CN113050189A - Method, device and equipment for reconstructing logging curve and storage medium - Google Patents

Method, device and equipment for reconstructing logging curve and storage medium Download PDF

Info

Publication number
CN113050189A
CN113050189A CN201911381262.6A CN201911381262A CN113050189A CN 113050189 A CN113050189 A CN 113050189A CN 201911381262 A CN201911381262 A CN 201911381262A CN 113050189 A CN113050189 A CN 113050189A
Authority
CN
China
Prior art keywords
logging
model
curve
reconstructed
reconstruction
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN201911381262.6A
Other languages
Chinese (zh)
Other versions
CN113050189B (en
Inventor
李丙龙
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Gridsum Technology Co Ltd
Original Assignee
Beijing Gridsum Technology Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Beijing Gridsum Technology Co Ltd filed Critical Beijing Gridsum Technology Co Ltd
Priority to CN201911381262.6A priority Critical patent/CN113050189B/en
Publication of CN113050189A publication Critical patent/CN113050189A/en
Application granted granted Critical
Publication of CN113050189B publication Critical patent/CN113050189B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • GPHYSICS
    • G01MEASURING; TESTING
    • G01VGEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
    • G01V9/00Prospecting or detecting by methods not provided for in groups G01V1/00 - G01V8/00

Landscapes

  • Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Life Sciences & Earth Sciences (AREA)
  • General Physics & Mathematics (AREA)
  • Geophysics (AREA)
  • Geophysics And Detection Of Objects (AREA)

Abstract

The method extracts the characteristics of the logging curve to be reconstructed, generates model characteristics according to the characteristics, inputs the model characteristics to a reconstruction model, and outputs the reconstruction model as the reconstructed logging curve. The reconstruction model is a machine model corresponding to the geological type, which is obtained according to the geological type of the logging location, obviously, the reconstruction model can be used for multidimensional nonlinear regression based on machine learning, and the problem that the traditional method is limited by multidimensional and nonlinear can be effectively solved. Thereby obtaining a reconstructed well logging curve with high accuracy. Furthermore, the reconstructed model in this embodiment is a model selected from multiple machine learning models and having the highest reconstruction accuracy for the logging curve of the geological type of the logging location, so that the quality of the reconstructed logging curve is further improved, that is, the reconstructed logging curve can better reflect the geological characteristics of the logging area.

