A method and system for real-time prediction of fracturing wellhead pressure

Abstract

The present invention discloses a method and system for real-time prediction of fracturing wellhead pressure. The method includes the following steps: collecting historical construction data of the fracturing wellhead, and performing data processing and normalization on the historical construction data of the fracturing wellhead to obtain standard input data of the fracturing wellhead; obtaining historical wellhead pressure characteristic curve data based on the standard input data of the fracturing wellhead; constructing a fracturing wellhead pressure prediction model, and training, verifying and testing the fracturing wellhead pressure prediction model to obtain a tested fracturing wellhead pressure prediction model; using the real-time wellhead pressure characteristic curve data and the tested fracturing wellhead pressure prediction model to obtain a real-time predicted value of the fracturing wellhead pressure and a corresponding working condition label, and visualizing the real-time predicted value of the fracturing wellhead pressure and the corresponding working condition label. The present invention can make preparations in advance to deal with emergencies in different working conditions based on the real-time predicted value of the fracturing wellhead pressure and the corresponding working condition label.

Description

Technical Field

[0001] The present invention relates to the technical field of fracturing wellhead pressure prediction, and in particular to a method and system for real-time prediction of fracturing wellhead pressure. Background Art

[0002] Existing methods for obtaining fracturing wellhead pressure typically rely on current equipment information to examine existing parameters. These parameters are then analyzed to determine the operating conditions and implement corresponding measures. This can lead to delayed measures and untimely remediation, resulting in serious consequences. Furthermore, existing methods for obtaining fracturing wellhead pressure do not utilize big data to predict potential construction curve trends, making it impossible to extract experience from construction data, deduce it, and feed it back into the construction process. Summary of the Invention

[0003] In view of the above-mentioned deficiencies in the prior art, the present invention provides a method and system for real-time prediction of fracturing wellhead pressure.

[0004] In order to achieve the above-mentioned object of the invention, the technical solution adopted by the present invention is:

[0005] A method for real-time prediction of fracturing wellhead pressure comprises the following steps:

[0006] S1. Collect historical construction data of the fracturing wellhead, and perform data processing and normalization on the historical construction data of the fracturing wellhead to obtain standard input data of the fracturing wellhead;

[0007] S2. Obtain a wellhead pressure curve and a wellhead displacement curve based on the standard input data of the fracturing wellhead, determine a working condition label based on the wellhead pressure curve and the wellhead displacement curve, and fuse the standard input data and the working condition label of the fracturing wellhead into the wellhead pressure curve to obtain historical wellhead pressure characteristic curve data;

[0008] S3. Constructing a fracturing wellhead pressure prediction model, and using historical wellhead pressure characteristic curve data to train, verify, and test the fracturing wellhead pressure prediction model to obtain a tested fracturing wellhead pressure prediction model;

[0009] S4. Collect real-time construction data of the fracturing wellhead, obtain real-time wellhead pressure characteristic curve data based on the real-time construction data of the fracturing wellhead, use the real-time wellhead pressure characteristic curve data and the tested fracturing wellhead pressure prediction model to obtain the real-time predicted value of the fracturing wellhead pressure and the corresponding working condition label, and visualize the real-time predicted value of the fracturing wellhead pressure and the corresponding working condition label.

[0010] Furthermore, in step S1, the historical construction data of the fracturing wellhead includes the stage type of the fracturing wellhead, wellhead pressure, wellhead displacement, total wellhead liquid volume, sand concentration, sand ratio, total sand delivery, and proppant type.

[0011] Furthermore, the historical construction data of the fracturing wellhead is processed. The specific process is: the historical construction data of the fracturing wellhead is divided into digital type data and non-digital type data, the stage types of the fracturing wellhead in the non-digital type data are sorted by serial number to obtain the stage type value of the fracturing wellhead, and the proppant types of the fracturing wellhead in the non-digital type data are numbered to obtain the proppant type value of the fracturing wellhead.

[0012] Furthermore, in step S2, the operating condition labels include pumping conditions, temporary plugging and fracturing conditions, pre-acid fracturing conditions, sand plugging conditions, and equipment abnormal conditions.

