Volcanic rock porosity prediction method and device

Abstract

The invention provides a volcanic porosity prediction method and device. According to the method, the relationship among the porosity, the lithofacies and the conventional logging curve is established based on the existing measured data through the LASSO algorithm, the method depends on the measured data of a research area instead of empirical parameters or core samples of previous specific areas, the prediction process is less in man-made interference, the lithofacies information is added in the prediction, and the prediction accuracy is improved. And the accuracy of porosity prediction of the research area can be effectively improved.

Description

Technical Field

[0001] The invention relates to the technical field of geophysical exploration, and in particular to a method and device for predicting the porosity of volcanic rocks. Background Art

[0002] The development degree of volcanic reservoirs is closely related to the porosity. Generally, reservoirs are developed in high-porosity areas, so porosity prediction is very important for volcanic reservoir exploration. Currently, there are two main methods for volcanic rock porosity prediction: inversion method and empirical formula method.

[0003] The inversion method refers to the inversion of the porosity distribution of rocks by combining the rock porosity inversion model with certain rock physical parameters (such as acoustic slowness, resistivity, etc.). The physical basis of this method is that different rock compositions and structures will cause changes in rock physical parameters. For example, Han et al. proposed a method for inverting porosity based on rock acoustic slowness by establishing the relationship between rock porosity and acoustic slowness. This method uses the Gassmann equation to describe the corresponding relationship between the modulus of solid rock and porosity, and combines the empirical relationship obtained from the experiment to establish a mathematical model between acoustic slowness and porosity. By iterating the model, the inversion result of rock porosity can be obtained. Peng et al. used the resistivity inversion model to study the corresponding relationship between rock porosity and resistivity. This method establishes a mathematical model for resistivity inversion porosity based on the geometric average relationship between rock resistivity and porosity, pore water resistivity and surrounding rock resistivity. However, the main problems faced by this inversion method in accurately predicting porosity are:

[0004] (1) The establishment of the inversion model requires a large number of core samples for calibration. The formation conditions in different regions are different, and the calibration results are not universal;

[0005] (2) There is a certain degree of uncertainty in the inversion model itself, which directly affects the accuracy of the inversion results.

[0006] For example, the model of Han et al. requires obtaining core samples for acoustic parameter determination, but the acoustic characteristics of rocks in different strata vary greatly. The model established is only applicable to the specific strata type of this paper and cannot be directly applied to other areas. In the resistivity inversion model of Peng et al., some empirical parameters also need to be determined based on the experience of different strata, which also limits the scope of application of the model.

[0007] The empirical formula method refers to the calculation of the porosity value of rocks by combining empirical formulas with some logging parameters (such as acoustic time difference, density, natural gamma, etc.). This method utilizes the empirical correspondence between different logging parameters and rock porosity. For example, Raymer et al. established an empirical formula for porosity calculation using the acoustic time difference-density cross diagram by comprehensive analysis of multiple logging parameters. This method determines the empirical relationship between acoustic time difference and density and porosity based on core statistical analysis, and establishes an empirical equation for calculating porosity. Liu et al. considered natural gamma logging and proposed a method for calculating porosity by combining multiple logging parameters. This method comprehensively considers the three logging parameters of acoustic time difference, density, and natural gamma, establishes a multivariate linear regression model, and determines the empirical correspondence between each parameter and porosity. However, the empirical formula method has the following problems:

[0008] (1) Under different strata, the empirical formula is not universal and needs to be re-established for different strata;

[0009] (2) Empirical formulas are often too dependent on experience, lack certainty, and the accuracy of the results is difficult to guarantee.

[0010] For example, Raymer et al.'s method is only based on the statistical analysis of well fields in a specific area and is not universal. Liu et al.'s method contains a large number of empirical coefficients that need to be manually specified, and the results obtained by different users may have large deviations.

[0011] In summary, the two main volcanic rock porosity prediction methods currently have certain limitations, and the accuracy of the prediction results needs to be improved. This is mainly reflected in:

[0012] (1) It depends on empirical parameters or core samples in a specific area, and the effects may vary greatly in different formations;

[0013] (2) It involves a large number of empirical factors, and the results are easily affected by subjective judgment, making it difficult to guarantee accuracy;

[0014] (3) The logging data resources of the existing well fields are not fully utilized, and predictions are made based on limited core samples or experience, which is not accurate enough for reference.

[0015] Therefore, how to establish a more universal, accurate and objective porosity prediction method for volcanic reservoirs is a key scientific issue that needs to be urgently addressed in this field.

[0016] At the same time, accurate prediction of the porosity of volcanic reservoirs has very important practical significance for the rational development and utilization of volcanic gas, including but not limited to:

[0017] (1) Guiding the selection of volcanic gas well locations

[0018] Based on the porosity prediction results, we can scientifically identify target areas with high reservoir development and high gas production potential, and optimize the well layout plan. If the porosity prediction error is too large, it may lead to slow development progress and poor results.

[0019] (2) Evaluation of reserves and production potential

[0020] Porosity is one of the key parameters for calculating reserves. Optimizing porosity prediction can more accurately assess reserves and guide development intensity. If porosity prediction is inaccurate, it will directly lead to a large deviation in reserve resource assessment.

[0021] (3) Optimize development plan

[0022] The optimization of fracturing parameters for reservoirs with different porosities will be different. Accurate prediction of porosity distribution will help to formulate more reasonable development plans. Otherwise, improper selection of development parameters may occur.