Description

Method, device and equipment for reconstructing logging curve and storage medium
Technical Field
The present application relates to the field of well logging technologies, and in particular, to a method, an apparatus, a device, and a storage medium for reconstructing a well logging curve.
Background
The well logging is an important branch of applied geophysics, and various physical parameters of underground stratum and well bore technical conditions are measured by various well logging instruments of electromagnetism, nuclear physics, acoustics and the like so as to solve various geological and engineering technical problems in oil field exploration and development.
The curves of various stratum parameters obtained by logging and continuously recorded according to the depth are collectively called logging curves. The types of well logging curves include various types, such as resistivity curves, natural potential curves, and sonic time difference curves. However, due to the influence of the logging tool or the geological factors, the logging curve is often distorted to different degrees, and therefore the logging curve needs to be reconstructed.
However, in the conventional reconstruction method, a logging curve is reconstructed by adopting linear regression, and the reconstructed logging curve is inaccurate due to the limitation of input parameter dimension and a regression method. The inaccurate logging curve cannot truly reflect the lithology, physical property or electrical property of the stratum, for example, the inaccurate acoustic moveout curve may cause the lithology recognition error of the stratum, resulting in the difference between the exploration result and the real geology.
Disclosure of Invention
In view of the above, the present invention provides an information processing method and apparatus that overcomes or at least partially solves the above problems.
A method of reconstructing a well log, comprising:
extracting the characteristics of a logging curve to be reconstructed;
acquiring a reconstruction model corresponding to the geological type of a logging location according to the corresponding relation between a preset geological type and the reconstruction model, wherein the logging is logging for acquiring the logging curve to be reconstructed; the reconstruction model corresponding to each geological type in the corresponding relation is a model with the highest reconstruction accuracy of the logging curve of the logging of the geological type in various machine learning models;
inputting the model characteristics into the reconstruction model to obtain a reconstructed logging curve output by the reconstruction model, wherein the model characteristics are determined according to the characteristics.
Optionally, extracting features of the well log to be reconstructed comprises:
acquiring a preset reference curve of the logging curve to be reconstructed, wherein the reference curve is a curve which represents the geological type of the logging curve to be reconstructed and has the same response as that of the logging curve to be reconstructed;
and calculating the reference curve of the well logging curve to be reconstructed to obtain the characteristics, wherein the calculation comprises at least one of derivation, product between curves and quotient between curves.
Optionally, before the extracting the features of the log to be reconstructed, the method further includes:
calibrating the depths of the reference curve and the logging curve to be reconstructed to obtain the calibrated reference curve and the logging curve to be reconstructed;
the obtaining the characteristics by operating the reference curve of the logging curve to be reconstructed comprises:
and calculating the corrected reference curve to obtain the characteristics.
Optionally, the determining process of the correspondence between the geological type and the reconstructed model includes:
obtaining a plurality of logs of the geological region as sample logs;
dividing the sample logs into a first log set and a second log set, wherein the first log set and the second log set do not have an intersection, and the number of sample logs in the first log set is greater than that of the sample logs in the second log set;
training various types of models by using the logging curves of the sample logging in the first logging set to obtain a plurality of models to be selected;
inputting the characteristics of the logging curves of the sample logging in the second logging set into each model to be selected respectively to obtain a reconstructed result logging curve output by each model to be selected;
and taking the output candidate model with the reconstructed result logging curve most similar to the preset sample logging curve as the reconstructed model, wherein the preset sample logging curve is the logging curve which belongs to the same logging and is of the same type as the logging curve of the sample logging in the second logging set.
Optionally, before the inputting the model characteristic into the reconstructed model to obtain a reconstructed log output by the reconstructed model, the method further includes:
and carrying out dimension reduction processing on the features to obtain the model features.
Optionally, the plurality of machine learning models comprises:
a logistic regression model, a k-nearest neighbor model, a gradient lifting tree model, a support vector machine model, a naive Bayes model, a decision tree model and a random forest model.
An apparatus for reconstructing a well log, comprising:
the characteristic extraction unit is used for extracting the characteristics of the logging curve to be reconstructed;
the model acquisition unit is used for acquiring a reconstruction model corresponding to the geological type of the logging location according to the corresponding relation between the preset geological type and the reconstruction model, and the logging is logging for acquiring the logging curve to be reconstructed; the reconstruction model corresponding to each geological type in the corresponding relation is a model with the highest reconstruction accuracy of the logging curve of the logging of the geological type in various machine learning models;
and the curve reconstruction unit is used for inputting the model characteristics into the reconstruction model to obtain a reconstructed logging curve output by the reconstruction model, and the model characteristics are determined according to the characteristics.
Optionally, the model obtaining unit is further configured to obtain a correspondence between the geological type and the reconstructed model according to the following manner:
obtaining a plurality of logs of the geological region as sample logs;
dividing the sample logs into a first log set and a second log set, wherein the first log set and the second log set do not have an intersection, and the number of sample logs in the first log set is greater than that of the sample logs in the second log set;
training various types of models by using the logging curves of the sample logging in the first logging set to obtain a plurality of models to be selected;
inputting the characteristics of the logging curves of the sample logging in the second logging set into each model to be selected respectively to obtain a reconstructed logging curve output by each model to be selected;
and taking the selected model which is most similar to the output reconstructed logging curve and the preset sample logging curve as the reconstructed model.
A storage medium having stored thereon a program which, when executed by a processor, carries out the steps of the method for reconstructing a well log as described above.
An apparatus for reconstructing a well log, the apparatus comprising at least one processor, and at least one memory, bus connected to the processor; the processor and the memory complete mutual communication through the bus; the processor is used for calling the program instructions in the memory so as to execute the reconstruction method of the logging curve.
