Shale reservoir fracturing underground risk early warning method, device, equipment and system

Abstract

The invention provides a shale reservoir fracturing underground risk early warning method, device, equipment and system.The method comprises the steps that monitoring data are obtained, and the monitoring data comprise pressure data, sand ratio data and fracturing fluid displacement data; and extracting underground working condition characteristics based on the monitoring data by using a pre-trained risk early warning model, and predicting a working condition risk type and a working condition risk level according to the working condition characteristics. According to the method, the abnormal working condition risk type and risk level can be accurately predicted by monitoring the pressure data, the sand ratio data and the fracturing fluid displacement data and further extracting the working condition characteristics.

Description

Technical Field

[0001] The present invention relates to the technical field of oil well fracturing, and particularly to a method, device, equipment and system for downhole risk warning of shale reservoir fracturing. Background Art

[0002] For unconventional oil well reservoirs, due to the large number of fracturing stages per well, the large number of fracturing stages in each stage, and the huge volume of fracturing data, the types of risk events involved in the fracturing process are diverse and complex, and many pieces of information cannot be fully interpreted and are difficult to effectively identify. Existing fracturing risk identification methods mainly rely on manual judgment. Although simple, there are as many as thousands of fracturing stages in the actual field fracturing, and there are as many as tens of thousands of data points in each stage of fracturing, resulting in the difficulty of giving full play to personal experience, low identification efficiency, and difficulty in ensuring identification accuracy. Summary of the Invention

[0003] Embodiments of the present disclosure provide a method, device, equipment and system for downhole risk warning of shale reservoir fracturing, which can efficiently and accurately predict the types and levels of abnormal working condition risks.

[0004] According to one aspect of the present disclosure, a method for downhole risk warning of shale reservoir fracturing is provided, including:

[0005] Obtaining monitoring data, where the monitoring data includes pressure data, sand ratio data, and fracturing fluid displacement data;

[0006] Using a pre-trained risk warning model, based on the monitoring data, extracting downhole working condition features, and predicting the types and levels of working condition risks according to the working condition features.

[0007] In one implementation, the obtaining of the monitoring data includes:

[0008] Obtaining pressure data collected by a pressure sensor;

[0009] Obtaining fracturing fluid displacement data collected by a flowmeter;

[0010] Obtaining sand ratio data according to the sand addition amount.

[0011] In one implementation, the extracting of the downhole working condition features based on the monitoring data includes:

[0012] Generating a fracturing curve based on the monitoring data, where the fracturing curve includes a pressure curve, a sand ratio curve, and a displacement curve;

[0013] Based on the pressure curve, pressure curve features, pressure change frequency features, and pressure change rate features are extracted; the pressure curve features include at least one of the peak and valley values of the pressure, the pressure change trend under multiple different time windows, the pressure fluctuation amplitude, and the slope of the pressure curve, the pressure change frequency features include at least one of the main pressure change frequency and the harmonic components, and the pressure change rate features include the rate of pressure increase or decrease;

[0014] Based on the sand ratio curve, stage sand addition amount features are extracted, and the stage sand addition amount features are the sand addition amounts and the sand addition amount change trend under multiple different time windows;

[0015] Based on the displacement curve, stage displacement features are extracted, and the stage displacement features include at least one of the maximum displacement, the average displacement, the displacement of the fracturing fluid under multiple different time windows, and the displacement change trend.

[0016] In one implementation, the monitoring data further includes at least one of the fracturing wellbore pressure data, the wellbore deformation data, and the wellbore temperature data; wherein,

[0017] Obtain the fracturing wellbore pressure data of the oil well monitored by the wellhead optoelectronic sensor;

[0018] Obtain the wellbore deformation data monitored by the downhole optoelectronic sensor;

[0019] Obtain the wellbore temperature data monitored by the temperature sensor;

[0020] The extracting of the downhole working condition features based on the monitoring data further includes:

[0021] Generate a fracturing wellbore pressure change curve based on the fracturing wellbore pressure data;

[0022] Generate a wellbore deformation curve based on the wellbore deformation data;

[0023] Generate a wellbore temperature change curve based on the wellbore temperature data.

[0024] In one implementation, the method further includes:

[0025] Screen the monitoring data according to the monitoring data and the preset threshold corresponding to the monitoring data.

[0026] In one implementation, the method further includes:

[0027] Based on the monitoring data, determine the working section types corresponding to multiple different time windows, and the working section types include: the bridge plug pushing stage and the main fracturing sand addition stage;

[0028] Predict the working condition risk type and the working condition risk level according to the described working condition characteristics and the type of the working section.

[0029] In one implementation, the predicting the working condition risk type and the working condition risk level according to the described working condition characteristics and the type of the working section includes:

[0030] Determine the target working condition characteristics corresponding to the type of the working section according to the type of the working section;

[0031] Predict the working condition risk type and the working condition risk level based on the target working condition characteristics.

[0032] In one implementation, the pre-trained risk early warning model is obtained through the following steps:

[0033] Construct an initial risk early warning model, and create the attributes of the initial risk early warning model according to the physical entity wellbore shape and the fracturing fluid state, where the attributes include geometric attributes, flow attributes, and functional attributes;

[0034] Train the initial risk early warning model with sample data to obtain the pre-trained risk early warning model, where the sample data includes: sensor sample data as the input data of the initial risk early warning model, and risk type label data and risk level label data as the output data of the initial risk early warning model, and the sensor sample data includes pressure sample data, sand ratio sample data, and fracturing fluid displacement sample data.

