A method and device for calculating shale free oil saturation

Abstract

The application provides a shale free oil saturation calculation method and device. Through obtaining the surface temperature, geothermal gradient and spectrum of shale, the logging interpretation data is calculated by using a functional relationship; then a training data set and a verification data set are established, the verification data set is sent to a machine learning algorithm model trained by using the training data set, and a data learning result is obtained; based on the data learning result and the spectrum without three-peak characteristics, free oil signal data in the spectrum without three-peak characteristics is determined, the free oil signal data is fitted, and a normal distribution expression of the free oil spectrum is obtained; and the ratio of the free oil porosity component and the effective porosity component is used to calculate the free oil saturation of the target shale, so that the problems of high data acquisition cost and large calculation error in the existing shale free oil saturation calculation method are solved.

Description

Technical Field

[0001] This application relates to the field of shale oil exploration and development technology, and in particular to a method and apparatus for calculating the saturation of free oil in shale. Background Technology

[0002] Shale oil is an important alternative resource for my country's oil and gas resources. Free oil, as the most easily extracted crude oil in shale oil reservoirs, is currently a research hotspot in the field of oil and gas exploration. Free oil saturation is a key factor determining the economic viability of shale oil exploration and development, and accurate calculation of free oil saturation is of great significance for improving the recovery rate of free oil in shale.

[0003] Currently, various models are mainly used to calculate the saturation of free oil in shale, such as saturation evaluation models and mathematical models based on the amount of free oil and adsorbed oil. In addition, geochemical or engineering parameters can be used to calculate the saturation of free oil in shale through functional relationships, and a method based on nuclear magnetic resonance logging combined with centrifuge experiments can be used to determine the free oil saturation in different throat regions.

[0004] Among the above calculation methods, the calculation method based on multiple models is quick and simple, but the accuracy of the calculation results is low; the calculation method using geochemical or engineering parameters has smaller calculation errors, but requires more types of parameters and a larger workload of measurement; in addition, it is difficult to obtain free oil information by nuclear magnetic resonance experiments, so it is difficult to apply it widely. Summary of the Invention

[0005] This application provides a method and apparatus for calculating the saturation of free oil in shale, in order to solve the problems of high data acquisition costs and large calculation errors in existing methods for calculating the saturation of free oil in shale.

[0006] On the one hand, this application provides a method for calculating the saturation of free oil in shale, including:

[0007] Obtain shale surface temperature, geothermal gradient, and lateral relaxation time. The spectrum is used to calculate well logging interpretation data using functional relationships;

[0008] Based on the well logging interpretation data and the... The algorithm establishes a training dataset and a validation dataset, sends the validation dataset to a machine learning algorithm model trained using the training dataset, and obtains data learning results, which include: The T2 value corresponding to the peak value of the free oil signal in the spectrum, among which, those without the three-peak characteristic If a third peak exists in the spectrum, the value corresponding to the peak value of the third peak shall be used. The T2 value corresponds to the peak value of the free oil signal;

[0009] From the above Obtain from the spectrum that do not have a three-peak characteristic The spectrum, based on the learning results of the data, and the spectrum that does not have a three-peak characteristic The correspondence between spectra determines those that do not have a three-peak characteristic. The free oil signal data in the spectrum is obtained, and the free oil signal data is fitted to obtain the normal distribution expression of the free oil spectrum;

[0010] Read each free oil from the free oil spectrum. The free oil porosity component is used to calculate the free oil saturation of the shale by using the ratio of the free oil porosity component to the effective porosity component.

[0011] On the other hand, this application provides a method and apparatus for calculating the saturation of free oil in shale, comprising:

[0012] The acquisition module is used to acquire the surface temperature, geothermal gradient, and lateral relaxation time of the shale. The spectrum is used to calculate well logging interpretation data using functional relationships;

[0013] The learning module is used to compare the well logging interpretation data with the... The algorithm establishes a training dataset and a validation dataset, sends the validation dataset to a machine learning algorithm model trained using the training dataset, and obtains data learning results, which include: The T2 value corresponding to the peak value of the free oil signal in the spectrum, among which, those without the three-peak characteristic If a third peak exists in the spectrum, the value corresponding to the peak value of the third peak shall be used. The T2 value corresponds to the peak value of the free oil signal;

[0014] Fitting module, used to obtain from the Obtain from the spectrum that do not have a three-peak characteristic The spectrum, based on the learning results of the data, and the spectrum that does not have a three-peak characteristic The correspondence between spectra determines those that do not have a three-peak characteristic. The free oil signal data in the spectrum is obtained, and the free oil signal data is fitted to obtain the normal distribution expression of the free oil spectrum;

[0015] The calculation module reads each free oil from the free oil spectrum. The free oil porosity component is used to calculate the free oil saturation of the shale by using the ratio of the free oil porosity component to the effective porosity component.