By means of the technical scheme, the embodiment of the application provides a well logging curve reconstruction method, the characteristics of a well logging curve to be reconstructed are extracted, model characteristics are generated according to the characteristics, the model characteristics are input into a reconstruction model, and the output of the reconstruction model is the reconstructed well logging curve. Obviously, based on the characteristics of machine learning, the reconstruction model can be used for multidimensional nonlinear regression, and the problem that the traditional method is limited by multidimensional and nonlinear factors can be effectively solved, so that the reconstructed well logging curve with high accuracy is obtained. In addition, the reconstruction model is a machine model corresponding to the geological type obtained according to the geological type of the logging location, and a model with the highest accuracy is selected from multiple machine learning models by combining the strong correlation between the logging curve and the geological type so as to determine the reconstruction model corresponding to the logging location type, so that the accuracy of the reconstruction curve can be further improved.
The foregoing description is only an overview of the technical solutions of the present invention, and the embodiments of the present invention are described below in order to make the technical means of the present invention more clearly understood and to make the above and other objects, features, and advantages of the present invention more clearly understandable.
Drawings
Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The drawings are only for purposes of illustrating the preferred embodiments and are not to be construed as limiting the invention. Also, like reference numerals are used to refer to like parts throughout the drawings. In the drawings:
FIG. 1 is a flow chart illustrating a method for reconstructing a well log provided by an embodiment of the present application;
FIG. 2 is a schematic structural diagram illustrating a device for reconstructing a log according to an embodiment of the present disclosure;
fig. 3 shows a schematic structural diagram of a reconstruction apparatus for a well log according to an embodiment of the present application.
Detailed Description
In the field of well logging technology, the raw data of a well logging curve is usually collected and recorded according to 1 meter and 8 sampling points, the data values of each sampling point are connected to form a well logging curve form, and the form of the well logging curve directly reflects the formation properties of the formation, such as lithology, physical properties or electrical properties. In practical applications, the types of raw data collected by different logging tools are different, for example, a sonic velocity logging tool transmits a sound wave through a probe downhole, the sound wave propagates from mud to the formation, and the recorded curve is the time Δ t of the sound wave required to pass through the formation by 1 meter, which varies with depth, and is the sound wave time difference curve. Since Δ t depends on lithology and porosity, rock porosity can be determined, lithology can be identified, strata can be compared, and gas layers can be judged according to the morphology of the sonic moveout curve.
It can be seen that the morphology of different types of well logs in the same region may reflect the geological characteristics of the region.
However, due to the influence of various factors, abnormal points often appear on the well logging curve obtained in the well logging process, which results in that the well logging curve cannot correctly reflect geological characteristics, so that the abnormal well logging curve needs to be reconstructed to repair the abnormal value part.
Taking the type of the logging curve as an acoustic time difference curve as an example, the independent variable of the acoustic time difference curve is the depth of a sampling point, and the dependent variable is the time delta t required by the acoustic wave acquired by the logging instrument to pass through the stratum of 1 meter. Due to the influence of factors such as the borehole diameter, the layer thickness or the cycle skip, an abnormal value of the acoustic wave time difference curve occurs, so that the abnormal value cannot represent the response of the actual stratum, and great influence is brought to subsequent application, and therefore, the reconstruction of the acoustic wave time difference curve to complete the repair of the abnormal value part is very critical and important.
In the prior art, the reconstruction method of the logging curve is to calculate a regression equation, and calculate the repair value of the abnormal part by using the regression equation. For example, when reconstructing the acoustic wave time difference curve by using the prior art, first, a regression equation with better correlation is obtained by performing correlation regression on other well logging curves in the same area with better correlation to the acoustic wave time difference curve and a normal acoustic wave time difference curve. Then, the regression equation is used for calculating the abnormal part in the time difference curve of the sound wave to be reconstructed, so that the sound wave time difference value which is really reflected relatively close to the stratum is obtained. In practice, however, there are different degrees of correlation between the various types of logs, but not all are linearly related. Therefore, the conventional method is limited by the dimension of the input parameter and the regression method, and a satisfactory result is difficult to obtain.
Based on the method, the logging curve reconstruction method based on the machine learning model is provided, the logging curve can be reconstructed, and the logging curve with higher accuracy is obtained.
Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.
Fig. 1 shows a flowchart of a method for reconstructing a well log according to an embodiment of the present application, and as shown in fig. 1, the method may specifically include the following steps:
s101, obtaining a preset reference curve of a logging curve to be reconstructed.
In particular, the well log to be reconstructed is a well log in which an anomaly exists, and the type of the well log can include various types. The reference curve is a curve representing the geological type of the log from which the log is obtained and having the same response as the log to be reconstructed. The response means that the curve reflects the geological characteristics, and the same response means that the geological characteristics (such as rock strata) reflected by the curve a are the same as the geological characteristics reflected by the curve B at any sampling point.
For example, if the type of log to be reconstructed is an acoustic log, then the reference curve is a log acquired at the log to be reconstructed or another log. Wherein the other logs are of the same geological type as the logs of the log to be reconstructed, and the type of the reference curve is correlated (i.e. the response is the same) with the type of the acoustic log. It should be noted that the other logs and the log of the log to be reconstructed may be different logs in the same area (ensuring that the geological features are the same) or in the same geological region. In order to ensure the accuracy of the reconstruction of the well log, the reference curve in this embodiment may be a well log formed by collecting data in a well log to which the well log to be reconstructed belongs.
S102, calibrating the depths of the reference curve and the logging curve to be reconstructed to obtain the corrected reference curve and the logging curve to be reconstructed.
In this embodiment, the number of the reference curves may include a plurality of reference curves, but due to the influence of the weight or tension change of the logging tool, the depth display of the same formation may be different when different logging tools collect data. Therefore, the present embodiment calibrates the depths of the log to be reconstructed and the reference curve thereof to ensure that each curve has a corresponding response at the same depth.
S103, operating the reference curve of the logging curve to be reconstructed after depth correction to obtain characteristics.
Wherein, the operation comprises at least one of derivation, product between curves and quotient between curves. The logging curves to be reconstructed after depth correction comprise a plurality of logging curves, the shape of each logging curve can reflect the geological characteristics of the logging curve, a plurality of characteristic curves can be obtained by operating the logging curves, and it can be understood that the characteristic curves can represent the curve characteristics of all the logging curves.
