Methods, systems, and equipment for real-time updating of lithology modeling during drilling geological steering.

Abstract

The present invention belongs to the field of geological analysis, and specifically relates to a method, system and device for real-time updating of lithology modeling in geological guidance while drilling, aiming to solve the problem of insufficient modeling accuracy of existing lithology model establishment methods. The present invention includes: obtaining an initial lithology model with added lithology interpretation based on a seismic data body, and obtaining an initial well trajectory; obtaining seismic data bodies of different frequency bands by a wavelet transform method, and obtaining first wave impedance data bodies of different frequency bands, matching the positions of the first wave impedance data bodies of different frequency bands with the initial well trajectory, and determining multiple first wave impedance curves; collecting wave impedance curves while drilling during drilling; determining a predicted curve of the wave impedance curve by a random forest model, so that the predicted curve matches the wave impedance curve while drilling; determining a second wave impedance data body by a random forest model corresponding to the current depth, and then determining a predicted lithology model. The present invention effectively improves the modeling accuracy of the lithology model.

Description

Background Art

[0002] In the related art, a method for obtaining a lithology model based on a seismic data volume is to collect seismic data volumes of all depths in advance, so as to learn the relationship between the seismic data volume and different lithologies through a neural network, and then to obtain a lithology model by modeling.

[0003] However, the method of obtaining lithology models based on seismic data volumes does not take into account the thickness of the formations. The different correspondences between formations with different heterogeneities and seismic attributes result in insufficient modeling accuracy. Summary of the invention

[0004] In order to solve the above-mentioned problem in the prior art, that is, the problem of insufficient modeling accuracy of the existing method for obtaining lithology model based on seismic data volume, the present invention provides a method for real-time updating of lithology modeling in geosteering while drilling, the method comprising:

[0005] Acquire an initial lithology model with added lithology interpretation based on the seismic data volume, and acquire an initial well trajectory based on the initial lithology model;

[0006] The seismic data volume is used to obtain a plurality of seismic data volumes of different frequency bands by using a wavelet transform method, and a plurality of first wave impedance data volumes of different frequency bands are obtained by using a wave impedance inversion algorithm;

[0007] Based on the first wave impedance data bodies of the plurality of different frequency bands, determining frequency domain wave impedance curves of the plurality of different frequency bands by matching with the positions of the initial well trajectory;

[0008] Controlling the drilling equipment to perform drilling based on the initial well trajectory;

[0009] Collect the wave impedance curve while drilling during the drilling process;

[0010] Based on frequency domain wave impedance curves of multiple different frequency bands, a prediction curve of the frequency domain wave impedance curves of different frequency bands is determined by a random forest model, so that the prediction curve matches the wave impedance curve while drilling, and a random forest model corresponding to the current depth is obtained;

[0011] Based on the first wave impedance data volumes of the plurality of different frequency bands, a wave impedance prediction data volume is determined by a random forest model corresponding to the current depth to determine a predicted lithology model.

[0012] In some preferred embodiments, before the random forest model corresponding to the current depth determines the wave impedance prediction data volume and then determines the predicted lithology model, the method further includes:

[0013] For the target area, the probability density function of the data points corresponding to different lithological interpretations is obtained through the collected frequency domain wave impedance curves of multiple different frequency bands and the corresponding lithological interpretations;

[0014] The first wave impedance data body based on the multiple different frequency bands is used to determine the wave impedance prediction data body through the random forest model corresponding to the current depth, and the predicted lithology model is determined, including:

[0015] For each lithology, matching the probability density function of the lithology with the first wave impedance data volumes of the multiple different frequency bands is performed, determining the wave impedance prediction data volume through the random forest model corresponding to the current depth, and confirming the matching result of the wave impedance prediction data volume in each lithology probability density function as the predicted lithology model of the lithology;

[0016] The predicted lithology model of all lithologies is determined as the predicted lithology model.

[0017] In some preferred embodiments, the method of acquiring the probability density function of the data points corresponding to the different lithological interpretations through the collected frequency domain wave impedance curves of multiple different frequency bands and the corresponding lithological interpretations includes:

[0018] The lithology interpretation is applied to the Gaussian distribution model to obtain probability density functions of data points corresponding to different lithology interpretations.

