Wellbore stability control method, device, equipment and storage medium for dynamically adjusting drilling fluid ratio

Abstract

The present invention relates to a field of drilling technology, and is a wellbore stability control method, device, equipment and storage medium for dynamically adjusting the drilling fluid ratio, including obtaining current drilling environment data and real-time drilling data, wherein the drilling environment data includes drilling fluid performance data and geological data, wherein the real-time drilling parameters include drilling pressure, rotation speed, and mud pump displacement; the current drilling environment data and real-time drilling data are input into a drilling fluid ratio control model group to obtain an optimal drilling fluid ratio and drilling parameter combination. The present invention uses the current drilling fluid performance data, geological data, and real-time drilling parameters to predict the optimal drilling fluid ratio and drilling parameter combination to adapt to the current geological conditions and drilling requirements, increase the pertinence of the drilling fluid, ensure the close fit and high adaptability of the drilling fluid to the wellbore, avoid the problem that the drilling fluid cannot effectively support the wellbore due to the low fit between the drilling fluid and the wellbore, and aggravate the problem of wellbore instability, thereby improving the stability of the wellbore.

Description

Technical Field

[0001] The invention relates to the technical field of drilling, and is a wellbore stability control method, device, equipment and storage medium for dynamically adjusting the ratio of drilling fluid. Background Art

[0002] With the continuous development of oil exploration and drilling technology, the requirements for wellbore stability are getting higher and higher. Wellbore instability not only affects drilling efficiency, but may also lead to serious safety accidents. Wellbore stability is a key factor in the oil and gas drilling process. Wellbore instability can lead to a series of problems, such as wellbore collapse, well leakage, and drill sticking, which not only increases drilling costs, but may also cause environmental pollution and safety accidents. Drilling fluid plays a vital role in maintaining wellbore stability because it can provide wellbore pressure, carry cuttings, cool the drill bit, and lubricate the drill string.

[0003] At present, the traditional drilling fluid ratio is often based on experience or fixed formulas, and cannot be adjusted in real time according to the actual state of the wellbore during the drilling process. This single ratio method cannot adapt to the complex and changeable underground environment, resulting in the performance of the drilling fluid being unable to meet the requirements of wellbore stability. Therefore, the mismatch of drilling fluid performance may aggravate wellbore instability, increase the risk of drilling accidents, and reduce drilling efficiency.

[0004] Specifically, different geological conditions, rock types and drilling depths have different requirements for the performance of drilling fluids. However, existing drilling fluid proportioning technologies are often based on experience or fixed formulas, lack specificity, and cannot ensure the close fit and high adaptability of the drilling fluid to the well wall. As a result, the low fit between the drilling fluid and the well wall will lead to the inability of the drilling fluid to effectively support the well wall, aggravating well wall instability. At the same time, insufficient performance of the drilling fluid will also affect drilling efficiency and increase drilling costs. Although intelligent technology has been widely used in the drilling field in recent years, the degree of intelligence in drilling fluid proportioning and well wall stability control still needs to be improved. There is a lack of real-time monitoring and dynamic adjustment capabilities, and the drilling fluid proportioning cannot be adjusted according to changes in the well wall state, resulting in the drilling fluid proportioning being unable to respond to changes in the well wall state in a timely manner and being unable to effectively prevent the occurrence of well wall instability.

[0005] The existing public patent document 1, with the publication number CN115438090A, discloses a method and system for analyzing and predicting the types of wellbore instability based on block drop image features, which belongs to the field of oil and gas drilling. The method first obtains various types of block drop image feature data, then performs data mining on the block drop image feature data to obtain a wellbore instability type analysis and prediction model, and finally uses the wellbore instability type analysis and prediction model to analyze and predict the wellbore instability type. The present invention forms a wellbore instability type analysis and prediction method based on block drop image feature data mining based on block drop image features by using advanced data analysis methods such as data mining, which solves the problem that the existing wellbore instability types are difficult to identify, provides a more accurate and rapid classification and prediction method for wellbore instability, and provides a basis for forming wellbore stability control countermeasures.

[0006] The second existing patent document, with the publication number CN104732064B, discloses a design method for oil-based drilling fluid parameters, which includes the following steps in order: testing the physical properties and structural properties of formation rocks through experiments; establishing a comprehensive characterization model of the wettability of formation rocks and oil-based drilling fluid systems by typical minerals, pore structures of formation rocks, and oil-based drilling fluid parameters according to the physical properties and structural properties of formation rocks, and drawing a relationship diagram between oil-based drilling fluid parameters and wettability; establishing a mechanical model of the wellbore stability of formation rocks and oil-based drilling fluid systems under wettability conditions, and drawing a relationship diagram between oil-based drilling fluid parameters and wellbore stability; determining oil-based drilling fluid parameters according to the relationship diagram between oil-based drilling fluid parameters and wellbore stability. The window of oil-based drilling fluid parameters designed by the present invention is accurate and reasonable, and can quantitatively evaluate the effect of oil-based drilling fluid on maintaining wellbore stability, prevent the occurrence of complex situations downhole, and reasonably control the cost of oil-based drilling fluid.

[0007] The above two patent documents do not involve a method for achieving wellbore stability control based on dynamic adjustment of drilling fluid ratio. Although the model is also used in patent document 2 to complete the design of oil-based drilling fluid parameters, the model and data used in the dynamic adjustment of drilling fluid process of the present invention are different. Summary of the invention

[0008] The present invention provides a wellbore stability control method, device, equipment and storage medium for dynamically adjusting the drilling fluid ratio, which overcomes the shortcomings of the above-mentioned prior art and can effectively solve the problem that the existing traditional drilling fluid ratio is often based on experience or a fixed formula, lacks specificity, and is prone to low adhesion between the drilling fluid and the wellbore wall, which easily leads to the drilling fluid being unable to effectively support the wellbore wall, thereby aggravating the problem of wellbore instability.

[0009] One of the technical solutions of the present invention is achieved by the following measures: A wellbore stability control method for dynamically adjusting the drilling fluid ratio comprises:

[0010] Obtain current drilling environment data and real-time drilling data, where the drilling environment data includes drilling fluid performance data and geological data, and the real-time drilling data includes drilling pressure, rotation speed, and mud pump displacement;

[0011] The current drilling environment data and real-time drilling data are input into the drilling fluid ratio control model group to obtain the optimal drilling fluid ratio and drilling parameter combination, wherein the drilling fluid ratio control model group includes an associated prediction model and a collaborative optimization model. The associated prediction model is obtained by deep learning the neural network using a number of first samples, each of which includes the identification of the drilling fluid ratio and drilling fluid performance data and geological data. The collaborative optimization model is obtained by deep learning the neural network using a number of second samples, each of which includes the identification of the optimal drilling fluid ratio and drilling parameter combination, and the combination of the drilling fluid ratio and any corresponding drilling parameter.

