Well logging curve layer prediction method and system based on optimal window wavelet transform

Abstract

The present invention belongs to the technical field of well logging data processing and interpretation, and discloses a well logging curve stratification prediction method and system based on optimal window wavelet transform, which pre-processes the well logging curves of coring wells and non-coring wells; performs artificial stratification and sedimentary microfacies analysis on natural gamma logging curves; performs wavelet transform on natural gamma logging curves and natural potential logging curves respectively to obtain the optimal window size; stratifies non-coring wells according to the optimal window size; and merges stratification points to obtain the best automatic stratification effect of the well logging curve. The present invention uses the mutual contrast constraint of artificial stratification and automatic stratification, uses the loss function to determine the optimal window size of the wavelet transform modulus maximum of the natural gamma GR curve and the natural potential SP curve, and uses the GR curve and the SP curve to perform stratification respectively, and finally performs stratification post-processing, effectively reducing missed classification and wrong classification, and obtaining the best automatic stratification effect.

Description

Technical Field

[0001] The invention belongs to the technical field of well logging data processing and interpretation, and in particular relates to a well logging curve hierarchical prediction method and system based on optimal window wavelet transform. Background Art

[0002] In the interpretation of well logging curves, especially in lithology interpretation, well logging phase classification, and reservoir division, well logging curve stratification is the primary task. Manual stratification methods are time-consuming and laborious, and traditional automatic stratification methods such as clustering and extreme variance are inefficient, and the classification results are highly arbitrary. Therefore, it is necessary to optimize the stratification parameters to adapt them to the specific work area and geological environment.

[0003] The automatic stratification algorithm of logging curves mainly includes mathematical statistics methods: including intra-layer difference method, ordered cluster analysis (optimal segmentation method), least squares change point analysis method and extreme variance clustering method, etc., and artificial intelligence methods: including fuzzy clustering and neural network methods. These methods have achieved certain application results in practice, but only use the characteristics of the logging curve itself without adding the constraints of the core, so the geological significance is not very clear and the stratification results are relatively arbitrary. In addition, there are also methods such as activity analysis method, differential extreme value method, multi-scale analysis method, etc., which have also been applied to the automatic stratification of logging curves.

[0004] Activity analysis and differential extreme value methods are simple and convenient to calculate, but they are generally used to analyze a single logging curve. The information is single and not representative, and it is easily affected by noise, which will lead to too many errors in the stratification interface and affect the stratification effect. As a powerful non-stationary signal analysis tool, wavelet transform has been widely used in geophysical exploration and logging analysis. Its good localization characteristics can improve the resolution of logging sequences. Generally speaking, the main factor causing the change of logging values ​​is the change of formation properties. Wavelet transform can detect signal mutations and describe the degree of signal mutation, so that it can more accurately capture the changes in formation properties. The multi-scale analysis of logging curves using low-pass wavelets, modulus maximum, etc. can quantitatively divide and distinguish the formation interfaces of different sedimentary periods at different scales, thereby achieving the purpose of detecting the position of the formation interface.

[0005] Therefore, using wavelet transform to stratify logging curves is an effective method, but the optimal window of wavelet transform modulus maximum is difficult to determine, and the window size plays a key role in the stratification results.

[0006] Through the above analysis, the problems and defects of the prior art are as follows:

[0007] (1) The existing automatic stratification algorithm for well logging curves only utilizes the characteristics of the well logging curves themselves without adding the constraints of the core. Therefore, the geological significance is not clear and the stratification results are relatively arbitrary.

[0008] (2) The activity analysis method and the derivative extreme value method are based on the analysis of a single logging curve. The information is single and not representative. Moreover, it is easily affected by noise, resulting in too many errors in the stratification interface and affecting the stratification effect. Summary of the invention

[0009] In order to overcome the problems existing in the related art, the disclosed embodiment of the present invention provides a method and system for hierarchical prediction of well logging curves based on optimal window wavelet transform, which involves classification and recognition, attribute extraction and information mining of well logging curves. The technical solution is as follows:

[0010] The present invention is implemented in this way: a logging curve stratification prediction method based on optimal window wavelet transform comprises:

[0011] S1, preprocessing the logging curves of the cored well and the non-cored well;

[0012] S2, manual stratification and automatic stratification parameter selection for the coring well;

[0013] S3, stratification of non-cored wells based on optimal window size;

[0014] S4, post-processing the stratification points to obtain the best automatic stratification effect.