[0013] Furthermore, in step S2, the operating condition label is determined according to the wellhead pressure curve and the wellhead displacement curve. The specific process is as follows: according to the pressure difference on the wellhead pressure curve and the displacement difference on the displacement curve, it is judged whether a pressure gradient is formed at the hole to determine the fracturing ball pumping condition; according to the first increase and then decrease feature of the wellhead pressure curve and the stable and unchanging feature of the displacement curve, the pre-acid fracturing condition is determined; according to the rising sudden rise feature of the wellhead pressure curve and the downward trend feature of the displacement curve, the sand plugging condition is determined; according to the sudden drop feature of the wellhead pressure curve and the first sudden drop and then flat feature of the displacement curve, the abnormal equipment condition is determined.

[0014] Furthermore, in step S3, the fracturing wellhead pressure prediction model includes an input layer, a fully connected layer, a position encoding layer, an encoder, a decoder and an output layer connected in sequence; the encoder includes a multi-head attention mechanism layer, a linear activation layer and a feedforward neural network module connected in sequence; the multi-head attention mechanism layer is used to map the wellhead pressure characteristic curve data output by the position encoding layer into a query vector, a key vector and a value vector respectively through a linear transformation, calculate the attention score according to the query vector and the key vector, and normalize the attention score using Softmax normalization to generate an attention weight.

[0015] Furthermore, the data processing process of the position encoding layer is expressed as:

[0016]

[0017] Input pos =X pos +PE pos

[0018] Among them: PE (pos,2i) is the encoding result of the wellhead pressure characteristic curve data in even dimensions, PE (pos,2i+1)is the encoding result of the wellhead pressure characteristic curve data in odd dimensions, sin is the sine function, pos is the current time step position, i is the feature dimension index, d is the dimension of the embedding space, coS is the cosine function, Input pos is the wellhead pressure characteristic curve data with position information, X pos is the wellhead pressure characteristic curve data, PE pos This is the encoding result of the wellhead pressure characteristic curve data.

[0019] Furthermore, the data processing process of the multi-head attention mechanism layer is expressed as:

[0020]

[0021] Where: Attention(Q, K, V) is the attention score obtained by the multi-head attention mechanism layer, Q is the query vector, K is the key vector, V is the value vector, softmax is the softmax activation function, d is the dimension of the embedding space, d k is the dimension of the key vector.

[0022] Furthermore, the number of attention heads in the multi-head attention mechanism layer is set to 4.

[0023] A system for real-time prediction of fracturing wellhead pressure using the above method, comprising a fracturing wellhead data acquisition unit, a fracturing wellhead data processing unit, a fracturing wellhead pressure real-time prediction unit, and a visualization unit;

[0024] The data acquisition unit of the fracturing wellhead is used to collect historical construction data and real-time construction data of the fracturing wellhead;

[0025] The data processing unit of the fracturing wellhead is used to process and normalize the historical construction data of the fracturing wellhead, obtain standard input data of the fracturing wellhead, obtain the wellhead pressure curve and the wellhead displacement curve based on the standard input data of the fracturing wellhead, determine the working condition label according to the wellhead pressure curve and the wellhead displacement curve, and integrate the standard input data and the working condition label of the fracturing wellhead into the wellhead pressure curve to obtain historical wellhead pressure characteristic curve data; obtain real-time wellhead pressure characteristic curve data based on the real-time construction data of the fracturing wellhead;

[0026] The real-time prediction unit for the fracturing wellhead pressure is used to construct a fracturing wellhead pressure prediction model, and use historical wellhead pressure characteristic curve data to train, verify and test the fracturing wellhead pressure prediction model to obtain a tested fracturing wellhead pressure prediction model; and use the real-time wellhead pressure characteristic curve data and the tested fracturing wellhead pressure prediction model to obtain a real-time predicted value of the fracturing wellhead pressure and a corresponding working condition label;

[0027] The visualization unit is used to visualize the real-time predicted value of the fracturing wellhead pressure and the corresponding working condition label.