[0012] The following are further optimizations and / or improvements to the above technical solutions:

[0013] The above-mentioned drilling fluid ratio control model group includes an association prediction model and a collaborative optimization model. The deep learning methods and neural networks used in the association prediction model and the collaborative optimization model are the same, but the sample types used in the deep learning of the two are different. The deep learning methods of the association prediction model include:

[0014] Acquire a number of first samples, and divide them into a training sample set, a verification sample set, and a test sample set according to a ratio, wherein each first sample includes an identifier of a drilling fluid ratio and drilling fluid performance data and geological data;

[0015] The multi-layer feedforward neural network is trained using the training set, and a loss function is introduced during the training. When the value of the loss function is stable, the training is terminated to obtain an associated prediction model, wherein the multi-layer feedforward neural network includes an input layer, a hidden layer, and an output layer. The input layer receives drilling fluid performance data and geological data as input data, the hidden layer includes one or more layers, nonlinearity is introduced through an activation function, the input data is processed and features are extracted, and the output layer outputs the predicted drilling fluid ratio;

[0016] The trained association prediction model is verified and tested using the validation set and test set respectively, the model parameters of the association prediction model are optimized, and an association prediction model that meets the test evaluation requirements is output.

[0017] The above also includes dynamically adjusting the drilling fluid ratio according to the optimal drilling fluid ratio and drilling parameter combination output by the drilling fluid ratio control model group, including:

[0018] Confirm whether the received drilling fluid ratio is complete and accurate and meets the requirements of the current drilling operation;

[0019] In response to this, the amount of various materials required for the drilling fluid is calculated according to the drilling fluid ratio and the current mud volume;

[0020] The drilling fluid performance test results are obtained, and the drilling fluid ratio is adjusted in reverse according to the performance test results until the performance test requirements are met.

[0021] The above-mentioned drilling fluid performance test results include the evaluation results of the filtration performance of the oil-based drilling fluid system, the evaluation results of the HTHP filtration performance of the oil-based drilling fluid, and the evaluation results of the rheological properties of the water-based drilling fluid.

[0022] The above also includes reverse optimization of the drilling fluid ratio control model group according to the actual implementation results after the drilling fluid ratio is dynamically adjusted, including:

[0023] Real-time monitoring and judgment of whether drilling fluid performance or downhole key parameters are abnormal, where drilling fluid performance or downhole key parameters are data of the dynamically adjusted drilling fluid used at the drilling site;

[0024] In response, the parameters of each model in the drilling fluid ratio control model group are adjusted, and the drilling fluid ratio control model group is deep-learned again.

[0025] The second technical solution of the present invention is achieved by the following measures: a wellbore stability control device for dynamically adjusting the drilling fluid ratio, comprising:

[0026] An acquisition unit, which acquires current drilling environment data and real-time drilling data, wherein the drilling environment data includes drilling fluid performance data and geological data, and the real-time drilling data includes drilling pressure, rotation speed, and mud pump displacement;

[0027] The ratio determination unit inputs the current drilling environment data and the real-time drilling data into the drilling fluid ratio control model group to obtain the optimal drilling fluid ratio and drilling parameter combination, wherein the drilling fluid ratio control model group includes an associated prediction model and a collaborative optimization model. The associated prediction model is obtained by deep learning the neural network using a plurality of first samples, each of which includes the identification of the drilling fluid ratio and the drilling fluid performance data and geological data. The collaborative optimization model is obtained by deep learning the neural network using a plurality of second samples, each of which includes the identification of the optimal drilling fluid ratio and drilling parameter combination, and the combination of the drilling fluid ratio and any corresponding drilling parameter.

[0028] The following are further optimizations and / or improvements to the above technical solutions:

[0029] The above-mentioned ratio determination unit comprises:

[0030] Model training module, drilling fluid ratio control model group includes association prediction model and collaborative optimization model. The deep learning methods and neural networks used in deep learning of association prediction model and collaborative optimization model are the same, but the sample types used in deep learning of the two are different. The deep learning methods of association prediction model include:

[0031] Acquire a number of first samples, and divide them into a training sample set, a verification sample set, and a test sample set according to a ratio, wherein each first sample includes an identifier of a drilling fluid ratio and drilling fluid performance data and geological data;

[0032] The multi-layer feedforward neural network is trained using the training set, and a loss function is introduced during the training. When the value of the loss function is stable, the training is terminated to obtain an associated prediction model, wherein the multi-layer feedforward neural network includes an input layer, a hidden layer, and an output layer. The input layer receives drilling fluid performance data and geological data as input data, the hidden layer includes one or more layers, nonlinearity is introduced through an activation function, the input data is processed and features are extracted, and the output layer outputs the predicted drilling fluid ratio;

[0033] Use the validation set and test set to verify and test the trained association prediction model, optimize the model parameters of the association prediction model, and output an association prediction model that meets the test evaluation requirements;

[0034] The ratio optimization module inputs the current drilling environment data and real-time drilling data into the drilling fluid ratio control model group to obtain the optimal drilling fluid ratio and drilling parameter combination.

[0035] The above also includes:

[0036] The dynamic ratio adjustment unit dynamically adjusts the drilling fluid ratio according to the optimal drilling fluid ratio and drilling parameter combination output by the drilling fluid ratio control model group, including:

[0037] Confirm whether the received drilling fluid ratio is complete and accurate and meets the requirements of the current drilling operation;

[0038] In response to this, the amount of various materials required for the drilling fluid is calculated according to the drilling fluid ratio and the current mud volume;

[0039] Obtain the drilling fluid performance test results, and adjust the drilling fluid ratio in reverse according to the performance test results until the performance test meets the requirements;

[0040] The model reverse optimization unit, after dynamically adjusting the drilling fluid ratio, reversely optimizes the drilling fluid ratio control model group according to the actual implementation results, including:

[0041] Real-time monitoring and judgment of whether drilling fluid performance or downhole key parameters are abnormal, where drilling fluid performance or downhole key parameters are data of the dynamically adjusted drilling fluid used at the drilling site;

[0042] In response, the parameters of each model in the drilling fluid ratio control model group are adjusted, and the drilling fluid ratio control model group is deep-learned again.