[0015] In step S1, the logging curves include: natural gamma logging curves and natural potential logging curves.

[0016] In step S1, the well logging curves of the coring well and the non-coring well are preprocessed, including: using an adaptive morphological filtering method based on sample entropy to remove small changes in the well logging curves, and the two basic operations of morphological filtering are dilation and erosion;

[0017] The expansion operation is:

[0018] ;

[0019] The corrosion operation is:

[0020] ;

[0021] In the formula, For expansion operation, For corrosion operations, is the logging curve value, is defined in The length is A discrete sequence of is a structural element, defined in The length is a discrete sequence of ; and ; are the serial numbers of the discrete sequences, To obtain the maximum value operation;

[0022] The Gaussian distribution is used to establish the structural element. The length of the structural element is the number of sampling points corresponding to the thickness of the first layer to be identified (here is 17). The standard deviation determines the maximum value of the Gaussian distribution. The larger the standard deviation, the smaller the maximum value of the Gaussian distribution. The sample entropy is calculated and limited to The reciprocal is taken to get the standard deviation, and the structural element is designed according to the standard deviation; the expression of the unnormalized structural element is:

[0023] ;

[0024] In the formula, is a discrete sequence number, is the standard deviation, For exponential calculation operations;

[0025] Normalized Structural Element for:

[0026] ;

[0027] Standard Deviation for:

[0028] ;

[0029] In the formula, The sample entropy of the well logging curve value in the selected window is used to calculate the normalized structural element Finally, the method of first dilation and then corrosion and first corrosion and then dilation and then averaging is adopted to remove the interference of small jagged changes in the logging curve.

[0030] In step S2, the coring well is manually stratified, including:

[0031] Based on the processed natural gamma-ray logging curves and core characteristics, including lithology, composition, combination law, color, structure, thickness and rhythmic characteristics, the high-resolution sequence stratigraphy principle is applied, combined with the logging phase characteristics, to divide the single well sedimentary microfacies within the isochronous stratigraphic framework; record the number of artificial stratification layers and layer point number , is the layer number, which is a known constraint for automatic layering parameter selection.

[0032] In step S2, automatic stratification parameter selection is performed on the coring well, including:

[0033] Perform wavelet transform on natural gamma logging curve:

[0034] ;

[0035] In the formula, is the calculated wavelet coefficient, is the scale factor, and the binary wavelet transform is selected. , is the scale coefficient; is a time shift, representing the sample points of the well logging curve at depth; is the real number domain of the integral space, is the natural gamma logging curve, is the wavelet function;

[0036] At a given window size Next, calculate the modulus maximum , and draw the modulus maximum curve;

[0037] ;

[0038] In the formula, To take the absolute value operation, For the window Find the maximum value inside;

[0039] Given a scale factor , calculate the Lipschitz coefficient , the expression is:

[0040] ;

[0041] In the formula, is the modulus maximum position, is the wavelet transform of adjacent scales;

[0042] The mutation point is found through the wavelet transform modulus maximum value to determine the stratum interface.

[0043] Furthermore, the natural gamma logging curve is subjected to continuous wavelet transform, the wavelet transform modulus maximum is calculated, and the window size is set to ;

[0044] Calculate the wavelet transform modulus maximum curve according to different window sizes, and calculate the Lipschitz coefficient, through the preset threshold Determine the stratification point, stratify the logging curve, and record the number of stratification layers and layer point number ,in, is the window size sequence number, is the order of the layered points, and the value of the loss function is calculated. The expression is:

[0045] ;

[0046] In the formula, For the window is The loss value when is the number of artificial stratification layers, according to and The number of Eliminate so that and The number is equal; When it is the smallest, the corresponding stratification point number is the optimal natural gamma stratification , the selected window size is the optimal window size .