[0028] The present invention has the following beneficial effects:

[0029] (1) The present invention constructs a fracturing wellhead pressure prediction model and uses real-time wellhead pressure characteristic curve data and the tested fracturing wellhead pressure prediction model to obtain a real-time predicted value of the fracturing wellhead pressure and a corresponding working condition label. This allows preparations to be made in advance to deal with emergencies in different working conditions based on the real-time predicted value of the fracturing wellhead pressure and the corresponding working condition label.

[0030] (2) The present invention obtains a wellhead pressure curve and a wellhead displacement curve according to standard input data of the fracturing wellhead, determines a working condition label according to the wellhead pressure curve and the wellhead displacement curve, and fuses the standard input data and the working condition label of the fracturing wellhead into the wellhead pressure curve to obtain historical wellhead pressure characteristic curve data. In the data processing stage, data of different working conditions are determined, and the wellhead pressure characteristic curve data with the working condition label is input into the fracturing wellhead pressure prediction model for separate training, so that the fracturing wellhead pressure prediction model can recognize the type of curve under different working conditions, thereby increasing the prediction accuracy of the fracturing wellhead pressure prediction model under different working conditions;

[0031] (3) The present invention proposes a system for real-time prediction of fracturing wellhead pressure, comprising a data acquisition unit for the fracturing wellhead, a data processing unit for the fracturing wellhead, a real-time prediction unit for pressure at the fracturing wellhead and a visualization unit; the data acquisition unit for the fracturing wellhead can be used to acquire historical construction data of the fracturing wellhead and real-time construction data of the fracturing wellhead; the data processing unit for the fracturing wellhead can be used to process and normalize the historical construction data of the fracturing wellhead, obtain standard input data of the fracturing wellhead, obtain a wellhead pressure curve and a wellhead displacement curve based on the standard input data of the fracturing wellhead, determine a working condition label according to the wellhead pressure curve and the wellhead displacement curve, and convert the standard input data and the wellhead displacement curve of the fracturing wellhead into working condition labels. The working condition label is integrated into the wellhead pressure curve to obtain historical wellhead pressure characteristic curve data; real-time wellhead pressure characteristic curve data is obtained based on the real-time construction data of the fracturing wellhead; the real-time prediction unit of the fracturing wellhead pressure can be used to construct a fracturing wellhead pressure prediction model, and the historical wellhead pressure characteristic curve data is used to train, verify and test the fracturing wellhead pressure prediction model to obtain the tested fracturing wellhead pressure prediction model; the real-time wellhead pressure characteristic curve data and the tested fracturing wellhead pressure prediction model are used to obtain the real-time prediction value of the fracturing wellhead pressure and the corresponding working condition label; the visualization unit can be used to visualize the real-time prediction value of the fracturing wellhead pressure and the corresponding working condition label. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1A flow chart of a method for real-time prediction of fracturing wellhead pressure;

[0033] Figure 2 This is a schematic diagram of the working condition label type;

[0034] Figure 3 A schematic diagram of the system structure for real-time prediction of fracturing wellhead pressure. DETAILED DESCRIPTION

[0035] The specific embodiments of the present invention are described below to facilitate understanding of the present invention by those skilled in the art. However, it should be clear that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the appended claims, these changes are obvious, and all inventions and creations utilizing the concepts of the present invention are protected.

[0036] like Figure 1 As shown, a method for real-time prediction of fracturing wellhead pressure includes steps S1-S4, which are specifically as follows:

[0037] S1. Collect historical construction data of the fracturing wellhead, and perform data processing and normalization on the historical construction data of the fracturing wellhead to obtain standard input data of the fracturing wellhead.

[0038] In an optional embodiment of the present invention, a data acquisition unit is constructed for a fracturing wellhead, and historical fracturing wellhead operation data is collected by the data acquisition unit. The historical fracturing wellhead operation data includes the fracturing wellhead stage type, wellhead pressure, wellhead displacement, total wellhead liquid volume, sand concentration, sand ratio, total sand delivery, and proppant type.