[0023] (4) Reduce resource waste

[0024] Reasonable porosity prediction can avoid ineffective development in areas with poor reservoir development and save investment costs. If the porosity prediction is inaccurate, a large amount of ineffective mining may be carried out in non-high-quality reservoir areas, resulting in a waste of resources.

[0025] (5) Increase natural gas production

[0026] Accurately predicting the distribution of high-porosity areas can effectively increase the single-well production of commercial gas wells and the gas production potential of the entire gas field. If the porosity prediction is wrong, it is impossible to accurately lock in the high-yield area, resulting in low single-well production;

[0027] (6) Reduce environmental impact

[0028] Reasonable prediction can avoid over-exploitation and reduce damage to the ecological environment, while large errors in porosity prediction may lead to blind and indiscriminate development and increase environmental burden.

[0029] In summary, the development of a new, efficient and accurate method for predicting the porosity of volcanic reservoirs will be able to better guide the exploration and development of volcanic gas resources, avoid waste of resources, and improve gas production efficiency, which has important economic value and practical significance. This also fully illustrates the importance and necessity of accurately and efficiently predicting the porosity of volcanic reservoirs. Summary of the invention

[0030] In view of this, the present invention aims to propose a technical solution for predicting porosity, which relies on measured data, effectively reduces human interference, and adds lithofacies information to porosity prediction, which can significantly improve the accuracy of porosity prediction.

[0031] According to one aspect of the present invention, a method for predicting the porosity of volcanic rocks is proposed, the method comprising:

[0032] Step 1, obtaining measured porosity curves, lithofacies interpretation result curves and logging data curves of multiple wells;

[0033] Step 2, selecting measured porosity curves, lithofacies interpretation result curves and logging data curves of some wells among the plurality of wells to construct a training data set;

[0034] Step 3, using the LASSO algorithm to learn the relationship between porosity and lithofacies interpretation results and logging data based on the constructed training data set, and obtaining an initial LASSO porosity prediction model;

[0035] Step 4, constructing a verification data set using the measured porosity curves, lithofacies interpretation result curves and logging data curves of the remaining wells in the plurality of wells;

[0036] Step 5, using the verification data set to verify the initial LASSO porosity prediction model to obtain the optimal LASSO porosity prediction model;

[0037] Step 6: Use the optimal LASSO porosity prediction model to predict the porosity of wells in the study area that do not have measured porosity curves.

[0038] In some embodiments, the logging data curves include acoustic wave time difference curves, natural gamma curves, and density curves.

[0039] In some embodiments, in step 3, the following J(L m ) is used as the objective function to construct the initial LASSO porosity prediction model:

[0040]

[0041]

[0042]

[0043] Where m represents the number of the well, M is the total number of the selected wells, and L m represents the collection of porosity curve, lithofacies interpretation result curve and logging data curve of well numbered m. is the L of M wells m The average value of , q is the adjustment factor.

[0044] In some embodiments, validating the LASSO porosity prediction model includes:

[0045] When the verification fails, the q value in the current LASSO porosity prediction model is modified to adjust the current LASSO porosity prediction model.

[0046] In some embodiments, the value range of q is 1, 2, 4, or 6.

[0047] In some embodiments, step 5 specifically includes:

[0048] Step 51, using the current LASSO porosity prediction model, predicting the porosity curves of the remaining wells based on the lithofacies interpretation result curves and the logging data curves of the remaining wells in the validation data set;

[0049] Step 52, comparing the predicted porosity curves and the measured porosity curves of the remaining wells;

[0050] Step 53, if the degree of agreement between the predicted porosity curves and the measured porosity curves of the remaining wells reaches the preset standard, then the verification is confirmed to be successful, and the current LASSO porosity prediction model is confirmed to be the optimal LASSO porosity prediction model; if the degree of agreement between the predicted porosity curves and the measured porosity curves of the remaining wells does not reach the preset standard, then the verification is confirmed to be unsuccessful, and the current LASSO porosity prediction model is adjusted, and then the process returns to step 51, and the porosity curves of the remaining wells are predicted again using the adjusted LASSO porosity prediction model, until the degree of agreement between the predicted porosity curves and the measured porosity curves of the remaining wells reaches the preset standard.

[0051] In some embodiments, the number of the partial wells used to construct the training data set accounts for 55% to 85% of the plurality of wells.

[0052] According to another aspect of the present invention, a volcanic rock porosity prediction device is also proposed, the device comprising:

[0053] A measured data acquisition unit, used to obtain measured porosity curves, lithofacies interpretation result curves and logging data curves of multiple wells;

[0054] A training data set construction unit, used for selecting measured porosity curves, lithofacies interpretation result curves and logging data curves of some wells among the plurality of wells to construct a training data set;

[0055] An initial model building unit is used to learn the relationship between porosity and lithofacies interpretation results and logging data based on the constructed training data set using the LASSO algorithm to obtain an initial LASSO porosity prediction model;

[0056] A verification data set construction unit, used to construct a verification data set using the measured porosity curves, lithofacies interpretation result curves and logging data curves of the remaining wells in the plurality of wells;

[0057] A model verification unit, used to verify the initial LASSO porosity prediction model using the verification data set to obtain an optimal LASSO porosity prediction model;

[0058] The porosity prediction unit is used to use the optimal LASSO porosity prediction model to predict the porosity of wells in the study area that have no measured porosity curves.

[0059] In some embodiments, the logging data curves include acoustic wave time difference curves, natural gamma curves, and density curves.

