阅读说明：本技术 一种层序约束下的岩性识别方法 (Lithology identification method under sequence constraint ) 是由 龙飞 郭海伟 王志锋 蒙玉平 杜志宏 耿洪涛 王海洋 易建锋 于 2019-11-25 设计创作，主要内容包括：本发明涉及一种层序约束下的岩性识别方法,属于油气勘探开发技术领域。该方法包括以下步骤：1)获取同一目的层序的设定数目的已知井的取心数据；2)根据所述取心数据,确定目的层序内发育的岩性类别,并建立目的层序内各岩性类别与自然伽马测井数据的对应关系；3)根据所述对应关系以及目的层序中未知井的自然伽马测井数据,识别未知井的岩性。本发明依据同一目的层序的已知井的取心数据进行识别,结果误差较小、精度较高；且利用的是已知井的取心数据,无需再获取未知井的取心数据,因此成本较低。(The invention relates to a lithology identification method under the sequence constraint, and belongs to the technical field of oil and gas exploration and development. The method comprises the following steps: 1) acquiring the coring data of a set number of known wells of the same target sequence; 2) determining lithology categories developing in the target sequence according to the coring data, and establishing a corresponding relation between each lithology category in the target sequence and natural gamma well logging data; 3) and identifying the lithology of the unknown well according to the corresponding relation and the natural gamma ray logging data of the unknown well in the target sequence. The method carries out identification according to the coring data of the known well of the same target sequence, and has small error and high precision; and moreover, the coring data of the known well are utilized, and the coring data of the unknown well do not need to be acquired, so the cost is low.)

1. A lithology recognition method under the sequence constraint is characterized by comprising the following steps:

1) acquiring the coring data of a set number of known wells of the same target sequence;

2) determining lithology categories developing in the target sequence according to the coring data, and establishing a corresponding relation between each lithology category in the target sequence and natural gamma well logging data;

3) and identifying the lithology of the unknown well according to the corresponding relation and the natural gamma ray logging data of the unknown well in the target sequence.

2. The method of claim 1, wherein the natural gamma well logging data is AmpGR, where AmpGR is GR-MinGR, GR is the original natural gamma well logging data, MinGR is the minimum original natural gamma well logging data corresponding to the rock with the lowest shale content in the target sequence, and AmpGR is the natural gamma well logging data after standardized processing.

3. The lithology recognition method under the sequence constraint of claim 1, wherein the process of establishing the corresponding relationship comprises:

a: according to the step 2), selecting a training data set by taking the lithology type and the corresponding natural gamma logging data as samples;

b: establishing the corresponding relationship by using a training data set, wherein the corresponding relationship comprises: the first corresponding relation between each lithology type in the target sequence and the upper boundary value of the natural gamma well logging data, and the second corresponding relation between each lithology type in the target sequence and the lower boundary value of the natural gamma well logging data.

4. The lithology recognition method under the sequence constraint of claim 3, wherein the method for calculating the upper boundary value and the lower boundary value corresponding to the lithology of a certain category comprises the following steps:

counting the mean value of natural gamma well logging data of each lithology type in a target sequence;

secondly, sequentially arranging the lithology categories from small to large according to the average value corresponding to each lithology category in the target sequence;

calculating the average value of the natural gamma well logging data of the lithology of a certain category and the lithology of a larger adjacent category, and recording the average value as an upper boundary value corresponding to the lithology of the category; and calculating the average value of the natural gamma well logging data of the lithology of a certain category and the lithology of a smaller adjacent category, and recording the average value as a lower boundary value corresponding to the lithology of the category.

5. The method of lithology recognition under sequence constraints of claim 3, further comprising: and according to the step 2), selecting a test data set by taking the lithology type and the corresponding natural gamma logging data as samples, and evaluating the corresponding relation by using the test data set.

6. The method of lithology recognition under sequence constraints of claim 5, wherein the training data set and the testing data set are non-overlapping and a ratio of training data set to testing data set sample numbers is 7: 3.

7. The method of lithology recognition under sequence constraints of claim 3, wherein the step 3) of recognizing the lithology of the unknown well comprises:

and judging whether the natural gamma well logging data of a certain depth point of the unknown well is between an upper boundary value and a lower boundary value corresponding to a certain category of lithology, and if so, judging that the depth point of the unknown well is the category of lithology.

