CN115796350A - Method and system for predicting total organic carbon content of hydrocarbon source rock in few well regions in sea area - Google Patents
Method and system for predicting total organic carbon content of hydrocarbon source rock in few well regions in sea area Download PDFInfo
- Publication number
- CN115796350A CN115796350A CN202211476535.7A CN202211476535A CN115796350A CN 115796350 A CN115796350 A CN 115796350A CN 202211476535 A CN202211476535 A CN 202211476535A CN 115796350 A CN115796350 A CN 115796350A
- Authority
- CN
- China
- Prior art keywords
- carbon content
- organic carbon
- total organic
- prediction model
- source rock
- 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.)
- Pending
Links
Images
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to a method and a system for predicting the total organic carbon content of hydrocarbon source rocks in a few well zones in a sea area. The method comprises the following steps: acquiring the total range of hydrocarbon source rock prediction in the sea of a research area, the source position of the research area and geological data; determining a total organic carbon content prediction model data table according to the total range, the position of the object source and the geological data; processing the total organic carbon content prediction model data table to obtain processed total organic carbon content prediction model data; training the processed total organic carbon content prediction model data to obtain a training optimized total organic carbon content prediction model; and predicting the total organic carbon content of the source rock at any position in the research area according to the trained and optimized total organic carbon content prediction model, and determining a high-quality source rock distribution area. The method can fully consider the influence of multiple factors and improve the prediction precision of the total organic carbon content.
Description
Technical Field
The invention relates to the field of total organic carbon content prediction, in particular to a method and a system for predicting the total organic carbon content of hydrocarbon source rock of a few well regions in a sea area.
Background
Compared with the world oil country, the onshore oil and gas resources of China are relatively deficient, and the crude oil yield can not meet the increasing demand for the long time. In 1993, china has become a crude oil net import country, in 2009, the external dependence of crude oil in China exceeds a warning line of 50%, and the dependence is increased year by year, in 2021, the external dependence of crude oil in China is up to 72%, the total sum of natural gas imports is increased by 19.9%, and the external dependence is up to 46%. These striking data indicate that the contradiction of insufficient energy supply in China is increasingly prominent, and increasing the crude oil yield per se in China is a main measure for building the resource foundation of energy safety in China. Although onshore oil and gas resources are deficient in China, ocean oil and gas resources are abundant in China, and how to efficiently exploit the oil and gas resources in the ocean is a main problem facing people. The exploration and development of oil and gas resources in the sea area are much more complex than those on land, and the oil and gas resource exploration and development system has the characteristics of high cost of a drilling platform, limited drilling range, rare drilling quantity, few drilling cores and the like. Therefore, the exploration and development of oil and gas resources in the sea area need accurate fixed points and efficient exploitation to obtain a good effect. If the distribution area of high-quality hydrocarbon source rock can be found, even the development depth of the development horizon of the high-quality hydrocarbon source rock can be determined, the method has great significance for the exploration and development of oil and gas resources in sea areas.
The method aims at the prediction of high-quality hydrocarbon source rocks in the hydrocarbon-bearing basin, and most researches select the abundance of organic matters to represent the quality of the hydrocarbon source rocks. While organic matter abundance is often characterized by total organic carbon content (TOC). Existing studies have all performed TOC predictions by building linear models based on well logs, seismic data and limited measured data. It has the following several drawbacks and disadvantages:
1. the error of the logging curve data and the seismic data is large, and the prediction precision is low.
2. The abundance of organic matters in the hydrocarbon source rock is controlled by a plurality of factors such as a source condition, a carrying condition, a hydrodynamic condition and the like, the influence of the factors is not fully considered in the existing research, a single-factor linear model is mostly established, and actually, the TOC and the influence of the factors are in a nonlinear relation.
3. The accuracy of the existing model depends on a large amount of measured data, the existing model usually belongs to a few wells and a well-free area in a sea area, the measured data are few, and a common linear model is difficult to apply.
Disclosure of Invention
The invention aims to provide a method and a system for predicting the total organic carbon content of hydrocarbon source rocks in a few well zones in a sea area, which can fully consider the influence of multiple factors and improve the prediction precision of the total organic carbon content.
In order to achieve the purpose, the invention provides the following scheme:
a method for predicting the total organic carbon content of hydrocarbon source rocks in a few well zones in a sea area comprises the following steps:
acquiring a total range of hydrocarbon source rock prediction in a sea area of a research area, wherein the total range comprises the number of well positions, well position coordinates and corresponding scales of the research area;
determining the position of an object source of the research area according to the well position coordinates and the corresponding scale;
acquiring geological data of a well position in a research area;
determining a total organic carbon content prediction model data table according to the total range, the position of the object source and geological data;
processing the total organic carbon content prediction model data table to obtain processed total organic carbon content prediction model data;
training the processed total organic carbon content prediction model data to obtain a training optimized total organic carbon content prediction model;
and predicting the total organic carbon content of the source rock at any position in the research area according to the trained and optimized total organic carbon content prediction model, and determining a high-quality source rock distribution area.