[0019] In some preferred embodiments, the method of obtaining a plurality of seismic data volumes of different frequency bands by using a wavelet transform method for the seismic data volume, and obtaining a plurality of first wave impedance data volumes of different frequency bands by using a wave impedance inversion algorithm, comprises:

[0020] The seismic data volume is subjected to wavelet transformation at the integration interval of a plurality of set frequency factors to obtain a plurality of first wave impedance data volumes of different frequency bands, wherein each frequency factor corresponds to a frequency band.

[0021] In some preferred embodiments, the random forest model is trained using the bi-norm as a loss function.

[0022] In some preferred embodiments, the frequency domain wave impedance curves of different frequency bands are determined based on the frequency domain wave impedance curves of multiple different frequency bands by using a random forest model, so that the predicted curve matches the wave impedance curve while drilling, and a random forest model corresponding to the current depth is obtained, including:

[0023] The frequency domain wave impedance curves of the multiple different frequency bands are divided into multiple batches by means of a sliding window, and the multiple batches are input into the random forest model in sequence.

[0024] In some preferred embodiments, sequentially inputting multiple batches into the random forest model comprises inputting multiple batches into the random forest model in order from shallow to deep.

[0025] In another aspect of the present invention, a system for real-time updating of lithology modeling in drilling location guidance is provided, the system comprising:

[0026] A preparation module is used to obtain an initial lithology model with added lithology interpretation based on the seismic data volume, and obtain an initial well trajectory based on the initial lithology model; obtain multiple seismic data volumes of different frequency bands by wavelet transform method from the seismic data volume, and obtain multiple first wave impedance data volumes of different frequency bands by wave impedance inversion algorithm; based on the multiple first wave impedance data volumes of different frequency bands, determine frequency domain wave impedance curves of multiple different frequency bands by matching with the position of the initial well trajectory;

[0027] A collection module, used to control the drilling equipment to drill based on the initial well trajectory; collect the wave impedance curve while drilling during the drilling process;

[0028] The updating module is used to determine the prediction curve of the frequency domain wave impedance curves of different frequency bands through the random forest model based on the frequency domain wave impedance curves of multiple different frequency bands, so that the prediction curve matches the wave impedance curve while drilling, and obtain the random forest model corresponding to the current depth; based on the first wave impedance data body of the multiple different frequency bands, determine the wave impedance prediction data body through the random forest model corresponding to the current depth, and determine the predicted lithology model.

[0029] According to a third aspect of the present invention, an electronic device is provided, comprising:

[0030] at least one processor; and

[0031] a memory communicatively connected to at least one of the processors; wherein,

[0032] The memory stores instructions executable by the processor, and the instructions are used to be executed by the processor to implement the above-mentioned method for real-time updating of lithology modeling in geosteering while drilling.

[0033] In a fourth aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, wherein the computer instructions are used to be executed by the computer to implement the above-mentioned method for real-time updating of lithology modeling in while-drilling geosteering.

[0034] Beneficial effects of the present invention:

[0035] The present invention collects the wave impedance curve while drilling and uses it as a reference to determine the relationship between the pre-collected seismic data and the wave impedance curve at the current depth. Based on the relationship, the pre-measured seismic data is regenerated into an updated wave impedance prediction data body, and then the predicted lithology model is determined. The method determines the relationship between seismic data and lithology at each depth by continuously updating the relationship between the seismic data body and the lithology model, thereby improving the modeling accuracy of the lithology model. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Other features, objects and advantages of the present application will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings:

[0037] Figure 1 It is a flow chart of a method for real-time updating of lithology modeling in geosteering while drilling according to an embodiment of the present invention.

[0038] Figure 2 It is a schematic diagram of the effect of obtaining a plurality of first wave impedance data bodies of different frequency bands in an embodiment of the present invention.

[0039] Figure 3 It is a schematic diagram of the principle of determining the prediction curve of the frequency domain wave impedance curve of different frequency bands by using the random forest model in an embodiment of the present invention.

[0040] Figure 4 is a schematic diagram simulating the simultaneous drilling of three wells in an embodiment of the present invention.

[0041] Figure 5 Schematic diagram of the lithology model predicted in the embodiment of the present invention. DETAILED DESCRIPTION

[0042] The present application will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the relevant invention, rather than to limit the invention. It is also necessary to explain that, for ease of description, only the parts related to the relevant invention are shown in the accompanying drawings.

[0043] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0044] In the related art, there are still three main problems in the method of determining lithology model through seismic data.