[0047] Furthermore, the natural gamma-ray logging curve is layered by wavelet transform modulus maximum value, and the optimal layering point number is obtained by using the GR curve optimal layering method. and the optimal window size .

[0048] In step S3, the non-cored wells are stratified, including:

[0049] Based on the optimal window size and , and the non-coring wells were stratified separately.

[0050] In step S4, the delamination points are post-processed, including:

[0051] According to and The obtained stratification points are merged. and The difference is no more than The dividing point of the sampling points is merged, where The number of sampling points representing the first layer of the area obtains the best automatic stratification effect using natural gamma logging curves and natural potential logging curves.

[0052] Another object of the present invention is to provide a logging curve stratification prediction system based on optimal window wavelet transform, which is used to control the logging curve stratification prediction method based on optimal window wavelet transform. The system includes:

[0053] Well logging curve preprocessing module, used for denoising and preprocessing the natural gamma logging curves and natural potential logging curves of coring wells and non-coring wells;

[0054] The coring well artificial stratification module is used to perform artificial stratification of natural gamma logging curves and sedimentary microfacies analysis based on processed natural gamma logging curves and core characteristics;

[0055] The automatic layering parameter selection module for coring wells is used to perform wavelet transform on the natural gamma logging curve and the natural potential logging curve to obtain the optimal window size. and ;

[0056] Non-cored well stratification module for optimizing window size and stratify non-cored wells;

[0057] The layer post-processing module is used to post-process the layer points to obtain the best layer effect of natural gamma logging curve and natural potential logging curve

[0058] In combination with all the above technical solutions, the beneficial effects of the present invention are as follows:

[0059] The present invention determines the optimal window size of the wavelet transform modulus maxima of the GR curve and the SP curve through the mutual comparison constraint of manual stratification and automatic stratification through the proposed loss function, and uses the GR curve and the SP curve for stratification respectively, and finally performs stratification post-processing, which can comprehensively consider the influence of the first layer thickness (the minimum thickness formation) and the second layer thickness (the maximum thickness formation), effectively reduce missed classification and wrong classification, and obtain the best automatic stratification effect. Conventional wavelet transform modulus maxima are used for automatic stratification of logging curves, and the window parameters can only be judged by experience, and the influence of window selection on the stratification results cannot be quantitatively characterized. The present invention uses the GR curve and the SP curve for wavelet transform modulus maxima stratification respectively, and calculates the loss function of manual stratification and automatic stratification according to the manual stratification results to obtain the optimal window size, thereby achieving the best stratification effect.

[0060] The present invention optimizes the stratification parameters according to the coring data to realize the automatic stratification of the logging curve, and specifically involves the classification and identification, attribute extraction and information mining of the logging curve. The present invention determines the optimal window parameters through manual stratification and automatic stratification of the coring well, and then applies it to the automatic stratification of the non-coring well, so as to improve the accuracy and efficiency of the stratification of the logging curve.

[0061] The present invention can greatly improve the efficiency of logging curve processing. For a large amount of logging data in the same measurement area, only manual stratification is required for typical coring wells to obtain the optimal parameters for automatic stratification. Applying it to non-coring wells can quickly and efficiently achieve manual stratification and be used for subsequent logging interpretation, microfacies identification, and core identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] The accompanying drawings herein are incorporated in and constitute a part of the specification, illustrate embodiments consistent with the present disclosure, and together with the description, serve to explain the principles of the present disclosure;

[0063] Figure 1It is a flow chart of a logging curve stratification prediction method based on optimal window wavelet transform provided by an embodiment of the present invention;

[0064] Figure 2 Schematic diagram of morphological filtering results of GR logging curves provided in an embodiment of the present invention (solid line is the original curve, and dotted line is the pre-processed curve)

[0065] Figure 3 1 is a schematic diagram of the wavelet transform modulus maximum GR layering provided by an embodiment of the present invention, wherein (a) is a schematic diagram before filtering, and (b) is a schematic diagram before filtering;

[0066] Figure 4 1 is a schematic diagram of wavelet transform modulus maximum SP layering provided by an embodiment of the present invention, wherein (a) is a schematic diagram before filtering, and (b) is a schematic diagram before filtering;