Optionally, the geological data comprises horizons, depths, lithology, rock colour, dephasing and paleotopographic grade of the source rock.
Optionally, the processing the total organic carbon content prediction model data table to obtain processed total organic carbon content prediction model data specifically includes:
and carrying out discrete feature coding, dimensionless processing and abnormal point processing on the data in the total organic carbon content prediction model data table to obtain processed total organic carbon content prediction model data.
Optionally, the training the processed data of the total organic carbon content prediction model to obtain a trained and optimized total organic carbon content prediction model specifically includes:
training by adopting three methods, namely a support vector machine, an artificial neural network and a random forest according to the processed total organic carbon content prediction model data to obtain a plurality of trained total organic carbon content prediction models;
and selecting a model with the highest precision from the trained total organic carbon content prediction models as the trained and optimized total organic carbon content prediction model.
Optionally, the predicting the total organic carbon content of the source rock at any position in the research area according to the trained total organic carbon content prediction model to determine the high-quality source rock distribution area specifically includes:
predicting the total organic carbon content of the hydrocarbon source rock at any position in the research area according to the trained total organic carbon content prediction model to obtain a plurality of prediction results;
and determining a high-quality hydrocarbon source rock distribution area according to each prediction result.
A system for predicting the total organic carbon content of hydrocarbon source rocks in a few well zones in a sea area comprises:
the overall range obtaining module is used for obtaining an overall range predicted by the hydrocarbon source rocks in the sea of the research area, and the overall range comprises the number of well positions, well position coordinates and corresponding scales of the research area;
the object source position determining module is used for determining the object source position of the research area according to the well position coordinates and the corresponding scale;
the geological data acquisition module is used for acquiring geological data of a well position in a research area;
the prediction model data table determining module is used for determining a total organic carbon content prediction model data table according to the total range, the position of the object source and geological data;
the prediction model data table processing module is used for processing the total organic carbon content prediction model data table to obtain processed total organic carbon content prediction model data;
the prediction model training module is used for training the processed total organic carbon content prediction model data to obtain a training optimized total organic carbon content prediction model;
and the high-quality source rock distribution area determination module is used for predicting the total organic carbon content of the source rock at any position in the research area according to the trained and optimized total organic carbon content prediction model and determining the high-quality source rock distribution area.
Optionally, the geological data comprises horizons, depths, lithology, rock colour, dephasing and paleotopographic grade of the source rock.
Optionally, the prediction model data table processing module specifically includes:
and the prediction model data table processing unit is used for carrying out discrete feature coding, dimensionless processing and abnormal point processing on the data in the total organic carbon content prediction model data table to obtain the processed total organic carbon content prediction model data.
Optionally, the prediction model training module specifically includes:
the prediction model training unit is used for training by adopting three methods, namely a support vector machine, an artificial neural network and a random forest according to the processed total organic carbon content prediction model data to obtain a plurality of trained total organic carbon content prediction models;
and the training optimized total organic carbon content prediction model determining unit is used for selecting a model with the highest precision from the plurality of trained total organic carbon content prediction models as the training optimized total organic carbon content prediction model.
Optionally, the module for determining the high-quality source rock distribution area specifically includes:
the prediction unit is used for predicting the total organic carbon content of the hydrocarbon source rock at any position in the research area according to the trained total organic carbon content prediction model to obtain a plurality of prediction results;
and the high-quality source rock distribution area determining unit is used for determining the high-quality source rock distribution area according to each prediction result.
According to the specific embodiment provided by the invention, the invention discloses the following technical effects:
the invention provides a method for predicting the total organic carbon content of hydrocarbon source rocks in a few well zones in a sea area, which takes the actual geological conditions of a hydrocarbon-bearing basin, such as the predicted total range, the position of a matter source, geological data and other factors as parameters, takes the TOC of the hydrocarbon source rocks as a label variable, trains by using a machine learning algorithm to establish a TOC prediction model of the hydrocarbon source rocks, comprehensively considers various geological factors influencing the TOC distribution of the hydrocarbon source rocks, and remarkably improves the accuracy of a prediction result. The method has good application effect on the TOC prediction of the hydrocarbon source rock of the few well zones in the sea area, can realize the TOC prediction of different depths and different lithologies of different well positions, and can complete the TOC plane distribution prediction of the same position in the whole research area.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required in the embodiments will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a flow chart of the method for predicting the total organic carbon content of hydrocarbon source rocks in a few well zones in a sea area according to the present invention;
FIG. 2 is a structural diagram of a system for predicting the total organic carbon content of hydrocarbon source rocks in a few well areas in a sea area according to the invention;
FIG. 3 is a schematic diagram of determining the location of the source of the hydrocarbon source rock in the sea area within the overall range and the location of the source of the study area;
FIG. 4 is a schematic view of an anomaly;
FIG. 5 is a schematic diagram of the accuracy of the optimal model;
FIG. 6 is a graph of support vector machine model results data;
FIG. 7 is a data diagram of the result of the artificial neural network model;
FIG. 8 is a graph of random forest model result data;
FIG. 9 is a schematic view of a prediction system software interface;
fig. 10 is a diagram showing the prediction result.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The invention aims to provide a method and a system for predicting the total organic carbon content of hydrocarbon source rocks in a few well regions in a sea area, which can fully consider multi-factor influence and improve the prediction accuracy of the total organic carbon content.