[0016] This application provides a method and apparatus for calculating the free oil saturation of shale, which obtains the surface temperature, geothermal gradient, and other parameters of the shale. The spectrum is used to calculate well logging interpretation data using functional relationships; then, training and validation datasets are established, and the validation dataset is sent to a machine learning algorithm model trained on the training dataset to obtain data learning results; based on these data learning results and those without three-peak characteristics... The spectrum was used to determine the spectrum that did not have a three-peak characteristic. The method involves analyzing free oil signal data from the shale spectrum and fitting this data to obtain a normal distribution expression for the free oil spectrum. The free oil saturation of the target shale is then calculated using the ratio of the free oil porosity component to the effective porosity component. Compared to existing methods for calculating shale free oil saturation, the method provided in this application combines the learning results of the machine learning algorithm model with the function expression, thus solving the problem of large calculation errors in existing methods. Attached Figure Description

[0017] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0018] Figure 1 This is a schematic diagram of the shale layer structure on which the embodiments of this application are based;

[0019] Figure 2 A flowchart illustrating a method for calculating the saturation of free oil in shale, provided as an embodiment of this application;

[0020] Figure 3 This is a schematic diagram of the signaling interaction of the shale free oil saturation calculation method provided in the embodiments of this application;

[0021] Figure 4 Provided for the embodiments of this application Spectral distribution diagram;

[0022] Figure 5 Provided for the embodiments of this application Spectral distribution diagram Figure 2 ;

[0023] Figure 6 This is a schematic diagram of the structure of the XGBoost algorithm model provided in the embodiments of this application;

[0024] Figure 7 Provided for the embodiments of this application Spectral distribution diagram Figure 3 ;

[0025] Figure 8 Provided for the embodiments of this application Spectral distribution diagram Figure 4 ;

[0026] Figure 9 This is a schematic diagram of the free oil saturation distribution provided in an embodiment of this application;

[0027] Figure 10 A structural block diagram of the shale free oil saturation calculation device provided in the embodiments of this application;

[0028] Figure 11 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0029] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0030] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0031] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of the relevant data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and corresponding operation entry points are provided for users to choose to authorize or refuse.

[0032] Shale oil is an important alternative resource for my country's oil and gas resources. Free oil, as the most easily extracted crude oil in shale oil reservoirs, is currently a research hotspot in the oil and gas exploration field. Free oil saturation is a key factor determining the economic viability of shale oil exploration and development, and accurate calculation of free oil saturation is of great significance for improving the recovery rate of free oil in shale. Currently, various models are mainly used to calculate shale free oil saturation, but the accuracy of the calculation results is low. Geochemical or engineering parameters can also be used to calculate shale free oil saturation through functional relationships, but this requires a variety of parameter types. In addition, a method based on nuclear magnetic resonance logging combined with centrifuge experiments can be used to determine free oil saturation in different throat intervals, but obtaining free oil information through nuclear magnetic resonance experiments is difficult and its widespread application is limited. Calculating shale free oil saturation remains a key problem that urgently needs to be solved in shale oil exploration and development.

[0033] Figure 1This is a schematic diagram of the shale layer structure on which the embodiments of this application are based. See also: Figure 1 As shown, the mineral composition of shale formations is complex, typically including shale oil, quartz 103, clay matter 104, carbonate minerals, pyrite, etc. Shale oil is further divided into free oil 101 and adsorbed oil 102. Free oil 101 mainly refers to petroleum in shale pores and fractures, while adsorbed oil 102 refers to petroleum resources in adjacent and interlayered layers of dense carbonate or clastic rocks within the shale formation.

