Heuristic convolutional neural network-based permeability prediction method

Abstract

The invention provides a permeability prediction method based on a heuristic convolutional neural network, and the method comprises the steps: carrying out the lithologic interpretation and analysis of geological exploration data, so as to obtain a sandstone sample set comprising a first rock physical elastic parameter and a porosity parameter; performing correlation analysis on the parameters of the sandstone sample set to obtain a high-porosity sandstone sample set comprising a second rock physical elastic parameter and a permeability parameter; performing correlation analysis on the second rock physical elastic parameter and the permeability to obtain a sensitive parameter combination set comprising a third rock physical elastic parameter, a shear Lee factor parameter and the permeability; carrying out cross analysis on the shearing Lee factor parameter and the permeability parameter of the sensitive parameter combination set, and obtaining a classification standard according to the relationship between the shearing Lee factor parameter and the permeability parameter; and inputting the classified samples into a permeability prediction model to obtain final permeability prediction data. By introducing the shearing Lee factor parameter, the permeability is convenient to predict.

Claims

1. A method for permeability prediction based on heuristic convolutional neural network, characterized in that: The method comprises: S1: Obtain a sandstone sample set including first rock physical elastic parameters and porosity parameters by performing lithology interpretation and analysis on geological exploration data; S2: performing correlation analysis on the first rock physical elastic parameter and the porosity to obtain a high-porosity sandstone sample set including a second rock physical elastic parameter and a permeability parameter; S3: Based on rock physics analysis, performing correlation analysis on the second rock physical elastic parameter and the permeability to obtain a sensitive parameter combination set including a third rock physical elastic parameter, a shear Lee factor parameter and the permeability; S4: performing intersection analysis on the shear Lee factor parameter and the permeability parameter of the sensitive parameter combination set, and obtaining a classification standard according to the relationship between the shear Lee factor parameter and the permeability parameter; S5: Based on the classification standard, the high-porosity sandstone sample set is classified according to the classification standard, and the classified samples are input into the prediction permeability model to obtain the final prediction permeability data.

2. The method for permeability prediction based on heuristic convolutional neural network according to claim 1 is characterized in that: The geological exploration data include permeability curves; The step S1 comprises: Cleaning the geological exploration data and taking the logarithm of the permeability curve to obtain a preprocessed well logging data set; The preprocessed logging data set is screened according to preset sandstone screening conditions to obtain a sandstone sample set including a first rock physical elastic parameter and a porosity parameter.

3. The method for permeability prediction based on heuristic convolutional neural network according to claim 1, characterized in that: The sensitive parameter combination set includes a first sensitive parameter combination set and a second sensitive parameter combination set; The step S3 comprises: Perform intersection analysis on the second rock physical elastic parameter and the permeability parameter to obtain the first sensitive parameter combination set; The porosity parameter and the density parameter in the first sensitive parameter combination set are calculated by using a preset rock physics model to obtain a shear Lee factor parameter; The shear Lee factor parameter and the first sensitive parameter combination set are combined together to obtain a second sensitive parameter combination set including a third rock physical elastic parameter, a shear Lee factor parameter and the permeability parameter.

4. The method for permeability prediction based on heuristic convolutional neural network according to claim 3 is characterized in that: The porosity parameter and density parameter in the first sensitive parameter combination set are calculated by using a preset rock physics model to obtain the shear Lee factor parameter, which includes: Obtain a preset rock physics model; Calculating the shear modulus parameters of the dry rock of the preset rock physics model according to the boundary theory method; Adaptively adjusting the initial shear compliance factor of the preset rock physics model according to the particle swarm algorithm to obtain a shear compliance factor parameter; Substituting the shear modulus parameter of the dry rock, the shear compliance factor parameter and the porosity parameter into the first formula to obtain the shear modulus parameter of the rock matrix; According to the Gassmann equation, the shear modulus parameter of the rock matrix is ​​converted into the modulus parameter of the fluid-saturated rock, and the shear wave velocity parameter is calculated; The shear modulus parameter of the rock matrix, the porosity parameter, the density parameter and the shear wave velocity parameter are substituted into the second formula to obtain the shear Lee factor.

