A drilling rate of penetration optimization method, system, device and medium based on MLP and efficient PSO

Abstract

A drilling mechanical drilling speed optimization method, system, device and medium based on MLP and efficient PSO, the method comprising: data acquisition; data preprocessing and feature extraction; establishing an MLP-based mechanical drilling speed ROP prediction model; training the model; establishing a prediction optimization model; predicting the model; evaluating and verifying the model; the system, device and medium are used to realize a drilling mechanical drilling speed optimization method based on MLP and efficient PSO; the MLP prediction model with high prediction accuracy is obtained through repeated training and testing; then the MLP model is further optimized by using the efficient PSO algorithm based on GPU, the drilling parameter combination that can make the mechanical drilling speed ROP as large as possible is found through the algorithm, and then the optimal drilling parameter combination is input into the trained MLP prediction model to obtain the optimal and most ideal mechanical drilling speed ROP; the drilling efficiency and accuracy can be greatly improved by applying the present application to the actual drilling process.

Description

Technical Field

[0001] This invention relates to the field of oil drilling technology, specifically to a drilling machinery drilling speed optimization design method based on MLP and high-efficiency PSO. Background Technology

[0002] As oil and gas exploration continues to expand into deeper, offshore, and unconventional areas, drilling difficulty and costs are constantly increasing, posing a significant challenge to drilling optimization. To address this challenge, drilling engineers are focusing on the Rate of Exploitation (ROP) of drilling machinery, as drilling costs are primarily determined by drilling time.

[0003] Optimizing the Rate of Drilling (ROP) involves two steps: a forward prediction model and a backward optimization model. The ROP prediction model is crucial for improving ROP and reducing drilling time. Researchers have evolved from traditional mathematical and geological information-based methods to applications based on machine learning and deep learning network frameworks, achieving numerous positive experimental results. Patent CN202011454345.6 describes a method for predicting and optimizing ROP in oil and gas drilling based on the CART algorithm. This method comprehensively considers the impact of eight drilling parameters on ROP, uses the CART algorithm to perform regression calculations on these parameters, and establishes a weighted model of the factors influencing ROP, thus enabling better prediction and optimization. However, in actual drilling, ROP is also affected by many other drilling parameters, including drilling pressure, rotational speed, and flow rate. How to establish intelligent data prediction and optimization models based on in-depth mining and analysis of historical drilling data, considering different formation conditions and cuttings characteristics, remains a pressing issue.

[0004] Multilayer Perceptron (MLP) neural network models in deep learning possess strong fitting capabilities and wide applicability. By optimizing their structural design and activation function selection, the performance and prediction accuracy of MLP models can be effectively improved. Particle Swarm Optimization (PSO) algorithms are characterized by simple design, strong global search capabilities, and fast convergence speed, and are commonly used to solve various complex optimization problems. Utilizing the particle-parallel optimization characteristics of PSO and combining it with a GPU architecture can significantly improve the performance of the PSO algorithm. Therefore, this invention proposes a drilling machinery drilling speed optimization design method based on MLP and efficient PSO. Through theoretical and practical case studies, the MLP-PSO model demonstrates high accuracy, reliability, and interpretability, providing a more reliable basis for optimizing drilling parameters in production, thereby facilitating better drilling operation planning and improving drilling efficiency. Summary of the Invention