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in further detail below.
Fig. 1 is a flowchart of a method for predicting the total organic carbon content of hydrocarbon source rocks in a few wells in a sea area, as shown in fig. 1, the method for predicting the total organic carbon content of hydrocarbon source rocks in a few wells in a sea area includes:
step 101: and acquiring a predicted overall range of the hydrocarbon source rock in the sea of the research area, wherein the overall range comprises the number of well positions, well position coordinates and corresponding scales of the research area.
Step 102: and determining the position of the object source of the research area according to the well position coordinates and the corresponding scale. Specifically, the position is taken as the origin of coordinates, the entire study area is drawn into a rectangular coordinate system, and the position of each well is expressed by the abscissa (X) and the ordinate (Y) with the origin as the starting point based on the scale and the well coordinates. Where X is the lateral distance from the origin and Y is the longitudinal distance from the origin.
Step 103: and acquiring geological data of the well position in the research area. The geological data comprises the horizon, depth, lithology, rock color, sedimentary facies and paleotopographic grade of the source rock.
Step 104: and determining a total organic carbon content prediction model data table according to the overall range, the position of the object source and the geological data. Data such as abscissa (X), ordinate (Y), horizon, depth, lithology, rock color, sedimentary facies, paleotopographic grade, TOC and the like of the source rock sample are integrated into a spreadsheet.
Step 105: processing the total organic carbon content prediction model data table to obtain processed total organic carbon content prediction model data, which specifically comprises the following steps:
and carrying out discrete feature coding, dimensionless processing and abnormal point processing on the data in the total organic carbon content prediction model data table to obtain processed total organic carbon content prediction model data.
Because the parameters selected by the prediction model are in the form of characters, the parameters need to be converted into data for establishing the prediction model. The invention selects One-Hot codes to preprocess data, thereby digitalizing the characteristics of the discrete text. Meanwhile, due to different magnitudes of different parameters, dimensionless processing needs to be performed on variable data, so that data standardization is completed, and the contrast is improved. The invention adopts Z-score standardization to carry out dimensionless processing on the data. After dimensionless processing is completed, abnormal values are extracted from each parameter data through quantiles based on a statistical method, and the influence of the abnormal values on the accuracy of the prediction model is reduced. Meanwhile, the sedimentary facies corresponding to the discrete feature codes are empirically assigned as one of the input features. Sedimentary facies belt empirical assignments are weighted assignments of sedimentary facies belts based on geologist experience. The higher the phase band TOC, the greater its value.
Step 106: training the processed total organic carbon content prediction model data to obtain a training optimized total organic carbon content prediction model, which specifically comprises the following steps:
training by adopting three methods, namely a support vector machine, an artificial neural network and a random forest according to the processed total organic carbon content prediction model data to obtain a plurality of trained total organic carbon content prediction models;
and selecting a model with the highest precision from the trained total organic carbon content prediction models as the trained and optimized total organic carbon content prediction model.
75% of all data were selected for training and the remaining 25% were tested. And selecting the model with the highest precision in each algorithm as the optimal model of the algorithm by continuously adjusting parameters. The input features of each algorithm may be trained by selecting discrete feature codes and data based on empirical assignments, normalized and denormalized data, and denormalized values. Meanwhile, the parameter adjusting range of the support vector machine algorithm is set to be (1) C adjustment, and gamma is default; (2) gamma is adjusted, and C is defaulted; (3) c, three kinds of gamma adjustment are carried out; setting the adjustment range of the parameters of the artificial neural network as (1) alpha adjustment, and setting a hiden _ layer default; (2) adjusting hiden _ layer, and defaulting alpha; (3) adjusting three types of alpha and hiden _ layer; setting the adjustment range of the random forest algorithm parameters as (1) max _depthadjustment, and setting minsamplesleef as default; (2) minsamplesleef adjustment, max _ depth default; (3) max _ depth, minsamplesleef. Thus, there are 18 ways for the input features of each algorithm, from which the highest accuracy model is preferred. The invention selects a 10 fold cross validation method (k fold cross validation). The training set is divided into 10 disjoint subsets, and 9 subsets are selected as the training set and the remaining 1 subset is selected as the verification set from the divided subsets. And putting the trained model on a verification set every time to obtain the accuracy. The final accuracy of the model was determined by calculating the average 10 times. The evaluation indexes of the prediction accuracy are RMS and R 2 . The calculation formula is as follows:
step 107: predicting the total organic carbon content of the source rock at any position in the research area according to the trained and optimized total organic carbon content prediction model, and determining a high-quality source rock distribution area, wherein the method specifically comprises the following steps:
predicting the total organic carbon content of the hydrocarbon source rock at any position in the research area according to the trained total organic carbon content prediction model to obtain a plurality of prediction results;
and determining a high-quality hydrocarbon source rock distribution area according to each prediction result.