8. The lithology recognition method under the sequence constraint of claim 3, wherein step B further comprises a process of correcting the established correspondence.

9. The method of identifying lithology under sequence constraints of claim 8, wherein the parameters used in the correction in step B include variance, standard deviation and skewness.

10. The method of lithology identification under sequence constraints of claim 1, wherein the lithology categories include: sandstone, siltstone, argillaceous siltstone, silty mudstone, and mudstone.

Technical Field

The invention relates to a lithology identification method under the sequence constraint, and belongs to the technical field of oil and gas exploration and development.

Background

Lithology recognition is a fundamental work in geological research and provides basic data for reservoir description. The existing lithology identification methods comprise the following steps:

1) a geological method.

And (4) identifying lithology by geologists through observing the rock core according to the sedimentary characteristics such as granularity, rhythm and the like of rock core components. The lithology identified by the method is the most accurate, but the coring cost is high, only a few coring wells have the feasibility of the method, and therefore the lithology identification covering the whole area cannot be carried out by the method.

2) A cluster analysis method.

The method is characterized in that logging data are used as basic data, clustering analysis is carried out on the logging data of a research area, and the type of lithology of the research area is divided according to lithology classification quantity. The method has the characteristics of simplicity and rapidness, but the cluster analysis is carried out on the logging data of the whole well section, and the analysis result is greatly influenced by the data of the non-target layer of the whole well section, so that a large error is generated.

3) A seismic method.

The method carries out lithology identification by extracting seismic attributes such as longitudinal and transverse wave velocity, amplitude and the like which can represent lithology types. The method is greatly restricted by factors such as seismic acquisition, data processing, personal experience of geologists and the like, so that the lithology identification error is large, and in addition, due to the restriction of seismic wavelength, the resolution of seismic data is low, and the thin lithology cannot be accurately identified.

In summary, the existing lithology identification method has the problems of high cost, large error and low precision.

Disclosure of Invention

The invention aims to provide a lithology identification method under the sequence constraint, which is used for solving the problems of high cost, large error and low precision in the conventional lithology identification method.

In order to achieve the purpose, the invention provides a lithology identification method under the sequence constraint, which comprises the following steps:

1) acquiring the coring data of a set number of known wells of the same target sequence;

2) determining lithology categories developing in the target sequence according to the coring data, and establishing a corresponding relation between each lithology category in the target sequence and natural gamma well logging data;

3) and identifying the lithology of the unknown well according to the corresponding relation and the natural gamma ray logging data of the unknown well in the target sequence.

The beneficial effects are that: according to the method, the corresponding relation between each lithology type in the target sequence and natural gamma well logging data is established by acquiring the coring data of the known well of the same target sequence, and the lithology of the unknown well can be identified according to the natural gamma well logging data of the unknown well in the target sequence according to the corresponding relation; the sedimentary bodies in the same sequence are isochronous sedimentary units with the same cause, and the result error of the identification according to the known well coring data of the same target sequence is smaller and the accuracy is higher; and moreover, the coring data of the known well are utilized, and the coring data of the unknown well do not need to be acquired, so the cost is low.

Further, the natural gamma logging data is AmpGR, where AmpGR is GR-MinGR, GR is original natural gamma logging data, MinGR is the minimum original natural gamma logging data corresponding to the rock with the lowest shale content in the target sequence, and AmpGR is the natural gamma logging data after the normalization processing.

The beneficial effects are that: the invention uses a formula to calculate the AmpGR (GR-MinGR), establishes the corresponding relation between the natural gamma logging data AmpGR after the standardization processing and each lithology type in the target sequence, can ensure that GR data measured by different wells and different instruments have comparability, and can solve the problem that different logging instruments have measurement difference to the same lithology type.

Further, the process of establishing the corresponding relationship includes:

a: according to the step 2), selecting a training data set by taking the lithology type and the corresponding natural gamma logging data as samples;

b: establishing the corresponding relationship by using a training data set, wherein the corresponding relationship comprises: the first corresponding relation between each lithology type in the target sequence and the upper boundary value of the natural gamma well logging data, and the second corresponding relation between each lithology type in the target sequence and the lower boundary value of the natural gamma well logging data.