[0039] The present invention constructs a data processing unit for a fracturing wellhead, and processes and normalizes historical construction data of the fracturing wellhead through the data processing unit to obtain standard input data for the fracturing wellhead. The present invention processes the historical construction data of the fracturing wellhead by specifically dividing the historical construction data of the fracturing wellhead into digital data and non-digital data, sorting the stage types of the fracturing wellhead in the non-digital data by serial number to obtain the stage type values ​​of the fracturing wellhead, and numbering the proppant types of the fracturing wellhead in the non-digital data to obtain the proppant type values ​​of the fracturing wellhead.

[0040] S2. Obtain a wellhead pressure curve and a wellhead displacement curve based on the standard input data of the fracturing wellhead, determine a working condition label based on the wellhead pressure curve and the wellhead displacement curve, and integrate the standard input data and working condition label of the fracturing wellhead into the wellhead pressure curve to obtain historical wellhead pressure characteristic curve data.

[0041] In an optional embodiment of the present invention, the present invention constructs a data processing unit for the fracturing wellhead, which can obtain the wellhead pressure curve and the wellhead displacement curve based on the standard input data of the fracturing wellhead, determine the working condition label according to the wellhead pressure curve and the wellhead displacement curve, and integrate the standard input data and the working condition label of the fracturing wellhead into the wellhead pressure curve to obtain historical wellhead pressure characteristic curve data.

[0042] like Figure 2 As shown, the operating condition labels include pumping condition, temporary plugging and fracturing condition, pre-acid fracturing condition, sand plugging condition and equipment abnormal condition.

[0043] The present invention determines the working condition label according to the wellhead pressure curve and the wellhead displacement curve. The specific process is as follows: according to the pressure difference on the wellhead pressure curve and the displacement difference on the displacement curve, whether a pressure gradient is formed at the hole is judged to determine the fracturing ball pumping working condition; according to the first increase and then decrease feature of the wellhead pressure curve and the stable and unchanging feature of the displacement curve, the pre-acid fracturing working condition is determined; according to the rising sudden rise feature of the wellhead pressure curve and the downward trend feature of the displacement curve, the sand plugging working condition is determined; according to the sudden drop feature of the wellhead pressure curve and the first sudden drop and then flattening feature of the displacement curve, the abnormal working condition of the equipment is determined.

[0044] S3. Construct a fracturing wellhead pressure prediction model, and use historical wellhead pressure characteristic curve data to train, verify and test the fracturing wellhead pressure prediction model to obtain a tested fracturing wellhead pressure prediction model.

[0045] In an optional embodiment of the present invention, a real-time fracturing wellhead pressure prediction unit is constructed. The real-time fracturing wellhead pressure prediction unit constructs a fracturing wellhead pressure prediction model and trains, verifies, and tests the fracturing wellhead pressure prediction model using historical wellhead pressure characteristic curve data to obtain a tested fracturing wellhead pressure prediction model.

[0046] The fracturing wellhead pressure prediction model includes a sequentially connected input layer, a fully connected layer, a position encoding layer, an encoder, a decoder, and an output layer. The encoder includes a sequentially connected multi-head attention mechanism layer, a linear activation layer, and a feedforward neural network module. The multi-head attention mechanism layer is used to map the wellhead pressure characteristic curve data output by the position encoding layer into a query vector, a key vector, and a value vector through a linear transformation. The attention score is calculated based on the query vector and the key vector, and the attention score is normalized using Softmax normalization to generate attention weights. The decoder includes a sequentially connected masked self-attention layer, a linear activation layer, a multi-head attention mechanism layer, and a feedforward neural network module. The masked self-attention layer is used to predict the next time series of fracturing wellhead pressure prediction values ​​based on the currently generated fracturing wellhead pressure prediction sequence values.

[0047] The present invention converts the wellhead pressure characteristic curve data into a space with a dimension of 64 through linear neurons through a fully connected layer, so as to retain the key features of the wellhead pressure characteristic curve data that affect the wellhead pressure prediction.

[0048] The present invention introduces the position information of the wellhead pressure characteristic curve data through the position encoding layer. Specifically, torch.sin is used in even dimensions and torch.cos is used in odd dimensions to perform position encoding on each time step (0, 1, 2, ...). The position information of each time step is calculated, and the obtained position encoding data is added to the time step feature in the input data, and then added to the wellhead pressure characteristic curve data.