Optionally, operations may be performed between the results obtained through the operations, and the obtained results may also be used as features.
And S104, performing dimension reduction processing on the features to obtain model features.
Because the reference curve may have interference data or repeated data, the step performs dimensionality reduction on the obtained characteristic curve, removes the interference curve or repeated curve, and obtains the characteristic curve after dimensionality reduction as the model characteristic.
The specific implementation of dimension reduction can be seen in the prior art.
And S105, acquiring a reconstruction model corresponding to the geological type of the logging location according to the preset corresponding relation between the geological type and the reconstruction model.
Wherein the logging is a logging for acquiring a logging curve to be reconstructed. And the reconstruction model corresponding to each geological type in the corresponding relation is the model with the highest reconstruction accuracy of the logging curve of the logging of the geological type in the multiple machine learning models.
Optionally, the plurality of machine learning models may include a logistic regression model, a k-nearest neighbor model, a gradient-boosted tree model, a support vector machine model, a naive bayes model, a decision tree model, and a random forest model. Because the learning abilities of each type of machine learning model to different model characteristics are different, the corresponding relation between the geological type and the reconstruction model can be obtained through training and testing, and therefore the model with the highest reconstruction accuracy of the logging curve is selected according to the geological type of the region where the logging is located.
It is understood that any one set of corresponding relations includes a geological type and a model with the highest reconstruction accuracy of a log of a well log of the geological type, and this embodiment is described as an example of determining one set of corresponding relations, and specific embodiments may include:
and S1, obtaining a plurality of logs in the logging region of the logging curve to be reconstructed, and taking the logs as sample logs.
In this embodiment, a plurality of logs in the same area as the log of the log to be reconstructed may be obtained as sample logs. It will be appreciated that the geological conditions of these logs are the same as the geological conditions of the logs from which the logs to be reconstructed were acquired. Therefore, the geological features of the logs acquired from these logs and the reconstructed log to be reconstructed react at the same depth, i.e., the response, should be the same.
And S2, dividing the sample logging into a first logging set and a second logging set.
In this embodiment, if the number of sample logs is N, the number of the first log set is N1, the number of the second log set is N2, and N1+ N2, and the first log set and the second log set do not intersect with each other, the number of sample logs in the first log set is greater than the number of sample logs in the second log set, that is, N1> N2. Note that in this embodiment, N1 may be preset to 80% N.
And S3, training multiple types of models by using the logging curves of the sample logging in the first logging set to obtain multiple models to be selected.
During the training, the well log of any sample well in the first well log set may include a training target curve and a training reference curve. Logging curves which are collected and recorded in the sample logging and have the same type as the logging curves to be reconstructed are recorded as training target curves; and other logging curves which are collected and recorded in the sample logging and have the same type as the reference curve of the logging curve to be reconstructed are recorded as training reference curves.
For example, if the type of the log to be reconstructed is an acoustic moveout curve, the training target curve in the log of any sample log in the first log set is: acoustic moveout curves are acquired and recorded in the sample log. The test reference curve in the well logs of any sample well log in the first well log set is: a log of the same type as the reference curve of the sonic moveout curve is acquired and recorded in the sample log.
According to the feature extraction method of S102 to S103, the model features of the training reference curve of any sample well logging in the first well logging set are extracted, the model features are respectively input to the multiple machine learning models, the training target curve of the sample well logging is output as the target of each model, the multiple machine learning models are respectively trained, and the trained multiple machine learning models are obtained. And carrying out the training process on each sample well log in the first well logging set to finally obtain a plurality of trained models to be selected.
And S4, testing the multiple models to be selected by using the logging curves of the sample logging in the second logging set, and selecting a reconstruction model corresponding to the geological type from the multiple models to be selected.
Specifically, during the testing, the well log of any sample well in the second well log set may include a test target curve and a test reference curve. Logging curves which are collected and recorded in the sample logging and have the same type as the logging curves to be reconstructed are recorded as testing target curves; and collecting other logging curves of the same type as the reference curve of the logging curve to be reconstructed in the sample logging, and recording the other logging curves as test reference curves.
According to the feature extraction method of S102-S103, the model features of the test target curve of any sample well logging in the second well logging set are extracted, and the model features are respectively input into each candidate model obtained in S3, so that a reconstructed result well logging curve output by each candidate model is obtained. And further, taking the selected model with the output result logging curve most similar to the preset sample logging curve as a reconstruction model. The preset sample logging curve is a logging curve which belongs to the same well and is of the same type as the logging curve of the sample logging in the second logging set. Alternatively, the sample log may be a test target curve of the sample log, or the sample log may be another curve set in advance.
It should be noted that, in this step, the output result log and the preset sample log may be subjected to similarity analysis. Optionally, a candidate model which can output a result well log curve with the highest similarity to the test target curve can be selected as a final reconstruction model. It should be noted that the similarity analysis of the well logs can be implemented in a manner similar to that of the prior art.
It will be appreciated that the sample logs in the second set of logs may comprise a plurality of sample logs, and that testing with each sample log may result in one reconstructed model, so that this step may result in a plurality of reconstructed models. Further, in a case where a plurality of models can be selected according to the selection principle, a reconstruction model with the highest reconstruction speed may be selected from the plurality of reconstruction models as a final reconstruction model.
The second set of logs including the first and second sample logs are used as an example for illustration.
And selecting a model to be selected which can output a reconstructed logging curve with the highest similarity with the test target curve as a final reconstructed model and recording the model as the first reconstructed model. And selecting a model to be selected which can output a reconstructed well logging curve with the highest similarity with the test target curve as a final reconstructed model and recording the model as a second reconstructed model.
And when the first reconstruction model and the second reconstruction model are the same reconstruction model, selecting the reconstruction model as a final reconstruction model.