[0035] In one implementation, the sample data further includes expert sample data, and the expert sample data includes:

[0036] Sample fracturing curves generated based on the sensor sample data, and risk type expert label data and risk level expert label data corresponding to the sample fracturing curves.

[0037] In one implementation, the sensor sample data further includes fracturing wellbore pressure sample data, wellbore wall deformation sample data, and wellbore temperature sample data.

[0038] In one implementation, the initial risk early warning model is a fracturing risk digital twin model.

[0039] According to another aspect of the present disclosure, there is provided a downhole risk early warning device for shale reservoir fracturing, including:

[0040] A data acquisition module for acquiring monitoring data, where the monitoring data includes pressure data, sand ratio data, and fracturing fluid displacement data;

[0041] A feature extraction module, configured to use a pre-trained risk warning model to extract downhole condition features based on the monitoring data;

[0042] A risk prediction module, configured to predict the type and level of working condition risks according to the condition features.

[0043] According to another aspect of the present disclosure, there is provided an electronic device, including:

[0044] At least one processor; and

[0045] At least one memory storing a computer program,

[0046] The processor calls the computer program to cause the processor to execute the shale reservoir fracturing downhole risk warning method according to one aspect of the present disclosure described above.

[0047] According to another aspect of the present disclosure, there is provided a shale reservoir fracturing downhole risk warning system, including:

[0048] A downhole fracturing integrated machine, on which a sensor group is distributedly installed, and the sensor group is used to collect monitoring data;

[0049] The electronic device according to another aspect of the present disclosure described above, configured to predict the type and level of working condition risks according to the monitoring data;

[0050] A server, the server is connected to the downhole fracturing integrated machine and the electronic device, and the server has a database for storing operation condition data and warning data.

[0051] The above technical features can be combined in various suitable ways or replaced by equivalent technical features as long as the purpose of the present invention can be achieved.

[0052] One or more technical solutions provided in the embodiments of the present disclosure can efficiently and accurately predict abnormal working condition risk types and risk levels. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In the following description of exemplary embodiments with reference to the drawings, more details, features and advantages of the present disclosure are disclosed. In the drawings:

[0054] Figure 1 A flowchart showing a shale reservoir fracturing downhole risk warning method according to an exemplary embodiment of the present disclosure;

[0055] Figure 2 A flowchart showing another shale reservoir fracturing downhole risk warning method according to an exemplary embodiment of the present disclosure;

[0056] Figure 3Shows an exemplary block diagram of the warning output of the working condition risk type in the fracturing downhole risk warning method of a shale reservoir according to an exemplary embodiment of the present disclosure;

[0057] Figure 4 Shows a schematic block diagram of a shale reservoir fracturing downhole risk warning device according to an exemplary embodiment of the present disclosure;

[0058] Figure 5 Shows a schematic block diagram of a shale reservoir fracturing downhole risk warning system according to an exemplary embodiment of the present disclosure. Detailed implementation manners

[0059] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although some embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. On the contrary, these embodiments are provided to more thoroughly and completely understand the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not used to limit the protection scope of the present disclosure.

[0060] It should be understood that the various steps recorded in the method embodiments of the present disclosure can be executed in different orders and / or executed in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present disclosure is not limited in this regard.

[0061] The term "including" and its variants used herein are open-ended, that is, "including but not limited to". The term "based on" is "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description. It should be noted that the concepts such as "first" and "second" mentioned in the present disclosure are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0062] It should be noted that the modifications of "one" and "plural" mentioned in the present disclosure are illustrative rather than restrictive. Those skilled in the art should understand that unless otherwise clearly stated in the context, it should be understood as "one or more".

[0063] The names of the messages or information exchanged between multiple devices in the embodiments of the present disclosure are only for illustrative purposes and are not used to limit the scope of these messages or information.

[0064] The solutions of the embodiments of the present disclosure are described below with reference to the accompanying drawings.

[0065] An embodiment of the present disclosure provides a method for early warning of downhole risks in shale reservoir fracturing. Referring to Figure 1 , the method includes:

[0066] S101, obtaining monitoring data, where the monitoring data includes pressure data, sand ratio data, and fracturing fluid displacement data;

[0067] S102, using a pre-trained risk early warning model, based on the monitoring data, extracting downhole working condition characteristics, and predicting the type and level of working condition risks according to the working condition characteristics.

[0068] An embodiment of the present disclosure uses a pre-trained risk early warning model, according to the monitored pressure data, sand ratio data, and fracturing fluid displacement data, and further extracts working condition characteristics, which can accurately predict the type and level of abnormal working condition risks.

[0069] In some embodiments of the present disclosure, obtaining the monitoring data includes: obtaining the pressure data collected by a pressure sensor; obtaining the fracturing fluid displacement data collected by a flowmeter, and obtaining the sand ratio data according to the sand addition amount.

[0070] During actual construction, a plurality of sensors can be fixedly installed on a downhole fracturing integrated machine to form a sensor group. The sensors such as pressure sensors, flowmeters, etc. are used to collect various monitoring data in real time, and the collected monitoring data is transmitted to the early warning integrated machine. The early warning integrated machine can include an electronic device, and the electronic device has a processor, which is responsible for performing intelligent analysis and calculation according to the monitoring data during the fracturing process to predict the type and level of working condition risks.

[0071] The pressure data can be collected by a pressure sensor, such as oil pressure data, casing pressure data, pump pressure data, etc. The fracturing fluid displacement data can be collected by a flowmeter in the pipeline. The sand ratio data can be directly obtained according to the sand addition amount. For example, the inlet fluid flow of the sand mixing equipment is detected by a flowmeter, and the rotation speed of the sand conveying auger is detected by an encoder or other sensors, and then the theoretical input amount of the input sand and other solid particles is calculated to calculate the sand ratio; the sand ratio can also be detected by installing a nuclear density meter at the outlet of the sand mixing equipment to detect the fluid density and then converted into the sand ratio.