[0067] Figure 5 are the stratified results of different window sizes of the GR curve provided by the embodiment of the present invention, where w=7 in (a), w=15 in (b), and w=25 in (c);

[0068] Figure 6 1 is the stratification result of different window sizes of the SP curve provided by the embodiment of the present invention, wherein w=7 in (a), w=15 in (b), and w=25 in (c);

[0069] Figure 7 is a schematic diagram of artificial stratification provided by an embodiment of the present invention;

[0070] Figure 8 is a loss function curve diagram of different window GR layers provided by an embodiment of the present invention;

[0071] Fig. 9 is a loss function curve diagram of different window SP layers provided by an embodiment of the present invention;

[0072] Fig.10 It is a schematic diagram of the best layering of the GR curve provided by an embodiment of the present invention;

[0073] Fig.11 It is a schematic diagram of the best layering of the SP curve provided by an embodiment of the present invention;

[0074] Fig.12 It is a comparison chart between the final automatic stratification result and the manual stratification provided by the embodiment of the present invention. DETAILED DESCRIPTION

[0075] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below in conjunction with the accompanying drawings. In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention can be implemented in many other ways different from those described herein, and those skilled in the art can make similar improvements without violating the connotation of the present invention, so the present invention is not limited by the specific implementation disclosed below.

[0076] The innovation of the present invention is that the present invention takes artificial stratification as a constraint, obtains the optimal parameters of automatic stratification for well logging by minimizing the loss function, and establishes an automatic stratification workflow for well logging.

[0077] Embodiment: The method for layered prediction of well logging curves based on optimal window wavelet transform provided by the embodiment of the present invention specifically comprises the following steps:

[0078] S1, preprocessing the logging curves of cored wells and non-cored wells;

[0079] Different rocks are in different physical and chemical environments during the diagenesis process, and have different rock components, so the types and concentrations of radionuclides they contain are also different. The natural gamma curve reflects the natural radioactivity intensity of the rock, which is closely related to the stratigraphic age, sedimentary environment and lithology combination. Therefore, the natural gamma logging curve can be used for stratigraphic division.

[0080] The natural potential logging curve plays an important role in the division of clastic rock formations, lithology identification, reservoir fluid identification and evaluation. The amplitude of the natural potential curve is related to the mud content, fluid type, and formation thickness, and can reflect the type of reservoir sand body. Therefore, the present invention mainly uses the natural gamma (GR) logging curve and the natural potential (SP) logging curve for stratification.

[0081] First, the logging curves of all wells (including coring wells and non-coring wells) are preprocessed, mainly including denoising and smoothing. The denoising method can be selected from median filtering, Gaussian filtering and morphological filtering, etc., in order to eliminate the tiny sawtooth of the logging curve and reduce the possibility of false formations in the layer. The denoising method selected by the present invention is the morphological filtering method. The morphological filtering results of the GR logging curve are shown in Figure 2 .

[0082] The two basic operations of morphological filtering are dilation and erosion:

[0083] The expansion operation is:

[0084] ;

[0085] The corrosion operation is:

[0086] ;

[0087] In the formula, For expansion operation, For corrosion operations, is the logging curve value, is defined in The length is A discrete sequence of is a structural element, defined in The length is a discrete sequence of ; and ; are the serial numbers of the discrete sequences, To obtain the maximum value operation;

[0088] The selection of structural elements directly affects the effect of morphological filtering. Sample entropy can reflect the local complexity of the logging curve, that is, the formation structure, and the structural elements can be adaptively determined according to the window complexity. High entropy areas usually contain more details or textures, while low entropy areas are relatively flat or uniform. Based on this, large structural elements are used for smoothing in high entropy areas, and small structural elements are used in low entropy areas to maintain features. The present invention uses Gaussian distribution to establish structural elements. The length of the structural element is the number of sampling points corresponding to the thickness of the first layer to be identified, and 11 is selected in the present invention. The standard deviation determines the maximum value of the Gaussian distribution. The larger the standard deviation, the smaller the maximum value of the Gaussian distribution. Therefore, the present invention first calculates the sample entropy, limits the sample entropy to between [1,10], then takes the reciprocal to obtain the standard deviation, and then designs the structural element based on the standard deviation.