[0060] In some embodiments, the initial model building unit is specifically used to: m ) is used as the objective function to construct the initial LASSO porosity prediction model:

[0061]

[0062]

[0063]

[0064] Where m represents the number of the well, M is the total number of the selected wells, and L m represents the collection of porosity curve, lithofacies interpretation result curve and logging data curve of well numbered m. is the L of M wells m The average value of , q is the adjustment factor.

[0065] In some embodiments, validating the LASSO porosity prediction model includes:

[0066] When the verification fails, the q value in the current LASSO porosity prediction model is modified to adjust the current LASSO porosity prediction model.

[0067] In some embodiments, the value range of q is 1, 2, 4, or 6.

[0068] In some implementations, the model adjustment confirmation unit specifically includes:

[0069] A porosity curve prediction subunit, used to predict the porosity curves of the remaining wells based on the lithofacies interpretation result curves and the logging data curves of the remaining wells in the validation data set by using the current LASSO porosity prediction model;

[0070] A comparison subunit, used for comparing the predicted porosity curves and the measured porosity curves of the remaining wells;

[0071] The best model confirmation subunit, if the degree of agreement between the predicted porosity curves and the measured porosity curves of the remaining wells reaches a preset standard, confirms that the verification has passed, and confirms that the current LASSO porosity prediction model is the best LASSO porosity prediction model; if the degree of agreement between the predicted porosity curves and the measured porosity curves of the remaining wells does not reach the preset standard, confirms that the verification has not passed, adjusts the current LASSO porosity prediction model, and then returns to step 51, using the adjusted LASSO porosity prediction model to predict the porosity curves of the remaining wells again, until the degree of agreement between the predicted porosity curves and the measured porosity curves of the remaining wells reaches the preset standard.

[0072] In some embodiments, the number of the partial wells used to construct the training data set accounts for 55% to 85% of the plurality of wells.

[0073] According to another aspect of the present invention, an electronic device is also provided, the electronic device comprising:

[0074] Memory, which stores executable instructions:

[0075] A processor runs the executable instructions in the memory to implement the volcanic rock porosity prediction method described above.

[0076] According to another aspect of the present invention, a computer-readable storage medium is provided. The computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the volcanic rock porosity prediction method described above is implemented.

[0077] The beneficial effects of the present invention include at least:

[0078] (1) Improved the accuracy of porosity prediction

[0079] The present invention establishes a relationship model between porosity and logging parameters through a machine learning method, which can effectively utilize the information of multiple logging parameters. It is more accurate and reliable than the traditional empirical method. The training of the machine learning model does not rely on subjective experience, the results are more objective, and the prediction error is smaller.

[0080] (2) Expanded the scope of application of the model

[0081] The present invention uses measured data from different wells in the same area for modeling, and does not rely on any core data. Therefore, the established model can be directly applied to other wells in the area without recalibration, and has a wider range of applicability.

[0082] (3) Improved work efficiency

[0083] The present invention realizes porosity prediction in a digital way, avoiding the low efficiency of manual interpretation. The prediction results can be quickly obtained by simply inputting the logging data, which greatly improves the work efficiency.

[0084] (4) Provides reliable support for later development

[0085] The highly accurate porosity prediction of the present invention provides reliable data support for later reservoir evaluation, production capacity prediction, development plan design, etc., and reduces development risks;

[0086] (5) Has prospects for promotion and application

[0087] The technical solution of the present invention can be extended to the porosity prediction of other types of reservoirs and has broad prospects for extension and application.

[0088] In summary, the technical solution of the present invention can significantly improve prediction accuracy, expand the scope of application, improve work efficiency, provide support for later development, and has important technological progress significance and application prospects.

[0089] The methods and apparatus of the present invention have other features and advantages that will be apparent from or will be described in detail in the accompanying drawings and subsequent detailed descriptions incorporated herein, which together serve to explain the specific principles of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0090] The above and other objects, features and advantages of the present invention will become more apparent through a more detailed description of exemplary embodiments of the present invention in conjunction with the accompanying drawings, wherein like reference numerals generally represent like components throughout the exemplary embodiments of the present invention.

[0091] Figure 1 A flow chart of a volcanic rock porosity prediction method according to an embodiment of the present invention is shown.

[0092] Figure 2 A graph showing measured porosity, petrographic interpretation results, and conventional well logging of multiple wells according to an exemplary embodiment of the present invention is shown.

[0093] Figure 3 A comparison diagram of a measured porosity curve and a predicted porosity curve according to an exemplary embodiment of the present invention is shown. DETAILED DESCRIPTION

[0094] The preferred embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although the preferred embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided to make the present invention more thorough and complete, and to fully convey the scope of the present invention to those skilled in the art.

[0095] Example 1

[0096] Figure 1 The flowchart of the volcanic rock porosity prediction method according to an embodiment of the present invention is shown. As shown in the figure, the example includes steps 1 to 6.

[0097] Step 1: Obtain measured porosity curves, lithofacies interpretation result curves and logging data curves of multiple wells.

[0098] In some embodiments, the logging data curve may include an acoustic wave time difference curve, a natural gamma curve, and a density curve.

[0099] Step 2: Select measured porosity curves, lithofacies interpretation result curves and logging data curves of some wells in the plurality of wells to construct a training data set.

[0100] In some embodiments, the number of the partial wells used to construct the training data set accounts for 55% to 85% of the plurality of wells.