[0072] The types of working condition risks, such as sand plugging risk, stability risk. Here, the stability risk means that the fracturing fluid exerts a huge pressure on the wellbore during the fracturing process, which may cause the wellbore to be unstable or the crack to expand, thus triggering safety accidents such as ground collapse and wellbore collapse.

[0073] Specifically corresponding to the monitoring data, when the fracturing fluid displacement remains unchanged and the pressure increases, it can be determined that it is a risk area. When the fracturing fluid displacement decreases and the pressure continues to increase, it proves that sand plugging has occurred. When the slope suddenly becomes very large, it can be determined that it is a risk area.

[0074] The working condition risk level, such as high risk, medium risk, low risk, no risk, or the risk level is divided by numbers, such as first-level risk, second-level risk, third-level risk, etc.

[0075] In some embodiments of the present disclosure, based on the monitoring data, downhole working condition characteristics are extracted, including:

[0076] Based on the monitoring data, a fracturing curve is generated, where the fracturing curve includes a pressure curve, a sand ratio curve, and a displacement curve;

[0077] Based on the pressure curve, pressure curve characteristics, pressure change frequency characteristics, and pressure change rate characteristics are extracted; the pressure curve characteristics include at least one of the peak and valley values of the pressure, the pressure change trend under multiple different time windows, the pressure fluctuation amplitude, and the slope of the pressure curve, the pressure change frequency characteristics include at least one of the main frequency of the pressure change and the harmonic component, and the pressure change rate characteristics include the rate of pressure rise or fall;

[0078] Based on the sand ratio curve, the stage sand addition amount characteristics are extracted, and the stage sand addition amount characteristics include the sand addition amount under multiple different time windows and the change trend of the sand addition amount;

[0079] Based on the displacement curve, the stage displacement characteristics are extracted, and the stage displacement characteristics include at least one of the maximum displacement, the average displacement, the fracturing fluid displacement under multiple different time windows, and the change trend of the displacement.

[0080] The fracturing curve is a curve with time as the abscissa and pressure, sand ratio, and displacement as the ordinates. The pressure curve can directly reflect the underground situation during the construction process. The sand ratio curve and the displacement curve are controllable, that is, by adjusting the sand addition amount, the sand ratio curve can be changed, and by adjusting the displacement of the fracturing fluid, the displacement curve can be changed. Combining the characteristics of the pressure curve with those of the sand ratio curve and the displacement curve makes the extracted working condition characteristics more comprehensive, and can analyze the construction situation, formation situation, and fracture characteristics, etc., predict the possible types and levels of working condition risks, and perform appropriate regulation to ensure the safe and smooth progress of the construction.

[0081] Specifically, for the pressure curve characteristics such as the pump pressure curve characteristics, the changing trend of the pump pressure, the peak and valley values of the pump pressure, the characteristic of the pump pressure fluctuation amplitude, the slope of the pump pressure curve, the average value of the pump pressure, etc. can be extracted. For the pressure change frequency characteristics such as the pump pressure change frequency characteristics, the corresponding frequency characteristics such as the main frequency and harmonic components can be extracted by analyzing the frequency and period of the pump pressure change. For the pressure change rate characteristics such as the pump pressure change rate characteristics, the rate indexes of the pump pressure rising or falling during the construction process can be extracted by calculating the change rate of the pump pressure, such as the pump pressure rising rate and the pump pressure falling rate. Different time windows such as the proppant addition section, the pressure application section, the pressure stabilization section, etc. can also be further divided into multiple different time windows for the proppant addition section, and the corresponding duration and change rules of different time windows can be concerned.

[0082] For the pressure change characteristics such as the pressure change frequency characteristics and the pressure change rate characteristics, they can be further divided into instantaneous pressure change and stage pressure change. The instantaneous pressure change refers to the instantaneous short-term change of the pressure, and this change is relatively short on the time scale. The instantaneous change of the pressure can be calculated in each window by setting a suitable time window, for example, using indicators such as the maximum value, the minimum value, and the average value to represent the instantaneous change characteristics of the pressure. The stage pressure change refers to the continuous change trend of the pressure over time under different working section types during the construction process. For each working section type, the change rate, the fluctuation range, the duration, etc. of the pressure can be calculated respectively to represent the stage change characteristics of the pressure.

[0083] In some embodiments of the present disclosure, the monitoring data may further include at least one of the fracturing wellbore pressure data, the wellbore wall deformation data, and the wellbore internal temperature data. Specifically, the fracturing wellbore pressure data of the oil well monitored by the wellhead optoelectronic sensor can be obtained; the wellbore wall deformation data monitored by the downhole optoelectronic sensor can be obtained; the wellbore internal temperature data monitored by the temperature sensor can be obtained.

[0084] The wellhead optoelectronic sensor can detect and control the fracturing wellbore pressure, which is convenient for judging whether the downhole working condition is normal. The downhole optoelectronic sensor can measure the deformation of the wellbore wall and regularly transmit the wellbore wall deformation data to the early warning integrated machine for the electronic device in the early warning integrated machine to analyze it, so as to ensure that corresponding measures can be taken before early warning or collapse occurs. The temperature sensor can monitor the temperature change inside the oil well, so as to realize the prediction of the temperature-pressure equation and provide a further data basis for the prediction of the downhole working condition risk.