[0005] In order to overcome the shortcomings of the prior art, the present invention aims to provide a drilling machinery drilling speed optimization method, system, equipment and medium based on MLP and high-efficiency PSO, so as to solve the problems of the existing drilling speed prediction and optimization intelligent algorithms being single, having low prediction accuracy and limited applicability.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A drilling machinery rate optimization method based on MLP and efficient PSO includes the following steps: Step 1: Data Acquisition of Historical Oil Drilling Parameters; Obtain the historical oil drilling parameters required for modeling, including well depth (DEPTH), bit rotation speed (RPMA), bit pressure (WOBA), mud flow rate (MFIA), mud density (MDIA), drilling fluid flow rate (HKLA), bit torque (TQA), drilling pump pressure (SPPA), mud penetration (MFOA), drilling fluid density (BDTI), bit speed (TVA), and mechanical drilling speed (ROP), and save the measured data in tables; Step 2: Preprocess and extract features from the historical oil drilling parameter data collected in Step 1. Remove or replace any unreasonable data with negative or null values ​​from the historical oil drilling parameter data collected in Step 1. To ensure all feature data are within the normal range, perform outlier deletion and replacement, nearest-neighbor filling, and outlier data format adjustment operations to normalize the data. Extract features from the preprocessed data. First, use the Random Forest algorithm to analyze the sensitivity of each drilling parameter and select highly sensitive feature parameters. Then, use the Pearson algorithm to detect the correlation between each feature parameter and the prediction target ROP, selecting highly correlated feature parameters for experimentation. Organize the preprocessed and feature-extracted drilling parameter data into a table as the dataset for subsequent model training and prediction. Step 3: Establish a drilling mechanical rate of penetration (ROP) prediction model based on MLP; using the preprocessed parameter data from Step 2 as input and the ROP as output, establish a drilling mechanical rate of penetration (ROP) prediction model based on MLP; the drilling mechanical rate of penetration (ROP) prediction model based on MLP is constructed using the multilayer perceptron (MLP) model framework, which consists of an input layer, a hidden layer, and an output layer; the model uses drilling parameter data as input and the ROP as output, the hidden layer uses the ReLU activation function, the output layer uses the tanh function, and the mean squared error is used as the loss function; Step 4: Train the prediction model obtained in Step 3. Use the drilling parameter data of one well as the training dataset, using 70% of the data as the training set for model training; use the remaining 30% of the data as the validation set for model tuning and selecting appropriate hyperparameters; evaluate the trained prediction model using the drilling parameter data of another well, compare the model's prediction results with the real data in the test set, and use three metrics—R², MSE, and relative error—to measure the model's performance, so as to obtain a high-accuracy MLP-based drilling mechanical rate of penetration (ROP) prediction model. Step 5: Establish a drilling mechanical drilling rate ROP prediction and optimization model based on MLP; optimize the above prediction model using the efficient PSO algorithm, set the condition for the PSO algorithm to end the iteration as reaching the specified accuracy, use the output of the MLP model obtained in step 4 as the objective function, and optimize the drilling parameter combination that maximizes the mechanical drilling rate ROP. Step 6: Input the drilling parameter combination obtained in Step 5 into Step 4 to obtain a highly accurate MLP prediction model and obtain the optimized drilling mechanical rate of penetration (ROP). Step 7: Compare the optimized ROP (Recovery Point Operating Rate) with the historical drilling parameters to determine if the set accuracy of 10 has been achieved. -6 The results are then compared to historical data to determine if there is any improvement. If the results do not meet the requirements, steps 5 and 6 are repeated until satisfactory results are obtained. If the given accuracy is met and there is an improvement, the final drilling mechanical rate of penetration (ROP) prediction optimization model based on MLP-PSO and the optimal combination of drilling parameters are obtained.

[0007] Step 5 includes the following two steps: Step 5.1 Establish a mechanical drilling rate of exploit (ROP) optimization model. Using the output of the MLP prediction model as the objective function, and limiting the range of values ​​for drill bit rotation speed (RPMA), drill bit pressure (WOBA), and mud flow rate (MFIA, where Q is the value), establish the optimization model equation that maximizes ROP. The objective equation F of the optimization model is as follows: (1) In equation (1), WOB represents the drill bit pressure, RPM represents the drill bit rotation speed, Q represents the mud flow rate, ROP represents the mechanical rotation speed, and N represents the amount of data input to the optimization equation; The three constraints are as follows: (2) (3) (4) Step 5.2 Use the efficient PSO algorithm to iteratively optimize the above optimization model.

[0008] The execution process of step 5.2 includes: Step 5.2.1 Data initialization; Step 5.2.2 Input the above optimization model and parameter constraints; Step 5.2.3 The particles update their velocity and position in parallel to find the optimal drilling parameters; Step 5.2.4 Calculate the mechanical drilling rate (ROP); Step 5.2.5 compares and obtains the optimal drilling parameter configuration and the optimal mechanical rate of penetration (ROP); If the termination condition is met in step 5.2.6, repeat steps 5.2.3-5.2.5. Step 5.2.6 Output the optimal drilling parameters and the optimal rate of penetration (ROP); In step 5.2.3, the formula for updating the particle's velocity and position is: (5) (6) In the formula, i = 1, 2, ..., M; d = 1, 2, ..., N. This represents the velocity of each particle. This indicates the current position of the particle. This represents the individual optimal position of the particle. ω represents the global optimal position of the particle swarm; ω is a non-negative number representing the inertia factor, the larger the value, the wider the range of particle leaps; C1 and C2 represent the acceleration factors of the particle swarm; r1 and r2 are any numbers within the range [0,1]; α is a constraint factor used to control the weight of the velocity.