And 107, selecting the optimal model with the highest precision as a TOC prediction model of the research area, packaging the model into corresponding software by using a Python platform, realizing the TOC prediction of the hydrocarbon source rock at the same layer position at any position in the research area, finally defining a high-quality hydrocarbon source rock distribution area, and guiding the exploration and development of oil and gas in the sea area.
Aiming at the defects of the existing hydrocarbon source rock TOC prediction technology, the invention provides a method for predicting the total organic carbon content of hydrocarbon source rock in a few well zones in a sea area. The method takes actual geological conditions of the hydrocarbon-bearing basin, such as well position, depth, terrain slope, lithology, rock color and deposition, and the like as parameters, takes the TOC of the hydrocarbon source rock as a label variable, and establishes a TOC prediction model of the hydrocarbon source rock by applying 3 algorithms of a support vector machine, an artificial neural network and a random forest in a machine learning algorithm. According to the method, various geological factors influencing the TOC distribution of the hydrocarbon source rock are comprehensively considered, and the accuracy of the prediction result is obviously improved. The method has good application effect on the TOC prediction of the hydrocarbon source rock in the few well areas in the sea area, can realize the TOC prediction of different depths and different lithologies of different well positions, can complete the TOC plane distribution prediction of the same layer in the whole research area, and can provide effective theoretical and technical support for selecting a high-quality hydrocarbon source rock development area and a well drilling well position in the sea area.
FIG. 2 is a structural diagram of a total organic carbon content prediction system for hydrocarbon source rocks in a few well zones in a sea area. As shown in fig. 2, a system for predicting the total organic carbon content of a hydrocarbon source rock in a region with few well zones in a sea area comprises:
the overall range obtaining module 201 is used for obtaining an overall range of hydrocarbon source rock prediction in the sea of the research area, wherein the overall range comprises the number of well positions, well position coordinates and corresponding scales of the research area.
And an object source position determining module 202, configured to determine an object source position of the research region according to the well position coordinates and the corresponding scale.
And the geological data acquisition module 203 is used for acquiring geological data of the well position in the research area. The geological data comprises the horizon, depth, lithology, rock color, sedimentary facies and paleotopographic grade of the source rock.
A prediction model data table determining module 204, configured to determine a prediction model data table of total organic carbon content according to the total range, the position of the object, and the geological data; and determining a total organic carbon content prediction model data table.
And the prediction model data table processing module 205 is configured to process the total organic carbon content prediction model data table to obtain processed total organic carbon content prediction model data.
And the prediction model training module 206 is configured to train the processed total organic carbon content prediction model data to obtain a total organic carbon content prediction model after training optimization.
And the high-quality hydrocarbon source rock distribution area determining module 207 is used for predicting the total organic carbon content of the hydrocarbon source rock at any position in the research area according to the trained and optimized total organic carbon content prediction model and determining the high-quality hydrocarbon source rock distribution area.
The prediction model data table processing module 205 specifically includes:
and the prediction model data table processing unit is used for carrying out discrete feature coding, dimensionless processing and abnormal point processing on the data in the total organic carbon content prediction model data table to obtain the processed total organic carbon content prediction model data.
The prediction model training module 206 specifically includes:
the prediction model training unit is used for training by adopting three methods, namely a support vector machine, an artificial neural network and a random forest according to the processed total organic carbon content prediction model data to obtain a plurality of trained total organic carbon content prediction models;
and the training optimized total organic carbon content prediction model determining unit is used for selecting a model with the highest precision from the plurality of trained total organic carbon content prediction models as the training optimized total organic carbon content prediction model.
The determination module 207 for the high-quality source rock distribution area specifically includes:
the prediction unit is used for predicting the total organic carbon content of the hydrocarbon source rock at any position in the research area according to the trained total organic carbon content prediction model to obtain a plurality of prediction results;
and the high-quality source rock distribution area determining unit is used for determining the high-quality source rock distribution area according to each prediction result.
The invention applies the artificial intelligence machine learning algorithm to the geological practice of oil and gas resource exploration and development, effectively improves the defects of the prior art, and specifically comprises the following steps:
1. the method comprehensively considers a plurality of geological factors influencing the TOC distribution of the hydrocarbon source rock, and corrects the conventional prediction method only considering a single factor, so that the accuracy of the prediction result is obviously improved.
2. The TOC prediction model of the hydrocarbon source rock is established by using a support vector machine, an artificial neural network and 3 random forest algorithms in a machine learning algorithm, so that the defects of the conventional linear model are overcome, and the establishment of the nonlinear prediction model is more practical.