[0085] In actual implementation, similar to the pressure sensor, the wellhead optoelectronic sensor, the downhole optoelectronic sensor, and the temperature sensor can be set to monitor various sensor data.

[0086] At this time, based on the monitoring data, downhole working condition characteristics are extracted, which further includes: generating a fracturing wellbore pressure change curve based on the fracturing wellbore pressure data, generating a wellbore deformation curve based on the wellbore deformation data, and generating a downhole temperature change curve based on the downhole temperature data.

[0087] The fracturing wellbore pressure change curve, the wellbore deformation curve, and the downhole temperature change curve can be integrated into the aforementioned fracturing curve, or a separate set of curves can be generated.

[0088] By comprehensively analyzing the fracturing wellbore pressure data, the wellbore deformation data, and the downhole temperature data and extracting the corresponding characteristics, the type and level of working condition risks can be predicted more accurately, thereby helping to monitor and control the pressure changes during the fracturing construction process, providing timely alarms and suggestions to ensure the safety and effectiveness of the construction.

[0089] In some embodiments of the present disclosure, after extracting the downhole working condition characteristics, the extracted downhole working condition characteristics can be further standardized, such as using Z-score standardization: for each characteristic, calculate its mean and standard deviation, then subtract the mean from each data point and divide by the standard deviation. This can make the mean of the characteristics be 0 and the standard deviation be 1. Through the standardization process, the dimensional differences between different characteristics can be eliminated, making them have the same scale, which is convenient for subsequent unified analysis and comparison.

[0090] In some embodiments of the present disclosure, in order to reduce the computational load, after obtaining the monitoring data, the monitoring data can also be filtered according to the monitoring data and the preset thresholds corresponding to the monitoring data.

[0091] The preset thresholds can be pre-stored in a database, which can be set on a server. The preset thresholds include, for example, oil pressure thresholds, casing pressure thresholds, sand ratio thresholds, fracturing fluid displacement thresholds, fracturing wellbore pressure thresholds, wellbore deformation thresholds, temperature thresholds, etc. The preset thresholds can be set based on the experience of historical data, or during the training phase of the risk warning model, during the process of continuously optimizing the model parameters, the recommended oil pressure thresholds, casing pressure thresholds, sand ratio thresholds, fracturing fluid displacement thresholds, fracturing wellbore pressure thresholds, wellbore deformation thresholds, temperature thresholds, etc. can be generated. For example, by collecting and training the on-site fracturing construction data, through the analysis and processing of a large amount of data and using data mining techniques, the thresholds under different working conditions can be automatically formed.

[0092] In some embodiments of the present disclosure, considering that different working sections may focus on different monitoring data, in order to further reduce the computational load and improve the processing speed of the equipment, as Figure 2 shown, the downhole risk warning method for shale reservoir fracturing may further include:

[0093] S201. Obtain monitoring data, where the monitoring data includes pressure data, sand ratio data, and fracturing fluid displacement data;

[0094] S202. Using a pre-trained risk warning model, based on the monitoring data, determine the working section types corresponding to multiple different time windows. The working section types include: bridge plug pushing stage, main fracturing and sand addition stage;

[0095] S203. Based on the monitoring data, extract downhole working condition characteristics;

[0096] S204. According to the working condition characteristics and the working section types, predict the working condition risk types and the working condition risk levels.

[0097] Step S201 and S203 are respectively similar to the aforementioned S101 and S102. For the relevant content, reference can be made to the description of S101 and S102 above, and will not be elaborated here.

[0098] The sequence of step S202 and step S203 can be interchanged, that is, S203 can also be executed first, and then S202.

[0099] Preferably, first use a pre-trained risk warning model, based on the monitoring data, to determine the working section types corresponding to multiple different time windows. The working section types include: bridge plug pushing stage, main fracturing and sand addition stage. Then, in combination with the working section types, when extracting the downhole working condition characteristics, for each working section, the monitoring data with low correlation with the working section can be excluded first, that is, only the characteristics corresponding to the monitoring data with high correlation with the working section are extracted to reduce the data processing volume.

[0100] Of course, when extracting the downhole working condition characteristics, the characteristics corresponding to all the monitoring data can also be extracted, and then when predicting, the working section types are taken into consideration, that is, according to the working section types, the target working condition characteristics corresponding to the working section types are determined. Specifically, the working condition characteristics with low correlation with the working section can be excluded to obtain the working condition characteristics with high correlation, and the working condition characteristics with high correlation are used as the target working condition characteristics. For example, in the bridge plug pushing stage, there is no sand addition, so the sand ratio data is the working condition characteristic with low correlation at this time; then, based on the target working condition characteristics, the working condition risk types and the working condition risk levels are predicted.

[0101] In the embodiments of the present disclosure, the pre-trained risk warning model plays a key role and is responsible for analyzing and performing intelligent calculations on complex fracturing working conditions. Specifically, the risk warning model can be trained through the following methods:

[0102] First, construct an initial risk warning model. According to the shape of the physical entity wellbore and the state of the fracturing fluid, create the attributes of the initial risk warning model, including geometric attributes, flow attributes, and functional attributes. That is, based on the wellbore shape and fluid state of the physical entity, define the geometric attributes, flow attributes, and functional attributes of the physical model. During specific implementation, software can be used to simulate and enhance the behavior of the downhole fracturing integrated machine to classify the risk types and risk levels of the working conditions. The functional attributes refer to the functional characteristics and behavior patterns of the downhole fracturing integrated machine, such as pressure sensing, pressure regulation, displacement regulation, etc.