[0089] The expression of the unnormalized structure element is:

[0090] ;

[0091] In the formula, is a discrete sequence number, is the standard deviation, For exponential calculation operations;

[0092] Normalized Structural Element for:

[0093] ;

[0094] Standard Deviation for:

[0095] ;

[0096] In the formula, The sample entropy of the well logging curve value in the selected window is used to calculate the normalized structural element Finally, the method of first dilation and then corrosion or first corrosion and then dilation and then averaging is used to remove the interference of small jagged changes in the logging curve.

[0097] S2, manual stratification and automatic stratification parameter selection for the coring well;

[0098] Core data is the most intuitive basis for the division of sedimentary microfacies. Its rock characteristics and sedimentary structural characteristics provide a direct basis for the division of sedimentary microfacies of a single well. Based on the processed natural gamma logging curve and core characteristics, including lithology, composition, combination law, color, structure, thickness and rhythmicity, the high-resolution sequence stratigraphy principle is applied, starting from the identification of the base level cycle interface, combined with the logging phase characteristics, the single well sedimentary microfacies is divided within the isochronous stratigraphic framework. Record the number of artificial stratification layers and layer point number ,in, is the layer number, which is a known constraint for automatic layering parameter selection.

[0099] According to the processed natural gamma ray GR curve and core characteristics, the GR curve is artificially layered and the sedimentary microfacies are analyzed, and the number of layers is recorded. and layer point number ( is the layer number), as a known constraint for automatic layering parameter selection.

[0100] S3, stratification of non-cored wells based on optimal window size;

[0101] Take the GR curve as an example and perform wavelet transform on it, where: represents the GR curve, is the scale factor, and here we choose the binary wavelet transform, so , is the scale factor, is the time shift, representing the sample point of the logging curve at depth, is a wavelet function, and the present invention selects a Gaussian wavelet function, then is the wavelet coefficient calculated. In a given window size Next, calculate the modulus maximum , and draw the modulus maximum curve. Given a scale Calculate the Lipschitz coefficient ,measure Point singularity, When it is 1, it indicates that the curve has no singularity. When it is 0, it indicates that the curve is discontinuous. The bigger the description exist The smoother the point, The smaller the description exist The more sudden the change, the more likely it is that the mutation point will change. Therefore, the maximum value of the modulus of wavelet transform can be used to find the mutation point, that is, the stratum interface.

[0102] ;

[0103] In the formula, is the calculated wavelet coefficient, is the scale factor, and the binary wavelet transform is selected. , is the scale coefficient; For time shift; is the real number domain of the integral space, is the natural gamma logging curve, is the wavelet function;

[0104] ;

[0105] In the formula, To take the absolute value operation, For the window Find the maximum value inside;

[0106] Given a scale factor , calculate the Lipschitz coefficient , the expression is:

[0107] ;

[0108] In the formula, is the modulus maximum position, is the wavelet transform of adjacent scales;

[0109] The wavelet transform modulus maximum increases with the scale increases with the increase of When fixed, the larger the maximum value of the wavelet transform modulus is, the higher the value is. The larger the scale factor, the sharper the change. (generally taking the finest scale), the maximum value of the wavelet transform modulus and the window size There is a direct relationship between the window size and the wavelet transform modulus maximum. Therefore, the key factor that determines the stratification result of the wavelet transform modulus maximum is the window size parameter. In order to obtain the optimal stratification result, the following operations are performed on the natural gamma GR curve and the natural potential SP curve respectively:

[0110] 3.1 First, perform wavelet transform on the natural gamma GR curve, calculate the wavelet transform modulus maximum, and set the window size range to In practice, the resolution of the logging curve is generally The window size determines the sensitivity to the mutation point. Wavelet transform is performed according to different window sizes, and the Lipschitz coefficient is calculated. , stratify the logging curve and record the number of stratification layers and layer point number ( is the window size sequence number, is the stratification point number), and the value of the loss function is calculated by the following formula:

[0111] ;

[0112] In general, and Not equal, resulting in and are not equal, so when calculating, we should base and The number of Eliminate so that and This formula combines the effect of the number of layers and the sequence number of the stratigraphic boundary points. The smaller the value, the better the stratification effect and the closer it is to the manual stratification result. Figure 7 . The stratification point number corresponding to the minimum is the optimal natural gamma stratification , the selected window size is the optimal window size The wavelet transform modulus maximum SP stratification results are shown in Figure 4 .