3. The method has a good application effect on the TOC prediction of the source rock of the few well regions in the sea area, can realize the TOC prediction of different depths and different lithologies at different well positions, and can complete the TOC plane distribution prediction of the same layer in the whole research area.
4. And the software matched with the system ensures that the TOC prediction becomes simple and easy to operate, and the efficiency and the precision are obviously improved.
Example 1:
taking a certain depression in a certain sea area of south China sea as an example, the method for predicting the total organic carbon content of the hydrocarbon source rock of the less well area in the sea area comprises the following steps:
step one, determining the total prediction range of the hydrocarbon source rock in the sea area according to the well position distribution and the well position coordinates.
Determining the position of an object source in a research area, taking the position as the origin of coordinates, dividing the whole research area into a rectangular coordinate system, and representing the position of each well by using the origin of coordinates of an abscissa (X) and an ordinate (Y) based on a scale and well coordinates, wherein the well coordinates are (X) n ,y n ). FIG. 3 is a schematic diagram of determining the location of the source of the hydrocarbon source rock in the sea area within the overall range and the location of the source of the research area.
And step three, collecting and calculating geological data of the well position of the research area, wherein the geological data specifically comprises the position, depth, lithology, rock color, sedimentary facies, paleotopographic gradient and the like of the source rock. Some of the data in this example are shown in Table 1.
Table 1 partial data
And step four, manufacturing a TOC prediction model data table. Integrating data such as X, Y, horizon, depth, lithology, rock color, sedimentary facies, paleotopographic grade, TOC and the like of the hydrocarbon source rock sample into a spreadsheet.
And fifthly, preprocessing the prediction model data. The method specifically comprises discrete feature coding, dimensionless processing and abnormal point processing. In the embodiment, one-Hot coding is selected to preprocess data, so that the discrete text is subjected to feature digitization, and part of data is shown in table 2. Meanwhile, because the magnitude of different parameters is different, the variable data needs to be subjected to dimensionless processing, in this embodiment, Z-score standardization is selected to perform dimensionless processing on the data, and the processing completion result is shown in table 3. After the dimensionless processing is completed, abnormal values are extracted from the parameter data through quantiles based on a statistical method, the result in this embodiment is shown in fig. 4 (the dots with darker colors in fig. 4 are abnormal points), and fig. 4 is a schematic diagram of the abnormal points. Meanwhile, the depositional facies corresponding to discrete feature codes are assigned empirically as an input feature. Sedimentary facies belt empirical assignments are weighted assignments of sedimentary facies belts based on geologist experience.
Table 2 partial preprocessing data
TABLE 3 results of processing completion
And step six, carrying out prediction model training. The method selects three algorithms of a support vector machine, an artificial neural network and a random forest to establish a prediction model. 75% of all data were selected for training and the remaining 25% were tested. And selecting the model with the highest precision in each algorithm as the optimal model of the algorithm through parameter adjustment. The input features of each algorithm may be trained by selecting discrete feature codes and data based on empirical assignments, normalized and denormalized data, and denormalized values. Meanwhile, setting the parameter adjusting range of the support vector machine algorithm as (1) adjusting, and setting gamma as default; (2) gamma is adjusted, and C is defaulted; (3) c, three kinds of gamma adjustment are carried out; setting the adjustment range of the parameters of the artificial neural network as (1) alpha adjustment, and setting a hiden _ layer default; (2) adjusting hiden _ layer, and defaulting alpha; (3) adjusting three types of alpha, hiden _ layer; setting the adjustment range of the random forest algorithm parameters as (1) max _depthadjustment, and setting minsamplesl eaf as default; (2) minsamplesleef adjustment, max _ depth default; (3) max _ depth, minsamplesl eaf. The model results are shown in fig. 6-8, wherein fig. 6 is a data diagram of the support vector machine model results, fig. 7 is a data diagram of the artificial neural network model results, and fig. 8 is a data diagram of the random forest model results. The invention selects a 10-fold cross validation method (k fold cross validation). The training set is divided into 10 disjoint subsets, and 9 subsets are selected as the training set and the remaining 1 subset is selected as the verification set from the divided subsets. And putting the trained model on a verification set every time to obtain the accuracy. The optimal model accuracy of the three algorithms is shown in fig. 5, and fig. 5 is a schematic diagram of the optimal model accuracy.
And step seven, after model training is finished, selecting the optimal model with the highest precision as the TO C prediction model of the research area, packaging the model into corresponding software by using a Python platform, realizing TOC prediction of the hydrocarbon source rock at the same level at any position in the research area, finally defining a high-quality hydrocarbon source rock distribution area, and guiding oil and gas exploration and development in the sea area. FIG. 9 is a schematic view of a prediction system software interface; fig. 10 is a diagram showing the prediction result.