[0103] Physical entities (such as wellbores, fluids, etc.) will undergo various state changes during the fracturing construction process. To better analyze and identify risks, geometric attributes, flow attributes, functional attributes, etc. of the physical model can be defined based on these state changes. At the initial stage of model construction, by considering the actual physical process and integrating different factors and combinations, more accurate characteristic information can be provided for the risk warning model. For example, factors such as wellbore size, tubing type, pressure, flow rate, sandstone type, etc. can be considered, and based on this, an initial risk warning model can be constructed, which can provide a more effective basis for the training and application of the risk warning model.

[0104] Then, input the sample data and the initial model parameters;

[0105] Again, use the sample data to train the initial risk warning model to obtain a pre-trained risk warning model. The sample data includes: sensor sample data as the input data of the initial risk warning model and risk type label data and risk level label data as the output data of the initial risk warning model. The sensor sample data includes pressure sample data, sand ratio sample data, and fracturing fluid displacement sample data.

[0106] The constructed initial risk warning model can be a fracturing risk digital twin model. Use Python to construct an LSTM neural network model based on sensor sample data such as pressure sample data, sand ratio sample data, and fracturing fluid displacement sample data to digitally model the downhole fracturing integrated machine. Then, extract the sample characteristics of the downhole working conditions according to the sensor sample data, and combine the risk type label data and risk level label data to perform iterative training and learning on the initial risk warning model to generate automatically optimized parameters.

[0107] During training, for sensor sample data such as pressure sample data, sand ratio sample data, and fracturing fluid displacement sample data, a filter can be used to screen the data first. For example, a sensor sample data threshold is preset, and sample data exceeding the sensor sample data threshold is removed to provide good sample data for model training and improve the quality of the sample data used for training. In this way, it can be avoided that sensor data that is significantly inconsistent with the corresponding type participates in the model training process and affects the accuracy of the model.

[0108] Similarly, the sensor sample data threshold can be an empirical value or generated during the training process of the initial risk warning model. The sensor sample data threshold can be stored in the database of the server.

[0109] The sample data can be directly obtained from the database of the server. Generally, the early fracturing construction condition data will be stored in the database of the server, such as the operation condition database, for easy retrieval and use when needed later.

[0110] During the model training process, the predicted risk type and predicted risk level will be obtained based on the input sample data. Then, the predicted risk type and predicted risk level will be compared with the true risk type label data and risk level label data, and combined with the generated error, the model parameters will be continuously corrected, optimized, and iteratively trained to obtain a pre-trained risk warning model. Here, the predicted risk type is the predicted value analyzed by the model based on the sample data, and the predicted risk level is similar. The pre-trained risk warning model refers to a risk warning model that has been pre-trained.

[0111] In some embodiments of the present disclosure, in order to further improve the accuracy of the risk prediction model, the sample data participating in the training can also include expert sample data. Among them, the expert sample data includes: a sample fracturing curve generated based on the sensor sample data, and the risk type expert label data and risk level expert label data corresponding to the sample fracturing curve.

[0112] Specifically, a sample fracturing curve can be generated in combination with the sensor sample data. Experts in this field can provide risk type expert labels and risk level expert labels based on the sample fracturing curve in combination with construction condition experience, and integrate the risk type expert labels and risk level expert labels into the model training process. When implemented, the sample fracturing curve generated based on the sensor sample data, as well as the risk type expert label data and risk level expert label data corresponding to the sample fracturing curve, can be sent into the model for training together, or only the risk type expert label data and risk level expert label data corresponding to the sensor sample data can be sent into the model for training.

[0113] In some embodiments of the present disclosure, the sensor sample data participating in model training may further include fracturing wellbore pressure sample data, wellbore deformation sample data, and downhole temperature sample data. By increasing the types of sample data, when the model is trained, it can comprehensively use multi-dimensional sensor data to adjust parameters and improve the accuracy of prediction.

[0114] Furthermore, in the early stage of model training, for the input sample data, according to the type of working section to which it belongs, the sample data can be divided according to the type of working section. For example, according to indicators such as construction time, type of injected liquid, and pressure change, the type of working section to which it belongs can be distinguished, and then divided. At this time, correspondingly, during model training, the type of working section will also be combined. For example, the model will associate the sample data with the type of working section, and for each type of working section, use the corresponding training set to train the corresponding sub-model of the working section; each sub-model trains and optimizes the data of each type of working section, so that the entire model can better understand and analyze the working condition characteristics of different working sections.

[0115] Combined with the trained risk warning model, the type of working condition risk and the level of working condition risk can be predicted based on the real-time collected monitoring data, providing a reference or basis for subsequent further construction control. Specifically, for the output method of the prediction result, it can be directly output the specific type of working condition risk in each working condition time period, such as sand plug risk, and the level of working condition risk, such as high risk. Of course, it can also output the type of working condition risk and the level of working condition risk in each working condition time period in the form of probability, such as Figure 3 , Figure 3 shows an exemplary block diagram of the output of the type of working condition risk warning in risk warning. Among them, after the risk prediction model obtains the monitoring data, it performs analysis and calculation, and can output the probability (P 1 , T 2 , T 3 , ……, Tn, where n is the number of working condition time periods) corresponding to the type of working condition risk for each working condition time period, and m is the number of risk types. The output probability can be in the form of a set or a matrix. 1 , P 2 , ……, Pm), where m is the number of risk types. The output probability can be in the form of a set or a matrix.