[0045] In the related art, the wave impedance parameter prediction model is inverted based on the pre-collected seismic data of multiple wells, thereby determining the lithology model. The current technology re-inverts the wave impedance based on the latest wave impedance logging curve collected and the existing wave impedance curve to achieve the prediction of the geological model. In other words, the wave impedance is recalculated every time new seismic data is collected.

[0046] In view of this, the embodiments of the present disclosure propose a method that can reconfirm the relationship between seismic data and formation lithology at each depth, so as to reduce the problem of insufficient accuracy of lithology modeling caused by differences in the relationship between seismic data and formation lithology at different depths.

[0047] In the first embodiment of the present disclosure, Figure 1 is a flow chart of a method for real-time updating of lithology modeling in geosteering while drilling according to an embodiment of the present invention. Figure 1 The method for real-time updating of lithology modeling in geosteering while drilling includes the following steps:

[0048] Step S11, acquiring an initial lithology model with added lithology interpretation based on the seismic data volume, and acquiring an initial well trajectory based on the initial lithology model;

[0049] The seismic data body usually includes seismic reflection time, seismic reflection waveform, seismic observation data, etc. Seismic data of multiple wells in the target area are collected in advance, and the relationship between the seismic data body and the lithology is learned by the neural network method for the data of the full depth section, so as to determine the initial lithology model. Among them, the neural network method can be, for example, a long short-term memory network (LSTM), a lightweight convolutional neural network (Rock-ShuffleNetV2) or a deep artificial neural network (ANN). The present disclosure does not specifically limit the neural network for generating the initial lithology model. The lithology interpretation can be, for example, mudstone, limestone and sandstone, or other lithology types. The present application does not specifically limit the specific lithology interpretation type. On the basis of clarifying the initial lithology model, the well trajectory can be determined according to any method or rule, and the well trajectory can be generated based on the predicted lithology model obtained in this application, or the initial well trajectory can be adjusted by the same or different rules as those for generating the initial well trajectory.

[0050] Step S12, obtaining multiple seismic data volumes of different frequency bands by wavelet transform method from the seismic data volume, and obtaining multiple first wave impedance data volumes of different frequency bands by wave impedance inversion algorithm;

[0051] Wavelet Transform is a mathematical method used to decompose a signal into components at different scales, which are called wavelet coefficients. Through the wavelet transform method, the seismic data volume of the target area can be decomposed into seismic data volumes of different frequency bands. Different frequency bands characterize the dominant frequency bands of the seismic data volume. In different heterogeneous strata, the dominant frequency bands of the seismic data volume can achieve better characterization effects than other frequency bands. The seismic data volumes of multiple different frequency bands obtained by wavelet transform are subjected to wave impedance inversion, and the first wave impedance data volumes of multiple different frequency bands obtained also correspond to different stratum heterogeneities, so that the frequency domain wave impedance data volumes of different frequency bands can better characterize the lithology under the corresponding stratum heterogeneity. The seismic data volumes of different frequency bands obtained by wavelet transform are shown in Figure 1. Figure 2 shown.

[0052] Step S13, obtaining multiple seismic data volumes of different frequency bands by wavelet transform method from the seismic data volume, and obtaining multiple first wave impedance data volumes of different frequency bands by wave impedance inversion algorithm, wherein the first wave impedance data volume represents the wave impedance data volume obtained by inverting the seismic data volume of a single frequency band;

[0053] The determined frequency domain wave impedance curves of multiple different frequency bands are used as reference materials for subsequent adjustment, and are ready for subsequent matching with the wave impedance curve while drilling. The three-dimensional spatial position of the initial well trajectory has been determined, and the values ​​of multiple first wave impedance data bodies corresponding to the three-dimensional spatial position of the initial well trajectory are extracted to obtain multiple first wave impedance curves.

[0054] Step S14, controlling the drilling equipment to perform drilling based on the initial well trajectory;

[0055] Step S15, collecting a while-drilling wave impedance curve during the drilling process;

[0056] The WWD wave impedance curve is obtained by calculation based on the WWD acoustic wave and WWD density.

[0057] Step S16, based on the multiple first wave impedance curves, determine a prediction curve of the wave impedance curve through a random forest model, so that the prediction curve matches the wave impedance curve while drilling, and obtain a random forest model corresponding to the current depth;

[0058] The processing principle of the random forest algorithm is as follows Figure 3 As shown, in Figure 3 In the present invention, the first wave impedance curves of multiple different frequency bands are input into the random forest model, and the prediction curve of the wave impedance curve can be output.