In the present specification, the embodiments are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. For the system disclosed by the embodiment, the description is relatively simple because the system corresponds to the method disclosed by the embodiment, and the relevant points can be referred to the description of the method part.
The principles and embodiments of the present invention have been described herein using specific examples, which are provided only to help understand the method and the core concept of the present invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, the specific embodiments and the application range may be changed. In view of the above, the present disclosure should not be construed as limiting the invention.
Claims (10)
1. A method for predicting the total organic carbon content of hydrocarbon source rocks of a few well zones in a sea area is characterized by comprising the following steps:
acquiring a total range predicted by the hydrocarbon source rock in the sea of the research area, wherein the total range comprises the number of well positions, well position coordinates and corresponding scales of the research area;
determining the position of an object source in the research area according to the well position coordinates and the corresponding scale;
acquiring geological data of a well position in a research area;
determining a total organic carbon content prediction model data table according to the total range, the position of the object source and geological data;
processing the total organic carbon content prediction model data table to obtain processed total organic carbon content prediction model data;
training the processed total organic carbon content prediction model data to obtain a training optimized total organic carbon content prediction model;
and predicting the total organic carbon content of the source rock at any position in the research area according to the trained and optimized total organic carbon content prediction model, and determining a high-quality source rock distribution area.
2. The method for predicting the total organic carbon content of the hydrocarbon source rock of the offshore region with few wells according to claim 1, wherein the geological data comprise horizons, depths, lithologies, rock colors, sedimentary facies and paleotopographic gradients of the hydrocarbon source rock.
3. The method for predicting the total organic carbon content of the hydrocarbon source rock of the offshore region less well region according to claim 1, wherein the step of processing the total organic carbon content prediction model data table to obtain processed total organic carbon content prediction model data specifically comprises the steps of:
and carrying out discrete feature coding, dimensionless processing and abnormal point processing on the data in the total organic carbon content prediction model data table to obtain the processed total organic carbon content prediction model data.
4. The method for predicting the total organic carbon content of the hydrocarbon source rock of the offshore region with less wells according to claim 1, wherein the training of the processed total organic carbon content prediction model data to obtain the trained and optimized total organic carbon content prediction model specifically comprises:
training by adopting three methods, namely a support vector machine, an artificial neural network and a random forest according to the processed total organic carbon content prediction model data to obtain a plurality of trained total organic carbon content prediction models;
and selecting a model with the highest precision from the trained total organic carbon content prediction models as the trained and optimized total organic carbon content prediction model.
5. The method for predicting the total organic carbon content of the hydrocarbon source rock of the well-less zone in the sea area according to the claim 1, wherein the method for predicting the total organic carbon content of the hydrocarbon source rock at any position in the research area according to the trained total organic carbon content prediction model to determine the high-quality hydrocarbon source rock distribution area specifically comprises the following steps:
predicting the total organic carbon content of the hydrocarbon source rock at any position in the research area according to the trained total organic carbon content prediction model to obtain a plurality of prediction results;
and determining a high-quality hydrocarbon source rock distribution area according to each prediction result.
6. A system for predicting the total organic carbon content of hydrocarbon source rocks in a few well zones in a sea area is characterized by comprising the following steps:
the overall range obtaining module is used for obtaining an overall range predicted by the hydrocarbon source rocks in the sea of the research area, and the overall range comprises the number of well positions, well position coordinates and corresponding scales of the research area;
the object source position determining module is used for determining the object source position of the research area according to the well position coordinates and the corresponding scale;
the geological data acquisition module is used for acquiring geological data of a well position in a research area;
the prediction model data table determining module is used for determining a total organic carbon content prediction model data table according to the total range, the position of the object source and geological data;
the prediction model data table processing module is used for processing the total organic carbon content prediction model data table to obtain processed total organic carbon content prediction model data;
the prediction model training module is used for training the processed total organic carbon content prediction model data to obtain a total organic carbon content prediction model after training optimization;
and the high-quality source rock distribution area determination module is used for predicting the total organic carbon content of the source rock at any position in the research area according to the trained and optimized total organic carbon content prediction model and determining the high-quality source rock distribution area.
7. The system of claim 6, wherein the geological data comprises horizons, depths, lithology, rock color, dephasing, and paleotopographic gradients of the source rock.
8. The system for predicting the total organic carbon content of the hydrocarbon source rock of the hydrocarbon well region in the sea area according to claim 6, wherein the prediction model data table processing module specifically comprises:
and the prediction model data table processing unit is used for carrying out discrete feature coding, dimensionless processing and abnormal point processing on the data in the total organic carbon content prediction model data table to obtain the processed total organic carbon content prediction model data.
9. The system for predicting the total organic carbon content of the hydrocarbon source rock of the hydrocarbon well region in the sea area according to claim 6, wherein the prediction model training module specifically comprises:
the prediction model training unit is used for training by adopting three methods of a support vector machine, an artificial neural network and a random forest according to the processed total organic carbon content prediction model data to obtain a plurality of trained total organic carbon content prediction models;
and the training optimized total organic carbon content prediction model determining unit is used for selecting a model with the highest precision from the plurality of trained total organic carbon content prediction models as the training optimized total organic carbon content prediction model.