[0113] The loss function curves of different window SP layers are shown in Fig. 9 ; The stratification results of SP curves with different window sizes are shown in Figure 6 , w=7 in (a), w=15 in (b), and w=25 in (c).

[0114] 3.2 The natural gamma (GR) curve is better at distinguishing the thickness of the first layer, while the natural potential (SP) curve is better at reflecting the effect of the thickness of the second layer. Therefore, repeat the steps of 3.1 for the natural potential curve (SP) to obtain the optimal stratification point number. and the optimal window size , wavelet transform modulus maximum GR stratification results are shown in Figure 3 The loss function curves of different window GR layers are shown in Figure 8 ; The stratification results of GR curves with different window sizes are shown in Figure 5 , w=7 in (a), w=15 in (b), and w=25 in (c).

[0115] S4, post-process the stratification points to obtain the best automatic stratification effect

[0116] Based on the optimal window size and , stratify the non-cored wells according to the process of 3.

[0117] Post-layer processing: and The obtained stratification points are merged. and The difference is no more than The boundary points of the sampling points are overlapped, where The number of sampling points representing the first layer of the area can obtain the best automatic stratification effect using the GR curve and SP curve. The best stratification result of the GR curve is shown in Fig.10 The best stratification results of SP curve are shown in Fig.11 .

[0118] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0119] To further demonstrate the positive effects of the above embodiment, the present invention conducts the following experiments based on the above technical solution. The final automatic stratification results are compared with the manual stratification results. Fig.12 .

[0120] from Fig.12 It can be seen that the GR curve and SP curve are used to perform the wavelet transform modulus maximum optimal window stratification, and basically all the layers divided by manual stratification are picked up, and only two small layers at 2638m and 2705m are omitted. After manual re-examination, the small layers divided by the manual stratification method miss the two small layers (2665m and 2690m) divided by the method of the present invention. Therefore, the present invention can greatly improve the efficiency of logging curve stratification on the basis of ensuring high accuracy and reliability, and can be used for subsequent sedimentary microfacies division, logging interpretation, etc.

[0121] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any modifications, equivalent substitutions and improvements made by any technician familiar with the technical field within the technical scope disclosed by the present invention and within the spirit and principles of the present invention should be covered within the protection scope of the present invention.

Claims

1. A well logging curve stratification prediction method based on optimal window wavelet transform, characterized in that: The method comprises the steps of: S1, preprocessing the logging curves of cored wells and non-cored wells; S2, performing manual stratification and automatic stratification parameter optimization on the coring well; S3, stratification of non-cored wells based on optimal window size; S4, post-processing the stratification points to obtain the automatic stratification effect; In step S2, automatic stratification parameter optimization is performed on the coring well, including: The natural gamma logging curve is subjected to wavelet transform, and the expression is: ； In the formula, is the calculated wavelet coefficient, is the scale factor, and the binary wavelet transform is selected. , is the scale coefficient; is a time shift, representing the sample points of the well logging curve at depth; is the real number domain of the integral space, is the natural gamma logging curve, is the wavelet function; At a given window size Next, calculate the modulus maximum , and draw the modulus maximum curve; ； In the formula, To take the absolute value operation, For the window Find the maximum value inside; Given a scale factor , calculate the Lipschitz coefficient , the expression is: ； In the formula, is the modulus maximum position, , is the wavelet transform of adjacent scales; Find the mutation point through the wavelet transform modulus maximum value and determine the stratum interface; Step 3.1, perform continuous wavelet transform on the natural gamma logging curve, calculate the wavelet transform modulus maximum, and set the window size range to ; Calculate the wavelet transform modulus maximum curve according to different window sizes, and calculate the Lipschitz coefficient, through the preset threshold Determine the stratification point, stratify the logging curve, and record the number of stratification layers and layer point number ,in, is the window size sequence number, is the stratification point number, and the value of the loss function is calculated as follows: ； In the formula, For the window is The loss value when is the number of artificial stratification layers, is the manual stratification point number, according to and The number of Eliminate so that and The number is equal; When it is the smallest, the corresponding stratification point number is the optimal natural gamma stratification , the selected window size is the optimal window size .

2. The logging curve layered prediction method based on optimal window wavelet transform according to claim 1 is characterized in that: In step S1, the logging curves include natural gamma logging curves and natural potential logging curves.

3. The logging curve layered prediction method based on optimal window wavelet transform according to claim 1 is characterized in that: In step S1, the well logging curves of the coring well and the non-coring well are preprocessed, including: using an adaptive morphological filtering method based on sample entropy to remove small changes in the well logging curves, and the operations of the morphological filtering are expansion and erosion; The expansion operation is: ； The corrosion operation is: ； In the formula, For expansion operation, For corrosion operations, is the logging curve value, is defined in The length is A discrete sequence of is a structural element, defined in The length is a discrete sequence of ; and ; , are all discrete sequence numbers, To obtain the maximum value operation; The Gaussian distribution is used to establish the structural element. The length of the structural element is the number of sampling points corresponding to the thickness of the thin layer to be identified. The standard deviation determines the maximum value of the Gaussian distribution. The larger the standard deviation, the smaller the maximum value of the Gaussian distribution. The sample entropy is calculated and limited to The reciprocal is taken to get the standard deviation, and the structural element is designed according to the standard deviation; the expression of the unnormalized structural element is: ； In the formula, is the standard deviation, For exponential calculation operations; Normalized Structural Element for: ； Standard Deviation for: ； In the formula, is the sample entropy of the well logging curve value in the selected window; Calculate the normalized structural element Finally, the method of first dilation and then corrosion or first corrosion and then dilation and then averaging is used to remove the interference of small jagged changes in the logging curve.

4. The method for layered prediction of well logging curves based on optimal window wavelet transform according to claim 1, characterized in that: In step S2, the coring well is manually stratified, including: Based on the processed natural gamma-ray logging curves and core characteristics, including lithology, composition, combination law, color, structure, thickness and rhythmic characteristics, the high-resolution sequence stratigraphy principle is applied, combined with the logging phase characteristics, to divide the single well sedimentary microfacies within the isochronous stratigraphic framework; record the number of artificial stratification layers and layer point number , is the layer number, which is a known constraint for automatic layering parameter optimization.

5. The logging curve layered prediction method based on optimal window wavelet transform according to claim 1 is characterized in that: In step S2, repeat step 3.1 for the natural potential logging curve to obtain the optimal layer point number and the optimal window size .

6. The method for layered prediction of well logging curves based on optimal window wavelet transform according to claim 5, characterized in that: In step S3, the non-coring wells are stratified, including: and , and the non-coring wells were stratified separately.

7. The method for layered prediction of well logging curves based on optimal window wavelet transform according to claim 5, characterized in that: In step S4, the delamination points are post-processed, including: According to and The obtained stratification points are merged. and The difference is no more than The dividing point of the sampling points is merged, where The number of sampling points representing the first layer of the area is used to obtain the automatic stratification effect using the natural gamma log curve and the natural potential log curve.

8. A logging curve stratification prediction system based on optimal window wavelet transform, characterized in that: The system is used to control the logging curve layered prediction method based on optimal window wavelet transform as described in any one of claims 1 to 7, and the system includes: Well logging curve preprocessing module, used for preprocessing natural gamma logging curves and natural potential logging curves of coring wells and non-coring wells; The coring well artificial stratification module is used to perform artificial stratification of natural gamma logging curves and sedimentary microfacies analysis based on processed natural gamma logging curves and core characteristics; The automatic stratification parameter optimization module for coring wells is used to perform wavelet transform on the natural gamma logging curve and natural potential logging curve to obtain the optimal window size. and ; Non-cored well stratification module for optimizing window size and stratify non-cored wells; The layered post-processing module is used to post-process the layered points to obtain the layered effects of the natural gamma logging curve and the natural potential logging curve.

Images