When the first reconstruction model and the second reconstruction model are different reconstruction models, the speed of the reconstructed logging curve output by the first reconstruction model and the speed of the reconstructed logging curve output by the second reconstruction model are compared, and the reconstruction model with the higher speed is selected as the final reconstruction model.
For example, the well logging curves of the first sample well logging include a well logging curve a, a well logging curve b, a well logging curve c and a well logging curve d, wherein the well logging curve a is a testing target curve, and the well logging curve b, the well logging curve c and the well logging curve d are testing reference curves. The method comprises the following steps of obtaining model characteristics of a logging curve a according to the logging curve b, the logging curve c and the logging curve d, inputting the model characteristics of the logging curve a into each model to be selected, and enabling each model to be selected to output a result logging curve. Further, similarity analysis is performed on each result logging curve and the logging curve a to obtain a result logging curve (marked as a1) most similar to the logging curve a, and the candidate model of the output result logging curve a1 is used as the first reconstruction model in the embodiment.
Referring to the method above, a second reconstructed model is obtained by performing a test using a log of the second sample log. Since the second reconstruction model and the first reconstruction model are different machine models, and the speed of the second reconstruction model output result log a2 is faster than the speed of the first reconstruction model output result log a1, the second reconstruction model is used as the final reconstruction model in the present embodiment.
It can be understood that, according to the above method, a plurality of sets of corresponding relationships of a plurality of geological types may be included in advance, and when the logging curve needs to be reconstructed, the reconstruction model corresponding to the geological type of the region where the logging curve is located may be directly selected.
And S106, inputting the model characteristics into the reconstruction model to obtain a reconstructed logging curve output by the reconstruction model.
In this embodiment, the model features obtained in S104 are input to the final reconstruction model to obtain a reconstructed logging curve. Because the training data and the test data of the reconstructed model are acquired from the logging in the same region as the logging curve to be reconstructed, the reconstructed logging curve can well reflect the stratum characteristics of the region, and the method has application value.
According to the technical scheme, the method for reconstructing the logging curve extracts the characteristics of the logging curve to be reconstructed, generates the model characteristics according to the characteristics, inputs the model characteristics into the reconstruction model, and outputs the reconstruction model as the reconstructed logging curve. The reconstructed model is a machine model corresponding to the geological type obtained according to the geological type of the logging location, and obviously, the reconstructed model can be used for multidimensional nonlinear regression based on machine learning, so that the problem that the traditional method is limited by multidimensional and nonlinear factors can be effectively solved, and a reconstructed logging curve with high accuracy can be obtained. And in addition, a model with the highest accuracy is selected from the multiple machine learning models by combining the strong correlation between the logging curve and the geological type so as to determine a reconstruction model corresponding to the logging location type, and the accuracy of the reconstruction curve can be further improved.
Further, the selection process of the reconstructed model comprises a training process and a testing process, wherein the training process is carried out according to the logging curves of the multiple sample logging at the logging location, and the testing process is carried out according to the logging curves of the multiple sample logging at the logging location, so that the reconstructed model in the embodiment is a model with the highest reconstruction accuracy of the logging curves of the logging of the types in the multiple machine learning models, the quality of the logging curves after reconstruction is further improved, and the output logging curves can better reflect the geological characteristics of the logging area.
It should be further noted that, in this embodiment, S102 and S104 are optional steps, and through depth correction or dimension reduction, the data quality of the reference curve is ensured, so that the characteristics of the characteristic curve to be reconstructed with high quality can be obtained, and the quality of the training data and the test data when the reconstructed model is obtained can be ensured, so that the accuracy of the reconstructed model is ensured.
The embodiment of the application also provides a reconstruction device of the logging curve, which is described below, and the reconstruction device of the logging curve described below and the reconstruction method of the logging curve described above can be referred to correspondingly.
Referring to fig. 2, a schematic structural diagram of a device for reconstructing a log according to an embodiment of the present application is shown, and as shown in fig. 2, the device may include:
a feature extraction unit 201, configured to extract features of a well log to be reconstructed;
the model obtaining unit 202 is configured to obtain a reconstruction model corresponding to the geological type of a logging location according to a preset correspondence between the geological type and the reconstruction model, where the logging is a logging for acquiring the logging curve to be reconstructed; the reconstruction model corresponding to each geological type in the corresponding relation is a model with the highest reconstruction accuracy of the logging curve of the logging of the geological type in various machine learning models;
and the curve reconstruction unit 203 is configured to input a model characteristic into the reconstruction model to obtain a reconstructed logging curve output by the reconstruction model, where the model characteristic is determined according to the characteristic.
Optionally, the feature extraction unit is configured to extract features of the log to be reconstructed, and includes: the feature extraction unit is specifically the same as:
acquiring a preset reference curve of the logging curve to be reconstructed, wherein the reference curve is a curve which represents the geological type of the logging curve to be reconstructed and has the same response as that of the logging curve to be reconstructed;
and calculating the reference curve of the well logging curve to be reconstructed to obtain the characteristics, wherein the calculation comprises at least one of derivation, product between curves and quotient between curves.
Optionally, the apparatus further comprises a calibration unit;
the calibration unit is used for calibrating the depths of the reference curve and the logging curve to be reconstructed before the characteristics of the logging curve to be reconstructed are extracted, so that the reference curve after depth correction and the logging curve to be reconstructed are obtained;
optionally, the feature extracting unit is configured to obtain the feature by performing an operation on a reference curve of the well log to be reconstructed, and includes:
the feature extraction unit is specifically configured to obtain the feature by performing operation on the calibrated reference curve.
Optionally, the model obtaining unit is configured to obtain the reconstructed model corresponding to the geological type of the logging location according to a preset correspondence between the geological type and the reconstructed model, and includes: the model obtaining unit is specifically used for obtaining the corresponding relation between the geological type and the reconstruction model;