10. The system for predicting the total organic carbon content of hydrocarbon source rocks of the offshore region with less wells according to claim 6, wherein the module for determining the distribution area of the high-quality hydrocarbon source rocks specifically comprises:
the prediction unit is used for predicting the total organic carbon content of the hydrocarbon source rock at any position in the research area according to the trained total organic carbon content prediction model to obtain a plurality of prediction results;
and the high-quality source rock distribution area determining unit is used for determining the high-quality source rock distribution area according to each prediction result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211476535.7A CN115796350A (en) | 2022-11-23 | 2022-11-23 | Method and system for predicting total organic carbon content of hydrocarbon source rock in few well regions in sea area |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211476535.7A CN115796350A (en) | 2022-11-23 | 2022-11-23 | Method and system for predicting total organic carbon content of hydrocarbon source rock in few well regions in sea area |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115796350A true CN115796350A (en) | 2023-03-14 |
Family
ID=85440668
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211476535.7A Pending CN115796350A (en) | 2022-11-23 | 2022-11-23 | Method and system for predicting total organic carbon content of hydrocarbon source rock in few well regions in sea area |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115796350A (en) |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107436452A (en) * | 2016-05-27 | 2017-12-05 | 中国石油化工股份有限公司 | Hydrocarbon source rock Forecasting Methodology and device based on probabilistic neural network algorithm |
| CN108717211A (en) * | 2018-06-01 | 2018-10-30 | 北京师范大学 | A kind of prediction technique of the Effective source rocks abundance in few well area |
| US10190998B1 (en) * | 2018-08-29 | 2019-01-29 | Research Institute Of Petroleum Exploration & Development, Dagang Oil Field Of Cnpc | Method and device for evaluating and predicting a shale oil enrichment areas of fault lacustrine basins |
| CN110501744A (en) * | 2019-08-26 | 2019-11-26 | 中海石油(中国)有限公司上海分公司 | Hydrocarbon source rock organic carbon geophysics quantitative forecasting technique, device, equipment and storage medium |
| CN110988997A (en) * | 2019-12-27 | 2020-04-10 | 中国海洋石油集团有限公司 | Hydrocarbon source rock three-dimensional space distribution quantitative prediction technology based on machine learning |
| CN111948718A (en) * | 2020-07-21 | 2020-11-17 | 中国石油天然气集团有限公司 | Method and device for predicting total organic carbon content of shale gas reservoir |
| CN113065279A (en) * | 2021-03-15 | 2021-07-02 | 中国石油大学(北京) | Method, device, equipment and storage medium for predicting total organic carbon content |
| CN113219553A (en) * | 2021-04-25 | 2021-08-06 | 中国石油大学(北京) | Method and device for predicting total organic carbon content of hydrocarbon source rock |
| CN113657548A (en) * | 2021-08-31 | 2021-11-16 | 平安医疗健康管理股份有限公司 | Medical insurance abnormity detection method and device, computer equipment and storage medium |
| US20210365808A1 (en) * | 2020-05-22 | 2021-11-25 | Landmark Graphics Corporation | Facilitating hydrocarbon exploration from earth system models |
| CN114235650A (en) * | 2021-12-16 | 2022-03-25 | 中海石油(中国)有限公司 | Method and system for predicting distribution of hydrocarbon source rocks |
| CN114721898A (en) * | 2022-03-16 | 2022-07-08 | 缀初网络技术(上海)有限公司 | Edge cloud server utilization rate prediction method and device based on boosting algorithm and storage medium |
-
2022
- 2022-11-23 CN CN202211476535.7A patent/CN115796350A/en active Pending
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107436452A (en) * | 2016-05-27 | 2017-12-05 | 中国石油化工股份有限公司 | Hydrocarbon source rock Forecasting Methodology and device based on probabilistic neural network algorithm |
| CN108717211A (en) * | 2018-06-01 | 2018-10-30 | 北京师范大学 | A kind of prediction technique of the Effective source rocks abundance in few well area |
| US10190998B1 (en) * | 2018-08-29 | 2019-01-29 | Research Institute Of Petroleum Exploration & Development, Dagang Oil Field Of Cnpc | Method and device for evaluating and predicting a shale oil enrichment areas of fault lacustrine basins |
| CN110501744A (en) * | 2019-08-26 | 2019-11-26 | 中海石油(中国)有限公司上海分公司 | Hydrocarbon source rock organic carbon geophysics quantitative forecasting technique, device, equipment and storage medium |
| CN110988997A (en) * | 2019-12-27 | 2020-04-10 | 中国海洋石油集团有限公司 | Hydrocarbon source rock three-dimensional space distribution quantitative prediction technology based on machine learning |