[0101] For example, in step 1, the measured porosity curves, lithofacies interpretation result curves and logging data curves of 10 wells in the study area are obtained. In step 2, the measured porosity curves, lithofacies interpretation result curves and logging data curves of 70% of the wells (i.e., 7 wells) can be selected to construct a training data set. The training data set includes 7 training subsets, and the mth training subset in the training data set can be expressed as L m ={p m ,litho m , ac m ,gr m ,den m}, m = 1, ... 7, where p m represents the measured porosity curve of well numbered m, litho m The lithofacies interpretation curve of well numbered m, ac m represents the acoustic time difference curve of well numbered m, gr m represents the natural gamma curve of well numbered m, den m Represents the density curve of the well numbered m.

[0102] Step 3: LASSO algorithm is used to learn the relationship between porosity and lithofacies interpretation results and logging data based on the constructed training data set to obtain the initial LASSO porosity prediction model.

[0103] The LASSO (Least Absolute Shrinkage and Selection Operator) algorithm is a regularized regression algorithm widely used in statistical analysis and machine learning.

[0104] The main idea of ​​the LASSO algorithm is to introduce the L1 norm regularization term into the loss function of the model to limit the size of the regression coefficient during regression analysis, thereby achieving regression coefficient compression and variable selection. Its objective function usually has the following form:

[0105] min RSS+λ*sum|βi|

[0106] Among them, RSS represents the residual sum of squares, λ is the regularization parameter, and βi is the regression coefficient.

[0107] The LASSO algorithm has the following two main characteristics:

[0108] (1) Compression regression coefficient

[0109] The regularization term shrinks the size of the regression coefficient, reduces the impact of unimportant variables, and simplifies the model;

[0110] (2) Variable selection

[0111] When λ is set to a certain value, the regression coefficients of some variables can be reduced to 0, thereby achieving automatic selection of variables.

[0112] Compared with linear regression, the LASSO algorithm can build a simpler and more interpretable model by shrinking and selecting variables, and can handle the collinearity problems that may exist in regression analysis.

[0113] In some embodiments, the following J(L m ) is used as the objective function to construct the initial LASSO porosity prediction model:

[0114]

[0115]

[0116]

[0117] Where m represents the number of the well, M is the total number of the selected wells, and L mrepresents the collection of porosity curve, lithofacies interpretation result curve and logging data curve of well numbered m. is the L of M wells m The average value of , q is the adjustment factor.

[0118] In some examples, an adjustment factor q=1 may be set in the initial LASSO porosity prediction model.

[0119] Step 4: construct a verification data set using the measured porosity curves, lithofacies interpretation result curves and logging data curves of the remaining wells in the plurality of wells.

[0120] In the above step 1, measured porosity curves, lithofacies interpretation result curves and logging data curves of multiple wells are obtained. The remaining wells here refer to the other wells among these multiple wells except for some wells used to construct the training data set.

[0121] As described above, in some embodiments, the number of the partial wells used to construct the training data set accounts for 55% to 85% of the multiple wells, and accordingly, the number of the remaining wells used to construct the verification data set accounts for 15% to 45% of the multiple wells. For example, in step 1, the measured porosity curves, lithofacies interpretation result curves and logging data curves of 10 wells in the study area are obtained. In step 2, the measured porosity curves, lithofacies interpretation result curves and logging data curves of 70% of the wells (i.e., 7 wells) can be selected to construct the training data set. In step 4, the measured porosity curves, lithofacies interpretation result curves and logging data curves of the remaining 3 wells are used to construct the verification data set.

[0122] Step 5: Use the verification data set to verify the initial LASSO porosity prediction model to obtain the optimal LASSO porosity prediction model.

[0123] In some implementations, step 5 may specifically include:

[0124] Step 51, using the current LASSO porosity prediction model, predicting the porosity curves of the remaining wells based on the lithofacies interpretation result curves and the logging data curves of the remaining wells in the validation data set;

[0125] Step 52, comparing the predicted porosity curves and the measured porosity curves of the remaining wells;

[0126] Step 53, if the degree of agreement between the predicted porosity curves and the measured porosity curves of the remaining wells reaches the preset standard, then the verification is confirmed to be successful, and the current LASSO porosity prediction model is confirmed to be the optimal LASSO porosity prediction model; if the degree of agreement between the predicted porosity curves and the measured porosity curves of the remaining wells does not reach the preset standard, then the verification is confirmed to be unsuccessful, and the current LASSO porosity prediction model is adjusted, and then the process returns to step 51, and the porosity curves of the remaining wells are predicted again using the adjusted LASSO porosity prediction model, until the degree of agreement between the predicted porosity curves and the measured porosity curves of the remaining wells reaches the preset standard.

[0127] In some embodiments, validating the LASSO porosity prediction model includes:

[0128] When the verification fails, the q value in the current LASSO porosity prediction model is modified to adjust the current LASSO porosity prediction model.

[0129] After in-depth research, the inventors found that the q value can be preferably adjusted within the range of 1, 2, 4, and 6. When q takes these values, it is beneficial to make the predicted porosity curve converge to the measured porosity curve as soon as possible.

[0130] The degree of agreement between the predicted porosity curve and the measured porosity curve can be measured by the following means:

[0131] (1) Correlation Coefficient can be used

[0132] The linear correlation coefficient between the predicted porosity and the measured porosity can be calculated. The closer the value is to 1, the stronger the linear correlation between the two is and the better the degree of agreement is. In this case, a correlation coefficient threshold can be set. When the calculated linear correlation coefficient is greater than the correlation coefficient threshold, it is judged that the agreement between the two has reached the preset standard. Otherwise, it is judged that the agreement between the two has not reached the preset standard.