[0049] The data processing process of the position encoding layer is expressed as:

[0050]

[0051] Input pos =X pos +PE pos

[0052] Among them: PE (pos,2i) is the encoding result of the wellhead pressure characteristic curve data in even dimensions, PE (pos,2i+1) is the encoding result of the wellhead pressure characteristic curve data in odd dimensions, sin is the sine function, pos is the current time step position, i is the feature dimension index, d is the dimension of the embedding space, coS is the cosine function, Input pos is the wellhead pressure characteristic curve data with position information, X pos is the wellhead pressure characteristic curve data, PE pos This is the encoding result of the wellhead pressure characteristic curve data.

[0053] Each input feature in the fracturing wellhead pressure prediction model is embedded into a 128-dimensional vector, so the dimension of the embedding space is 128. The position index represents each element in the position encoding vector, which ranges from 0 to 127 in this model.

[0054] The data processing process of the multi-head attention mechanism layer is expressed as:

[0055]

[0056] Where: Attention(Q, K, V) is the attention score obtained by the multi-head attention mechanism layer, Q is the query vector, K is the key vector, V is the value vector, softmax is the softmax activation function, d is the dimension of the embedding space, d k is the dimension of the key vector.

[0057] The number of attention heads in the multi-head attention mechanism layer is set to 4. By setting the number of attention heads in the multi-head attention mechanism layer to 4, the present invention can capture different information in the input sequence from multiple different angles and reduce computational complexity while ensuring model accuracy.

[0058] The present invention can capture the interdependence between various time points in the input wellhead pressure characteristic curve data through a multi-head attention mechanism layer, and predict the upcoming fluctuation trend (xt+1, xt+2, xt+3..., xt+30) by learning the interdependence between historical data points (x1, x2, x3,..., xt).

[0059] The feedforward neural network module, consisting of two linear transformation layers and a ReLU activation function, utilizes this module to strengthen the nonlinear relationship between processing progress and operating parameters, thereby improving prediction accuracy. Specifically, after the output wellhead pressure characteristic curve data is transformed through the linear activation layer, the attention output is processed by the feedforward network and mapped to the target space to obtain output data. The output data is then converted into a readable feature value through a decoder for reduction processing, and the wellhead fracturing prediction value and operating condition label are provided as the final output result.

[0060] The present invention uses historical wellhead pressure characteristic curve data to train, verify and test the fracturing wellhead pressure prediction model. In this process, the loss function selects nn.SmoothL1Loss function and CurvatureLoss, and the optimizer uses AdamW.

[0061] S4. Collect real-time construction data of the fracturing wellhead, obtain real-time wellhead pressure characteristic curve data based on the real-time construction data of the fracturing wellhead, use the real-time wellhead pressure characteristic curve data and the tested fracturing wellhead pressure prediction model to obtain the real-time predicted value of the fracturing wellhead pressure and the corresponding working condition label, and visualize the real-time predicted value of the fracturing wellhead pressure and the corresponding working condition label.

[0062] In an optional embodiment of the present invention, the present invention collects real-time construction data of the fracturing wellhead through a data acquisition unit at the fracturing wellhead, then uses a data processing unit at the fracturing wellhead to obtain real-time wellhead pressure characteristic curve data, and finally uses a real-time pressure prediction unit at the fracturing wellhead to obtain a real-time predicted value of the fracturing wellhead pressure and a corresponding working condition label, and uses a visualization unit to visualize the real-time predicted value of the fracturing wellhead pressure and the corresponding working condition label.

[0063] like Figure 3As shown, a system for real-time prediction of fracturing wellhead pressure using the above method includes a fracturing wellhead data acquisition unit, a fracturing wellhead data processing unit, a fracturing wellhead pressure real-time prediction unit and a visualization unit.

[0064] In an optional embodiment of the present invention, the data acquisition unit of the fracturing wellhead is used to acquire historical construction data of the fracturing wellhead and real-time construction data of the fracturing wellhead.