| US20210365808A1 (en) * | 2020-05-22 | 2021-11-25 | Landmark Graphics Corporation | Facilitating hydrocarbon exploration from earth system models |
| CN111948718A (en) * | 2020-07-21 | 2020-11-17 | 中国石油天然气集团有限公司 | Method and device for predicting total organic carbon content of shale gas reservoir |
| CN113065279A (en) * | 2021-03-15 | 2021-07-02 | 中国石油大学(北京) | Method, device, equipment and storage medium for predicting total organic carbon content |
| CN113219553A (en) * | 2021-04-25 | 2021-08-06 | 中国石油大学(北京) | Method and device for predicting total organic carbon content of hydrocarbon source rock |
| CN113657548A (en) * | 2021-08-31 | 2021-11-16 | 平安医疗健康管理股份有限公司 | Medical insurance abnormity detection method and device, computer equipment and storage medium |
| CN114235650A (en) * | 2021-12-16 | 2022-03-25 | 中海石油(中国)有限公司 | Method and system for predicting distribution of hydrocarbon source rocks |
| CN114721898A (en) * | 2022-03-16 | 2022-07-08 | 缀初网络技术(上海)有限公司 | Edge cloud server utilization rate prediction method and device based on boosting algorithm and storage medium |
Non-Patent Citations (4)
| Title |
|---|
| 冯明刚;严伟;葛新民;朱林奇;: "利用随机森林回归算法预测总有机碳含量", 矿物岩石地球化学通报, no. 03 * |
| 李正勇;郑姝;彭瀚霖;梁家驹;章珂;王恒;: "安岳气田下寒武统海相烃源岩总有机碳含量测井预测", 测井技术, no. 04 * |
| 蔡冬梅;赵弟江;彭靖松;付立;曾金昌;程耀清;: "优质烃源岩识别及其多属性反演技术定量评价――以渤海海域辽东南洼陷为例", 石油地球物理勘探, no. 02 * |
| 郭双;谢锐杰;邓勇;王伟;: "致密油藏烃源岩有机碳含量预测方法研究", 中国锰业, no. 04 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101930082B (en) | Method for distinguishing reservoir fluid type by adopting resistivity data | |
| CN105275456B (en) | A kind of method utilizing well-log information identification high-quality mud shale | |
| CN109653725A (en) | A layer water flooding degree log interpretation method is stored up based on sedimentary micro and the mixed of rock phase | |
| CN103670388A (en) | Method for evaluating organic carbon content of shale | |
| CN105022098B (en) | A kind of nonmarine deposit body identification based on section inter-layer information and Forecasting Methodology | |
| Tan et al. | Quantitative evaluation methods for water-flooded layers of conglomerate reservoir based on well logging data | |
| CN103628816B (en) | The method of seismic properties Guidance Levels well guiding and device thereof | |
| US20230160304A1 (en) | Method and system for predicting relative permeability curve based on machine learning | |
| CN113886989A (en) | Petroleum drilling parameter optimization method and system based on machine learning | |
| CN116305850A (en) | Stratum thermal conductivity prediction method based on random forest model | |
| CN104965237A (en) | Method of quantitatively and optimally selecting unconventional oil-gas favorable area | |
| Lothe et al. | Storage resources for future European CCS deployment; A roadmap for a Horda CO2 storage hub, offshore Norway | |
| CN106199754B (en) | Oil gas drilling target integrates optimizing evaluation method | |
| CN107818518A (en) | A kind of method in quantization signifying oil shale in-situ exploitation constituency | |
| CN110118994B (en) | Continental facies hydrocarbon source rock quantitative prediction method based on seismic inversion and machine learning | |
| CN115796350A (en) | Method and system for predicting total organic carbon content of hydrocarbon source rock in few well regions in sea area | |
| Wang et al. | Joint rough sets and Karhunen-Loève transform approach to seismic attribute selection for porosity prediction in a Chinese sandstone reservoir | |
| Edmundson et al. | An empirical approach to estimating hydrocarbon column heights for improved predrill volume prediction in hydrocarbon exploration | |
| CN116433059A (en) | Intelligent evaluation method and device for shale oil dessert | |
| Ramesh et al. | Seismic Image Enhancement from Principal Component Analysis: A Case Study from KG Basin | |
| CN117407841B (en) | Shale layer seam prediction method based on optimization integration algorithm | |
| CN110568149A (en) | Fine and rapid quantitative simulation method for hydrocarbon generation and discharge history of sedimentary basin hydrocarbon source rock | |
| CN111325427B (en) | Exploration result determining method and device for exploratory well | |
| Zhao et al. | Machine Learning for Productivity Prediction in Heterogeneous Carbonate Gas Reservoirs, Central Sichuan Basin, China | |
| Li et al. | Seismic attributes blending and borehole image facies analysis for reservoir prediction: Case study for a sand-conglomerate reservoir in Pinghu oil field, East China Sea Basin |
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 |