[0133] (2) The coefficient of determination (R^2) can be used

[0134] It can reflect the degree of explanation of the predicted porosity to the measured porosity change. Its value is between 0 and 1. The larger the value, the higher the degree of explanation and the better the consistency. In this case, a determination coefficient threshold can be set. When the calculated determination coefficient is greater than the determination coefficient, it is judged that the consistency between the two reaches the preset standard. Otherwise, it is judged that the consistency between the two does not reach the preset standard.

[0135] (3) You can use the mean absolute error (MAE)

[0136] The average absolute value of the difference between the predicted porosity and the measured porosity can be calculated. The smaller the value, the smaller the prediction error and the better the degree of agreement. In this case, the average absolute error threshold can be set. When the calculated average absolute error is less than the average absolute error threshold, it is judged that the agreement between the two has reached the preset standard. Otherwise, it is judged that the agreement between the two has not reached the preset standard.

[0137] (4) Mean absolute percentage error (MAPE)

[0138] Calculate the average value of the percentage error between the predicted value and the measured value. The smaller the value, the better the consistency. In this case, a mean absolute percentage error threshold can be set. When the calculated mean absolute percentage error is less than the mean absolute percentage error threshold, it is judged that the consistency between the two reaches the preset standard. Otherwise, it is judged that the consistency between the two does not reach the preset standard.

[0139] (5) Draw a cross-validation curve

[0140] Through cross-validation, a comparison chart of the predicted sample and the measured sample can be obtained to observe the degree of fit and determine whether the fit between the predicted porosity curve and the measured porosity curve meets the preset standard;

[0141] (6) t-test

[0142] It can be used to test whether there are significant differences between the two sets of data. The larger the P value, the better the consistency between the two sets of data. In this case, a P value threshold can be set. When the calculated P value is greater than the P value threshold, it is judged that the consistency between the two sets of data meets the preset standard. Otherwise, it is judged that the consistency between the two sets of data does not meet the preset standard.

[0143] The above statistical indicators and / or visualization methods may also be used in combination to evaluate the agreement between the predicted porosity curve and the measured porosity curve, which is not limited in the present invention.

[0144] In summary, this embodiment discloses a method for predicting the porosity of volcanic rocks. This embodiment uses the LASSO algorithm to establish the relationship between porosity and lithofacies and conventional logging curves based on existing measured data. This method relies on the measured data of the study area rather than empirical parameters or core samples in a specific area in the past. In addition, the prediction process of this embodiment has less human interference, and lithofacies information is added to the prediction, which can effectively improve the accuracy of porosity prediction in the study area.

[0145] Example 2

[0146] According to one embodiment of the present invention, a volcanic rock porosity prediction device is disclosed. The device comprises:

[0147] A measured data acquisition unit, used to obtain measured porosity curves, lithofacies interpretation result curves and logging data curves of multiple wells;

[0148] A training data set construction unit, used for selecting measured porosity curves, lithofacies interpretation result curves and logging data curves of some wells among the plurality of wells to construct a training data set;

[0149] An initial model building unit is used to learn the relationship between porosity and lithofacies interpretation results and logging data based on the constructed training data set using the LASSO algorithm to obtain an initial LASSO porosity prediction model;

[0150] A verification data set construction unit, used to construct a verification data set using the measured porosity curves, lithofacies interpretation result curves and logging data curves of the remaining wells in the plurality of wells;

[0151] A model verification unit, used to verify the initial LASSO porosity prediction model using the verification data set to obtain an optimal LASSO porosity prediction model;

[0152] The porosity prediction unit is used to use the optimal LASSO porosity prediction model to predict the porosity of wells in the study area that have no measured porosity curves.

[0153] In some embodiments, the logging data curves include acoustic wave time difference curves, natural gamma curves, and density curves.

[0154] In some embodiments, the initial model building unit is specifically as follows: m ) is used as the objective function to construct the initial LASSO porosity prediction model:

[0155]

[0156]

[0157]

[0158] Where m represents the number of the well, M is the total number of the selected wells, and L m represents the collection of porosity curve, lithofacies interpretation result curve and logging data curve of well numbered m. is the L of M wells m The average value of , q is the adjustment factor.

[0159] In some embodiments, validating the LASSO porosity prediction model includes:

[0160] When the verification fails, the q value in the current LASSO porosity prediction model is modified to adjust the current LASSO porosity prediction model.

[0161] In some embodiments, the value range of q is 1, 2, 4, or 6.

[0162] In some implementations, the model adjustment confirmation unit specifically includes:

[0163] A porosity curve prediction subunit, used to predict the porosity curves of the remaining wells based on the lithofacies interpretation result curves and the logging data curves of the remaining wells in the validation data set by using the current LASSO porosity prediction model;

[0164] A comparison subunit, used for comparing the predicted porosity curves and the measured porosity curves of the remaining wells;

[0165] The best model confirmation subunit is used to confirm that the verification has passed if the degree of fit between the predicted porosity curves and the measured porosity curves of the remaining wells reaches a preset standard, and confirm that the current LASSO porosity prediction model is the best LASSO porosity prediction model; if the degree of fit between the predicted porosity curves and the measured porosity curves of the remaining wells does not reach the preset standard, then confirm that the verification has not passed, adjust the current LASSO porosity prediction model, and return the adjusted LASSO porosity prediction model to the porosity curve prediction subunit, so that the porosity curve prediction subunit uses the adjusted LASSO porosity prediction model to predict the porosity curves of the remaining wells again until the degree of fit between the predicted porosity curves and the measured porosity curves of the remaining wells reaches the preset standard.