[0009] A drilling machinery rate of penetration optimization system based on MLP and high-efficiency PSO includes: Data acquisition module: used for collecting historical parameter data of oil drilling; Data processing module: used to process and extract features from the acquired historical oil drilling parameter data; The model training module is used to train the established drilling speed prediction model based on the historical parameter dataset of oil drilling to obtain the trained drilling speed prediction model. The drilling speed prediction module is used to input real-time drilling data of the block to be predicted into the drilling speed prediction model for prediction, and obtain the mechanical drilling speed.

[0010] A processing device, comprising at least a processor and a memory, wherein the memory stores a computer program, and the processor, when running the computer program, implements a drilling machinery drilling speed optimization method based on MLP and high-efficiency PSO as described above.

[0011] A computer storage medium storing computer-readable instructions that can be executed by a processor to implement a drilling machinery rate of penetration optimization method based on MLP and high-efficiency PSO according to any one of the preceding claims.

[0012] Compared with the prior art, the advantages of the present invention are as follows: 1. High accuracy and efficiency in prediction: As a commonly used artificial neural network model, MLP can accurately predict the drilling speed of drilling machinery under different operating conditions; by utilizing the parallel computing characteristics of PSO and combining it with GPU architecture, it can quickly search for the global optimal solution in the process of optimizing the drilling speed of drilling machinery.

[0013] 2. Iterability and adaptability: The drilling speed optimization design method based on MLP and efficient PSO combines the advantages of MLP and PSO models, enabling it to continuously adjust the model and optimize parameters in a timely manner according to the actual working conditions of the drilling machinery, so as to adapt to the actual drilling speed requirements under various working conditions.

[0014] In summary, compared with existing technologies, this invention combines the machine learning algorithm MLP with the efficient intelligent optimization algorithm PSO, which enables high-precision prediction and optimization of the mechanical drilling rate (ROP) and drilling parameters in oil drilling, and has the advantages of significantly improving oil drilling efficiency and reducing drilling costs. Attached Figure Description

[0015] Figure 1 This is a flowchart of the method of the present invention.

[0016] Figure 2 A partial diagram illustrating the modeling process for the Random Forest algorithm.

[0017] Figure 3 This is a feature analysis diagram of the Pearson algorithm.

[0018] Figure 4 This is a graph showing the ROP (Recovery Point of Exploration) results for mechanical drilling speed based on MLP (Mechanical Drilling Rate).

[0019] Figure 5 This is a flowchart of an efficient PSO algorithm.

[0020] Figure 6 The figure shows the prediction and optimization results of ROP (Recovery Point Operation) based on MLP-PSO. Detailed Implementation