The beneficial effects are that: the method specifically establishes a first corresponding relation between each lithology type in the target sequence and the upper boundary value of the natural gamma well logging data and a second corresponding relation between each lithology type in the target sequence and the lower boundary value of the natural gamma well logging data, and lays a foundation for accurately judging the lithology of the unknown well.

Further, the method for calculating the upper boundary value and the lower boundary value corresponding to the lithology of a certain category comprises the following steps:

counting the mean value of natural gamma well logging data of each lithology type in a target sequence;

secondly, sequentially arranging the lithology categories from small to large according to the average value corresponding to each lithology category in the target sequence;

calculating the average value of the natural gamma well logging data of the lithology of a certain category and the lithology of a larger adjacent category, and recording the average value as an upper boundary value corresponding to the lithology of the category; and calculating the average value of the natural gamma well logging data of the lithology of a certain category and the lithology of a smaller adjacent category, and recording the average value as a lower boundary value corresponding to the lithology of the category.

The beneficial effects are that: the invention provides a method for calculating a first corresponding relation between each lithology type in a target sequence and a boundary value on natural gamma well logging data, and a method for calculating a second corresponding relation between each lithology type in the target sequence and a lower boundary value on the natural gamma well logging data, so that the accuracy of the unknown well lithology judgment of the invention is improved.

Further, the method also comprises the following steps: and according to the step 2), selecting a test data set by taking the lithology type and the corresponding natural gamma logging data as samples, and evaluating the corresponding relation by using the test data set.

The beneficial effects are that: the invention also comprises a step of evaluating the established corresponding relation, thereby realizing the effective evaluation of the correctness of the established corresponding relation.

Further, the training data set and the test data set do not overlap, and the ratio of the number of samples of the training data set to the number of samples of the test data set is 7: 3.

The beneficial effects are that: the training data set and the test data set samples are divided according to the proportional relation of 7:3, so that the accurate establishment of the corresponding relation and the accurate evaluation of the evaluation process can be ensured.

Further, the method for identifying the lithology of the unknown well in the step 3) comprises the following steps: and judging whether the natural gamma well logging data of a certain depth point of the unknown well is between an upper boundary value and a lower boundary value corresponding to a certain category of lithology, and if so, judging that the depth point of the unknown well is the category of lithology.

The beneficial effects are that: based on the upper boundary value and the lower boundary value corresponding to different types of lithologies, the accurate identification of the lithologies of the unknown well is realized according to the size of the natural gamma well logging data of a certain depth point of the unknown well.

Further, step B further includes a process of correcting the established correspondence.

The beneficial effects are that: the accuracy of the established corresponding relation is improved, and the rock property of the unknown well is accurately identified.

Further, the parameters used for the correction in step B include variance, standard deviation and skewness.

The beneficial effects are that: the accuracy of the established corresponding relation is improved based on the mathematical parameters, and the accurate identification of the lithology of the unknown well is realized.

Further, the lithology categories include: sandstone, siltstone, argillaceous siltstone, silty mudstone, and mudstone.

Drawings

FIG. 1 is a flow chart of a lithology identification method in an embodiment of the lithology identification method under sequence constraint of the invention;

FIG. 2 is a comparison graph of amplitude-average values AmpGR of the argillaceous siltstones GR before and after skewness correction in an embodiment of the lithology recognition method under sequence constraint;

FIG. 3 is a comparison diagram of lithology and actual lithology identified in a test data set in an embodiment of the lithology identification method under the sequence constraint of the invention.

Detailed Description

First embodiment of the lithology identification method under the sequence constraint:

in this embodiment, under the guidance of the high-resolution sequence theory that "the depositional bodies in the same sequence are isochronous depositional units with the same cause", gamma logging data GR is used as a research object, and a lithology identification method under the constraint of the sequence is proposed, as shown in fig. 1, specifically including the following steps:

s1: selecting coring data of 150 wells drilled in the target sequence FI 1 in the research area, and extracting lithology of development in the FI 1 sequence as follows: sandstone, siltstone, argillaceous siltstone, siltstone mudstone, as shown in table 1.

TABLE 1

S2: GR well log data for 150 wells drilled in the study area encountering the fi 1 sequence were normalized using the formula AmpGR ═ GR-MinGR. The GR is original natural gamma well logging data of each depth point in the F I1 sequence, the MinGR is a minimum GR value corresponding to sandstone in the F I1 sequence, and the AmpGR is a GR amplitude value of other lithological GR relatively pure sandstone in the F I1 sequence.