[0043] The third technical solution of the present invention is achieved through the following measures: an electronic device, including a processor and a memory, wherein the memory stores a computer program, and the computer program is loaded and executed by the processor to implement the steps in the wellbore stability control method of dynamically adjusting the drilling fluid ratio.

[0044] The fourth technical solution of the present invention is achieved through the following measures: a storage medium, on which is stored a computer program that can be read by a computer, and the computer program is configured to execute the steps in the wellbore stability control method for dynamically adjusting the drilling fluid ratio when running.

[0045] The present invention utilizes current drilling fluid performance data, geological data, and real-time drilling data to dynamically predict the optimal drilling fluid ratio and drilling parameter combination to adapt to current geological conditions and drilling requirements, thereby increasing the pertinence of the drilling fluid, ensuring a close fit and high adaptability between the drilling fluid and the well wall, avoiding the problem of low fit between the drilling fluid and the well wall resulting in the drilling fluid being unable to effectively support the well wall, aggravating the problem of well wall instability, improving well wall stability, and reducing the risk of well wall instability. Furthermore, the present invention uses the correlation prediction model and the collaborative optimization model to make targeted adjustments to the drilling fluid ratio. The two are obtained by deep learning of the neural network through different samples. Therefore, the model can be continuously optimized and updated according to different geological conditions and drilling requirements, so that it has a high degree of adaptability and stability, and can more accurately and dynamically adjust the drilling fluid ratio. The correlation prediction model outputs the predicted optimal drilling fluid ratio, and the collaborative optimization model can analyze the optimal drilling fluid ratio and drilling parameter combination, reduce unnecessary drilling fluid consumption, and make the drilling process smoother and more efficient. It can not only reduce the time loss of stopping drilling and repairing wells, but also reduce drilling costs and improve economic benefits. In addition, the present invention can also dynamically adjust the drilling fluid ratio according to the optimal drilling fluid ratio and drilling parameter combination output by the drilling fluid ratio control model group, combined with the corresponding drilling fluid performance, so that the drilling fluid ratio is more in line with the working conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Attached Figure 1 A schematic diagram of an implementation environment provided for an embodiment of the present invention.

[0047] Attached Figure 2 A schematic flow chart of a method for adjusting the ratio of drilling fluid provided in one embodiment of the present invention.

[0048] Attached Figure 3 A flowchart of a model deep learning method provided for one embodiment of the present invention.

[0049] Attached Figure 4 A schematic flow chart of a method for dynamically adjusting the drilling fluid ratio provided in one embodiment of the present invention.

[0050] Attached Figure 5 A schematic flow chart of a method for reversely optimizing a drilling fluid ratio control model group according to actual implementation results provided in one embodiment of the present invention.

[0051] Attached Figure 6 A schematic diagram of the structure of a device provided in one embodiment of the present invention.

[0052] Attached Figure 7 A schematic diagram of the structure of a device provided in yet another embodiment of the present invention. DETAILED DESCRIPTION

[0053] The present invention is not limited by the following embodiments, and specific implementation methods can be determined based on the technical solution of the present invention and actual conditions.

[0054] Those skilled in the art will understand that, unless otherwise stated, a "module" or "unit" in the embodiments of the present invention refers to a computer program or a part of a computer program with a predetermined function, and works together with other related parts to achieve a predetermined goal, and can be implemented in whole or in part by using software, hardware (such as a processing circuit or memory) or a combination thereof. Similarly, a processor (or multiple processors or memories) can be used to implement one or more modules or units. In addition, each module or unit can be part of an overall module or unit that includes the function of the module or unit.

[0055] In addition, the term “plurality” in the embodiments of the present invention refers to two or more than two, and “first” and “second” etc. are used for distinguishing descriptions and should not be understood as implying relative importance.

[0056] At present, the traditional drilling fluid ratio is often based on experience or fixed formulas, and cannot be adjusted in real time according to the actual state of the wellbore during drilling. This single ratio method cannot adapt to the complex and changeable underground environment, resulting in the performance of the drilling fluid being unable to meet the requirements of wellbore stability. Therefore, the mismatch of drilling fluid performance may aggravate wellbore instability, increase the risk of drilling accidents, and reduce drilling efficiency.

[0057] Specifically, different geological conditions, rock types and drilling depths have different performance requirements for drilling fluids. However, existing drilling fluid ratio technology is often based on experience or fixed formulas, lacks specificity, and cannot ensure the close fit and high adaptability of the drilling fluid to the well wall, so that the low fit between the drilling fluid and the well wall will lead to the drilling fluid being unable to effectively support the well wall, aggravating the well wall instability. At the same time, the insufficient performance of the drilling fluid will also affect the drilling efficiency and increase the drilling cost. Although intelligent technology has been widely used in the field of drilling in recent years, the degree of intelligence in drilling fluid ratio and well wall stability control still needs to be improved. There is a lack of real-time monitoring and dynamic adjustment capabilities, and the drilling fluid ratio cannot be adjusted according to changes in the well wall state, resulting in the drilling fluid ratio being unable to respond to changes in the well wall state in a timely manner, and unable to effectively prevent the occurrence of well wall instability. Therefore, in order to achieve dynamic adjustment of the drilling fluid ratio, an embodiment of the present invention provides a well wall stability control method for dynamically adjusting the drilling fluid ratio.

[0058] The embodiment of the present invention provides a wellbore stability control method for dynamically adjusting the drilling fluid ratio. By deep learning a neural network, a drilling fluid ratio control model group is obtained. The drilling fluid ratio control model group includes an associated prediction model and a collaborative optimization model. The associated prediction model can explore the correlation between the drilling fluid ratio and the wellbore stability. The optimal drilling fluid ratio is obtained by predicting the current drilling fluid performance data and geological data. The collaborative optimization model collaboratively optimizes the drilling fluid ratio and drilling parameters. The optimal drilling fluid ratio and drilling parameter combination are obtained by comparing the optimal drilling fluid ratio output by the associated prediction model with real-time drilling data, and the drilling fluid ratio is dynamically adjusted.

[0059] Among them, the method provided by the embodiment of the present invention may involve artificial intelligence (AI) technology, and can be implemented based on artificial intelligence technology, for example, using a deep learning approach and using sample training to obtain a corresponding model.

[0060] Among them, Machine Learning (ML) is a multi-disciplinary interdisciplinary subject involving probability theory, statistics, approximation theory, convex analysis, algorithm complexity theory and other disciplines. It specializes in studying how computers simulate or implement human learning behavior to acquire new knowledge or skills and reorganize existing knowledge structures to continuously improve their performance. Machine learning is the core of artificial intelligence and the fundamental way to make computers intelligent. Its applications are spread across all areas of artificial intelligence.

[0061] Among them, deep learning (DL) specifically refers to machine learning based on deep neural network models and methods. It is developed on the basis of statistical machine learning, artificial neural network and other algorithm models, combined with the development of contemporary big data and large computing power. The most important technical feature of deep learning is the ability to automatically extract features.

[0062] The above-mentioned machine learning and deep learning usually include technologies such as neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and self-learning.

[0063] Among them, the loss function is in the process of training the neural network. Because we hope that the output of the neural network is as close as possible to the value we really want to predict, we can compare the predicted value of the current network with the target value, and then update the weight vector of each layer of the neural network according to the difference between the two (there is usually an initialization process before the first update, that is, pre-configuring parameters for each layer in the neural network), until the neural network can predict the target value or a value very close to the target value. Therefore, in deep learning, it is necessary to pre-define "how to compare the difference between the predicted value and the target value", which is the loss function.

[0064] As attached Figure 1 As shown, it shows a schematic diagram of an implementation environment provided by an embodiment of the present invention. The implementation environment may include: training equipment and use equipment.

[0065] Both the training device and the use device are computer devices; optionally, the computer device is a terminal device, such as a mobile phone, a tablet computer, a PC (Personal Computer) and other electronic devices; or, the computer device is a server, which can be a single server, a server cluster composed of multiple servers, or a cloud computing service center, which is not limited to the embodiments of the present invention.

[0066] The training device refers to a computer device with the ability to train and learn neural networks. Optionally, the training device has the ability to obtain neural networks, and train and learn them according to application requirements. For example, the training device obtains neural networks from other devices through the network, and then trains them through training samples according to application requirements, so that the neural network has the ability to obtain the optimal drilling fluid ratio and drilling parameter combination; Optionally, the training device has the ability to construct neural networks, and it can construct neural networks by itself according to application requirements, and then train and learn them. For example, in order to achieve the optimal drilling fluid ratio and drilling parameter combination based on drilling environment data and real-time drilling data, the training device constructs a neural network by itself, and then trains and learns it through samples according to application requirements.

[0067] The device used refers to a computer device that has the requirements for using a neural network. Optionally, the device uses a neural network according to application requirements and obtains the neural network from other devices through a network. For example, if the device uses a neural network that has the requirements for predicting the optimal drilling fluid ratio and drilling parameter combination, it can obtain a neural network that has completed training and learning to predict the optimal drilling fluid ratio and drilling parameter combination from other devices through a network, and use the neural network to predict the optimal drilling fluid ratio and drilling parameter combination.

[0068] Based on this, the technical solution of the present invention will be introduced and explained in combination with several embodiments and drawings.

[0069] Embodiment 1: As attached Figure 2 As shown, the embodiment of the present invention discloses a wellbore stability control method for dynamically adjusting the drilling fluid ratio, comprising:

[0070] Step S110, obtaining current drilling environment data and real-time drilling data, wherein the drilling environment data includes drilling fluid performance data and geological data, and the real-time drilling data includes drilling pressure, rotation speed, and mud pump displacement.

[0071] Specifically, the process of obtaining the current drilling environment data and real-time drilling data is as follows:

[0072] (1) Determine the testing environment, locate the wellbore environment to be tested, and determine the current drilling fluid type, drilling equipment, drilling environment, and geological data based on drilling data. The geological data includes formation type, rock mechanics parameters, etc.

[0073] (2) Use monitoring equipment installed at the drilling site (such as performance testing instruments and various sensors) to obtain drilling fluid performance data and real-time drilling data. The drilling fluid performance data includes density, viscosity, API filtration loss, etc., and the real-time drilling data includes temperature, drilling pressure, rotation speed, mud pump displacement, etc.

[0074] It should be noted that all data obtained here need to be associated with wellbore stability.

[0075] The sensors mentioned above may include but are not limited to:

[0076] Temperature sensor: Utilizes the property of materials that change their resistance, capacitance or potential with temperature to convert temperature into electrical signal output. Common temperature sensors include thermocouples, thermistors, etc., which monitor the temperature changes underground in real time.

[0077] Pressure sensor: It measures pressure based on piezoresistive effect, piezoelectric effect or capacitive effect. Piezoresistive pressure sensor reflects pressure change by measuring the change of resistance value of piezoresistive element; piezoelectric pressure sensor uses the uneven distribution of charge generated by piezoelectric material after force to generate voltage signal; capacitive pressure sensor reflects pressure change by measuring the change of capacitance value, monitoring the pressure change in the well, such as bottom hole pressure, well wall pressure, etc.

[0078] Vibration sensor: reflects the vibration state of an object by measuring the frequency, amplitude and other parameters of the object's vibration. Common vibration sensors include piezoelectric and eddy current types, which monitor the vibration of downhole equipment in real time, such as drilling machinery and pumps.

[0079] In actual applications, it is necessary to select and adjust according to specific needs and scenarios. At the same time, in order to achieve real-time monitoring of well wall stability, it is also necessary to combine other sensors and monitoring equipment, such as liquid level sensors, pH sensors, etc.

[0080] Step S120, input the current drilling environment data and real-time drilling data into the drilling fluid ratio control model group to obtain the optimal drilling fluid ratio and drilling parameter combination, wherein the drilling fluid ratio control model group includes an associated prediction model and a collaborative optimization model, the associated prediction model is obtained by deep learning the neural network using a plurality of first samples, each of which includes an identification of the drilling fluid ratio and drilling fluid performance data and geological data, and the collaborative optimization model is obtained by deep learning the neural network using a plurality of second samples, each of which includes an identification of the optimal drilling fluid ratio and drilling parameter combination, and a combination of the drilling fluid ratio and any corresponding drilling parameter.

[0081] In this embodiment, the drilling fluid ratio control model group includes an association prediction model and a collaborative optimization model. Although the sample types used in the deep learning process of the association prediction model and the collaborative optimization model are different, the deep learning method and neural network used are the same. Therefore, in this embodiment, the deep learning method of the association prediction model is taken as an example to illustrate the deep learning process. Figure 3 As shown, including:

[0082] Step S121, obtaining a number of first samples, and dividing them into a training sample set, a verification sample set, and a test sample set according to a ratio, wherein each first sample includes an identifier of a drilling fluid ratio and drilling fluid performance data and geological data;

[0083] Each of the above-mentioned first samples includes the identification of the drilling fluid ratio and the drilling fluid performance data and geological data, wherein the specific drilling fluid performance data and geological data are screened from the historical drilling fluid performance data and historical geological data corresponding to each different drilling fluid ratio, and the data closely related to the wellbore stability are screened out. In this embodiment, the drilling fluid performance data may include but is not limited to the drilling fluid density, viscosity, API filtration loss, and the geological data include the formation type and rock mechanical parameters; after screening, all the data are cleaned, sorted, outliers and missing values ​​are removed, and standardized or normalized, and further according to the input data requirements of the multi-layer feedforward neural network, the corresponding data encoding and conversion are performed.

[0084] Step S122, using the training set to train the multi-layer feedforward neural network, introducing a loss function during training, and ending the training when the value of the loss function is stable to obtain an associated prediction model, wherein the multi-layer feedforward neural network includes an input layer, a hidden layer, and an output layer, the input layer receives drilling fluid performance data and geological data as input data, the hidden layer includes one or more layers, nonlinearity is introduced through an activation function, the input data is processed and feature extracted, and the output layer outputs a predicted drilling fluid ratio.

[0085] The structure of the multi-layer feedforward neural network used in the above training is as follows:

[0086] The multi-layer feedforward neural network includes an input layer, a hidden layer, and an output layer;

[0087] The input layer receives drilling fluid performance data and geological data as input data;

[0088] Hidden layer (depending on the complexity of the problem, multiple hidden layers can be designed, each hidden layer contains multiple neurons), which introduces nonlinearity through activation functions to process input data and extract features; the output of each neuron usually generates nonlinearity through an activation function (such as ReLU, Sigmoid or tanh), and a function is selected when using it;

[0089] The output layer outputs the predicted drilling fluid ratio.

[0090] Among them, the multi-layer feedforward neural network formula is as follows:

[0091]

[0092] In the formula, is an activation function, such as ReLU; is the first neurons to the current layer The weights of the neurons; is the first The output of a neuron; The current layer The bias term of each neuron; is the number of neurons in the input layer or the previous hidden layer.

[0093] Input Layer Represents drilling fluid performance data and geological data, output layer represents the predicted drilling fluid ratio.

[0094] When the multi-layer feedforward neural network is trained using the training set as mentioned above, optimization algorithms such as back propagation algorithm and gradient descent can also be used to train the multi-layer feedforward neural network, and a suitable loss function, such as mean square error (MSE), can be set to measure the gap between the model prediction value and the true value.

[0095] Step S123, using the validation set and the test set to respectively validate and test the trained association prediction model, optimize the model parameters of the association prediction model, and output the association prediction model that meets the test evaluation requirements.

[0096] The above-mentioned validation set is used to validate the association prediction model, adjust the model parameters of the association prediction model (i.e., optimize the weights and bias parameters of the multi-layer feedforward neural network), use the test set to test the association prediction model, and evaluate and analyze the test results. The evaluation and analysis methods include mean square error index (MSE), mean absolute error index (MAE), root mean square error index (RMSE), and determination coefficient index (R2). When the evaluation result meets the test evaluation requirements (i.e., the threshold condition of the evaluation and analysis index, and the specific threshold is set according to the actual situation), all training steps are terminated. When the evaluation result does not meet the test evaluation requirements, the weights and bias parameters of the multi-layer feedforward neural network can be optimized and then re-trained.

[0097] In this embodiment, the collaborative optimization model balances the optimization requirements between different drilling parameters through multi-objective optimization (such as weighted loss function) to achieve collaborative optimization of drilling fluid ratio and drilling parameters. The specific process of using a number of second samples to perform deep learning on the neural network to obtain the collaborative optimization model is the same as step S121 to step S122, with the difference that each second sample includes an identifier of the optimal drilling fluid ratio and drilling parameter combination, and a combination of the drilling fluid ratio and any corresponding drilling parameter, wherein the combination of the drilling fluid ratio and any corresponding drilling parameter is the optimal combination of the drilling fluid ratio and any corresponding drilling parameter, and the corresponding combination of any drilling parameters here is the drilling parameters that are close to the drilling parameters in the optimal combination of drilling fluid ratio and drilling parameters.

[0098] The present invention discloses a wellbore stability control method for dynamically adjusting the drilling fluid ratio. The method utilizes current drilling fluid performance data, geological data, and real-time drilling data to dynamically predict the optimal drilling fluid ratio and drilling parameter combination to adapt to current geological conditions and drilling requirements, thereby increasing the pertinence of the drilling fluid, ensuring the close fit and high adaptability of the drilling fluid to the wellbore, avoiding the problem of the drilling fluid being unable to effectively support the wellbore due to low fit between the drilling fluid and the wellbore, aggravating the wellbore instability, improving the wellbore stability, and reducing the risk of wellbore instability. Furthermore, the present invention uses an associated prediction model and a collaborative optimization model to make targeted adjustments to the drilling fluid ratio. The two are obtained by deep learning of the neural network through different samples. Therefore, the model can be continuously optimized and updated according to different geological conditions and drilling requirements, so that it has a high degree of adaptability and stability, and can more accurately and dynamically adjust the drilling fluid ratio. The associated prediction model outputs the predicted optimal drilling fluid ratio, and the collaborative optimization model can analyze the optimal drilling fluid ratio and drilling parameter combination, reduce unnecessary drilling fluid consumption, and make the drilling process smoother and more efficient. It can not only reduce the time loss such as stopping drilling and repairing wells, but also reduce drilling costs and improve economic benefits.

[0099] Embodiment 2: As attached Figure 4 As shown, the embodiment of the present invention discloses a wellbore stability control method for dynamically adjusting the drilling fluid ratio, which is a further optimization of the above embodiment, and further includes dynamically adjusting the drilling fluid ratio according to the optimal drilling fluid ratio and drilling parameter combination output by the drilling fluid ratio control model group, including:

[0100] Step S210, confirm whether the received drilling fluid ratio is complete and accurate and meets the requirements of the current drilling operation.

[0101] Step S220, in response to which, according to the drilling fluid ratio and the current mud volume, the amount of various materials required for the drilling fluid is calculated.

[0102] Specifically, if it is confirmed that the received drilling fluid ratio is complete and accurate and meets the needs of the current drilling operation, the batching work will be carried out. That is, after the material preparation and batching equipment inspection are completed, the drilling fluid ratio control model group outputs the optimal drilling fluid ratio and the current mud volume, and calculates the amount of various materials required for the drilling fluid.