[0021] The technical solution of the present invention will now be described in detail with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0022] A drilling machinery rate optimization method based on MLP and efficient PSO includes the following steps: Step 1: Data Acquisition of Historical Oil Drilling Parameters; Obtain the historical oil drilling parameters required for modeling, including well depth (DEPTH), bit rotation speed (RPMA), bit pressure (WOBA), mud flow rate (MFIA), mud density (MDIA), drilling fluid flow rate (HKLA), bit torque (TQA), drilling pump pressure (SPPA), mud penetration (MFOA), drilling fluid density (BDTI), bit speed (TVA), and mechanical drilling speed (ROP), and save the measured data in tables; In step 1, historical drilling parameter data for two different oil wells in the same area were obtained, as shown in Table 1: 3156.2 60 11.4 4089.98 10.14 176.39 20.26 24.49 10.2 0.01 255.14 40.43 3156.4 60 11.1 4090.24 10.14 176.76 19.31 24.38 10.19 0.01 254.6 40.43 3156.6 60 11.3 4090.24 10.14 176.52 19.51 24.41 10.2 0.01 254.83 54.39 3156.8 61 12.8 4089.83 10.14 174.98 19.29 24.49 10.2 0.01 254.89 54.39 3157 60 11.3 4090.24 10.14 176.59 18.53 24.24 10.2 0.01 254.63 52.24 3157.2 58 12.8 4090.04 10.14 175.06 22.2 24.59 10.2 0.01 254.81 52.24 3157.4 59 12.3 4090.07 10.14 175.52 20.43 24.51 10.19 0.01 254.17 52.24 3157.6 61 11.8 4090.04 10.14 176.05 19.18 24.61 10.2 0.01 254.72 41.19 3157.8 59 12.5 4090.24 10.14 175.28 21.56 24.66 10.19 0.01 254.96 41.19 Table 1 Table 1 shows some historical drilling parameter data. As can be seen from the table, the data distribution is messy and irregular, so the raw data needs to be preprocessed. Step 2, see Figure 2 The process involves preprocessing and feature extraction of the historical oil drilling parameter data collected in step 1; removing or replacing unreasonable data with negative or null values ​​from the historical oil drilling parameter data collected in step 1; and standardizing the data by performing outlier deletion and replacement, neighboring value filling, and outlier data format adjustment to ensure that all feature data are within the normal range. Feature extraction is then performed on the preprocessed data. The random forest algorithm is first used to analyze the sensitivity of each drilling parameter, and highly sensitive feature parameters are selected to analyze the contribution score of each feature to the prediction target. The higher the score, the higher the sensitivity to the prediction target. The modeling process of the random forest algorithm is then described. See Figure 3 Next, the Pearson algorithm is used to analyze the correlation between the feature parameters and the prediction target. Figure 3 The darker the color, the greater the correlation. Analysis shows that the 11 selected feature parameters are all highly correlated with the predicted target ROP and can be used as subsequent model training and prediction data. Step 3: Establish a drilling mechanical rate of penetration (ROP) prediction model based on MLP; using the preprocessed parameter data from Step 2 as input and the ROP as output, establish a drilling mechanical rate of penetration (ROP) prediction model based on MLP; the drilling mechanical rate of penetration (ROP) prediction model based on MLP is constructed using a multilayer perceptron (MLP) model framework, which consists of an input layer, a hidden layer, and an output layer; the model uses drilling parameter data as input and the ROP as output, the hidden layer uses the ReLU activation function, the output layer uses the tanh function, and the mean squared error is used as the loss function; A mechanical drilling rate of exploit (ROP) prediction model based on MLP was established. The model consists of four layers, and the specific parameters of each layer are shown in Table 2.

[0023] Table 2 Table 2 shows the parameters of each layer in the MLP prediction model. As can be seen from the table, we have constructed a 4-layer MLP model with an input dimension of 11 and an output dimension of 1. The dimensions of each hidden layer are 1024, 512, 256, 256, and 128, respectively, and the activation function is ReLU. Step 4, see Figure 4 The drilling prediction model obtained in step 3 was trained using drilling parameter data from one well as the training dataset. 70% of the data was used as the training set for model training, while the remaining 30% was used as the validation set for model tuning and selecting appropriate hyperparameters. The model was trained 2000 times. The training prediction model was then evaluated using drilling parameter data from another well. The model's prediction results were compared with the real data in the test set. R², MSE, and relative error were used to measure the model's performance. The results showed an R² of 0.91, an MSE of 11.78, and a relative error of only 2.8%. These results indicate that this MLP prediction model can effectively characterize the intrinsic relationship between drilling engineering parameters and mechanical drilling rate (ROP), achieving an accuracy improvement of approximately 46% compared to existing technologies. Step 5: Establish an MLP-based drilling mechanical rate of penetration (ROP) prediction and optimization model; optimize the above prediction model using the efficient PSO algorithm, setting the PSO algorithm iteration termination condition to reach a specified accuracy, and using the output of the MLP model obtained in Step 4 as the objective function to optimize and search for the drilling parameter combination that maximizes the mechanical rate of penetration (ROP); Step 5 includes the following two steps: 5.1 Establish a mechanical drilling rate of exploit (ROP) optimization model. Using the output of the MLP prediction model as the objective function, and limiting the ranges of drill bit rotation speed (RPMA), drill bit pressure (WOBA), and mud flow rate (MFIA, where Q is the value), an optimization model equation maximizing ROP is established. The objective equation F of the optimization model is as follows: (1) In equation (1), WOB represents drill bit pressure, RPM represents drill bit rotation speed, Q represents mud flow rate, ROP represents machine rotation speed, and N represents the amount of data input to the optimization equation. The three constraints are as follows: (2) (3) (4) 5.2 See also Figure 5 The above optimization model is iteratively optimized using an efficient PSO algorithm. The steps of this process are as follows: Figure 5 In this embodiment, the optimization period is 100m, and the optimization results are as follows: Figure 6 ;include: 5.2.1 Data initialization; 5.2.2 Input the above optimization model and parameter constraints; 5.2.3 Particles update their velocity and position in parallel to find the optimal drilling parameters; 5.2.4 Calculate the mechanical drilling rate of return (ROP); 5.2.5 The optimal drilling parameter configuration and the optimal mechanical rate of penetration (ROP) are obtained through comparison; 5.2.6 If the termination condition is met, repeat steps S523-S525; 5.2.6 Output optimal drilling parameters and optimal rate of penetration (ROP); In section 5.2.3, the formulas for updating the particle's velocity and position are as follows: (5) (6) In the formula, i = 1, 2, ..., M; d = 1, 2, ..., N. This represents the velocity of each particle. This indicates the current position of the particle. This represents the individual optimal position of the particle. ω represents the global optimal position of the particle swarm; ω is a non-negative number representing the inertia factor, the larger the value, the wider the range of particle leaps; C1 and C2 represent the acceleration factors of the particle swarm; r1 and r2 are any numbers within the range [0,1]; α is a constraint factor used to control the weight of the velocity. Step 6, see Figure 6 The drilling parameter combination optimized in step 5 is then input into step 4 to obtain a highly accurate MLP prediction model, resulting in the optimized drilling mechanical rate of penetration (ROP). Step 7: Next, compare the optimized drilling mechanical rate of penetration (ROP) with the data in the historical drilling parameters to determine whether the set accuracy of 10 has been achieved. -6 The results are then compared to historical data to determine if there is any improvement. If the results do not meet the requirements, steps 5 and 6 are repeated until satisfactory results are obtained. If the given accuracy is met and there is an improvement, the final drilling mechanical rate of penetration (ROP) prediction optimization model based on MLP-PSO and the optimal combination of drilling parameters are obtained.