[0059] Step S17, based on the first wave impedance data bodies of the multiple different frequency bands, determine the wave impedance prediction data body through the random forest model corresponding to the current depth to determine the predicted lithology model.

[0060] As the depth increases, it will be found that the wave impedance curve corresponding to the wave impedance data body collected in advance in the target area has already differed from the wave impedance curve while drilling. Therefore, it can be predicted that the pre-predicted wave impedance curve after the current depth may have a large error. It can be determined that the lithology model determined by the seismic data body in the target area has already had a decrease in accuracy. Therefore, it is necessary to use the wave impedance curve while drilling as a reference so that the wave impedance curve of deeper strata can be predicted using the data that has been collected, that is, the prediction curve corresponding to the second wave impedance data body. Through the random forest model, the relationship between the first wave impedance curves of multiple different frequency bands and the real wave impedance curve while drilling can be determined. Although the relationship between seismic data and lithology will change with different depths, it will not change dramatically, but will change slowly and smoothly. Therefore, it can be determined that the relationship between seismic data and lithology can predict a reliable wave impedance prediction data body within a distance after the current depth. Therefore, based on multiple first wave impedance data bodies, the accurate second wave impedance data body within a distance after the current depth can be determined through this relationship, that is, the lithology can be determined through the second wave impedance data body.

[0061] The Random Forest algorithm is an ensemble learning method that builds multiple decision trees for classification or regression prediction.

[0062] Through the above-mentioned embodiment, the wave impedance curve while drilling is collected during the drilling process, and it is used as a reference to determine the relationship between the pre-collected seismic data and the wave impedance curve while drilling at the current depth. Based on the relationship, the pre-recorded seismic data is regenerated into an updated wave impedance prediction data body, and then the predicted lithology model is determined. This method determines the relationship between seismic data and lithology at each depth by continuously updating the relationship between the seismic data body and the lithology model, thereby improving the modeling accuracy of the lithology model.

[0063] We simulated the simultaneous drilling of three wells. Before the well trajectory landed, the lithology distribution data of the first 200 meters of the drill bit could be obtained through the actual drilling data. Figure 4 This process is illustrated, showing the trend of the loss function, the accuracy of the test dataset, and the accuracy of the validation dataset over three rounds of model updates. In the first two model updates, the correlation with the validation dataset reached 85%, and the third model update further improved the correlation to over 90%. This result shows that the more actual drilling data is available, the better the prediction performance of the model. It demonstrates the feasibility of the neural network model in actual drilling applications.

[0064] like Figure 5 As shown, a model update is performed every 50 meters of depth. In the first four model updates, the dynamic model can actively predict the inner layer between 2600 meters and 3000 meters of depth. These predictions can capture the changes in impedance values ​​of more than 1000. For the layer where the impedance value drops sharply at a depth of about 3000 meters, the fifth model update can intuitively reflect this change, which is about 200 meters deep from the interface. Through the above analysis, it is obvious that this model update workflow can provide reliable lithological interpretation for the area below 200 meters from the drill bit.

[0065] In the embodiments of the present disclosure, in the formation lithology model, there may be a situation where a certain area belongs to multiple lithology types, and there are also situations where the probabilities of belonging to the multiple lithology types are different. In order to provide the analyzability of the predicted lithology model, the present disclosure proposes a method for generating lithology models for different lithologies separately. The second embodiment of the present disclosure further illustrates the method for generating lithology models for different lithologies separately in the present disclosure.

[0066] In the second embodiment of the present disclosure, the method for generating lithology models for different lithologies separately includes the following steps:

[0067] For the target area, the probability density function of the data points corresponding to different lithological interpretations is obtained through the collected frequency domain wave impedance curves of multiple different frequency bands and the corresponding lithological interpretations;

[0068] The first wave impedance data body based on the plurality of different frequency bands is used to determine the wave impedance prediction data body through the random forest model corresponding to the current depth, and the predicted lithology model is determined, comprising:

[0069] The first wave impedance data body based on the multiple different frequency bands is used to determine the second wave impedance prediction data body through the random forest model corresponding to the current depth; for each lithology, the probability density function of the lithology is matched with the data points in the second wave impedance data body to confirm the predicted lithology model of the lithology;

[0070] The predicted lithology model of all lithologies is determined as the predicted lithology model.