[0116] The embodiments of the present disclosure also provide a downhole risk warning device for shale reservoir fracturing. Referring to Figure 4 , Figure 4 shows a schematic block diagram of a downhole risk warning device for shale reservoir fracturing according to an exemplary embodiment of the present disclosure. The device includes:

[0117] A data acquisition module 401, configured to acquire monitoring data, where the monitoring data includes pressure data, sand ratio data, and fracturing fluid displacement data;

[0118] The feature extraction module 402 is used to extract downhole working condition features based on the monitoring data by using a pre-trained risk warning model;

[0119] The risk prediction module 403 is used to predict the type and level of the working condition risk according to the working condition features.

[0120] The embodiment of the present disclosure uses a pre-trained risk warning model, according to the monitored pressure data, sand ratio data and fracturing fluid displacement data, and further extracts the working condition features, and can accurately predict the type and level of abnormal working condition risks.

[0121] In some embodiments of the present disclosure, the data acquisition module 402 acquires the monitoring data including: acquiring the pressure data collected by the pressure sensor; acquiring the fracturing fluid displacement data collected by the flowmeter, and acquiring the sand ratio data according to the sand addition amount.

[0122] During actual construction, a plurality of sensors can be fixedly installed on the downhole fracturing integrated machine to form a sensor group. The sensors such as pressure sensors, flowmeters, etc. are used to collect various monitoring data in real time, and the collected monitoring data is transmitted to the warning integrated machine. The warning integrated machine can include an electronic device, and the electronic device has a processor or the risk warning device, which is responsible for performing intelligent analysis and calculation according to the monitoring data during the fracturing process to predict the type and level of the working condition risk.

[0123] The pressure data can be collected by a pressure sensor, such as the oil pressure data, casing pressure data, pump pressure data, etc. The fracturing fluid displacement data can be collected by a flowmeter in the pipeline. The sand ratio data can be directly obtained through the sand addition amount. For example, the inlet fluid flow of the sand mixing equipment is detected by a flowmeter, and the rotation speed of the sand conveying auger is detected by an encoder or other sensors, and then the theoretical input amount of the input sand and other solid particles is calculated to calculate the sand ratio; the sand ratio can also be calculated by installing a nuclear densitometer at the outlet of the sand mixing equipment to detect the fluid density and then converting it.

[0124] The type of working condition risk, such as sand plug risk, stability risk. Here, the stability risk refers to that the fracturing fluid exerts a huge pressure on the wellbore during the fracturing process, which may cause the wellbore to be unstable or the fracture to expand, thus triggering safety accidents such as ground collapse and wellbore collapse.

[0125] The level of the working condition risk, such as high risk, medium risk, low risk, no risk, or the risk level is divided by numbers, such as first-level risk, second-level risk, third-level risk, etc.

[0126] In some embodiments of the present disclosure, the feature extraction module 402 is used to: generate a fracturing curve based on the monitoring data, where the fracturing curve includes a pressure curve, a sand ratio curve, and a displacement curve;

[0127] Based on the pressure curve, extract the pressure curve features, pressure change frequency features, and pressure change rate features; the pressure curve features include at least one of the peak and valley values of the pressure, the pressure change trend under multiple different time windows, the pressure fluctuation amplitude, and the pressure curve slope, the pressure change frequency features include at least one of the main pressure change frequency and harmonic components, and the pressure change rate features include the rate of pressure rise or fall;

[0128] Based on the sand ratio curve, extract the stage sand addition amount features, which include the sand addition amount under multiple different time windows and the sand addition amount change trend;

[0129] Based on the displacement curve, extract the stage displacement features, which include at least one of the maximum displacement, average displacement, the fracturing fluid displacement under multiple different time windows, and the displacement change trend.

[0130] Combining the features of the pressure curve with those of the sand ratio curve and displacement curve makes the extracted working condition features more comprehensive, enabling the analysis of construction conditions, formation conditions, fracture features, etc., predicting the possible types and levels of working condition risks, and performing appropriate regulation to ensure the safe and smooth progress of the construction.

[0131] In some embodiments of the present disclosure, the data acquisition module 401 can also be used to: acquire the fracturing wellbore pressure data of the oil well monitored by the wellhead optoelectronic sensor; acquire the wellbore deformation data monitored by the downhole optoelectronic sensor; acquire the wellbore temperature data monitored by the temperature sensor. Correspondingly, the feature extraction module 402 can also be used to: generate a fracturing wellbore pressure change curve based on the fracturing wellbore pressure data; generate a wellbore deformation curve based on the wellbore deformation data; generate a wellbore temperature change curve based on the wellbore temperature data.

[0132] The wellhead optoelectronic sensor can detect and control the fracturing wellbore pressure of the oil well, facilitating the judgment of whether the downhole working conditions are normal. The downhole optoelectronic sensor can measure the deformation of the wellbore wall and regularly transmit the wellbore deformation data to the early warning integrated machine for analysis by the electronic device in the early warning integrated machine, ensuring that corresponding measures can be taken before early warning or collapse occurs. The temperature sensor can monitor the temperature change inside the oil well, thereby realizing the prediction of the temperature-pressure equation and providing a further data basis for the prediction of downhole working condition risks.

[0133] By comprehensively integrating the fracturing wellbore pressure data, wellbore deformation data, and wellbore temperature data and extracting the corresponding features, the types and levels of working condition risks can be predicted more accurately.

[0134] In some embodiments of the present disclosure, considering that the monitored data concerned in different working stages may vary, in order to further reduce the computational amount and improve the device processing speed, the downhole risk early warning device for shale reservoir fracturing may further include: a working stage type determination module, configured to determine the working stage types corresponding to multiple different time windows based on the monitored data, and the working stage types include: the bridge plug pushing stage and the main fracturing sand adding stage. At this time, the risk prediction module 403 may further be configured to predict the working condition risk type and the working condition risk level according to the working condition characteristics and the working stage type.