the model obtaining unit is specifically configured to, when obtaining a corresponding relationship between the geological type and the reconstructed model:
obtaining a plurality of logs of the geological region as sample logs;
dividing the sample logs into a first log set and a second log set, wherein the first log set and the second log set do not have an intersection, and the number of sample logs in the first log set is greater than that of the sample logs in the second log set;
training various types of models by using the logging curves of the sample logging in the first logging set to obtain a plurality of models to be selected;
inputting the characteristics of the logging curves of the sample logging in the second logging set into each model to be selected respectively to obtain a reconstructed result logging curve output by each model to be selected;
and taking the output candidate model with the reconstructed result logging curve most similar to the preset sample logging curve as the reconstructed model, wherein the preset sample logging curve is the logging curve which belongs to the same logging and is of the same type as the logging curve of the sample logging in the second logging set.
Optionally, the apparatus further comprises a dimension reduction unit;
and the dimension reduction unit is used for performing dimension reduction processing on the characteristics to obtain the model characteristics before inputting the model characteristics into the reconstruction model to obtain a reconstructed logging curve output by the reconstruction model.
Optionally, the plurality of machine learning models comprises:
a logistic regression model, a k-nearest neighbor model, a gradient lifting tree model, a support vector machine model, a naive Bayes model, a decision tree model and a random forest model.
The device for reconstructing the logging curve comprises a processor and a memory, wherein the characteristic extraction unit, the model acquisition unit, the curve reconstruction unit and the like are stored in the memory as program units, and the processor executes the program units stored in the memory to realize corresponding functions.
The processor comprises a kernel, and the kernel calls the corresponding program unit from the memory. One or more kernels can be set, and the accuracy of the reconstruction curve is improved by adjusting the kernel parameters.
An embodiment of the present invention provides a storage medium on which a program is stored, which, when executed by a processor, implements the method for reconstructing a well log.
The embodiment of the invention provides a processor, which is used for running a program, wherein the program executes the reconstruction method of the logging curve during running.
The embodiment of the invention provides a device, and fig. 3 shows a schematic structural diagram of a device for reconstructing a well log provided by the embodiment of the application, wherein the device (30) comprises at least one processor 301, at least one memory 302 connected with the processor, and a bus 303; the processor and the memory complete mutual communication through a bus; the processor is used for calling the program instructions in the memory so as to execute the reconstruction method of the logging curve. The device herein may be a server, a PC, a PAD, a mobile phone, etc.
The present application further provides a computer program product adapted to perform a program for initializing the following method steps when executed on a data processing device:
a method of reconstructing a well log, comprising:
extracting the characteristics of a logging curve to be reconstructed;
acquiring a reconstruction model corresponding to the geological type of a logging location according to the corresponding relation between a preset geological type and the reconstruction model, wherein the logging is logging for acquiring the logging curve to be reconstructed; the reconstruction model corresponding to each geological type in the corresponding relation is a model with the highest reconstruction accuracy of the logging curve of the logging of the geological type in various machine learning models;
inputting the model characteristics into the reconstruction model to obtain a reconstructed logging curve output by the reconstruction model, wherein the model characteristics are determined according to the characteristics.
Optionally, extracting features of the well log to be reconstructed comprises:
acquiring a preset reference curve of the logging curve to be reconstructed, wherein the reference curve is a curve which represents the geological type of the logging curve to be reconstructed and has the same response as that of the logging curve to be reconstructed;
and calculating the reference curve of the well logging curve to be reconstructed to obtain the characteristics, wherein the calculation comprises at least one of derivation, product between curves and quotient between curves.
Optionally, before the extracting the features of the log to be reconstructed, the method further includes:
calibrating the depths of the reference curve and the logging curve to be reconstructed to obtain the calibrated reference curve and the logging curve to be reconstructed;
the obtaining the characteristics by operating the reference curve of the logging curve to be reconstructed comprises:
and calculating the corrected reference curve to obtain the characteristics.
Optionally, the determining process of the correspondence between the geological type and the reconstructed model includes:
obtaining a plurality of logs of the geological region as sample logs;
dividing the sample logs into a first log set and a second log set, wherein the first log set and the second log set do not have an intersection, and the number of sample logs in the first log set is greater than that of the sample logs in the second log set;
training various types of models by using the logging curves of the sample logging in the first logging set to obtain a plurality of models to be selected;
inputting the characteristics of the logging curves of the sample logging in the second logging set into each model to be selected respectively to obtain a reconstructed result logging curve output by each model to be selected;
and taking the output candidate model with the reconstructed result logging curve most similar to the preset sample logging curve as the reconstructed model, wherein the preset sample logging curve is the logging curve which belongs to the same logging and is of the same type as the logging curve of the sample logging in the second logging set.
Optionally, before the inputting the model characteristic into the reconstructed model to obtain a reconstructed log output by the reconstructed model, the method further includes:
and carrying out dimension reduction processing on the features to obtain the model features.
Optionally, the plurality of machine learning models comprises:
a logistic regression model, a k-nearest neighbor model, a gradient lifting tree model, a support vector machine model, a naive Bayes model, a decision tree model and a random forest model.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
In a typical configuration, a device includes one or more processors (CPUs), memory, and a bus. The device may also include input/output interfaces, network interfaces, and the like.
The memory may include volatile memory in a computer readable medium, Random Access Memory (RAM) and/or nonvolatile memory such as Read Only Memory (ROM) or flash memory (flash RAM), and the memory includes at least one memory chip. The memory is an example of a computer-readable medium.
Computer-readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), Read Only Memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), Digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information that can be accessed by a computing device. As defined herein, a computer readable medium does not include a transitory computer readable medium such as a modulated data signal and a carrier wave.
It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in the process, method, article, or apparatus that comprises the element.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The above are merely examples of the present application and are not intended to limit the present application. Various modifications and changes may occur to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims (10)