[0166] In some embodiments, the number of the partial wells used to construct the training data set accounts for 55% to 85% of the plurality of wells.

[0167] The present embodiment discloses a volcanic rock porosity prediction device. The present embodiment establishes the relationship between porosity and lithofacies and conventional logging curves based on existing measured data through the LASSO algorithm. The present embodiment relies on the measured data of the study area rather than empirical parameters or core samples of a specific area in the past. In addition, the prediction process of the present embodiment has less human interference, and lithofacies information is added to the prediction, which can effectively improve the accuracy of porosity prediction in the study area.

[0168] For other detailed descriptions and advantages of this embodiment, reference may be made to the corresponding descriptions in the aforementioned embodiments, which will not be repeated here.

[0169] Example 3

[0170] According to another aspect of the present invention, an electronic device is provided. The electronic device comprises:

[0171] Memory, which stores executable instructions:

[0172] A processor is used to execute the executable instructions in the memory to implement the volcanic rock porosity prediction method according to the present invention.

[0173] The method comprises the following steps:

[0174] Step 1, obtaining measured porosity curves, lithofacies interpretation result curves and logging data curves of multiple wells;

[0175] Step 2, selecting measured porosity curves, lithofacies interpretation result curves and logging data curves of some wells among the plurality of wells to construct a training data set;

[0176] Step 3, using the LASSO algorithm to learn the relationship between porosity and lithofacies interpretation results and logging data based on the constructed training data set, and obtaining an initial LASSO porosity prediction model;

[0177] Step 4, constructing a verification data set using the measured porosity curves, lithofacies interpretation result curves and logging data curves of the remaining wells in the plurality of wells;

[0178] Step 5, using the verification data set to verify the initial LASSO porosity prediction model to obtain the optimal LASSO porosity prediction model;

[0179] Step 6: Use the optimal LASSO porosity prediction model to predict the porosity of wells in the study area that do not have measured porosity curves.

[0180] In some embodiments, the logging data curves include acoustic wave time difference curves, natural gamma curves, and density curves.

[0181] In some embodiments, in step 3, the following J(L m ) is used as the objective function to construct the initial LASSO porosity prediction model:

[0182]

[0183]

[0184]

[0185] Where m represents the number of the well, M is the total number of the selected wells, and L m represents the collection of porosity curve, lithofacies interpretation result curve and logging data curve of well numbered m. is the L of M wells m The average value of , q is the adjustment factor.

[0186] In some embodiments, validating the LASSO porosity prediction model includes:

[0187] When the verification fails, the q value in the current LASSO porosity prediction model is modified to adjust the current LASSO porosity prediction model.

[0188] In some embodiments, the value range of q is 1, 2, 4, or 6.

[0189] In some embodiments, step 5 specifically includes:

[0190] Step 51, using the current LASSO porosity prediction model, predicting the porosity curves of the remaining wells based on the lithofacies interpretation result curves and the logging data curves of the remaining wells in the validation data set;

[0191] Step 52, comparing the predicted porosity curves and the measured porosity curves of the remaining wells;

[0192] Step 53, if the degree of agreement between the predicted porosity curves and the measured porosity curves of the remaining wells reaches the preset standard, then the verification is confirmed to be successful, and the current LASSO porosity prediction model is confirmed to be the optimal LASSO porosity prediction model; if the degree of agreement between the predicted porosity curves and the measured porosity curves of the remaining wells does not reach the preset standard, then the verification is confirmed to be unsuccessful, and the current LASSO porosity prediction model is adjusted, and then the process returns to step 51, and the porosity curves of the remaining wells are predicted again using the adjusted LASSO porosity prediction model, until the degree of agreement between the predicted porosity curves and the measured porosity curves of the remaining wells reaches the preset standard.

[0193] In some embodiments, the number of the partial wells used to construct the training data set accounts for 55% to 85% of the plurality of wells.

[0194] Specifically, the memory may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, a random access memory (RAM) and / or a cache memory (cache), etc. The non-volatile memory may include, for example, a read-only memory (ROM), a hard disk, a flash memory, etc.

[0195] The processor may be a central processing unit (CPU) or other forms of processing units with data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device to perform desired functions. In one embodiment of the present invention, the processor is used to run the computer-readable instructions stored in the memory.

[0196] The present embodiment discloses a volcanic rock porosity prediction device. The present embodiment establishes the relationship between porosity and lithofacies and conventional logging curves based on existing measured data through the LASSO algorithm. The present embodiment relies on the measured data of the study area rather than empirical parameters or core samples of a specific area in the past. In addition, the prediction process of the present embodiment has less human interference, and lithofacies information is added to the prediction, which can effectively improve the accuracy of porosity prediction in the study area.

[0197] For detailed description of this embodiment, reference may be made to the corresponding descriptions in the aforementioned embodiments, which will not be repeated here.

[0198] Example 4

[0199] According to another aspect of the present invention, a computer-readable storage medium is provided. The computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the method for predicting the porosity of volcanic rocks according to the present invention is implemented.