[0034] Considering the deep location and complex structure of shale formations, the analytical data obtained using conventional logging techniques is insufficient, and the cost of combining geochemical or engineering parameters with nuclear magnetic resonance (NMR) logging is high. Furthermore, shale has a source-reservoir integrated characteristic, and calculating shale free oil saturation using only machine learning algorithms or formulas results in significant errors. Therefore, this application provides a method and apparatus for calculating shale free oil saturation using only NMR logging data, addressing the problems of high data acquisition costs and large calculation errors in existing shale free oil saturation calculation methods.

[0035] Figure 2 A flowchart illustrating a method for calculating the saturation of free oil in shale, provided in an embodiment of this application, is shown below. Figure 2 As shown in the embodiments of this application, the method for calculating the saturation of free oil in shale includes:

[0036] S201. Obtain the surface temperature, geothermal gradient, and lateral relaxation time of the shale. The spectrum is used to calculate well logging interpretation data using functional relationships.

[0037] Before calculating the free oil saturation of shale, researchers use nuclear magnetic resonance logging technology to measure the target shale layer, thereby obtaining the surface temperature, geothermal gradient, and lateral relaxation time of the target shale. Spectrum. The spectrum is porosity about The curve showing the change of [value]. This can be understood as, in a coordinate system, the horizontal axis being [value]. The vertical axis represents porosity, and each Each has a corresponding porosity component.

[0038] Specifically, the well logging interpretation data includes: Geometric mean, effective porosity, and formation temperature profile.

[0039] calculate The functional relationship of the geometric mean is:

[0040] (1)

[0041] In the above functional relationship, yes Geometric mean, dimensionless; It is the lateral relaxation time, measured in milliseconds (ms). yes The time component, in milliseconds (ms); f i They are different The corresponding porosity component, in percentage.

[0042] The functional relationship for calculating effective porosity is:

[0043] (2)

[0044] In the above functional relationship, Effective porosity calculated with a starting time of 3ms, expressed as a percentage (%). yes The maximum value of the time component, in milliseconds (ms); They are different The corresponding porosity component, in percentage.

[0045] The functional relationship for calculating the formation temperature curve is:

[0046] (3)

[0047] In the above functional relationship, For depth The temperature of the formation at that time, in degrees Celsius. ; This represents the geothermal gradient, expressed in degrees Celsius per meter. ; The surface temperature is expressed in degrees Celsius. .

[0048] In one implementation, researchers want to obtain the effective porosity of the target shale, see [link to relevant documentation]. Figure 4 As shown, Figure 4 Provided for the embodiments of this application Schematic diagram of spectral distribution. From the diagram, we can read... For 3000ms, different The porosity component corresponding to the component can also be directly read from the graph. For example, the porosity component corresponding to 3ms is approximately 0.2%, therefore the effective porosity is the porosity component relative to the porosity component. The definite integral between 3ms and 3000ms.

[0049] In another approach, researchers aim to obtain the formation temperature curve of the target shale. If the measured surface temperature of the target shale region is 20℃ and the geothermal gradient is 2.35℃ / 100m, then the functional relationship for calculating the formation temperature curve is... Therefore, the formation temperature curve of the target shale can be obtained as follows: In the formula For depth The temperature of the formation at that time.

[0050] This method uses specialized measuring instruments to obtain the surface temperature, geothermal gradient, and other parameters of the target shale. The instrument offers high accuracy and speed in spectral analysis. Furthermore, it can be integrated with a microprocessor to automate measurement, fault diagnosis, recording, data processing, and analysis.

[0051] S202, based on well logging interpretation data and The algorithm establishes training and validation datasets, sends the validation dataset to the machine learning algorithm model trained on the training dataset, and obtains the data learning results.

[0052] Obtaining data with a three-peak characteristic from well logging interpretation data The effective porosity corresponding to the spectrum, the three peaks Geometric mean and three-peak formation temperature curve, and with three peaks The spectrum is packaged into data to create a training dataset; data without a three-peak characteristic are extracted from well logging interpretation data. The non-trimodal effective porosity corresponding to the spectrum, non-trimodal Geometric mean and non-trimodal formation temperature curves were obtained, and the data were packaged to create a validation dataset.

[0053] The results of data learning can be understood as: The T2 value corresponding to the peak value of the free oil signal in the spectrum, among which, those without the three-peak characteristic If a third peak exists in the spectrum, the value corresponding to the peak value of the third peak shall be used. The T2 value corresponds to the peak value of the free oil signal.