5. The method for permeability prediction based on heuristic convolutional neural network according to claim 3 is characterized in that: Described step S4 comprises: The high-porosity sandstone samples are divided into high-porosity and high-permeability samples and high-porosity and low-permeability samples according to the preset high-permeability and low-permeability boundaries; The cross analysis of the shear Lee factor and the permeability of the sensitive parameter combination set, and the classification criteria obtained according to the relationship between the shear Lee factor and the permeability include: Based on a preset threshold, the high-porosity sandstone sample set is divided into high-porosity and high-permeability samples and high-porosity and low-permeability samples; The third rock physical elastic parameter and the shear Lee factor in the second sensitive parameter combination set are subjected to intersection processing with the permeability parameter to determine a first classification threshold and a second classification threshold; Using a particle swarm algorithm to dynamically optimize the boundary between the high-porosity and high-permeability sample and the high-porosity and low-permeability sample within a preset floating range, with the goal of minimizing the permeability prediction error, to obtain a first floating threshold and a second floating threshold; Obtaining a first classification standard according to the first classification threshold and the first floating threshold; A second classification standard is obtained according to the second classification threshold and the second floating threshold.

6. The method for permeability prediction based on heuristic convolutional neural network according to claim 5 is characterized in that: The method of dynamically optimizing the boundary between the high porosity and high permeability sample and the high porosity and low permeability sample within a preset floating range by using a particle swarm algorithm to obtain a first floating threshold and a second floating threshold comprises: Step 1: According to the shear Lee factor, the parameters in the second sensitive parameter combination set are classified and cross-analyzed; Step 2: adjusting the intersection origin and the rotation angle so that the dividing line distinguishes the high-porosity and high-permeability sample from the high-porosity and low-permeability sample to determine the first dividing line and the second dividing line; Step 3: Adjust the first dividing line and the second dividing line to be perpendicular to each other to obtain the first parameter and the second parameter; Step 4: linearly fit the first parameter or the second parameter to the permeability to obtain a first permeability prediction error, wherein the first permeability prediction error is a permeability prediction error obtained based on a coordinate axis rotation method; Step 5: Dynamically optimize the boundary line within a preset floating range by using a particle swarm algorithm, and obtain the first floating threshold and the second floating threshold according to the principle of minimizing the permeability prediction error.

7. The method for permeability prediction based on heuristic convolutional neural network according to claim 5, characterized in that: The performing linear fitting of the first parameter or the second parameter with the permeability to obtain a first permeability prediction error comprises: When the third rock physical elastic parameter and the shear Lee factor are negatively correlated with the permeability parameter, a first permeability prediction error is obtained by linearly fitting the second parameter and the permeability; When the third rock physical elastic parameter and the shear Lee factor are positively correlated with the permeability parameter, a first permeability prediction error is obtained by linearly fitting the first parameter and the permeability.

8. The method for permeability prediction based on heuristic convolutional neural network according to claim 1, characterized in that: The step S5 comprises: Classifying the high-porosity sandstone sample set according to the classification standard to obtain sandstone subsets of different categories; the sandstone subsets include the third rock physical elastic parameter and the shear Lee factor; The third rock physical elastic parameter and the shear Lee factor in the sandstone subset are used as a training set and a test set; Using the training set to train a predicted permeability model to obtain a trained predicted permeability model; The test set obtains predicted permeability based on the trained prediction permeability model.

9. The method for permeability prediction based on heuristic convolutional neural network according to claim 8, characterized in that: The predicted permeability model includes a convolutional neural network; The step of training the prediction permeability model with the training set to obtain the trained prediction permeability model comprises: Step 1: Input the training set into the convolutional neural network to calculate the actual output value; Step 2: Using a preset loss function, calculate the error between the actual output value and the expected output value; Step 3: Repeat steps 1 to 2 until the preset loss function converges or reaches a preset number of training rounds.

10. The method for permeability prediction based on heuristic convolutional neural network according to claim 4, characterized in that: The first formula includes: Where: μ d is the shear modulus parameter of dry rock; μ m is the shear modulus parameter of the rock matrix; φ is the porosity; γ μ is the shear flexibility factor; The second formula includes: Where: c μ is the shear Lee factor; v S is the shear wave velocity; φ is the porosity; ρ s is the density; μ m is the shear modulus parameter of the rock matrix.

Images