[0200] The method comprises the following steps:

[0201] Step 1, obtaining measured porosity curves, lithofacies interpretation result curves and logging data curves of multiple wells;

[0202] Step 2, selecting measured porosity curves, lithofacies interpretation result curves and logging data curves of some wells among the plurality of wells to construct a training data set;

[0203] Step 3, using the LASSO algorithm to learn the relationship between porosity and lithofacies interpretation results and logging data based on the constructed training data set, and obtaining an initial LASSO porosity prediction model;

[0204] Step 4, constructing a verification data set using the measured porosity curves, lithofacies interpretation result curves and logging data curves of the remaining wells in the plurality of wells;

[0205] Step 5, using the verification data set to verify the initial LASSO porosity prediction model to obtain the optimal LASSO porosity prediction model;

[0206] Step 6: Use the optimal LASSO porosity prediction model to predict the porosity of wells in the study area that do not have measured porosity curves.

[0207] In some embodiments, the logging data curves include acoustic wave time difference curves, natural gamma curves, and density curves.

[0208] In some embodiments, in step 3, the following J(L m ) is used as the objective function to construct the initial LASSO porosity prediction model:

[0209]

[0210]

[0211]

[0212] Where m represents the number of the well, M is the total number of the selected wells, and L m represents the collection of porosity curve, lithofacies interpretation result curve and logging data curve of well numbered m. is the L of M wells m The average value of , q is the adjustment factor.

[0213] In some embodiments, validating the LASSO porosity prediction model includes:

[0214] When the verification fails, the q value in the current LASSO porosity prediction model is modified to adjust the current LASSO porosity prediction model.

[0215] In some embodiments, the value range of q is 1, 2, 4, or 6.

[0216] In some embodiments, step 5 specifically includes:

[0217] Step 51, using the current LASSO porosity prediction model, predicting the porosity curves of the remaining wells based on the lithofacies interpretation result curves and the logging data curves of the remaining wells in the validation data set;

[0218] Step 52, comparing the predicted porosity curves and the measured porosity curves of the remaining wells;

[0219] Step 53, if the degree of agreement between the predicted porosity curves and the measured porosity curves of the remaining wells reaches the preset standard, then the verification is confirmed to be successful, and the current LASSO porosity prediction model is confirmed to be the optimal LASSO porosity prediction model; if the degree of agreement between the predicted porosity curves and the measured porosity curves of the remaining wells does not reach the preset standard, then the verification is confirmed to be unsuccessful, and the current LASSO porosity prediction model is adjusted, and then the process returns to step 51, and the porosity curves of the remaining wells are predicted again using the adjusted LASSO porosity prediction model, until the degree of agreement between the predicted porosity curves and the measured porosity curves of the remaining wells reaches the preset standard.

[0220] In some embodiments, the number of the partial wells used to construct the training data set accounts for 55% to 85% of the plurality of wells.

[0221] The present embodiment discloses a volcanic rock porosity prediction device. The present embodiment establishes the relationship between porosity and lithofacies and conventional logging curves based on existing measured data through the LASSO algorithm. The present embodiment relies on the measured data of the study area rather than empirical parameters or core samples of a specific area in the past. In addition, the prediction process of the present embodiment has less human interference, and lithofacies information is added to the prediction, which can effectively improve the accuracy of porosity prediction in the study area.

[0222] The computer-readable storage medium according to the embodiment of the present invention stores non-transitory computer-readable instructions, and when the non-transitory computer-readable instructions are executed by a processor, all or part of the steps of the above-mentioned methods of the embodiments of the present invention are executed.

[0223] The above-mentioned computer-readable storage media include, but are not limited to: optical storage media (e.g., CD-ROM and DVD), magneto-optical storage media (e.g., MO), magnetic storage media (e.g., magnetic tape or mobile hard disk), media with built-in rewritable non-volatile memory (e.g., memory card) and media with built-in ROM (e.g., ROM box).

[0224] Those skilled in the art should be able to understand that in order to solve the technical problem of how to obtain a good user experience, the present embodiment may also include well-known structures such as a communication bus and an interface, and these well-known structures should also be included in the protection scope of the present invention.

[0225] For detailed description of this embodiment, reference may be made to the corresponding descriptions in the aforementioned embodiments, which will not be repeated here.

[0226] Example 5

[0227] The present invention selects 7 wells in a certain study block, uses 5 wells as learning wells to establish the relationship between porosity and lithofacies interpretation results and conventional logging curves for porosity prediction, and illustrates the results by comparing the measured porosity curves of the remaining 2 uninvolved wells with the predicted porosity curves.

[0228] Five wells M1, M2, M3, M4, and M5 can be randomly selected, and a training data set Lm={p m ,litho m , ac m ,gr m ,den m}(m=1~5), Figure 2 The graphs of measured porosity, lithofacies interpretation results and conventional logging of these five wells are shown.

[0229] The data of the remaining two wells can be used as the validation data set Pn = {p m ,litho n , ac n ,gr n ,den n}(n=1~2), the prediction accuracy of the LASSO porosity prediction model obtained based on the training data set was verified.

[0230] When selecting data, the porosity, lithofacies interpretation results, sonic time difference, gamma and density values ​​at the same depth point were selected for all seven wells.

[0231] According to the present invention, the training data set is used to train the LASSO model, an initial LASSO porosity prediction model is established, and then the {litho n , ac n ,gr n ,den n} as input for calculation, and the predicted porosity curves of the remaining two wells are obtained, which are compared with the measured porosity curves, such as Figure 3 As shown in the figure, it can be observed that the predicted porosity curves of these two wells are highly consistent with the measured porosity curves. Therefore, there is no need to adjust the initial LASSO porosity prediction model in this example.