[0065] In an optional embodiment of the present invention, the data processing unit of the fracturing wellhead is used to perform data processing and normalization on the historical construction data of the fracturing wellhead, obtain standard input data of the fracturing wellhead, obtain a wellhead pressure curve and a wellhead displacement curve based on the standard input data of the fracturing wellhead, determine the working condition label according to the wellhead pressure curve and the wellhead displacement curve, and integrate the standard input data and working condition label of the fracturing wellhead into the wellhead pressure curve to obtain historical wellhead pressure characteristic curve data; obtain real-time wellhead pressure characteristic curve data according to the real-time construction data of the fracturing wellhead.

[0066] In an optional embodiment of the present invention, the real-time prediction unit for the fracturing wellhead pressure is used to construct a fracturing wellhead pressure prediction model, and use historical wellhead pressure characteristic curve data to train, verify and test the fracturing wellhead pressure prediction model to obtain the tested fracturing wellhead pressure prediction model; use the real-time wellhead pressure characteristic curve data and the tested fracturing wellhead pressure prediction model to obtain the real-time prediction value of the fracturing wellhead pressure and the corresponding working condition label.

[0067] In an optional embodiment of the present invention, the visualization unit is used to visualize the real-time predicted value of the fracturing wellhead pressure and the corresponding working condition label.

[0068] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0069] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0070] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 The steps for the function specified in one or more boxes.

[0071] Specific embodiments are used in the present invention to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core ideas. At the same time, for those skilled in the art, according to the ideas of the present invention, there may be changes in the specific implementation methods and application scopes. In summary, the contents of this specification should not be understood as limiting the present invention.

[0072] Those skilled in the art will appreciate that the embodiments described herein are intended to help readers understand the principles of the present invention, and it should be understood that the scope of protection of the present invention is not limited to such specific descriptions and embodiments. Those skilled in the art can make various other specific variations and combinations based on the technical teachings disclosed in the present invention without departing from the essence of the present invention, and such variations and combinations are still within the scope of protection of the present invention.

Claims

1. A method for real-time prediction of fracturing wellhead pressure, characterized in that: The following steps are involved: S1. Collect historical construction data of the fracturing wellhead, and perform data processing and normalization on the historical construction data of the fracturing wellhead to obtain standard input data of the fracturing wellhead; S2. Obtain a wellhead pressure curve and a wellhead displacement curve based on the standard input data of the fracturing wellhead, determine a working condition label based on the wellhead pressure curve and the wellhead displacement curve, and fuse the standard input data and the working condition label of the fracturing wellhead into the wellhead pressure curve to obtain historical wellhead pressure characteristic curve data; The working condition label is determined based on the wellhead pressure curve and the wellhead displacement curve. The specific process is as follows: Based on the pressure difference on the wellhead pressure curve and the displacement difference on the displacement curve, it is judged whether a pressure gradient is formed at the hole to determine the fracturing ball pumping condition; based on the first increase and then decrease characteristics of the wellhead pressure curve and the stable and unchanging characteristics of the displacement curve, the pre-acid fracturing working condition is determined; based on the rising and sudden rise characteristics of the wellhead pressure curve and the downward trend characteristics of the displacement curve, the sand plugging working condition is determined; based on the sudden drop characteristics of the wellhead pressure curve and the first sudden drop and then flattening characteristics of the displacement curve, the abnormal equipment working condition is determined; S3. Constructing a fracturing wellhead pressure prediction model, and using historical wellhead pressure characteristic curve data to train, verify, and test the fracturing wellhead pressure prediction model to obtain a tested fracturing wellhead pressure prediction model; The fracturing wellhead pressure prediction model includes an input layer, a fully connected layer, a position encoding layer, an encoder, a decoder, and an output layer connected in sequence. The encoder includes a multi-head attention mechanism layer, a linear activation layer, and a feedforward neural network module connected in sequence. The multi-head attention mechanism layer is used to map the wellhead pressure characteristic curve data output by the position encoding layer into a query vector, a key vector, and a value vector through linear transformation. The attention score is calculated based on the query vector and the key vector, and the attention score is normalized using Softmax normalization to generate an attention weight. S4. Collect real-time construction data of the fracturing wellhead, obtain real-time wellhead pressure characteristic curve data based on the real-time construction data of the fracturing wellhead, use the real-time wellhead pressure characteristic curve data and the tested fracturing wellhead pressure prediction model to obtain the real-time predicted value of the fracturing wellhead pressure and the corresponding working condition label, and visualize the real-time predicted value of the fracturing wellhead pressure and the corresponding working condition label.