[0054] See Figure 4 As shown, Figure 4 Provided for the embodiments of this application The spectral distribution diagram clearly shows three prominent peaks, hence this... The spectrum is considered to have a three-peak characteristic. Spectrum, selected as the training dataset; see [link / reference] Figure 5 As shown, Figure 5 Provided for the embodiments of this application Spectral distribution diagram Figure 2The image clearly shows only two peaks, therefore this... The spectrum is considered to lack the characteristic of a three-peak pattern. The spectrum was selected as the validation dataset.

[0055] The machine learning algorithm model can be an SVM algorithm model, a decision tree algorithm model, a kernel function algorithm model, an XGBoost algorithm model, or other algorithm models, without any limitation.

[0056] In one implementation, after establishing training and validation datasets, researchers train the XGBoost algorithm model using the training dataset and input the validation dataset into the trained model, allowing it to learn autonomously and obtain data-driven learning results. See also... Figure 6 As shown, Figure 6 The diagram shows the structure of the XGBoost algorithm model provided in the embodiments of this application. The XGBoost algorithm model is an algorithm framework that linearly combines multiple base learners to generate a strong learner. Only base learner 1, base learner 2 and base learner N are shown in the figure. It can be understood that there may also be base learner N+1 and base learner N+N. The number of base learners is not limited here.

[0057] After obtaining the training dataset, the XGBoost algorithm first performs PCA dimensionality reduction on the training dataset, then sends the dimensionality-reduced training dataset to the training set section. The XGBoost algorithm model is then trained using the Cart tree constructed within the XGBoost algorithm model. The function expression of the XGBoost algorithm model is as follows:

[0058] (4)

[0059] In the above function expression, These are the model's predicted values; The number of trees to build for the model; For the first The weight of each tree; For the first The predicted value for each tree.

[0060] The construction of multiple Cart trees within the XGBoost algorithm model should employ a greedy approach, building each tree in a manner that maximizes gain. To minimize the loss function of the training data, optimization should be performed before... The results for the current Cart tree are known, so we only need to optimize each leaf node of the current Cart tree. To prevent overfitting, an L2 regularization term is added to the objective function. The objective function is expressed as follows:

[0061] (5)

[0062] In the above function expression, The objective function is... This represents the number of leaf nodes in the current learner. To fall into the first The sum of the first derivatives of all samples within each leaf node; To fall into the first The sum of the second derivatives of all samples within each leaf node; and This is a hyperparameter.

[0063] When the objective function is optimized, the function expression for the leaf node values ​​is:

[0064] (6)

[0065] In the above function expression, For the first leaf node values; To fall into the first The sum of the first derivatives of all samples within each leaf node; To fall into the first The sum of the second derivatives of all samples within each leaf node; This is a hyperparameter.

[0066] This method uses a machine learning algorithm model to model conditions that do not exhibit trimodal characteristics. The spectrum assumes the existence of a third peak, and can only utilize those with three-peak characteristics. In terms of spectrum calculation, it improves the performance of spectra that do not have a three-peak characteristic. The utilization rate of the spectrum provides data support for subsequent calculations.

[0067] S203, from Obtain from the spectrum that do not have a three-peak characteristic Spectrum, based on data learning results and without trimodal characteristics The correspondence between spectra determines those that do not have a three-peak characteristic. The free oil signal data in the spectrum is obtained, and the free oil signal data is fitted to obtain the normal distribution expression of the free oil spectrum.

[0068] The learning results of the data obtained in S202 and those without the three-peak characteristic The free oil signal data in the spectrum are packaged to obtain a fitted dataset; then the fitted dataset is sent to the normal model to obtain the normal distribution expression of the free oil spectrum.

[0069] When building a fitted dataset, it is necessary to consider the data learning results and those without trimodal characteristics. The last peak in the spectrum Determine whether the data learning result is greater than the final peak. If it is greater than a preset threshold, obtain the first value corresponding to when the porosity component is less than a preset threshold. Select from the data learning results to the first The porosity components corresponding to the intervals are the fitted dataset; if they are less than or equal to the last peak, the data is selected from the last peak. To the first The porosity components corresponding to these values ​​are the fitted dataset.

[0070] The functional expression of the normal model is:

[0071] (7)

[0072] In the above function expression, It is the lateral relaxation time component corresponding to the fitted dataset, in milliseconds (ms). Is with The corresponding porosity component, in percentages; The peak value of the fitted normal distribution is dimensionless. The mean of the fitted normal distribution is dimensionless. is the variance of the fitted normal distribution, which is dimensionless.

[0073] In one implementation, see [link to implementation details]. Figure 5 As shown, Figure 5 Provided for the embodiments of this application Spectral distribution diagram Figure 2 In the picture The results of the data learning are in milliseconds (ms). for The last peak in the spectrum The unit is milliseconds (ms). At this time... The data learning result is less than or equal to the final peak. If the preset threshold value corresponding to the porosity component The lateral relaxation time of the fitted dataset is 1000ms. arrive The first one greater than 1000ms The porosity components corresponding to the intervals.

[0074] In another implementation, see Figure 7 As shown, Figure 7 Provided for the embodiments of this application Spectral distribution diagram Figure 3 In the picture The results of the data learning are in milliseconds (ms). for The last peak in the spectrum The unit is milliseconds (ms). At this time... The data learning result is greater than the final peak. ,but The porosity component corresponding to the location is very small, and can be regarded as having almost no porosity component. Therefore, no data is used for fitting at this time.

[0075] In another implementation method, see Figure 8 As shown, Figure 8 Provided for the embodiments of this application Spectral distribution diagram Figure 4 In the picture The results of the data learning are in milliseconds (ms). for The last peak in the spectrum The unit is milliseconds (ms). At this time... The data learning result is greater than the final peak. If the preset threshold value corresponding to the porosity component The time is 1000ms, and the fitted dataset is... yes arrive The first one greater than 1000ms The porosity components corresponding to the intervals.

[0076] This method takes into account that the porosity component with small values ​​is a weak signal, and the presence of weak signals has little impact on the fitting results. Therefore, a threshold is set for the porosity component, which can effectively filter out the porosity component with small values, thereby reducing unnecessary data and improving computational efficiency and accuracy.

[0077] S204. Read each free oil from the free oil spectrum. The free oil porosity component is used to calculate the free oil saturation of the target shale by using the ratio of the free oil porosity component to the effective porosity component.

[0078] Reading free oil from the free oil spectrum Maximum and minimum values; calculate the normal distribution expression for free oil. In free oil Calculate the definite integral between the maximum and minimum values, and then calculate the free oil porosity component with respect to the free oil. In free oil The definite integral between the maximum and minimum values ​​is used to apply the normal distribution expression to free oil. The definite integral of the free oil porosity component with respect to the free oil The definite integral is used as the quotient to obtain the free oil saturation of the shale.

[0079] The functional expression for calculating the free oil saturation of the target shale, using the ratio of the free oil porosity component to the effective porosity component, is as follows:

[0080] (8)

[0081] In the above function expression, The free oil saturation of the target shale is expressed as a percentage (%). The expression for the normal distribution is obtained from S203; yes The maximum value of the component, in milliseconds (ms); yes The minimum value of the time component, in milliseconds (ms); They are different The corresponding porosity component, in percentage.

[0082] In one implementation, see [link to implementation details]. Figure 9 As shown, Figure 9 This is a schematic diagram of the free oil saturation distribution provided in an embodiment of this application. The first channel in the diagram does not exhibit a three-peak characteristic. The graph shows the spectrum; the second channel represents the normal distribution expression calculated from S203, i.e., the free oil spectrum; the third channel represents the shale free oil saturation calculated based on the free oil spectrum in the second channel. In the third channel, it can be seen that the calculated shale free oil saturation is related to... Figure 9 The ratio of the area of ​​the third peak to the total peak area shown is consistent with the requirements of well logging evaluation.

[0083] This method calculates the free oil saturation of shale using only nuclear magnetic resonance logging data, solving the problem of high data acquisition costs in existing shale free oil saturation calculation methods. At the same time, it combines the machine algorithm model learning results with the function expression, solving the problem of large calculation errors in existing shale free oil saturation calculation methods.

[0084] Figure 3 This is a schematic diagram of the signaling interaction for the shale free oil saturation calculation method provided in this application embodiment. (See attached diagram.) Figure 3 As shown, combined with Figure 2 The method for calculating the saturation of free oil in shale provided in this application includes the following steps:

[0085] S301, After the measuring instrument measures the target shale, it will... The spectrum is sent to the measurement server.

[0086] S302, Measurement server from Reading from the spectrum Maximum value, each The corresponding porosity component.