[0232] from Figure 3 It can be seen that the porosity prediction results for wells without measured porosity curves according to the present invention are highly accurate, providing better data support for subsequent reservoir analysis.

[0233] For other detailed descriptions of this exemplary embodiment, reference may be made to the corresponding descriptions in the aforementioned embodiments, which will not be repeated here.

[0234] The embodiments of the present invention have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The terms used herein are selected to best explain the principles of the embodiments, practical applications, or technical improvements to the technology in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. A method for predicting volcanic rock porosity, characterized in that: The method comprises: Step 1, obtaining measured porosity curves, lithofacies interpretation result curves and logging data curves of multiple wells; Step 2, selecting measured porosity curves, lithofacies interpretation result curves and logging data curves of some wells among the plurality of wells to construct a training data set; Step 3, using the LASSO algorithm, based on the constructed training data set, the relationship between porosity and lithofacies interpretation results and logging data is learned to obtain the initial LASSO porosity prediction model; Step 4, constructing a verification data set using the measured porosity curves, lithofacies interpretation result curves and logging data curves of the remaining wells in the plurality of wells; Step 5, using the verification data set to verify the initial LASSO porosity prediction model to obtain the optimal LASSO porosity prediction model; Step 6: Use the optimal LASSO porosity prediction model to predict the porosity of wells in the study area that do not have measured porosity curves.

2. The method according to claim 1, characterized in that The logging data curves include an acoustic time difference curve, a natural gamma curve and a density curve.

3. The method according to claim 3, characterized in that: In step 3, the following J(L m ) is used as the objective function to construct the initial LASSO porosity prediction model: Where m represents the number of the well, M is the total number of the selected wells, and L m represents the collection of porosity curve, lithofacies interpretation result curve and logging data curve of well numbered m. is the L of M wells m The average value of , q is the adjustment factor.

4. The method according to claim 3, characterized in that The validation of the LASSO porosity prediction model includes: When the verification fails, the q value in the current LASSO porosity prediction model is modified to adjust the current LASSO porosity prediction model.

5. The method according to claim 4, characterized in that The value range of q is 1, 2, 4, and 6.

6. The method according to claim 1, characterized in that The step 5 specifically includes: Step 51, using the current LASSO porosity prediction model, predicting the porosity curves of the remaining wells based on the lithofacies interpretation result curves and the logging data curves of the remaining wells in the validation data set; Step 52, comparing the predicted porosity curves and the measured porosity curves of the remaining wells; Step 53, if the degree of agreement between the predicted porosity curves and the measured porosity curves of the remaining wells reaches the preset standard, then the verification is confirmed to be successful, and the current LASSO porosity prediction model is confirmed to be the optimal LASSO porosity prediction model; if the degree of agreement between the predicted porosity curves and the measured porosity curves of the remaining wells does not reach the preset standard, then the verification is confirmed to be unsuccessful, and the current LASSO porosity prediction model is adjusted, and then the process returns to step 51, and the porosity curves of the remaining wells are predicted again using the adjusted LASSO porosity prediction model, until the degree of agreement between the predicted porosity curves and the measured porosity curves of the remaining wells reaches the preset standard.

7. The method according to claim 1, characterized in that The number of the partial wells used to construct the training data set accounts for 55% to 85% of the plurality of wells.

8. A volcanic rock porosity prediction device, characterized in that: The device comprises: A measured data acquisition unit, used to obtain measured porosity curves, lithofacies interpretation result curves and logging data curves of multiple wells; A training data set construction unit, used for selecting measured porosity curves, lithofacies interpretation result curves and logging data curves of some wells among the plurality of wells to construct a training data set; An initial model building unit is used to adopt the LASSO algorithm to learn the relationship between porosity and lithofacies interpretation results and logging data based on the constructed training data set, and obtain an initial LASSO porosity prediction model; A verification data set construction unit, used to construct a verification data set using the measured porosity curves, lithofacies interpretation result curves and logging data curves of the remaining wells in the plurality of wells; A model verification unit, used to verify the initial LASSO porosity prediction model using the verification data set to obtain an optimal LASSO porosity prediction model; The porosity prediction unit is used to use the optimal LASSO porosity prediction model to predict the porosity of wells in the study area that have no measured porosity curves.

9. The device according to claim 8, characterized in that The model adjustment confirmation unit specifically includes: A porosity curve prediction subunit, used to predict the porosity curves of the remaining wells based on the lithofacies interpretation result curves and the logging data curves of the remaining wells in the validation data set by using the current LASSO porosity prediction model; A comparison subunit, used for comparing the predicted porosity curves and the measured porosity curves of the remaining wells; The best model confirmation subunit is used to confirm that the verification has passed if the degree of fit between the predicted porosity curves and the measured porosity curves of the remaining wells reaches a preset standard, and confirm that the current LASSO porosity prediction model is the best LASSO porosity prediction model; if the degree of fit between the predicted porosity curves and the measured porosity curves of the remaining wells does not reach the preset standard, then confirm that the verification has not passed, adjust the current LASSO porosity prediction model, and return the adjusted LASSO porosity prediction model to the porosity curve prediction subunit, so that the porosity curve prediction subunit uses the adjusted LASSO porosity prediction model to predict the porosity curves of the remaining wells again until the degree of fit between the predicted porosity curves and the measured porosity curves of the remaining wells reaches the preset standard.

10. An electronic device, characterized in that: The electronic device comprises: Memory, which stores executable instructions: A processor, wherein the processor runs the executable instructions in the memory to implement the method according to any one of claims 1 to 7.

Images