S3: establishing a one-to-one corresponding relation between lithology types and electrical properties AmpGR of known lithology wells encountering an FI 1 sequence in a research area by utilizing the depth equality relation between the AmpGR data and coring data of all depth points of 150 wells encountering the FI 1 sequence in the research area, taking the lithology types and the AmpGR electrical property values of the 150 wells as sample data, and performing the following steps according to the ratio of 7: and 3, taking 105 wells of data as a training data set for establishing the lithology model, and taking 45 wells of data as a test data set for evaluating the identification effect of the lithology model.

And (3) performing multiple rounds of correction on the skewness (sk) of the AmpGR of the sandstone, siltstone, argillaceous siltstone, siltstone mudstone and mudstone in the training data set, and calculating the following parameters of the AmpGR values of various lithologies in the target sequence by using a statistical formula: mean value of () Variance (δ), standard deviation(s) data.

Rejecting outlier data by skewness (sk) correction: the skewness (sk) of the sample data reflects the distribution characteristics of the sample, and the sample with the normal distribution characteristics has statistics. And eliminating abnormal data such as a few maximum values and minimum values of the AmpGR in the target sequence stratum by correcting the skewness (sk) of the AmpGR for multiple times, so that the AmpGR values of various lithologies in the target sequence stratum are normally distributed, and after correction, the average values of the AmpGR of various lithologies in the same target sequence stratum accord with geological knowledge increased according to the increase of argillaceous content. The amplitude mean value AmpGR of the argillaceous siltstone GR before and after correction is shown in fig. 2.

TABLE 2

Lithology	Maximum of	Minimum size	Mean value	Variance (variance)	Deflection degree	Type (B)	Layering
								Sandstone	32.373	0	24.056	35.21065	-0.6543	GR	FⅠ1
Siltstone	140.513	16.011	32.05648	3.47338	1.79631	GR	FⅠ1
								Argillaceous siltstone	45.263	32.992	40.28634	11.29982	-6.85344	GR	FⅠ1
Silty mudstone	55.694	31.987	43.54692	71.72768	-46.161	GR	FⅠ1
								Mudstone	55.36	45.235	50.30063	8.63781	-0.82447	GR	FⅠ1

As can be seen from Table 2: the average AmpGR values of various lithological properties in the same target sequence stratum accord with the geological understanding increased according to the increase of the argillaceous content.

S4: on the basis of correcting skewness (sk) data of the AmpGR in the previous step, establishing the following parameters capable of representing sandstone, siltstone, argillaceous siltstone, siltstone mudstone and mudstone in the F I1 sequence stratum by using a training data set: the AmpGR mean, the AmpGR upper boundary and the AmpGR lower boundary are shown in Table 3, and the corresponding relation is the lithology model established by the invention. Wherein the upper boundary of AmpGR is obtained from the average of AmpGR means of adjacent lithologies according to the direction of AmpGR increase, such as the AmpGR upper boundary of siltstone (32.05648+ 40.28634)/2; the lower boundary of AmpGR is obtained from the average of AmpGR means of adjacent lithologies in the direction of AmpGR reduction, for example, siltstone AmpGR lower boundary ═ (32.05648+ 24.056)/2.

TABLE 3

S5: AmpGR of 45 wells drilled in the research area and meeting the F I1 sequence test data set is matched with AmpGR upper boundaries and AmpGR lower boundaries of sandstone, siltstone, silty mudstone and mudstone in the rock character model under the F I1 sequence stratum established in the step S4, the rock characters corresponding to AmpGR of each depth point in the 45-well F I1 sequence stratum are identified, and then the identified rock characters of each depth point in the 45-well F I1 sequence are compared with the actual rock characters to evaluate the effect of the rock character model, as shown in figure 3.

S6: AmpGR of 200 rock-character unknown wells drilled in the F I1 sequence in the research area is matched with AmpGR upper boundaries and AmpGR lower boundaries of sandstone, siltstone, argillaceous siltstone, silty mudstone and mudstone in the rock-character model in the F I1 sequence established in the step S4, and the lithology corresponding to the AmpGR of each depth point in the F I1 sequence of the 200 wells is identified, and the table 4 shows.