[0087] S303, the measurement server will read Maximum value, each The corresponding porosity component and the received The spectrum is sent to the computing server.

[0088] S304. After measuring the target shale, the measuring instrument sends the surface temperature and geothermal gradient to the computing server.

[0089] S305, The computing server calculates based on the received data. Geometric mean, effective porosity, and formation temperature profile.

[0090] from Read each from the spectrum The corresponding porosity component is calculated for each The sum of the products of the corresponding porosity components is used to obtain the product of the product of the product of each porosity component. The logarithm of the geometric mean, which is then obtained by taking the exponent. Geometric mean.

[0091] from Reading from the spectrum The maximum value of the porosity component is calculated with respect to the porosity component. The effective porosity is obtained by definite integrals between 3 milliseconds and the maximum value.

[0092] Calculate the product of the geothermal gradient and the formation depth, and sum it with the surface temperature to obtain the formation temperature curve between formation temperature and formation depth.

[0093] S306. The computing server creates training and validation datasets.

[0094] The computing server has a three-peak characteristic. Geometric mean, effective porosity, formation temperature profile and The spectrum is packaged into data to create a training dataset.

[0095] The computing server does not have a three-peak characteristic. Geometric mean, effective porosity, and formation temperature curves were packaged together to create a validation dataset.

[0096] S307. The computing server sends the verification dataset to the machine learning model trained using the training dataset to obtain the data learning results.

[0097] The machine learning algorithm model can be an SVM algorithm model, a decision tree algorithm model, a kernel function algorithm model, an XGBoost algorithm model, or other algorithm models, and there is no limitation here.

[0098] The learning results from this data include: The T2 value corresponding to the peak value of the free oil signal in the spectrum, among which, those without the three-peak characteristic If a third peak exists in the spectrum, the value corresponding to the peak value of the third peak shall be used. The T2 value corresponds to the peak value of the free oil signal.

[0099] S308, the computing server utilizes features that do not have a three-peak characteristic. Spectrum, based on data learning results and without trimodal characteristics The correspondence between spectra determines those that do not have a three-peak characteristic. The free oil signal data in the spectrum is obtained, and the free oil signal data is fitted to obtain the normal distribution expression of the free oil spectrum.

[0100] The results of data learning and those that do not have a three-peak characteristic The free oil signal data in the spectrum are packaged to obtain a fitted dataset; the fitted dataset is sent to the normal model to obtain the normal distribution expression of the free oil spectrum.

[0101] Based on the data learning results and those that do not have a three-peak characteristic The last peak in the spectrum Determine whether the data learning result is greater than the final peak. If so, obtain the first value corresponding to when the porosity component is less than a preset threshold. Select from the data learning results to the first The porosity components corresponding to the intervals are the fitted dataset; otherwise, select the data from the last peak. To the first The porosity corresponding to the values ​​between these values ​​is the fitted dataset.

[0102] S309. The calculation server substitutes the porosity component and the normal distribution expression of the free oil spectrum into the corresponding functional relationship to calculate the free oil saturation.

[0103] Reading free oil from the free oil spectrum Maximum and minimum values; calculate the normal distribution expression for free oil. In free oil Calculate the definite integral between the maximum and minimum values, and then calculate the free oil porosity component with respect to the free oil. In free oil The definite integral between the maximum and minimum values; using the normal distribution expression for free oil The definite integral of the free oil porosity component with respect to the free oil The definite integral is used as the quotient to obtain the free oil saturation of the shale.

[0104] Figure 10 The structural block diagram of the shale free oil saturation calculation device provided in the embodiments of this application is shown. For ease of explanation, only the parts relevant to the embodiments of this application are shown. See also Figure 10 As shown, the shale free oil saturation calculation device provided in this application embodiment includes: an acquisition module 1001, a learning module 1002, a fitting module 1003, and a calculation module 1004.