[0024] As can be seen from the embodiments, compared with the prior art, the present invention adopts a drilling mechanical drilling rate optimization method based on MLP and high-efficiency PSO, which can accurately and efficiently predict and optimize the mechanical drilling rate ROP. The mechanical drilling rate ROP is improved by 162% compared with historical data, and the rate of real-time optimization of drilling parameters is improved by about 600%.

Claims

1. A drilling machinery rate of penetration optimization method based on MLP and high-efficiency PSO, characterized in that, Includes the following steps: Step 1: Data Acquisition of Historical Oil Drilling Parameters; Acquire historical oil drilling parameters required for modeling, including well depth (DEPTH), bit rotation speed (RPM), bit pressure (WOB), mud flow rate (MFIA), mud density (MDIA), drilling fluid flow rate (HKLA), bit torque (TQA), drilling pump pressure (SPPA), mud permeability (MFOA), drilling fluid density (BDTI), bit speed (TVA), and mechanical drilling speed (ROP), and save the measured data in tables; Step 2: Preprocess and extract features from the historical oil drilling parameter data collected in Step 1; remove or replace some unreasonable data with negative or empty values ​​in the historical oil drilling parameter data collected in Step 1; in order to ensure that all feature data are within the normal range, perform outlier deletion and replacement, neighboring value filling and outlier data format adjustment operations to standardize the data. Feature extraction is performed on the preprocessed data. First, the sensitivity of each drilling parameter is analyzed using the random forest algorithm to select the feature parameters with higher sensitivity. Then, the Pearson algorithm is used to detect the correlation between each feature parameter and the prediction target ROP, so as to select the feature parameters with high correlation for experimentation. The drilling parameter data after data preprocessing and feature extraction are organized into a table as the dataset for subsequent model training and prediction. Step 3: Establish a drilling mechanical rate of penetration (ROP) prediction model based on Multilayer Perceptron (MLP). Using the preprocessed parameter data from Step 2 as input and the ROP as output, establish an MLP-based ROP prediction model. The MLP-based ROP prediction model is constructed using a multilayer perceptron (MLP) framework, consisting of an input layer, a hidden layer, and an output layer. The model uses drilling parameter data as input and the ROP as output. The hidden layer uses the ReLU activation function, the output layer uses the tanh function, and the mean squared error is used as the loss function. Step 4: Train the MLP-based ROP prediction model obtained in Step 3. Use the drilling parameter data of one well as the training dataset, using 70% of the data as the training set for model training; use the remaining 30% as the validation set for model tuning and selecting appropriate hyperparameters; evaluate the trained prediction model using the drilling parameter data of another well, comparing the model's prediction results with the real data in the test set, and using three metrics—R², MSE, and relative error—to measure the model's performance, thereby obtaining a high-accuracy MLP-based ROP prediction model. Step 5: Establish a drilling mechanical drilling rate ROP prediction and optimization model based on MLP; optimize the above prediction model using the efficient PSO algorithm, set the condition for the PSO algorithm to end the iteration as reaching the specified accuracy, use the output of the MLP model obtained in step 4 as the objective function, and optimize the drilling parameter combination that maximizes the mechanical drilling rate ROP. Step 5 includes the following two steps: Step 5.1 Establish a mechanical drilling rate of exploit (ROP) optimization model. Using the output of the MLP prediction model as the objective function, and limiting the ranges of drill bit rotation speed (RPM), drill bit pressure (WOB), and mud flow rate (MFIA) as constraints, establish the optimization model equation that maximizes ROP. The objective equation F of the optimization model is as follows: （1） In equation (1), WOB represents drill bit pressure, RPM represents drill bit rotation speed, Q represents mud flow rate, ROP represents machine rotation speed, and N represents the amount of data input to the optimization equation; the three constraints are as follows: （2） （3） （4） Step 5.2 Use the efficient PSO algorithm to iteratively optimize the above optimization model; Step 6: Input the drilling parameter combination obtained in Step 5 into Step 4 to obtain a highly accurate MLP prediction model and obtain the optimized drilling mechanical rate of penetration (ROP). Step 7: Compare the optimized ROP (Recovery Point Operating Rate) with the historical drilling parameters to determine if the set accuracy of 10 has been achieved. -6 The results are then compared to historical data to determine if there is any improvement. If the results do not meet the requirements, steps 5 and 6 are repeated until satisfactory results are obtained. If the given accuracy is met and there is an improvement, the final drilling mechanical rate of penetration (ROP) prediction optimization model based on MLP-PSO and the optimal combination of drilling parameters are obtained.