TABLE 4

Depth of top bound	Depth of bottom boundary	Name of lithology	Layering
				1450	1453.59	Mudstone	FⅠ1
1453.59	1454.3	Silty mudstone	FⅠ1
				1454.3	1462.2	Mudstone	FⅠ1
1462.2	1463.18	Siltstone	FⅠ1
				1463.18	1463.59	Argillaceous siltstone	FⅠ1
1463.59	1465.18	Silty mudstone	FⅠ1

The embodiment of the lithology identification method under the sequence constraint is as follows:

the lithology recognition method under the sequence constraint proposed in this embodiment is the same as steps S1 and S2 in the first embodiment, and is not described here again.

The difference between the lithology identification method under the sequence constraint proposed in this embodiment and the step S3 in the first embodiment is as follows: the method comprises the steps of establishing a one-to-one corresponding relation between lithology types and electrical properties AmpGR of known lithology wells drilled in a research area and encountering an F I1 sequence by utilizing the depth equality relation between the AmpGR data and coring data of all depth points of 150 wells drilled in the research area and encountering the F I1 sequence, and taking the lithology types and AmpGR electrical property values of the 150 wells as sample data which are used as training data sets for establishing a lithology model.

Sandstone, siltstone, argillaceous siltstone and powder with concentrated training dataAnd (3) performing multiple rounds of correction on the skewness (sk) of the AmpGR of the sandy mudstone and the mudstone, and calculating the following parameters of the AmpGR values of various lithologies in the target sequence by using a statistical formula: mean value of () Variance (δ), standard deviation(s) data.

S4: on the basis of correcting skewness (sk) data of the AmpGR in the previous step, establishing the following parameters capable of representing sandstone, siltstone, argillaceous siltstone, siltstone mudstone and mudstone in the F I1 sequence stratum by using a training data set: the corresponding relation among the AmpGR mean value, the AmpGR upper boundary and the AmpGR lower boundary is the lithology model established by the invention. Wherein the upper boundary of the AmpGR is obtained by the average value of the AmpGR mean values of adjacent lithologies according to the increasing direction of the AmppGR; the lower bound of AmpGR is obtained from the average of AmpGR means of adjacent lithologies according to AmpGR reduction direction.

S5: and matching the AmpGR of the 200-opening rock-character unknown well drilling in the F I1 sequence in the research area with the AmpGR upper boundary and the AmpGR lower boundary of sandstone, siltstone, argillaceous siltstone, silty mudstone, mudstone in the rock-character model in the F I1 sequence established in the step S4, and identifying the lithology corresponding to the AmpGR of each depth point in the F I1 sequence of the 200-opening well.

The third embodiment of the lithology identification method under the sequence constraint:

the difference between the lithology recognition method under the sequence constraint provided in this embodiment and the first embodiment is that the target sequence studied in this embodiment is changed, and the extracted developed lithology does not include sandstone, so MinGR is the minimum original natural gamma logging data corresponding to the rock with the lowest shale content in the sequence, and AmpGR is a GR amplitude value of other lithologies GR in the sequence relative to the rock with the lowest shale content in the sequence. The rest of the method process is the same as the first embodiment, and is not described herein.

The fourth embodiment of the lithology identification method under the sequence constraint:

the difference between the lithology recognition method under the sequence constraint provided in this embodiment and the first embodiment is that, in this embodiment, when the skewness (sk) of AmpGR of sandstone, siltstone, argillaceous siltstone, siltstone mudstone, and mudstone in the training data set is corrected, the correction is performed by using a filtering algorithm or the like in a manner different from the first embodiment. The rest of the method process is the same as the first embodiment, and is not described herein.

Fifth embodiment of the lithology identification method under the sequence constraint:

the difference between the lithology recognition method under the sequence constraint provided by this embodiment and the first embodiment is that the present embodiment utilizes a neural network to train sample data, and further establishes a corresponding relationship between each lithology type in the target sequence and natural gamma well logging data. The rest of the method process is the same as the first embodiment, and is not described herein.

Sixth embodiment of the lithology identification method under the sequence constraint:

the difference between the lithology identification method under the sequence constraint provided by the embodiment and the first embodiment is that the corresponding relation between each lithology type in the target sequence and natural gamma well logging data is established by using Fourier training sample data. The rest of the method process is the same as the first embodiment, and is not described herein.