[0105] Module 1001 is used to acquire the surface temperature, geothermal gradient, and lateral relaxation time of shale. The spectrum is used to calculate well logging interpretation data using functional relationships;

[0106] Learning module 1002 is used to interpret well logging data and... The spectrum is used to create training and validation datasets. The validation dataset is then sent to a machine learning algorithm model trained on the training dataset to obtain the data learning results, which include: The T2 value corresponding to the peak value of the free oil signal in the spectrum, among which, those without the three-peak characteristic If a third peak exists in the spectrum, the value corresponding to the peak value of the third peak shall be used. The T2 value corresponds to the peak value of the free oil signal;

[0107] Fitting module 1003, used to fit from Obtain from the spectrum that do not have a three-peak characteristic Spectrum, based on data learning results and without trimodal characteristics The correspondence between spectra determines those that do not have a three-peak characteristic. The free oil signal data in the spectrum is obtained, and the free oil signal data is fitted to obtain the normal distribution expression of the free oil spectrum;

[0108] Calculation module 1004 is used to read each free oil from the free oil spectrum. The free oil porosity component is used to calculate the free oil saturation of the shale by using the ratio of the free oil porosity component to the effective porosity component.

[0109] The shale free oil saturation calculation device provided in this application obtains the shale surface temperature, geothermal gradient, and... The spectrum is used to calculate well logging interpretation data using functional relationships; then, training and validation datasets are established, and the validation dataset is sent to a machine learning algorithm model trained on the training dataset to obtain data learning results; based on these data learning results and those without three-peak characteristics... The spectrum was used to determine the spectrum that did not have a three-peak characteristic. The method involves analyzing free oil signal data from the shale spectrum and fitting this data to obtain a normal distribution expression for the free oil spectrum. Using the free oil porosity component and the normal distribution expression, the free oil saturation of the target shale is calculated. Compared to existing methods for calculating shale free oil saturation, the method provided in this application combines the learning results of the machine learning algorithm model with the function expression, thus solving the problem of large calculation errors in existing methods.

[0110] Figure 11 See the schematic diagram of the electronic device provided in the embodiments of this application. Figure 11 As shown, the electronic device includes: a memory 1101, a processor 1102, and a computer program; wherein the computer program is stored in the memory 1101 and configured to be executed by the processor 1102. Figures 2 to 9 The processor 1102 is used to implement each step. Figure 10 Each module.

[0111] The memory 1101 and the processor 1102 are connected via a bus 1103.

[0112] For relevant instructions, please refer to the corresponding text. Figures 2 to 10 The relevant descriptions and effects of the steps in the corresponding embodiments are understood, and will not be elaborated on here.

[0113] This application also provides a computer-readable storage medium including computer code that, when run on a computer, causes the computer to perform actions such as... Figures 2 to 11 The method provided by any of the corresponding implementation methods.

[0114] This application also provides a computer program product, including program code, which, when a computer runs the computer program product, executes as follows: Figures 2 to 11 The method provided by any of the corresponding implementation methods.

[0115] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this application are indicated by the following claims.

[0116] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.

Claims

1. A method for calculating the saturation of free oil in shale, characterized in that, include: Obtain shale surface temperature, geothermal gradient, and lateral relaxation time. The spectrum is used to calculate well logging interpretation data using functional relationships; Based on the well logging interpretation data and the... The algorithm establishes a training dataset and a validation dataset, sends the validation dataset to a machine learning algorithm model trained using the training dataset, and obtains data learning results, which include: The T2 value corresponding to the peak value of the free oil signal in the spectrum, among which, those without the three-peak characteristic If a third peak exists in the spectrum, the value corresponding to the peak value of the third peak shall be used. The T2 value corresponds to the peak value of the free oil signal; From the above Obtain from the spectrum that do not have a three-peak characteristic The spectrum is used to determine the non-trimodal characteristics corresponding to the data learning results. Spectrum; Never has a three-peak characteristic In the spectrum, free oil signal data corresponding to the data learning results are obtained, and the free oil signal data is fitted to obtain the normal distribution expression of the free oil spectrum; Read each free oil from the free oil spectrum. The free oil porosity component corresponding to the free oil porosity component is used to calculate the free oil saturation of the shale by using the ratio of the free oil porosity component to the effective porosity component. The well logging interpretation data and the Spectrum, establishing training and validation datasets, including: Obtain the three-peak characteristic from the well logging interpretation data. The spectrum corresponds to the three-peak effective porosity, the geometric mean of the three-peak transverse relaxation time, and the three-peak formation temperature curve, and is compared with the three-peak... The spectrum is packaged into data to create a training dataset; Obtain data from the well logging interpretation data that do not exhibit a three-peak characteristic. The non-trimodal effective porosity corresponding to the spectrum, non-trimodal Geometric mean and non-trimodal formation temperature curves were obtained, and the data were packaged to create a validation dataset.