[0135] Further, after determining the working stage types corresponding to multiple different time windows, before the feature extraction module 402 extracts the downhole working condition characteristics, for each working stage, the monitored data with low correlation with the working stage may be removed first, that is, only the characteristics corresponding to the monitored data with high correlation with the working stage are extracted to reduce the data processing amount.

[0136] Certainly, the removal of the monitored data with low correlation may also be performed before the risk prediction module 403 makes a prediction. Specifically, the risk prediction module 403 may remove the working condition characteristics with low correlation with the working stage according to the working stage type to obtain the working condition characteristics with high correlation; based on the working condition characteristics with high correlation, the working condition risk type and the working condition risk level are predicted.

[0137] In some embodiments of the present disclosure, the pre-trained risk early warning model is obtained through the following steps:

[0138] First, an initial risk early warning model is constructed. According to the physical entity wellbore shape and the fracturing fluid state, the attributes of the initial risk early warning model are created, and the attributes include geometric attributes, flow attributes, and functional attributes. That is, according to the wellbore shape and the fluid state of the physical entity, the geometric attributes, flow attributes, and functional attributes of the physical model are defined. During specific implementation, software can be used to simulate and enhance the behavior mode of the downhole integrated fracturing machine to classify the risk types and risk levels of the working condition risks. The functional attribute refers to the functional characteristics and behavior mode of the downhole integrated fracturing machine, such as pressure sensing, pressure regulation, displacement regulation, etc.

[0139] Physical entities (such as wellbores, fluids, etc.) will experience various state changes during the fracturing construction process. In order to better analyze and identify risks, the geometric attributes, flow attributes, functional attributes, etc. of the physical model can be defined based on these state changes. At the initial stage of model construction, by considering the actual physical process and integrating different factors and combinations, more accurate feature information can be provided for the risk early warning model. For example, factors such as the wellbore size, pipe string type, pressure, flow rate, sandstone type, etc. can be considered, and on this basis, an initial risk early warning model is constructed, which can provide a more effective basis for the training and application of the risk early warning model.

[0140] Then, the initial risk warning model is trained using the sample data to obtain a pre-trained risk warning model. The sample data includes: sensor sample data as the input data of the initial risk warning model, and risk type label data and risk level label data as the output data of the initial risk warning model. The sensor sample data includes pressure sample data, sand ratio sample data, and fracturing fluid displacement sample data.

[0141] Regarding the construction and training of the initial risk warning model, reference can be made to the relevant descriptions of the shale reservoir fracturing downhole risk warning method in the foregoing embodiments, which will not be elaborated herein.

[0142] In some embodiments of the present disclosure, in order to further improve the accuracy of the risk prediction model, the sample data participating in the training may further include expert sample data. Among them, the expert sample data includes: sample fracturing curves generated based on the sensor sample data, and risk type expert label data and risk level expert label data corresponding to the sample fracturing curves.

[0143] In some embodiments of the present disclosure, the sensor sample data participating in the model training may further include fracturing wellbore pressure sample data, wellbore deformation sample data, and wellbore temperature sample data. By increasing the types of sample data, when the model is trained, it can comprehensively use multi-dimensional sensor data to adjust parameters and improve the prediction accuracy.

[0144] Furthermore, in the early stage of model training, for the input sample data, according to the type of the working section to which it belongs, the sample data can be divided by the type of the working section. For example, according to indicators such as construction time, type of injected liquid, and pressure change, the type of the working section to which it belongs can be distinguished, and then divided. At this time, correspondingly, during model training, the type of the working section will also be combined. For example, the model will associate the sample data with the type of the working section, and for each type of the working section, use the corresponding training set to train the corresponding sub-model of the working section; each sub-model trains and optimizes the data of each type of the working section, so that the entire model can better understand and analyze the working condition characteristics of different working sections.

[0145] The relevant content of the shale reservoir fracturing downhole risk warning device provided in the embodiments of the present disclosure corresponds to the shale reservoir fracturing downhole risk warning method in the foregoing exemplary embodiments. For the details not described herein, reference can be made to the relevant descriptions of the shale reservoir fracturing downhole risk warning method in the foregoing, which will not be elaborated herein.

[0146] The exemplary embodiments of the present disclosure further provide an electronic device, including: at least one processor; and a memory communicatively connected to the at least one processor. The memory stores a computer program executable by the at least one processor, and when the computer program is executed by the at least one processor, it is configured to cause the electronic device to execute the method according to the embodiments of the present disclosure as described above.

[0147] The exemplary embodiments of the present disclosure further provide a risk early warning system for a shale reservoir fracturing wellbore, as Figure 5 shown. The risk early warning system includes: a downhole fracturing integrated machine 501, on which a sensor group 502 is distributively installed for collecting monitoring data; an electronic device 503 according to the exemplary embodiments of the present disclosure, configured to predict the type and level of working condition risks based on the monitoring data; a server 505, which is connected to the downhole fracturing integrated machine 501 and the electronic device 503, and the server has a database for storing operation condition data and early warning data. The risk early warning model 504 may be embedded in the electronic device.

[0148] The database in the server 505 may store any data related to the fracturing working condition, such as operation condition data, fracturing data, fracturing risk early warning data, model parameters, real-time monitoring data, fault data, etc.