[0103] For example, the optimal drilling fluid ratio output by the drilling fluid ratio control model group includes water, bentonite, soda ash and other additives, and their respective proportions in the drilling fluid, wherein the other additives include additive 1 and additive 2; then the corresponding process of calculating the amount of various materials required for the drilling fluid according to the drilling fluid ratio and the current mud volume is as follows:

[0104] The drilling fluid ratio is: water: bentonite: soda ash: additive 1: additive 2 = 45:50:2:1:2 (just an example);

[0105] The current mud volume is V;

[0106] The quantities of various materials required for the drilling fluid are as follows:

[0107]

[0108] Step S230, obtaining the drilling fluid performance test result, and adjusting the drilling fluid ratio in reverse according to the performance test result until the performance test meets the requirements.

[0109] Specifically, the performance test of the drilling fluid is set according to the needs, for example, including the test of solid content, specific gravity, funnel viscosity, water loss, mud cake thickness, sand content, pH value and other indicators. The process of performing the performance test of the drilling fluid is an existing well-known technology. The drilling fluid performance test results obtained in this embodiment may include but are not limited to the evaluation results of the filtration performance of the oil-based drilling fluid system, the filtration performance evaluation results of the oil-based drilling fluid HTHP, and the rheological properties evaluation results of the water-based drilling fluid, which are specifically selected according to the type of drilling fluid (oil-based or water-based).

[0110] It should be noted that the evaluation process corresponding to the evaluation results of the base drilling fluid system filtration performance, the evaluation results of the HTHP filtration performance of the oil-based drilling fluid, and the evaluation results of the rheological properties of the water-based drilling fluid is as follows:

[0111] (1) Evaluation of the filtration performance of the oil-based drilling fluid system: The filtration volume of the oil-based drilling fluid containing the polymer anti-multivalent cation treatment agent prepared by the solution method was measured using a medium-pressure filtration instrument. The specific process is as follows:

[0112] Pour 150 mL of prepared drilling fluid into the medium-pressure filtration meter, put the gasket and special filter paper in place, cover the lid and tighten the seal, and record the filtrate volume within 30 minutes at an output pressure of 0.69 MPa, which is the medium-pressure API filtration loss of the drilling fluid in mL.

[0113] (2) Evaluation of HTHP filtration performance of oil-based drilling fluid: Pour 150 mL of the prepared drilling fluid into the mud cup, install the special filter paper for testing HTHP filtration loss, seal it, and place it in a high-temperature and high-pressure filtration tester (MJ-71, Modern Petroleum Technology Development Co., Ltd.). After heating to 150°C, connect the upper pressure valve stem and the back pressure valve stem, keep the upper pressure at 4.2 MPa and the back pressure at 0.7 MPa, record the filtrate volume after filtration for 30 minutes, and multiply the filtrate volume by 2 to get the HTHP filtration loss.

[0114] (3) Evaluation of rheological properties of water-based drilling fluid: The reference data of rheological properties of drilling fluid include apparent viscosity, plastic viscosity, dynamic shear force and static shear force. After the water-based drilling fluid was stirred at a high-frequency speed of 12,000 rpm for 20 min, the values ​​of Φ600 and Φ300 were measured using a six-speed rotational viscometer in accordance with the standard GB / T16782-1997, as well as the values ​​of Φ3 after stirring at 600 rpm for 1 min and then standing for 10 s and 10 min respectively. The apparent viscosity (AV), plastic viscosity (PV) and dynamic shear force (YP) of the water-based drilling fluid were calculated using the following formula.

[0115] (a) Apparent viscosity AV (unit: mPa·s) = 0.511*Φ600

[0116] (b) Plastic viscosity PV (unit: mPa·s) = Φ600-Φ300

[0117] (c) Dynamic shear force (unit: Pa) YP=0.511*(Φ300-PV)

[0118] (d) Initial cutting (τ initial) = 0.511*Φ3 (Standing for 10s)

[0119] Final cut (τfinal) = 0.511*Φ3 (Standing for 10 minutes)

[0120] Embodiment 3: As attached Figure 5 As shown, the embodiment of the present invention discloses a wellbore stability control method for dynamically adjusting the drilling fluid ratio, which is a further optimization of the above embodiment, and further includes reversely optimizing the drilling fluid ratio control model group according to the actual implementation results after the drilling fluid ratio is dynamically adjusted, including:

[0121] Step S310, real-time monitoring and judging whether the drilling fluid performance or downhole key parameters are abnormal, wherein the drilling fluid performance or downhole key parameters are data of the dynamically adjusted drilling fluid used at the drilling site.

[0122] It should be noted that real-time monitoring and judgment of whether drilling fluid performance or downhole key parameters are abnormal is achieved by communicating with existing drilling monitoring equipment. After the drilling fluid and drilling parameters matching the optimal drilling fluid ratio obtained by the drilling fluid ratio control model group and dynamic adjustment processing according to this embodiment are put into use, the drilling monitoring equipment obtains real-time drilling fluid performance and downhole key parameters.

[0123] Step S320, if the response is yes, adjust the parameters of each model in the drilling fluid ratio control model group, and re-perform deep learning on the drilling fluid ratio control model group; if the response is no, return to step S310 to continue monitoring.

[0124] It should be understood that the process of adjusting the parameters of each model in the drilling fluid ratio control model group and re-performing deep learning on the drilling fluid ratio control model group in the embodiment of the present invention is the same as the corresponding model deep learning method in the previous text, and will not be repeated here.

[0125] Embodiment 4: As attached Figure 6 As shown, the embodiment of the present invention discloses a wellbore stability control device for dynamically adjusting the drilling fluid ratio, comprising:

[0126] The acquisition unit acquires current drilling environment data and real-time drilling data, wherein the drilling environment data includes drilling fluid performance data and geological data, and the real-time drilling data includes drilling pressure, rotation speed, and mud pump displacement.

[0127] The ratio determination unit inputs the current drilling environment data and the real-time drilling data into the drilling fluid ratio control model group to obtain the optimal drilling fluid ratio and drilling parameter combination, wherein the drilling fluid ratio control model group includes an associated prediction model and a collaborative optimization model. The associated prediction model is obtained by deep learning the neural network using a plurality of first samples, each of which includes the identification of the drilling fluid ratio and the drilling fluid performance data and geological data. The collaborative optimization model is obtained by deep learning the neural network using a plurality of second samples, each of which includes the identification of the optimal drilling fluid ratio and drilling parameter combination, and the combination of the drilling fluid ratio and any corresponding drilling parameter.