2. The method according to claim 1, characterized in that, The well logging interpretation data obtained by calculating using functional relationships includes: From the above Read each from the spectrum The corresponding porosity component is used to calculate each of the following: The sum of the products of the corresponding porosity components is used to obtain the product of the product of the product of each porosity component and the quotient of the sum of the products of each porosity component. The logarithm of the geometric mean, which is then obtained by taking the exponent. Geometric mean; From the above Reading from the spectrum The maximum value of the porosity component is calculated with respect to the porosity component. The effective porosity is obtained by definite integral between 3 milliseconds and the maximum value. The product of the geothermal gradient and the formation depth is calculated and summed with the surface temperature to obtain the formation temperature curve between formation temperature and formation depth.

3. The method according to claim 1 or 2, characterized in that, The fitting of the free oil signal data to obtain the normal distribution expression of the free oil spectrum includes: The learning results of the data and the data that do not have the three-peak characteristic The free oil signal data in the spectrum are packaged to obtain a fitted dataset; The fitted dataset is sent to the normal model to obtain the normal distribution expression of the free oil spectrum.

4. The method according to claim 3, characterized in that, The learning results of the data and the data that do not have a three-peak characteristic The free oil signal data in the spectrum are packaged to obtain a fitted dataset, including: Based on the data learning results and the fact that it does not have a three-peak characteristic The last peak corresponding to the last peak in the spectrum Determine whether the data learning result is greater than the final peak. ; If so, obtain the first value corresponding to when the porosity component is less than a preset threshold. Select from the learning results of the data to the first The porosity components corresponding to these intervals are the fitted dataset; If not, select from the last peak transverse relaxation time to the first... The porosity components corresponding to these values ​​are the fitted dataset.

5. The method according to claim 4, characterized in that, The calculation of the free oil saturation of the shale using the ratio of the free oil porosity component to the effective porosity component specifically includes: Read the free oil from the free oil spectrum. Maximum value, minimum value; Calculate the normal distribution expression with respect to free oil. In free oil The definite integral between the maximum and minimum values ​​is calculated, and the free oil porosity component is calculated with respect to the free oil. In free oil The definite integral between the maximum and minimum values; Using the normal distribution expression for free oil The definite integral of the free oil porosity component with respect to the free oil The free oil saturation of the shale is obtained by dividing the definite integral by the quotient.

6. A device for calculating the saturation of free oil in shale, comprising: The acquisition module is used to acquire the surface temperature, geothermal gradient, and lateral relaxation time of the shale. The spectrum is used to calculate well logging interpretation data using functional relationships; The learning module is used to compare the well logging interpretation data with the... The algorithm establishes a training dataset and a validation dataset, sends the validation dataset to a machine learning algorithm model trained using the training dataset, and obtains data learning results, which include: The T2 value corresponding to the peak value of the free oil signal in the spectrum, among which, those without the three-peak characteristic If a third peak exists in the spectrum, the value corresponding to the peak value of the third peak shall be used. The T2 value corresponds to the peak value of the free oil signal; Fitting module, used to obtain from the Obtain from the spectrum that do not have a three-peak characteristic The spectrum is used to determine the non-trimodal characteristics corresponding to the data learning results. Spectrum; never exhibiting a three-peak characteristic In the spectrum, free oil signal data corresponding to the data learning results are obtained, and the free oil signal data is fitted to obtain the normal distribution expression of the free oil spectrum; The calculation module is used to read each free oil from the free oil spectrum. The free oil porosity component corresponding to the free oil porosity component is used to calculate the free oil saturation of the shale by using the ratio of the free oil porosity component to the effective porosity component. The learning module is specifically used to extract data with three-peak characteristics from the well logging interpretation data. The spectrum corresponds to the three-peak effective porosity, the geometric mean of the three-peak transverse relaxation time, and the three-peak formation temperature curve, and is compared with the three-peak... The spectrum is packaged into data to create a training dataset; data without a three-peak characteristic are extracted from the well logging interpretation data. The non-trimodal effective porosity corresponding to the spectrum, non-trimodal Geometric mean and non-trimodal formation temperature curves were obtained, and the data were packaged to create a validation dataset.

7. An electronic device, characterized in that, include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory to implement the method as described in any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the shale free oil saturation calculation method as described in any one of claims 1-5.

Images