric bridge plug intelligent blocking warning device, and the device and method embodiments in the present application are based on the same inventive concept. The device includes:

[0071] The real-time status data acquisition module 11 is used to acquire the real-time status data of the target electric bridge plug, including the acquisition of downhole environment data, fluid status data, electrical status data, and chemical status data.

[0072] The multi-channel data processing module 12 is used to establish a multivariate data processing channel. Based on the real-time state data, the multi-channels respectively perform data feature extraction and trend analysis to obtain a multivariate state feature set and a multivariate state trend set.

[0073] The multi-source data fusion module 13 is used to perform multi-source data fusion on the multi-state feature set and the multi-state trend set, respectively, to obtain a multi-source data fusion result.

[0074] The blocking state prediction network construction module 14 is used to obtain the historical fault records of the target electric bridge plug, combine the fault record data of electric bridge plugs of the same specification, perform data fusion analysis and learning training, and construct a blocking state prediction network.

[0075] The blocking state prediction module 15 is used to perform blocking state prediction according to the multi-source data fusion result and in combination with the blocking state prediction network to obtain a blocking state prediction result.

[0076] The comprehensive risk assessment and early warning module 16 is used to perform a comprehensive assessment based on the blocking state prediction result, obtain multiple risk level coefficients, and perform early warning processing according to the multiple risk level coefficients.

[0077] Furthermore, the real-time status data acquisition module 11 is also used to perform the following steps:

[0078] According to the working characteristics of the target electric bridge plug, key state factors are obtained, and the key state factors include but are not limited to environmental factors, fluid state factors, electrical state factors and chemical state factors; the basic structure data of the target electric bridge plug is obtained, and the key monitoring points are extracted in combination with the key state factors and historical maintenance records to obtain multiple key monitoring points; based on the multiple key monitoring points, a monitoring sensor module is deployed, and real-time state monitoring is performed through the monitoring sensor module to obtain the real-time state data.

[0079] Furthermore, the multi-channel data processing module 12 is further configured to perform the following steps:

[0080] Based on the key state factors, sample feature data sets are collected for training and a multivariate data processing channel is constructed, wherein the multivariate data processing channel includes an environmental data processing channel, a fluid data processing channel, an electrical data processing channel and a chemical data processing channel; using the multivariate data processing channel, multi-channel synchronous feature extraction is performed on the real-time state data to obtain a multivariate state feature set, wherein each state feature set corresponds to a key state factor; based on the multivariate state feature set, time series analysis is performed and smoothing is performed to obtain the multivariate state trend set.

[0081] Furthermore, the multi-source data fusion module 13 is also used to perform the following steps:

[0082] After the multivariate state feature set and the multivariate state trend set are standardized, correlation analysis is performed between different features and trends to obtain multiple correlation coefficients; based on the multiple correlation coefficients, multiple groups of strongly correlated features are identified and obtained; based on the multiple groups of strongly correlated features, data fusion processing is performed to obtain multi-source data fusion results.

[0083] Furthermore, the multi-source data fusion module 13 is also used to perform the following steps:

[0084] A deep learning method is used to construct a data fusion model, wherein the data fusion model includes an attention mechanism layer; through the data fusion model, data fusion processing is performed on the multiple groups of strongly correlated features to obtain the multi-source data fusion result, wherein the multi-source data fusion result includes key prediction features.

[0085] Furthermore, the blocking state prediction network construction module 14 is also used to perform the following steps:

[0086] The historical fault records of the target electric bridge plug and the fault record data of the electric bridge plugs with the same specifications are collected, and basic fault samples are obtained after data preprocessing; data fusion is performed based on the basic fault samples to extract multiple groups of fault event data, and the multiple groups of fault event data include multivariate feature data before the fault of different fault events, and corresponding fault feature data; for the multiple groups of fault event data, data clustering is performed with the fault type as the clustering target to obtain multiple sample fault data sets; supervised training is performed based on the multiple sample fault data sets to construct the blocking state prediction network.

[0087] Furthermore, the comprehensive risk assessment and early warning module 16 is also used to perform the following steps:

[0088] A comprehensive evaluation framework is set, wherein the comprehensive evaluation framework includes multiple evaluation indicators and multiple indicator weights; based on the blocking state prediction result, a predicted fault type and a corresponding predicted fault trend are obtained; according to the comprehensive evaluation framework, a risk level assessment is performed in combination with the predicted fault type and the predicted fault trend to obtain multiple risk level coefficients, wherein the multiple risk level coefficients correspond to different fault types; and risk warning processing is performed based on the multiple risk level coefficients.