1. A method of reconstructing a well log, comprising:
extracting the characteristics of a logging curve to be reconstructed;
acquiring a reconstruction model corresponding to the geological type of a logging location according to the corresponding relation between a preset geological type and the reconstruction model, wherein the logging is logging for acquiring the logging curve to be reconstructed; the reconstruction model corresponding to each geological type in the corresponding relation is a model with the highest reconstruction accuracy of the logging curve of the logging of the geological type in various machine learning models;
inputting the model characteristics into the reconstruction model to obtain a reconstructed logging curve output by the reconstruction model, wherein the model characteristics are determined according to the characteristics.
2. The method of claim 1, wherein extracting features of the log to be reconstructed comprises:
acquiring a preset reference curve of the logging curve to be reconstructed, wherein the reference curve is a curve which represents the geological type of the logging curve to be reconstructed and has the same response as that of the logging curve to be reconstructed;
and calculating the reference curve of the well logging curve to be reconstructed to obtain the characteristics, wherein the calculation comprises at least one of derivation, product between curves and quotient between curves.
3. The method of claim 2, further comprising, prior to said extracting features of the well log to be reconstructed:
calibrating the depths of the reference curve and the logging curve to be reconstructed to obtain the calibrated reference curve and the logging curve to be reconstructed;
the obtaining the characteristics by operating the reference curve of the logging curve to be reconstructed comprises:
and calculating the corrected reference curve to obtain the characteristics.
4. The method according to claim 1, wherein the determining of the correspondence between the geological type and the reconstructed model comprises:
obtaining a plurality of logs of the geological region as sample logs;
dividing the sample logs into a first log set and a second log set, wherein the first log set and the second log set do not have an intersection, and the number of sample logs in the first log set is greater than that of the sample logs in the second log set;
training various types of models by using the logging curves of the sample logging in the first logging set to obtain a plurality of models to be selected;
inputting the characteristics of the logging curves of the sample logging in the second logging set into each model to be selected respectively to obtain a reconstructed result logging curve output by each model to be selected;
and taking the output candidate model with the reconstructed result logging curve most similar to the preset sample logging curve as the reconstructed model, wherein the preset sample logging curve is the logging curve which belongs to the same logging and is of the same type as the logging curve of the sample logging in the second logging set.
5. The method of claim 1, further comprising, prior to said inputting said model features into said reconstructed model to obtain a reconstructed log output by said reconstructed model,:
and carrying out dimension reduction processing on the features to obtain the model features.
6. The method of any of claims 1-5, wherein the plurality of machine learning models comprises:
a logistic regression model, a k-nearest neighbor model, a gradient lifting tree model, a support vector machine model, a naive Bayes model, a decision tree model and a random forest model.
7. An apparatus for reconstructing a well log, comprising:
the characteristic extraction unit is used for extracting the characteristics of the logging curve to be reconstructed;
the model acquisition unit is used for acquiring a reconstruction model corresponding to the geological type of the logging location according to the corresponding relation between the preset geological type and the reconstruction model, and the logging is logging for acquiring the logging curve to be reconstructed; the reconstruction model corresponding to each geological type in the corresponding relation is a model with the highest reconstruction accuracy of the logging curve of the logging of the geological type in various machine learning models;
and the curve reconstruction unit is used for inputting the model characteristics into the reconstruction model to obtain a reconstructed logging curve output by the reconstruction model, and the model characteristics are determined according to the characteristics.
8. The apparatus of claim 7, wherein the model obtaining unit is further configured to obtain the correspondence between the geological type and the reconstructed model in the following manner:
obtaining a plurality of logs of the geological region as sample logs;
dividing the sample logs into a first log set and a second log set, wherein the first log set and the second log set do not have an intersection, and the number of sample logs in the first log set is greater than that of the sample logs in the second log set;
training various types of models by using the logging curves of the sample logging in the first logging set to obtain a plurality of models to be selected;
inputting the characteristics of the logging curves of the sample logging in the second logging set into each model to be selected respectively to obtain a reconstructed logging curve output by each model to be selected;
and taking the selected model which is most similar to the output reconstructed logging curve and the preset sample logging curve as the reconstructed model.
9. A storage medium having a program stored thereon, wherein the program, when executed by a processor, performs the steps of the method of reconstructing a well log according to any of claims 1 to 6.
10. An apparatus for reconstructing a well log, the apparatus comprising at least one processor, and at least one memory, bus connected to the processor; the processor and the memory complete mutual communication through the bus; the processor is used for calling the program instructions in the memory to execute the reconstruction method of the logging curve of any one of claims 1-6.
CN201911381262.6A 2019-12-27 2019-12-27 Reconstruction method, device and equipment of logging curve and storage medium Active CN113050189B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911381262.6A CN113050189B (en) 2019-12-27 2019-12-27 Reconstruction method, device and equipment of logging curve and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911381262.6A CN113050189B (en) 2019-12-27 2019-12-27 Reconstruction method, device and equipment of logging curve and storage medium

