Drilling rate modeling method based on multi-distance generation strategy and improved snow ablation optimization algorithm

Abstract

The invention relates to the field of deep geological exploration, in particular to a drilling rate modeling method based on a multi-distance generation strategy and an improved snow ablation optimization algorithm. Based on data obtained in an actual drilling site, new data equal to the obtained data sample number is generated according to a multi-distance data generation method, the new data is randomly divided into a plurality of parts, each part of data is combined with data in a training set, and a modeling precision result is obtained according to a verification set; based on a multi-metric data screening method, selecting one part of new data corresponding to the optimal precision, and adding the new data into a training set; and constructing a drilling speed model by adopting a support vector regression method, and optimizing and adjusting parameters of the drilling speed model by adopting an improved snow ablation optimization algorithm with improved weight, direction enhancement and a Levy flight-based disturbance strategy to obtain a final drilling speed model. According to the method, the problem of insufficient precision caused by insufficient data samples can be effectively solved, then a reliable drilling speed model is established, and a foundation is laid for drilling process control.

Description

Technical Field

[0001] The invention relates to the field of geological exploration, and in particular to a drilling speed modeling method based on a multi-distance generation strategy and an improved snow melting optimization algorithm. Background Art

[0002] The significance of drilling speed modeling in deep geological drilling is particularly important, especially when facing complex formations and rock mechanical environments. As one of the key indicators in the drilling process, drilling speed directly affects the efficiency and cost of drilling. Therefore, accurately predicting and controlling drilling speed not only helps to optimize the time and resources of drilling operations, but also effectively reduces equipment wear and energy consumption, thereby improving the economic benefits of the entire exploration project.

[0003] With the continuous development of deep drilling technology, the challenges faced by the drilling process are becoming more and more complex. In the geological environment of high pressure, high temperature and high stress, the traditional drilling rate modeling method faces insufficient data, and the high-dimensional changes in model parameters caused by the complex nonlinear changes in the downhole environment further aggravate the difficulty of modeling. Therefore, it is necessary to establish a reliable drilling rate model for the deep complex geological environment to effectively deal with the modeling difficulties in complex environments. Summary of the invention

[0004] In order to solve the problems of insufficient data samples and high-dimensional changes in model parameters faced in the deep geological drilling process, the present invention provides a drilling speed modeling method based on a multi-distance generation strategy and an improved snow melting optimization algorithm. Aiming at the deep complex geological environment, a suitable and reliable drilling speed modeling method is designed to effectively improve the drilling speed modeling accuracy and provide a reliable drilling speed model for the optimization control of the subsequent drilling process. The method mainly includes the following steps:

[0005] S1: Based on the data obtained at the actual drilling site, new data with the same number of data samples as the obtained data is generated according to the multi-distance data generation method, which is used to establish the drilling speed model;

[0006] S2: The generated new data is randomly divided into several parts, and the data obtained from the actual drilling site is divided into a training set, a validation set, and a test set. Each new data is combined with the data in the training set to train the drilling speed model. According to the modeling accuracy results obtained from the validation set, a multi-metric data screening method is used to select one new data corresponding to the best accuracy and add it to the training set;

[0007] S3: The support vector regression method is used to construct the drilling rate model. According to the improved snow melting optimization algorithm with improved weight, directional enhancement and perturbation strategy based on Levy flight, the parameters of the drilling rate model are optimized and adjusted to obtain the final drilling rate model.

[0008] A computer device includes a memory, a processor and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the drilling speed modeling method based on the multi-distance generation strategy and the improved snow melting optimization algorithm.

[0009] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the drilling speed modeling method based on a multi-distance generation strategy and an improved snow melting optimization algorithm.

[0010] A computer program product includes a computer program or an instruction. When the program or the instruction is executed by a processor, the steps of the drilling speed modeling method based on the multi-distance generation strategy and the improved snow melting optimization algorithm are implemented.

[0011] The technical solution provided by the present invention has the following beneficial effects: the present invention can effectively expand the drilling data sample through the designed multi-distance data generation method; through the designed multi-metric data screening method, it can effectively screen out reliable data that improves the model accuracy to expand the sample; based on the drilling speed model of support vector regression, the present invention designs an improved snow melting optimization algorithm with improved weights, directional enhancement and perturbation strategy based on Levy flight, which can determine the optimal values ​​of the drilling speed model parameters based on support vector regression. The present invention effectively solves the problem of high-dimensional parameter changes faced by drilling speed modeling, realizes the effective improvement of drilling speed accuracy, and provides a good foundation for the optimization and control of the subsequent drilling process. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The present invention will be further described below with reference to the accompanying drawings and embodiments, in which:

[0013] Figure 1 It is a structural diagram of a drilling speed modeling method based on a multi-distance generation strategy and an improved snow melting optimization algorithm in an embodiment of the present invention;

[0014] Figure 2 : is a simulation effect diagram of IEEE CEC2017 test function 9 in an embodiment of the present invention;

[0015] Figure 3 1 is a simulation effect diagram of IEEE CEC2017 test function 12 in an embodiment of the present invention;

[0016] Figure 4 It is a comparison diagram between the drilling speed modeling result and the actual drilling speed in the embodiment of the present invention. DETAILED DESCRIPTION

[0017] In order to have a clearer understanding of the technical features, purposes and effects of the present invention, specific embodiments of the present invention are now described in detail with reference to the accompanying drawings.

[0018] Example 1

[0019] Please refer to Figure 1 , Figure 1 : is a structural diagram of a drilling speed modeling method based on a multi-distance generation strategy and an improved snow melting optimization algorithm in an embodiment of the present invention. The method specifically includes the following steps:

[0020] S1: Based on the data obtained from the actual drilling site, a multi-distance data generation method is designed to generate new data with the same number of acquired data samples. The specific implementation process of S1 is as follows:

[0021] S1.1: The data obtained at the actual drilling site include drilling pressure, rotation speed, pump volume, density, drilling speed and depth. Calculate the sample size N of the actual drilling data.

[0022] S1.2: Use conditional generative adversarial networks to generate drilling data with a sample size of 2N. Based on Mahalanobis distance, Euclidean distance and Chebyshev distance, a new distance factor is determined as a measurement factor. The details are as follows:

[0023]

[0024] in, is the normalized Mahalanobis distance, is the normalized Chebyshev distance, Normalized Euclidean distance.

[0025] S1.3: Based on metric factors According to their size, the first N groups of samples are selected from large to small as the generated new data.

[0026] S2: randomly divide the generated new data into 10 parts, combine each part of the data with the data in the training set, test the accuracy of the model in the validation set, and design a multi-metric data screening method based on the accuracy results to select the best 1 part of the generated data into the training set. The specific implementation process of S2 is as follows:

[0027] S2.1: Randomly divide the newly generated N groups of data into 10 parts, and the number of samples in each part is 0.1*N.

[0028] S2.2: The N sets of data obtained from the actual drilling site are divided into training set, validation set and test set in a ratio of 8:1:1. The 10 sets of generated new data are respectively integrated with the training set data to train the drilling speed model based on support vector regression. For the model trained with the combined data each time, a multi-metric data screening factor is used to evaluate the modeling accuracy of the model in the validation set. The multi-metric data screening factor The specific calculation formula is as follows:

[0029]

[0030] in, is the root mean square error of the k-th training model in the validation set, is the maximum absolute error of the model trained for the kth time in the validation set, The mean absolute error of the model trained for the kth time on the validation set.

[0031] S2.3: Select the smallest The corresponding new data is used as the screened data and integrated into the training set.

[0032] 600 sets of data from a geological exploration drilling site were collected for modeling and analysis of the drilling process. The actual field data were divided into training set, validation set and test set in a ratio of 8:1:1. First, the multi-distance data generation method was used to generate 600 sets of new data, and the new data were randomly divided into 10 parts. Each piece of data was added to the training set, and the multi-metric data screening method was used to screen out the best 1 piece of data and add it to the training set.

[0033] S3: The support vector regression method is used to construct a drilling rate model, and an improved snow melting optimization algorithm with improved weight, directional enhancement and perturbation strategy based on Levy flight is designed to optimize and adjust the parameters of the support vector regression drilling rate model to obtain the final drilling rate model. The specific implementation process of S3 is as follows:

[0034] S3.1: Improved weights, as follows:

[0035]

[0036] in, is a random value between 0 and 1, ν is a random value that satisfies the normal distribution, t is the current iteration number, t max is the maximum number of iterations, ω sem is the final improved weight.

[0037] S3.2: Directional enhancement, as follows:

[0038]

[0039] in, is the position of the index1th individual at the tth iteration, is the speed of the index1th individual at the tth iteration, and are all random values ​​between 0 and 1, E li (·) is the elite pool, which contains four elite individuals, ξ is a random integer between 1 and 4, X *is the global optimal solution, and F(·) is the fitness function value.

[0040] S3.3: The perturbation strategy based on Levy flight is as follows:

[0041]

[0042] Among them, X chaos is the individual after Levy flight disturbance, θ is a constant, is the position of the i-th individual at the t-th iteration, and levy(·) is the perturbation based on the Levy flight.

[0043] The calculation of γ and e is as follows:

[0044] γ=4*γ*(1-γ)

[0045] e=2*κ(1,d)-1

[0046] Among them, κ(1,d) is a d-dimensional vector, and each element in the vector is a random number between 0 and 1.

[0047] Based on the multi-distance generation strategy and the improved snow melting optimization algorithm, the drilling speed model was constructed in combination with the support vector regression method. Before modeling, the algorithm performance was analyzed using test functions 9 and 12 in the IEEE CEC2017 benchmark test function. The experimental results are shown in Figure 2. Figure 2 and Figure 3 shown. Figure 2 and Figure 3 The results in the paper show that the improved snow melting optimization algorithm can converge to the optimal value in the early iteration, indicating that it has good global search and can effectively determine the optimal value of the model parameters. 600 sets of data from a geological exploration drilling site were collected for modeling analysis, of which 60 sets of data were used for test comparison. The comparison results are shown in the figure. Figure 4 The comparison results show that the drilling speed model after sample expansion can more accurately follow the changes in drilling speed, indicating that the generated samples can effectively expand the sample size and diversity of the data, which is conducive to dealing with the drilling speed modeling problem when the data is relatively insufficient.

[0048] Example 2

[0049] A computer device includes a memory, a processor and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the drilling speed modeling method based on the multi-distance generation strategy and the improved snow melting optimization algorithm.

[0050] Example 3

[0051] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the drilling speed modeling method based on a multi-distance generation strategy and an improved snow melting optimization algorithm.

[0052] Example 4

[0053] A computer program product includes a computer program or an instruction. When the program or the instruction is executed by a processor, the steps of the drilling speed modeling method based on the multi-distance generation strategy and the improved snow melting optimization algorithm are implemented.

[0054] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A drilling speed modeling method based on a multi-distance generation strategy and an improved snow melting optimization algorithm, characterized in that: The steps include: S1: Based on the data obtained at the actual drilling site, new data with the same number of data samples as the obtained data is generated according to the multi-distance data generation method, which is used to establish the drilling speed model; S2: The generated new data is randomly divided into several parts, and the data obtained from the actual drilling site is divided into a training set, a validation set, and a test set. Each new data is combined with the data in the training set to train the drilling speed model. According to the modeling accuracy results obtained from the validation set, a multi-metric data screening method is used to select one new data corresponding to the best accuracy and add it to the training set; S3: The support vector regression method is used to construct the drilling rate model. According to the improved snow melting optimization algorithm with improved weight, directional enhancement and perturbation strategy based on Levy flight, the parameters of the drilling rate model are optimized and adjusted to obtain the final drilling rate model.

2. A drilling speed modeling method based on a multi-distance generation strategy and an improved snow melting optimization algorithm as claimed in claim 1, characterized in that: In S1, the data obtained at the actual drilling site include drilling pressure, rotation speed, pump volume, density, drilling speed and depth.

3. The drilling speed modeling method based on multi-distance generation strategy and improved snow melting optimization algorithm as claimed in claim 1, characterized in that: In S1, the process of generating new data with the same number of acquired data samples according to the multi-distance data generation method is as follows: S1.1: Calculate the sample size N of actual drilling data; S1.2: Use conditional generative adversarial networks to generate drilling data with a sample size of 2N. Based on Mahalanobis distance, Euclidean distance and Chebyshev distance, a new distance factor is determined as a measurement factor. in, is the normalized Mahalanobis distance, is the normalized Chebyshev distance, Normalized Euclidean distance; S1.3: Based on the measurement factor The size of , select the first N groups of samples from large to small as the generated new data.

4. The drilling speed modeling method based on multi-distance generation strategy and improved snow melting optimization algorithm as claimed in claim 1, characterized in that: The specific implementation process of S2 is: S2.1: Randomly divide the generated N sets of new data into several parts, and the number of samples in each part is 0.1*N; S2.2: The N sets of data obtained from the actual drilling site are divided into training set, validation set and test set according to the set ratio; the several sets of new data obtained in S2.1 are combined with the training set data to train the drilling speed model. For the model trained with the combined data each time, the multi-metric data screening factor is used to evaluate the modeling accuracy of the model in the validation set. The multi-metric data screening factor The details are as follows: in, is the root mean square error of the k-th training model in the validation set, is the maximum absolute error of the model trained for the kth time in the validation set, The mean absolute error of the model trained for the kth time in the validation set; S2.3: Select the smallest The corresponding new data is used as the screened data and integrated into the training set.

5. The drilling speed modeling method based on multi-distance generation strategy and improved snow melting optimization algorithm as claimed in claim 1, characterized in that: In S3, the improved weights are: oh sem =0.75+0.15*θ1+0.4*ν+0.25*((2*tt max )-(tt max ) 2 ) Among them, θ1 is a random value between 0 and 1, ν is a random value that satisfies the normal distribution, t is the current iteration number, and t max is the maximum number of iterations, ω sem is the final improved weight.

6. The drilling speed modeling method based on multi-distance generation strategy and improved snow melting optimization algorithm as claimed in claim 1, characterized in that: In S3, the direction is enhanced as follows: in, is the position of the index1th individual at the tth iteration, is the speed of the index1th individual at the tth iteration, θ2, θ3 and θ4 are all random values ​​between 0 and 1, E li (·) is the elite pool, which contains four elite individuals, ξ is a random integer between 1 and 4, X * is the global optimal solution, and F(·) is the fitness function value.

7. The drilling speed modeling method based on multi-distance generation strategy and improved snow melting optimization algorithm as claimed in claim 1, characterized in that: In S3, the perturbation strategy based on Levy flight is: Among them, X chaos is the individual after Levy flight disturbance, θ is a constant, is the position of the i-th individual at the t-th iteration, levy(·) is the perturbation based on the Levy flight; The calculation of γ and e is as follows: γ=4*γ*(1-γ) e=2*κ(1,d)-1 Among them, κ(1,d) is a d-dimensional vector, and each element in the vector is a random number between 0 and 1.

8. A computer device comprising a memory, a processor and a computer program stored in the memory, characterized in that: The processor executes the computer program to implement the steps of the drilling speed modeling method based on the multi-distance generation strategy and the improved snow ablation optimization algorithm as described in any one of claims 1 to 7.

9. A computer-readable storage medium, characterized in that: A computer program is stored, and when the program is executed by a processor, the steps of the drilling speed modeling method based on a multi-distance generation strategy and an improved snow ablation optimization algorithm as described in any one of claims 1 to 7 are implemented.

10. A computer program product, characterized in that The method comprises a computer program or an instruction, which, when executed by a processor, implements the steps of the drilling speed modeling method based on a multi-distance generation strategy and an improved snow ablation optimization algorithm as described in any one of claims 1 to 7.

Images