[0149] Although the present invention has been described herein with reference to specific embodiments, it should be understood that these embodiments are merely examples of the principles and applications of the present invention. Therefore, it should be understood that many modifications can be made to the exemplary embodiments, and other arrangements can be designed, as long as they do not deviate from the spirit and scope of the present invention defined by the appended claims. It should be understood that different dependent claims and the features described herein can be combined in a manner different from that described in the original claims. It should also be understood that the features described in connection with a single embodiment can be used in other described embodiments.

Claims

1. A shale reservoir fracturing downhole risk early warning method, characterized in that: include: Acquiring monitoring data, wherein the monitoring data includes pressure data, sand ratio data, and fracturing fluid displacement data; The pre-trained risk warning model is used to extract downhole operating condition characteristics based on the monitoring data, and the operating condition risk type and operating condition risk level are predicted based on the operating condition characteristics.

2. The method according to claim 1, characterized in that: The acquisition of monitoring data comprises: Obtain pressure data collected by a pressure sensor; Obtaining fracturing fluid displacement data collected by the flow meter; Get sand ratio data based on the amount of sand added.

3. The method according to claim 1, characterized in that The extracting downhole working condition characteristics based on the monitoring data includes: Based on the monitoring data, a fracturing curve is generated, wherein the fracturing curve includes a pressure curve, a sand ratio curve and a displacement curve; Based on the pressure curve, extract pressure curve features, pressure change frequency features and pressure change rate features; the pressure curve features include at least one of pressure peak and valley values, pressure change trends corresponding to multiple different time windows, pressure fluctuation amplitude, and pressure curve slope; the pressure change frequency features include at least one of the main frequency and harmonic components of pressure change; and the pressure change rate features include the rate of pressure rise or fall; Based on the sand ratio curve, the characteristics of the amount of sand added in each stage are extracted, and the characteristics of the amount of sand added in each stage include the amount of sand added in multiple different time windows and the change trend of the amount of sand added; Based on the displacement curve, stage displacement characteristics are extracted, and the stage displacement characteristics include at least one of a maximum displacement, an average displacement, a fracturing fluid displacement in a plurality of different time windows, and a displacement change trend.

4. The method according to claim 1, characterized in that: The monitoring data also includes at least one of the following: fracture wellbore pressure data, well wall deformation data, and well temperature data; wherein: Obtaining the fracturing wellbore pressure data of the oil well monitored by the wellhead photoelectric sensor; Obtain the well wall deformation data monitored by downhole photoelectric sensors; Acquire the well temperature data monitored by the temperature sensor; The extracting downhole working condition characteristics based on the monitoring data also includes: Based on the fracturing wellbore pressure data, a fracturing wellbore pressure change curve is generated; Based on the wellbore deformation data, a wellbore deformation curve is generated; Based on the well temperature data, a well temperature change curve is generated.

5. The method according to claim 1, characterized in that The method further comprises: The monitoring data is screened according to the monitoring data and a preset threshold value corresponding to the monitoring data.

6. The method according to claim 1, characterized in that The method further comprises: Based on the monitoring data, determining the types of working sections corresponding to a plurality of different time windows, the types of working sections including: a bridge plug pushing stage, a main fracturing and sand adding stage; According to the operating condition characteristics and the working section type, the operating condition risk type and the operating condition risk level are predicted.

7. The method according to claim 6, characterized in that The predicting of the operating condition risk type and the operating condition risk level according to the operating condition characteristics and the working section type includes: According to the working section type, determining a target working condition characteristic corresponding to the working section type; Based on the target operating condition characteristics, the operating condition risk type and operating condition risk level are predicted.

8. The method according to claim 1, characterized in that The pre-trained risk warning model is obtained by training through the following steps: Constructing an initial risk warning model, and creating attributes of the initial risk warning model according to the physical entity wellbore shape and the fluid state of the fracturing fluid, wherein the attributes include geometric attributes, flow attributes, and functional attributes; The initial risk warning model is trained using sample data to obtain the pre-trained risk warning model, the sample data including: sensor sample data as input data of the initial risk warning model and risk type label data and risk level label data as output data of the initial risk warning model, the sensor sample data including pressure sample data, sand ratio sample data, and fracturing fluid displacement sample data.

9. The method according to claim 8, characterized in that The sample data also includes expert sample data, and the expert sample data includes: A sample fracturing curve generated based on the sensor sample data and risk type expert label data and risk level expert label data corresponding to the sample fracturing curve.

10. The method according to claim 8, characterized in that The sensor sample data also includes fracturing wellbore pressure sample data, well wall deformation sample data, and well temperature sample data.

11. The method according to claim 8, characterized in that The initial risk warning model is a fracturing risk digital twin model.

12. A shale reservoir fracturing downhole risk warning device, characterized in that: include: A data acquisition module, used to acquire monitoring data, wherein the monitoring data includes pressure data, sand ratio data and fracturing fluid displacement data; A feature extraction module, for extracting downhole working condition features based on the monitoring data using a pre-trained risk warning model; The risk prediction module is used to predict the operating condition risk type and operating condition risk level according to the operating condition characteristics.

13. An electronic device, characterized in that: include: at least one processor; as well as at least one memory storing a computer program, The processor calls the computer program to enable the processor to execute the method according to any one of claims 1 to 11.

14. A shale reservoir fracturing downhole risk early warning system, characterized in that: include: A downhole fracturing integrated machine, wherein the downhole fracturing integrated machine is provided with a sensor group, and the sensor group is used to collect monitoring data; The electronic device according to claim 13, used to predict the operating condition risk type and the operating condition risk level based on the monitoring data; A server is connected to the downhole fracturing integrated machine and the electronic device, and the server has a database for storing operating condition data and early warning data.

Images