2. The drilling machinery rate optimization method based on MLP and high-efficiency PSO according to claim 1, characterized in that: The execution process of step 5.2 includes: Step 5.2.1 Data initialization; Step 5.2.2 Input the above optimization model and parameter constraints; Step 5.2.3 The particles update their velocity and position in parallel to find the optimal drilling parameters; Step 5.2.4 Calculate the mechanical drilling rate (ROP); Step 5.2.5 compares and obtains the optimal drilling parameter configuration and the optimal mechanical rate of penetration (ROP); If the termination condition is met in step 5.2.6, repeat steps S523-S525. Step 5.2.6 Output the optimal drilling parameters and the optimal rate of penetration (ROP); In section 5.2.3, the formulas for updating the particle's velocity and position are as follows: （5） （6） In the formula, i = 1, 2, ..., M; d = 1, 2, ..., N. This represents the velocity of each particle. This indicates the current position of the particle. This represents the individual optimal position of the particle. ω represents the global optimal position of the particle swarm; ω is a non-negative number representing the inertia factor, the larger the value, the wider the range of particle leaps; C1 and C2 represent the acceleration factors of the particle swarm; r1 and r2 are any numbers within the range [0,1]; α is a constraint factor used to control the weight of the velocity.

3. A drilling machinery rate of penetration optimization system based on MLP and high-efficiency PSO, used to implement the method of claim 1, characterized in that, include: Data acquisition module: used for collecting historical parameter data of oil drilling; Data processing module: used to process and extract features from the acquired historical parameters of oil drilling; The model training module is used to train the established drilling rate prediction model based on the historical parameter dataset of oil drilling to obtain the trained drilling rate prediction model. The drilling speed prediction module is used to input real-time drilling data of the block to be predicted into the drilling speed prediction model for prediction, and obtain the mechanical drilling speed.

4. A processing apparatus, the processing apparatus comprising at least a processor and a memory, characterized in that, The memory stores a computer program, and when the processor runs the computer program, it implements the drilling machinery drilling speed optimization method based on MLP and high-efficiency PSO as described in any one of claims 1 to 2.

5. A computer storage medium, characterized in that, The storage medium stores computer-readable instructions that can be executed by a processor to implement a drilling machinery speed optimization method based on MLP and high-efficiency PSO according to any one of claims 1 to 2.

Images