[0128] The ratio determination unit includes:

[0129] (1) Model training module. The drilling fluid ratio control model group includes an association prediction model and a collaborative optimization model. The deep learning methods and neural networks used in the association prediction model and the collaborative optimization model are the same, but the sample types used in the deep learning of the two models are different. The deep learning methods of the association prediction model include:

[0130] Acquire a number of first samples, and divide them into a training sample set, a verification sample set, and a test sample set according to a ratio, wherein each first sample includes an identifier of a drilling fluid ratio and drilling fluid performance data and geological data;

[0131] The multi-layer feedforward neural network is trained using the training set, and a loss function is introduced during the training. When the value of the loss function is stable, the training is terminated to obtain an associated prediction model, wherein the multi-layer feedforward neural network includes an input layer, a hidden layer, and an output layer. The input layer receives drilling fluid performance data and geological data as input data, the hidden layer includes one or more layers, nonlinearity is introduced through an activation function, the input data is processed and features are extracted, and the output layer outputs the predicted drilling fluid ratio;

[0132] The trained association prediction model is verified and tested using the validation set and test set respectively, the model parameters of the association prediction model are optimized, and an association prediction model that meets the test evaluation requirements is output.

[0133] It should be understood that in the embodiment of the present invention, a number of first samples and a number of second samples are used in the model training module, and the association prediction model and the collaborative optimization model are respectively trained by the same deep learning method described above. Therefore, the deep learning process of the collaborative optimization model will not be repeated here.

[0134] (2) The ratio optimization module inputs the current drilling environment data and real-time drilling data into the drilling fluid ratio control model group to obtain the optimal drilling fluid ratio and drilling parameter combination.

[0135] Specifically, the current drilling fluid performance data and geological data are input into the correlation prediction model to obtain the optimal drilling fluid ratio, and the optimal drilling fluid ratio and real-time drilling data are input into the collaborative optimization model to obtain the optimal drilling fluid ratio and drilling parameter combination.

[0136] Embodiment 5: As attached Figure 7 As shown, the embodiment of the present invention discloses a wellbore stability control device for dynamically adjusting the drilling fluid ratio, comprising:

[0137] The acquisition unit acquires current drilling environment data and real-time drilling data, wherein the drilling environment data includes drilling fluid performance data and geological data, and the real-time drilling data includes drilling pressure, rotation speed, and mud pump displacement.

[0138] The ratio determination unit inputs the current drilling environment data and the real-time drilling data into the drilling fluid ratio control model group to obtain the optimal drilling fluid ratio and drilling parameter combination, wherein the drilling fluid ratio control model group includes an associated prediction model and a collaborative optimization model. The associated prediction model is obtained by deep learning the neural network using a plurality of first samples, each of which includes the identification of the drilling fluid ratio and the drilling fluid performance data and geological data. The collaborative optimization model is obtained by deep learning the neural network using a plurality of second samples, each of which includes the identification of the optimal drilling fluid ratio and drilling parameter combination, and the combination of the drilling fluid ratio and any corresponding drilling parameter.

[0139] The dynamic ratio adjustment unit dynamically adjusts the drilling fluid ratio according to the optimal drilling fluid ratio and drilling parameter combination output by the drilling fluid ratio control model group, including:

[0140] Confirm whether the received drilling fluid ratio is complete and accurate and meets the requirements of the current drilling operation;

[0141] In response to this, the amount of various materials required for the drilling fluid is calculated according to the drilling fluid ratio and the current mud volume;

[0142] The drilling fluid performance test results are obtained, and the drilling fluid ratio is adjusted in reverse according to the performance test results until the performance test requirements are met.

[0143] The model reverse optimization unit, after dynamically adjusting the drilling fluid ratio, reversely optimizes the drilling fluid ratio control model group according to the actual implementation results, including:

[0144] Real-time monitoring and judgment of whether drilling fluid performance or downhole key parameters are abnormal, where drilling fluid performance or downhole key parameters are data of the dynamically adjusted drilling fluid used at the drilling site;

[0145] In response, the parameters of each model in the drilling fluid ratio control model group are adjusted, and the drilling fluid ratio control model group is deep-learned again.

[0146] Embodiment 6: The embodiment of the present invention discloses a storage medium, on which a computer program readable by a computer is stored, and the computer program is configured to execute a wellbore stability control method for dynamically adjusting the drilling fluid ratio when running.

[0147] The above storage medium may include, but is not limited to: a USB flash drive, a read-only memory, a mobile hard disk, a magnetic disk or an optical disk, and other media that can store computer programs.

[0148] Embodiment 7: The embodiment of the present invention discloses an electronic device, including a processor and a memory, wherein the memory stores a computer program, and the computer program is loaded and executed by the processor to implement a wellbore stability control method for dynamically adjusting the drilling fluid ratio.

[0149] The processor may be a central processing unit (CPU), a general purpose processor, a digital signal processor (DSP), an ASIC, an FPGA or other programmable logic device, a transistor logic device, a hardware component or any combination thereof. It may implement or execute various exemplary logic blocks, modules and circuits described in conjunction with the disclosure of the present invention. It may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, and the like. The memory may include, but is not limited to, various media that can store computer programs, such as a USB flash drive, a read-only memory, a mobile hard disk, a magnetic disk or an optical disk.

[0150] It should be understood by those skilled in the art that the embodiments of the present invention may provide methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes. The solutions in the embodiments of the present invention may be implemented in various computer languages, for example, object-oriented programming language Java and interpreted scripting language JavaScript, etc.

[0151] 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 flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, 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 process in the flowchart. 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.

[0152] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate 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 A function specified in one or more boxes.

[0153] The above content is only a specific implementation mode of the present invention, which has strong adaptability and implementation effect, but the protection scope of the present invention is not limited to this. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered within the protection scope of the present invention. Therefore, equivalent changes made according to the claims of the present invention are still within the scope covered by the present invention.