[0089] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0090] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

[0091] This specification and the drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. An electric bridge plug intelligent blocking early warning method, characterized in that: The method comprises: Collect real-time status data of the target electric bridge plug, including downhole environment data, fluid status data, electrical status data, and chemical status data; Establish a multivariate data processing channel, and based on the real-time state data, perform data feature extraction and trend analysis in multiple channels to obtain a multivariate state feature set and a multivariate state trend set; Perform multi-source data fusion on the multi-state feature set and the multi-state trend set to obtain a multi-source data fusion result; Obtaining historical fault records of the target electric bridge plug, combining fault record data of electric bridge plugs of the same specification, performing data fusion analysis and learning training, and building a plugging state prediction network; According to the multi-source data fusion result, combined with the blocking state prediction network, a blocking state prediction is performed to obtain a blocking state prediction result; Perform a comprehensive assessment based on the blocking state prediction result, obtain multiple risk level coefficients, and perform early warning processing according to the multiple risk level coefficients; The multi-state feature set and the multi-state trend set are subjected to multi-source data fusion respectively to obtain a multi-source data fusion result, including: After the multivariate state feature set and the multivariate state trend set are standardized, correlation analysis between different features and trends is performed to obtain multiple correlation coefficients; According to the multiple correlation coefficients, identifying and acquiring multiple groups of strongly correlated features; Based on multiple groups of strongly correlated features, data fusion processing is performed to obtain multi-source data fusion results; Among them, based on multiple groups of strongly correlated features, data fusion processing is performed to obtain multi-source data fusion results, including: Using a deep learning method, constructing a data fusion model, wherein the data fusion model includes an attention mechanism layer; The data fusion model is used to perform data fusion processing on the multiple groups of strongly correlated features to obtain the multi-source data fusion result, which contains key prediction features.

2. The electric bridge plug intelligent blocking early warning method according to claim 1, characterized in that: Collect real-time status data of the target electric bridge plug, including downhole environment data, fluid status data, electrical status data, and chemical status data, including: According to the working characteristics of the target electric bridge plug, key state factors are obtained, wherein the key state factors include but are not limited to environmental factors, fluid state factors, electrical state factors and chemical state factors; Obtain the basic structural data of the target electric bridge plug, extract key monitoring points in combination with the key status factors and historical maintenance records, and obtain multiple key monitoring points; Based on the multiple key monitoring points, monitoring sensor modules are deployed, and real-time status monitoring is performed through the monitoring sensor modules to obtain the real-time status data.

3. The electric bridge plug intelligent blocking early warning method according to claim 2, characterized in that: Establish a multivariate data processing channel, and perform data feature extraction and trend analysis on multiple channels based on the real-time status data, including: Based on the key state factors, sample feature data sets are collected for training, and a multivariate data processing channel is constructed, wherein the multivariate data processing channel includes an environmental data processing channel, a fluid data processing channel, an electrical data processing channel, and a chemical data processing channel; Using the multivariate data processing channel, perform multi-channel synchronous feature extraction on the real-time state data to obtain a multivariate state feature set, wherein each state feature set corresponds to a key state factor; Based on the multivariate state feature set, time series analysis and smoothing processing are performed to obtain the multivariate state trend set.

4. The electric bridge plug intelligent blocking early warning method according to claim 1, characterized in that: Obtain the historical fault records of the target electric bridge plug, combine the fault record data of electric bridge plugs of the same specification, perform data fusion analysis and learning training, and build a plugging state prediction network, including: Collect historical fault records of the target electric bridge plug and fault record data of electric bridge plugs of the same specification, and obtain basic fault samples after data preprocessing; Performing data fusion based on the basic fault samples to extract multiple groups of fault event data, wherein the multiple groups of fault event data include multivariate feature data before the fault of different fault events and corresponding fault feature data; For the multiple groups of fault event data, data clustering is performed with fault type as a clustering target to obtain multiple sample fault data sets; Supervised training is performed based on the multiple sample fault data sets to construct the blocking state prediction network.

5. The electric bridge plug intelligent blocking early warning method according to claim 1, characterized in that: Comprehensively evaluating the blocking state prediction result, obtaining multiple risk level coefficients, and performing early warning processing according to the multiple risk level coefficients include: Setting a comprehensive evaluation framework, wherein the comprehensive evaluation framework includes a plurality of evaluation indicators and a plurality of indicator weights; Based on the blocking state prediction result, obtaining a predicted fault type and a corresponding predicted fault trend; According to the comprehensive assessment framework, a risk level assessment is performed in combination with the predicted fault type and the predicted fault trend to obtain a plurality of risk level coefficients, wherein the plurality of risk level coefficients correspond to different fault types; Risk warning processing is performed based on the multiple risk level coefficients.

6. An electric bridge plug intelligent blocking warning device, characterized in that: For executing the method according to any one of claims 1 to 5, the device comprises: A real-time status data acquisition module, which is used to collect real-time status data of the target electric bridge plug, including collecting downhole environment data, fluid status data, electrical status data, and chemical status data; A multi-channel data processing module, wherein the multi-channel data processing module is used to establish a multivariate data processing channel, and based on the real-time state data, the multi-channels respectively perform data feature extraction and trend analysis to obtain a multivariate state feature set and a multivariate state trend set; A multi-source data fusion module, wherein the multi-source data fusion module is used to perform multi-source data fusion on the multi-state feature set and the multi-state trend set, respectively, to obtain a multi-source data fusion result; A plugging state prediction network construction module, which is used to obtain the historical fault records of the target electric bridge plug, combine the fault record data of electric bridge plugs of the same specification, perform data fusion analysis and learning training, and construct a plugging state prediction network; A blocking state prediction module, wherein the blocking state prediction module is used to perform blocking state prediction according to the multi-source data fusion result and in combination with the blocking state prediction network to obtain a blocking state prediction result; A comprehensive risk assessment and early warning module is used to perform a comprehensive assessment based on the blocking status prediction result, obtain multiple risk level coefficients, and perform early warning processing according to the multiple risk level coefficients.

Images