2. The method for real-time prediction of fracturing wellhead pressure according to claim 1, characterized in that: In step S1 , the historical construction data of the fracturing wellhead include the stage type of the fracturing wellhead, wellhead pressure, wellhead displacement, total wellhead liquid volume, sand concentration, sand ratio, total sand delivery, and proppant type.

3. The method for real-time prediction of fracturing wellhead pressure according to claim 1, characterized in that: The historical construction data of the fracturing wellhead are processed. The specific process is as follows: the historical construction data of the fracturing wellhead is divided into digital type data and non-digital type data, the stage types of the fracturing wellhead in the non-digital type data are sorted by serial number to obtain the stage type value of the fracturing wellhead, and the proppant types of the fracturing wellhead in the non-digital type data are numbered to obtain the proppant type value of the fracturing wellhead.

4. The method for real-time prediction of fracturing wellhead pressure according to claim 1, characterized in that: In step S2, the operating condition labels include pumping conditions, temporary plugging and fracturing conditions, pre-acid fracturing conditions, sand plugging conditions, and equipment abnormal conditions.

5. The method for real-time prediction of fracturing wellhead pressure according to claim 1, characterized in that: The data processing process of the position encoding layer is expressed as: , in: is the encoding result of the wellhead pressure characteristic curve data in even dimensions, is the encoding result of the wellhead pressure characteristic curve data in odd dimensions, is a sine function, is the current time step position, is the feature dimension index, is the dimension of the embedding space, is the cosine function, is the wellhead pressure characteristic curve data with position information, is the wellhead pressure characteristic curve data, This is the encoding result of the wellhead pressure characteristic curve data.

6. The method for real-time prediction of fracturing wellhead pressure according to claim 1, characterized in that: The data processing process of the multi-head attention mechanism layer is expressed as: in: is the attention score obtained by the multi-head attention mechanism layer, is the query vector, is the key vector, is a value vector, for activation function, is the dimension of the embedding space, is the dimension of the key vector.

7. The method for real-time prediction of fracturing wellhead pressure according to claim 1, characterized in that: The number of attention heads in the multi-head attention mechanism layer is set to 4.

8. A system for real-time prediction of fracturing wellhead pressure using the method according to any one of claims 1 to 7, characterized in that: It includes a data acquisition unit at the fracturing wellhead, a data processing unit at the fracturing wellhead, a real-time pressure prediction unit at the fracturing wellhead, and a visualization unit; The data acquisition unit of the fracturing wellhead is used to collect historical construction data and real-time construction data of the fracturing wellhead; The data processing unit of the fracturing wellhead is used to process and normalize the historical construction data of the fracturing wellhead, obtain standard input data of the fracturing wellhead, obtain the wellhead pressure curve and the wellhead displacement curve based on the standard input data of the fracturing wellhead, determine the working condition label according to the wellhead pressure curve and the wellhead displacement curve, and integrate the standard input data and the working condition label of the fracturing wellhead into the wellhead pressure curve to obtain historical wellhead pressure characteristic curve data; obtain real-time wellhead pressure characteristic curve data based on the real-time construction data of the fracturing wellhead; The real-time prediction unit for the fracturing wellhead pressure is used to construct a fracturing wellhead pressure prediction model, and use historical wellhead pressure characteristic curve data to train, verify and test the fracturing wellhead pressure prediction model to obtain a tested fracturing wellhead pressure prediction model; and use the real-time wellhead pressure characteristic curve data and the tested fracturing wellhead pressure prediction model to obtain a real-time predicted value of the fracturing wellhead pressure and a corresponding working condition label; The visualization unit is used to visualize the real-time predicted value of the fracturing wellhead pressure and the corresponding working condition label.

Images