Publications (2)

Publication Number Publication Date
CN113050189A true CN113050189A (en) 2021-06-29
CN113050189B CN113050189B (en) 2024-06-21

Family

ID=76506968

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911381262.6A Active CN113050189B (en) 2019-12-27 2019-12-27 Reconstruction method, device and equipment of logging curve and storage medium

Country Status (1)

Country Link
CN (1) CN113050189B (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113790759A (en) * 2021-09-16 2021-12-14 滁州职业技术学院 Multi-sensor fusion type building measurement robot and control method thereof
CN114200524A (en) * 2021-10-29 2022-03-18 五季数据科技(北京)有限公司 Logging density curve correction method based on artificial intelligence deep learning
CN116070390A (en) * 2021-10-29 2023-05-05 大庆油田有限责任公司 Oilfield geological data processing method and processing device

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5828981A (en) * 1995-05-11 1998-10-27 Texaco Inc. Generating pore types and synthetic capillary pressure curves from wireline logs using neural networks
CN201747364U (en) * 2010-06-01 2011-02-16 中国石油天然气集团公司 Interval transit time curve reconstruction equipment
CN108138555A (en) * 2015-02-23 2018-06-08 奈克森能量无限责任公司 Method, system and the equipment of predicting reservoir property
CN108875122A (en) * 2018-04-25 2018-11-23 杭州迅美科技有限公司 The artificial intelligence approach and system of geologic parameter are calculated using well logging data
US20190080122A1 (en) * 2017-09-14 2019-03-14 Saudi Arabian Oil Company Subsurface reservoir model with 3d natural fractures prediction
WO2019075522A1 (en) * 2017-10-19 2019-04-25 Medibio Limited Risk indicator
CN109736790A (en) * 2018-12-28 2019-05-10 新疆工程学院 Recognition of Weil Logging Lithology method based on deepness belief network
CN109919184A (en) * 2019-01-28 2019-06-21 中国石油大学(北京) A kind of more well complex lithology intelligent identification Methods and system based on log data

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5828981A (en) * 1995-05-11 1998-10-27 Texaco Inc. Generating pore types and synthetic capillary pressure curves from wireline logs using neural networks
CN201747364U (en) * 2010-06-01 2011-02-16 中国石油天然气集团公司 Interval transit time curve reconstruction equipment
CN108138555A (en) * 2015-02-23 2018-06-08 奈克森能量无限责任公司 Method, system and the equipment of predicting reservoir property
US20190080122A1 (en) * 2017-09-14 2019-03-14 Saudi Arabian Oil Company Subsurface reservoir model with 3d natural fractures prediction
WO2019075522A1 (en) * 2017-10-19 2019-04-25 Medibio Limited Risk indicator
CN108875122A (en) * 2018-04-25 2018-11-23 杭州迅美科技有限公司 The artificial intelligence approach and system of geologic parameter are calculated using well logging data
CN109736790A (en) * 2018-12-28 2019-05-10 新疆工程学院 Recognition of Weil Logging Lithology method based on deepness belief network
CN109919184A (en) * 2019-01-28 2019-06-21 中国石油大学(北京) A kind of more well complex lithology intelligent identification Methods and system based on log data

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113790759A (en) * 2021-09-16 2021-12-14 滁州职业技术学院 Multi-sensor fusion type building measurement robot and control method thereof
CN113790759B (en) * 2021-09-16 2023-09-29 滁州职业技术学院 Multi-sensor fusion type building measurement robot and control method thereof
CN114200524A (en) * 2021-10-29 2022-03-18 五季数据科技(北京)有限公司 Logging density curve correction method based on artificial intelligence deep learning
CN116070390A (en) * 2021-10-29 2023-05-05 大庆油田有限责任公司 Oilfield geological data processing method and processing device
CN116070390B (en) * 2021-10-29 2023-11-21 大庆油田有限责任公司 Oilfield geological data processing method and processing device

Also Published As

Publication number Publication date
CN113050189B (en) 2024-06-21

Similar Documents

Publication Publication Date Title
CN113050189B (en) Reconstruction method, device and equipment of logging curve and storage medium
CN113919219A (en) Stratum evaluation method and system based on logging big data
JP7259969B2 (en) Seismic observation device, seismic observation method, program, and configuration program
NO342669B1 (en) Speed model for well time-depth conversion in seismic
US20110166842A1 (en) Layer stripping method
CN112487604A (en) Long-time nonlinear drift compensation method for output data of marine gravimeter
Maranò et al. Fitting earthquake spectra: Colored noise and incomplete data
CN112990320A (en) Lithology classification method and device, electronic equipment and storage medium
CN110927789B (en) Method and device for predicting shale plane distribution based on loss data
CN114444393A (en) Logging curve construction method and device based on time convolution neural network
Alkhalaf et al. Machine learning approach to classify water cut measurements using DAS fiber optic data
Andersen et al. Machine learning based prediction of porosity and water saturation from Varg field reservoir well logs
CN111580179A (en) Method, device and system for determining organic carbon content
Zhang et al. First break of the seismic signals in oil exploration based on information theory
Asedegbega et al. Application of machine learniing for reservoir facies classification in port field, Offshore Niger Delta
CN115577881A (en) Prediction method for deep coal gas content based on support vector machine regression model
CN112861061A (en) Geological stratification method and system and processing equipment
CN112363243A (en) Method, device, computer equipment and medium for predicting high-quality reservoir of shale oil
CN113267809A (en) Method and device for predicting I-type shale reservoir
US20230367031A1 (en) Building scalable geological property models using machine learning algorithms
CN111856571B (en) Method, device and equipment for converting logging frequency band speed into seismic frequency band speed
CN112987122A (en) Method and device for calculating mud content, electronic equipment and storage medium
Ghamdi et al. Emerging Techniques in Measuring Capillary Pressure and Permeability Using NMR and AI
Ahmadov Utilizing data-driven models to predict brittleness in Tuscaloosa marine shale: a machine learning approach
Dastan et al. A New Look at Nonlinear Regression in Well Test Interpretation

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant