Drilling rate modeling method based on improved bat algorithm and support vector regression

Abstract

The invention relates to the field of geological exploration, in particular to a drilling speed modeling method based on an improved bat algorithm and support vector regression. The method comprises the following steps: firstly, identifying and correcting abnormal data according to an interval anomaly detection method; then, a support vector regression method is adopted, and a drilling speed model is constructed; and finally, designing an improved bat algorithm, and determining the optimal parameter value of the drilling speed model. By the adoption of the drilling speed modeling method, drilling abnormal data can be effectively removed, drilling characteristics can be guaranteed, meanwhile, a high-precision drilling speed model can be constructed, accurate prediction of the drilling speed is achieved, and a reliable basis is provided for operation and adjustment of the drilling process.

Description

Technical Field

[0001] The invention relates to the field of geological exploration, and in particular to a drilling speed modeling method based on an improved bat algorithm and support vector regression. Background Art

[0004] In the process of developing geological drilling process control technology, due to the complex formations encountered in deep geological drilling, the rock types are varied, and there are complex mechanical environments with high geotemperature, high osmotic pressure, high geostress and mining disturbances, which leads to abnormal data in the drilling process. At the same time, the complex external rock mechanics environment makes the entire drilling process complex and nonlinear, which further leads to high-dimensional changes in model parameters. These nonlinear and high-dimensional change problems make it difficult to build a high-precision drilling speed model, and also make it impossible to provide effective guidance for subsequent drilling process operations. Summary of the invention

[0005] In order to solve the problem that it is difficult to construct a high-precision drilling speed prediction model due to the nonlinear and high-dimensional changes in the drilling process, the present invention provides a drilling speed modeling method based on an improved bat algorithm and support vector regression, which mainly includes:

[0006] S1: Analyze the interval of drilling data, identify outliers in the data, and replace them with the nearest normal data;

[0007] S2: Based on the replaced drilling sample data, a drilling speed prediction model is constructed based on the support vector regression method. The model takes the current drilling pressure, rotation speed, depth and drilling speed as input and the drilling speed at the next moment as output;

[0008] S3: Design an improved bat algorithm based on partial population update, iterative search and chaotic perturbation mechanism to optimize the hyperparameters of the drilling speed model, and use the optimized drilling speed model to accurately predict the drilling speed at the next moment.

[0009] 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 improved bat algorithm and support vector regression.

[0010] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the above-mentioned drilling speed modeling method steps based on an improved bat algorithm and support vector regression.

[0011] 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 improved bat algorithm and support vector regression are implemented.

[0012] The beneficial effects brought about by the technical solution provided by the present invention are as follows: the present invention takes into account the existence of abnormal data in the drilling data, and starting from the identification of abnormal data, effectively identifies the abnormal data, and replaces the abnormal data with normal data. In this way, the abnormal drilling data can be effectively eliminated to ensure the drilling characteristics. The data characteristics and model accuracy are guaranteed to a certain extent. At the same time, the nonlinear key existing in the drilling data is considered, and the support vector regression method is used to construct the drilling speed model. In the process of constructing the drilling speed model, the present invention designs an improved bat algorithm, designs the parameters of the model, determines the optimal model parameter values, realizes the effective prediction of the drilling speed, and effectively improves the modeling accuracy of the model. By constructing a high-precision and high-reliability drilling speed model, it is helpful to clarify the complex relationship between drilling parameters and drilling speed, and provide effective guidance for the subsequent operation of the drilling process. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] The present invention will be further described below with reference to the accompanying drawings and embodiments, in which:

[0014] Figure 1 It is a structural diagram of a drilling speed modeling method based on an improved bat algorithm and support vector regression in an embodiment of the present invention;

[0015] Figure 2 It is a simulation effect diagram of IEEE CEC2005 test function 1 in an embodiment of the present invention;

[0016] Figure 3 It is a simulation effect diagram of IEEE CEC2005 test function 2 in an embodiment of the present invention;

[0017] Figure 4 1 is a comparison chart of the drilling speed prediction result and the actual result in the embodiment of the present invention. DETAILED DESCRIPTION

[0018] 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.

[0019] Example 1

[0020] Please refer to Figure 1 , Figure 1 : is a structural diagram of a drilling speed modeling method based on an improved bat algorithm and support vector regression in an embodiment of the present invention, the method specifically includes:

[0021] S1: Analyze the interval of drilling data, identify outliers in the data, and replace them with the nearest normal data; specifically:

[0022] S1.1: Given a discrete sequence of drilling speed, drilling pressure, rotation speed and depth data, g1, g2, ..., g n , n is the number of samples, the upper limit of the statistical sample is U OUB With the lower limit U OLB :

[0023] U OLB =min(g1,g2,...,g n )

[0024] U OUB =max(g1,g2,...,g n )

[0025] S1.2: The interval (U OLB ,U OUB ) are taken as new sequence samples, and the new interval (U ILB ,U IUB ), where U ILB is the lower bound of the new interval, U IUB is the upper bound of the new interval;

[0026] S1.3: Calculate the distance d from the remaining samples to the midpoint of the new interval b , if d b ≥3|U IUB -U ILB |, then the sample is judged to be an abnormal sample; the new interval (U ILB ,U IUB ) is the sample closest to the abnormal sample and replaces the abnormal sample.

[0027] In this embodiment, the processing of the drilling site data is as follows:

[0028] A total of 760 sets of data were collected from a vertical well drilling site. Among these 760 sets of data, drilling pressure, depth, rotation speed, drilling fluid density and drilling speed were used as model inputs, and the drilling speed at the next moment was used as the model output. Before building the model, an outlier interval detection method was used to identify and correct abnormal data.

[0029] S2: Based on the replaced drilling sample data, a drilling speed prediction model is constructed based on the support vector regression method. The model takes the current drilling pressure, rotation speed, depth and drilling speed as input and the drilling speed at the next moment as output; specifically:

[0030] S2.1: Assume that a drilling data set P is given. Each set of samples in P contains four dimensions: drilling pressure, rotation speed, depth and drilling fluid density. Each set of samples corresponds to a model output q.

[0031] S2.2: Use high-dimensional mapping to map sample input to high-dimensional space:

[0032] q i =λφ(p i1 ,p i2 ,...,p i4 )+θ

[0033] Among them, q i is the model output corresponding to the i-th sample, λ is the weight, θ is the bias, φ(p i1 ,p i2 ,...,p i4 ) is a function that maps sample input to high-dimensional space, p i1 ,p i2 ,...,p i4 is the first dimension, the second dimension, ..., the fourth dimension in the i-th sample;

[0034] S2.3: Based on the data in the high-dimensional space obtained by mapping, slack variables are introduced to transform the nonlinear mapping problem into a constrained optimization problem:

[0035]

[0036] Among them, the hyperparameter C is the penalty factor, which is used to balance the model complexity and model accuracy, n is the number of samples, ξ i , is a slack variable, and ε is a small positive number.

[0037] S2.4: To fit the nonlinear mapping φ(p i1 ,p i2 ,…,p i4 ), using Gaussian kernel function K(x,x′)=exp(-||xx′|| 2 / 2σ 2 ) is used to fit the nonlinear mapping, σ is the kernel function bandwidth, and x and x′ are data samples.

[0038] In this example, 730 groups of the 760 groups of drilling data were randomly selected to train the drilling speed model, and the remaining 30 groups of data were used for testing. In both the training and testing drilling data, drilling pressure, depth, rotation speed and drilling fluid density were used as the input of the model, and the drilling speed at the next moment was used as the model output. The support vector regression method was further used to construct the drilling speed model.

[0039] S3: An improved bat algorithm based on partial population update, iterative search and chaotic perturbation mechanism is designed to optimize the hyperparameters of the drilling speed model and use the optimized drilling speed model to accurately predict the drilling speed at the next moment. Specifically:

[0040] S3.1: When the population iteration reaches 80% of the maximum number of iterations, the 20% of the population with the largest fitness function value is updated, and the update method of 15% of the population is:

[0041]

[0042] Among them, z * is the global optimal solution in the current bat population, γ1 and γ2 are random positive numbers between 0 and 1;

[0043] The remaining 5% of the population is updated as follows:

[0044]

[0045] Among them, p * is the individual optimal solution of the current bat population, t is the current iteration number, M is the maximum iteration number, γ3, γ4 and γ5 are random positive numbers between 0 and 1;

[0046] S3.2: At the end of each iteration, the global optimal solution is iteratively searched and updated:

[0047] z * =z * +z * *(0.02*γ6 11.5 -0.01*γ7)

[0048] Among them, γ6 and γ7 are random positive numbers between 0 and 1;

[0049] S3.3: During the iteration process, chaotic perturbation X is applied to individual bats. ch , as follows:

[0050]

[0051] Among them, ω ub is the upper limit of the weight, ω lb is the lower limit of the weight, X i is the bat individual that undergoes chaotic disturbance, and γ8 is a random positive number between 0 and 1.

[0052] In this example, based on the constructed drilling speed model, the penalty factor C and kernel function bandwidth σ in the model are taken as decision variables, the root mean square error between the model output and the actual value is taken as the optimization target, and the improved bat optimization algorithm is used to determine the optimal drilling speed penalty factor C and kernel function bandwidth σ.

[0053] The IEEE CEC2005 test function set is used to evaluate the performance of the optimization algorithm. The effects of test function 1 and test function 2 in the test function set are as follows: Figure 2 and Figure 3Further, based on the constructed drilling speed model and the determined penalty factor C and kernel function bandwidth σ value, the output of the drilling speed model in the test data is as follows: Figure 4 shown.

[0054] Example 2

[0055] A computer device / equipment / system 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 improved bat algorithm and support vector regression.

[0056] Example 3

[0057] 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 the improved bat algorithm and support vector regression.

[0058] Example 4

[0059] 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 improved bat algorithm and support vector regression are implemented.

[0060] 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 improved bat algorithm and support vector regression, characterized in that: include: S1: Analyze the interval of drilling data, identify outliers in the data, and replace them with the nearest normal data; S2: Based on the replaced drilling sample data, a drilling speed prediction model is constructed based on the support vector regression method. The model takes the current drilling pressure, rotation speed, depth and drilling fluid density as input and the drilling speed at the next moment as output; S3: Design an improved bat algorithm based on partial population update, iterative search and chaotic perturbation mechanism to optimize the hyperparameters of the drilling speed model, and use the optimized drilling speed model to accurately predict the drilling speed at the next moment.

2. A drilling speed modeling method based on improved bat algorithm and support vector regression as claimed in claim 1, characterized in that: The identification and replacement of abnormal data in S1 includes the following steps: S1.1: Given a discrete sequence of drilling speed, drilling pressure, rotation speed and depth data, g1, g2, ..., g n , n is the number of samples, the upper limit of the statistical sample is U OUB With the lower limit U OLB : U OLB =min(g1,g2,...,g n ) U OUB =max(g1,g2,...,g n ) S1.2: The interval (U OLB ,U OUB ) is used as the first preset percentage of samples far away from the interval boundary as the new sequence samples, and the new interval (U ILB ,U IUB ), where U ILB is the lower bound of the new interval, U IUB is the upper bound of the new interval; S1.3: Calculate the distance d from the remaining samples to the midpoint of the new interval b , if d b ≥3|U IUB -U ILB |, then the sample is judged to be an abnormal sample; the new interval (U ILB ,U IUB ) is the sample closest to the abnormal sample and replaces the abnormal sample.

3. A drilling speed modeling method based on improved bat algorithm and support vector regression as claimed in claim 1, characterized in that: The process of constructing the drilling speed prediction model based on the support vector regression method in S2 is as follows: S2.1: Assume that a drilling data set P is given. Each set of samples in P contains four dimensions: drilling pressure, rotation speed, depth and drilling fluid density. Each set of samples corresponds to a model output q. S2.2: Use high-dimensional mapping to map sample input to high-dimensional space: q i =λφ(p i1 ,p i2 ,...,p i4 )+θ Among them, q i is the model output corresponding to the i-th sample, λ is the weight, θ is the bias, φ(p i1 ,p i2 ,...,p i4 ) is a function that maps sample input to high-dimensional space, p i1 ,p i2 ,...,p i4 is the first dimension, the second dimension, ..., the fourth dimension in the i-th sample; S2.3: Based on the data in the high-dimensional space obtained by mapping, slack variables are introduced to transform the nonlinear mapping problem into a constrained optimization problem: Among them, the hyperparameter C is the penalty factor, which is used to balance the model complexity and model accuracy, n is the number of samples, ξ i , is the slack variable, ε is a positive number; S2.4: To fit the nonlinear mapping φ(p i1 ,p i2 ,...,p i4 ), using Gaussian kernel function K(x,x′)=exp(-||xx′|| 2 / 2σ 2 ) is used to fit the nonlinear mapping, σ is the kernel function bandwidth, and x and x′ are data samples.

4. The drilling speed modeling method based on improved bat algorithm and support vector regression as claimed in claim 1, characterized in that: The design process of the improved bat algorithm S3 based on partial population update, iterative search and chaotic perturbation mechanism is as follows: S3.1: When the population iteration reaches the second preset percentage of the maximum number of iterations, the population with the largest third preset percentage of fitness function value is updated, wherein the updating method of a part of the third preset percentage of the population is: Among them, z * is the global optimal solution in the current bat population, γ1 and γ2 are random positive numbers between 0 and 1; The population update method for the remaining part of the third preset percentage is: Among them, p * is the individual optimal solution of the current bat population, t is the current iteration number, M is the maximum iteration number, γ3, γ4 and γ5 are random positive numbers between 0 and 1; S3.2: At the end of each iteration, the global optimal solution is iteratively searched and updated: z * =z * +z * *(0.02*γ6 11.5 -0.01*γ7) Among them, γ6 and γ7 are random positive numbers between 0 and 1; S3.3: During the iteration process, chaotic perturbation X is applied to individual bats. ch , as follows: Among them, ω ub is the upper limit of the weight, ω lb is the lower limit of the weight, X i is the bat individual that undergoes chaotic disturbance, and γ8 is a random positive number between 0 and 1.

5. A drilling speed modeling method based on improved bat algorithm and support vector regression as claimed in claim 4, characterized in that: The second preset percentage is 80%, and the third preset percentage is 20%.

6. 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 improved bat algorithm and support vector regression as described in any one of claims 1 to 5.

7. 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 the improved bat algorithm and support vector regression as described in any one of claims 1 to 5 are implemented.

8. A computer program product, characterized in that It includes a computer program or instruction. When the program or instruction is executed by a processor, it implements the steps of the drilling speed modeling method based on the improved bat algorithm and support vector regression as described in any one of claims 1 to 5.

Images