Claims

1. A wellbore stability control method for dynamically adjusting the drilling fluid ratio, characterized in that: include: Obtain current drilling environment data and real-time drilling data, where the drilling environment data includes drilling fluid performance data and geological data, and the real-time drilling data includes drilling pressure, rotation speed, and mud pump displacement; Input the current drilling environment data and real-time drilling data into the drilling fluid ratio control model group to obtain the optimal drilling fluid ratio and drilling parameter combination, wherein the drilling fluid ratio control model group includes an associated prediction model and a collaborative optimization model, the associated prediction model is obtained by deep learning the neural network using a plurality of first samples, each of which includes the identification of the drilling fluid ratio and the drilling fluid performance data and geological data, and the collaborative optimization model is obtained by deep learning the neural network using a plurality of second samples, each of which includes the identification of the optimal drilling fluid ratio and drilling parameter combination, and the combination of the drilling fluid ratio and any corresponding drilling parameter; Among them, the deep learning methods and neural networks used in the association prediction model and the collaborative optimization model are the same, but the sample types used in the deep learning of the two are different. The deep learning methods of the association prediction model include: Acquire a number of first samples, and divide them into a training sample set, a verification sample set, and a test sample set according to a ratio, wherein each first sample includes an identifier of a drilling fluid ratio and drilling fluid performance data and geological data; The multi-layer feedforward neural network is trained using the training set, and a loss function is introduced during the training. When the value of the loss function is stable, the training is terminated to obtain an associated prediction model, wherein the multi-layer feedforward neural network includes an input layer, a hidden layer, and an output layer. The input layer receives drilling fluid performance data and geological data as input data, the hidden layer includes one or more layers, nonlinearity is introduced through an activation function, the input data is processed and features are extracted, and the output layer outputs the predicted drilling fluid ratio; The trained association prediction model is verified and tested using the validation set and test set respectively, the model parameters of the association prediction model are optimized, and an association prediction model that meets the test evaluation requirements is output.

2. The wellbore stability control method for dynamically adjusting the drilling fluid ratio according to claim 1 is characterized in that: It also includes dynamically adjusting the drilling fluid ratio according to the optimal drilling fluid ratio and drilling parameter combination output by the drilling fluid ratio control model group, including: Confirm whether the received drilling fluid ratio is complete and accurate and meets the requirements of the current drilling operation; In response to this, the amount of various materials required for the drilling fluid is calculated according to the drilling fluid ratio and the current mud volume; The drilling fluid performance test results are obtained, and the drilling fluid ratio is adjusted in reverse according to the performance test results until the performance test requirements are met.

3. The wellbore stability control method for dynamically adjusting the drilling fluid ratio according to claim 2 is characterized in that: The drilling fluid performance test results include the evaluation results of the filtration performance of the oil-based drilling fluid system, the evaluation results of the HTHP filtration performance of the oil-based drilling fluid, and the evaluation results of the rheological properties of the water-based drilling fluid.

4. The wellbore stability control method for dynamically adjusting the drilling fluid ratio according to claim 3 is characterized in that: It also includes reverse optimization of the drilling fluid ratio control model group according to the actual implementation results after the drilling fluid ratio is dynamically adjusted, including: Real-time monitoring and judgment of whether drilling fluid performance or downhole key parameters are abnormal, where drilling fluid performance or downhole key parameters are data of the dynamically adjusted drilling fluid used at the drilling site; In response, the parameters of each model in the drilling fluid ratio control model group are adjusted, and the drilling fluid ratio control model group is deep-learned again.

5. A wellbore stability control device for dynamically adjusting the drilling fluid ratio using the method as claimed in any one of claims 1 to 4, characterized in that: include: An acquisition unit, which acquires current drilling environment data and real-time drilling data, wherein the drilling environment data includes drilling fluid performance data and geological data, and the real-time drilling data includes drilling pressure, rotation speed, and mud pump displacement; The ratio determination unit inputs the current drilling environment data and the real-time drilling data into the drilling fluid ratio control model group to obtain the optimal drilling fluid ratio and drilling parameter combination, wherein the drilling fluid ratio control model group includes an associated prediction model and a collaborative optimization model, the associated prediction model is obtained by deep learning the neural network using a plurality of first samples, each of which includes the identification of the drilling fluid ratio and the drilling fluid performance data and geological data, and the collaborative optimization model is obtained by deep learning the neural network using a plurality of second samples, each of which includes the identification of the optimal drilling fluid ratio and the drilling parameter combination, and the combination of the drilling fluid ratio and any corresponding drilling parameter; Wherein, the ratio determination unit includes: Model training module, drilling fluid ratio control model group includes association prediction model and collaborative optimization model. The deep learning methods and neural networks used in deep learning of association prediction model and collaborative optimization model are the same, but the sample types used in deep learning of the two are different. The deep learning methods of association prediction model include: Acquire a number of first samples, and divide them into a training sample set, a verification sample set, and a test sample set according to a ratio, wherein each first sample includes an identifier of a drilling fluid ratio and drilling fluid performance data and geological data; The multi-layer feedforward neural network is trained using the training set, and a loss function is introduced during the training. When the value of the loss function is stable, the training is terminated to obtain an associated prediction model, wherein the multi-layer feedforward neural network includes an input layer, a hidden layer, and an output layer. The input layer receives drilling fluid performance data and geological data as input data, the hidden layer includes one or more layers, nonlinearity is introduced through an activation function, the input data is processed and features are extracted, and the output layer outputs the predicted drilling fluid ratio; Use the validation set and test set to verify and test the trained association prediction model, optimize the model parameters of the association prediction model, and output an association prediction model that meets the test evaluation requirements; The ratio optimization module inputs the current drilling environment data and real-time drilling data into the drilling fluid ratio control model group to obtain the optimal drilling fluid ratio and drilling parameter combination.

6. The wellbore stability control device for dynamically adjusting the drilling fluid ratio according to claim 5 is characterized in that: Also includes: The dynamic ratio adjustment unit dynamically adjusts the drilling fluid ratio according to the optimal drilling fluid ratio and drilling parameter combination output by the drilling fluid ratio control model group, including: Confirm whether the received drilling fluid ratio is complete and accurate and meets the requirements of the current drilling operation; In response to this, the amount of various materials required for the drilling fluid is calculated according to the drilling fluid ratio and the current mud volume; Obtain the drilling fluid performance test results, and adjust the drilling fluid ratio in reverse according to the performance test results until the performance test meets the requirements; The model reverse optimization unit, after dynamically adjusting the drilling fluid ratio, reversely optimizes the drilling fluid ratio control model group according to the actual implementation results, including: Real-time monitoring and judgment of whether drilling fluid performance or downhole key parameters are abnormal, where drilling fluid performance or downhole key parameters are data of the dynamically adjusted drilling fluid used at the drilling site; In response, the parameters of each model in the drilling fluid ratio control model group are adjusted, and the drilling fluid ratio control model group is deep-learned again.

7. An electronic device, characterized in that: The method comprises a processor and a memory, wherein a computer program is stored in the memory, and the computer program is loaded and executed by the processor to implement the steps in the wellbore stability control method for dynamically adjusting the drilling fluid ratio as described in any one of claims 1 to 4.

8. A storage medium, characterized in that: The storage medium stores a computer program that can be read by a computer, and the computer program is configured to execute the steps of the wellbore stability control method for dynamically adjusting the drilling fluid ratio as described in any one of claims 1 to 4 when running.

Images