[0071] Through the above embodiment, the probability density distribution of data points of different lithologies in the target area is determined by using the measured seismic data of the target area, and the wave impedance data bodies of different lithologies can be distinguished in the process of predicting the implicit model, and a separate predicted lithology model can be obtained for each lithology. By observing the predicted lithology model of each lithology individually or in combination, it is beneficial to determine the well trajectory and analyze the formation, and provides a data basis for geological research.

[0072] The method of acquiring the probability density functions of the data points corresponding to the different lithological interpretations through the collected frequency domain wave impedance curves of multiple different frequency bands and the corresponding lithological interpretations includes:

[0073] The lithology interpretation is applied to the Gaussian distribution model to obtain probability density functions of data points corresponding to different lithology interpretations.

[0074] It is understandable that by processing seismic signals through wavelet transform, signals of multiple frequency bands are extracted to analyze different frequency components in seismic data. Since the original seismic signal is coupled and superimposed by multiple frequency components, it cannot correspond to a specific tuned formation thickness. Sometimes thin formations cannot be characterized by low-frequency signals, and thick formations cannot be characterized by high-frequency signals. In order to take into account the seismic response characteristics of different formations, this application uses wavelet transform to obtain seismic waveforms of different frequencies for characterizing formations of different thicknesses. The third embodiment of the present disclosure further illustrates the method of wavelet transform in the present disclosure:

[0075] The method of obtaining a plurality of seismic data volumes of different frequency bands by using a wavelet transform method for the seismic data volume, and obtaining a plurality of first wave impedance data volumes of different frequency bands by using a wave impedance inversion algorithm, comprises:

[0076] The seismic data volume is subjected to wavelet transformation at the integration interval of a plurality of set frequency factors to obtain a plurality of first wave impedance data volumes of different frequency bands, wherein each frequency factor corresponds to a frequency band.

[0077] The input data is For example, input data correspond ,in Corresponding to 8 different frequency bands, t represents time. The 8 frequency bands are divided according to experience, and the range of each frequency band is predetermined. The seismic data volume is decomposed into 8 different frequency bands using continuous wavelet transform, 8 seismic data volumes of different frequency bands. The formula of continuous wavelet transform is defined in the time domain as follows:

[0078] ;

[0079] in, is the translation factor, Used to determine the spatial or temporal information of wavelet transform, represents the mother wavelet function, and is obtained by transforming the mother wavelet function Perform scaling and translation of the frequency factor a The complex conjugate obtained by translation; the translation factor can also be called position information.

[0080] The formula for the continuous wavelet transform is defined in the time domain, and for each frequency band extracted The subset of wavelet coefficients is reconstructed using inverse wavelet transform, and the signal of each frequency band is obtained as follows:

[0081] ;

[0082] 8 frequency bands of signals can be obtained , Figure 2 Schematic diagram of signals in 8 frequency bands in an embodiment of the present disclosure.

[0083] For the signals of 8 frequency bands, the seismic data volume of 8 different frequency bands is determined by spectrum decomposition method. , the seismic data of each different frequency band is dimensional data vector By performing wave impedance inversion on seismic data of eight different frequency bands in sequence, eight wave impedance inversion results with geological significance can be obtained.

[0084] Through the above-mentioned embodiments, wavelet transform obtains seismic data volumes of different frequencies, which are used to characterize strata of different thicknesses, and can take into account the seismic response characteristics of different strata, thereby improving prediction accuracy.

[0085] In the embodiment of the present disclosure, the Random Forest model is an integrated learning method, which performs classification by constructing multiple decision trees. In the classification algorithm, the Random Forest usually adopts the loss function of the Gini coefficient or information gain to improve the completion of the classification task. However, in the present application, the Random Forest model is used to generate a second wave impedance data body from multiple first wave impedance data bodies, and in the training process, it is to generate a prediction curve matching the while drilling wave impedance curve based on the first wave impedance curve, so the loss function of the Gini coefficient or information gain cannot be applied. The fourth embodiment of the present disclosure further explains the training method of the Random Forest model in the present disclosure.

[0086] In the fourth embodiment, the random forest model is trained using the bi-norm as the loss function.

[0087] In the disclosed embodiment, the essence of the random forest model in this embodiment is the prediction point. The random forest model inputs 8 values, which are the scatter points of the wave impedance values ​​corresponding to different frequency bands. That is, the input of the random forest model is for the same spatial position, and its state is described by 8 scatter points. The random forest model outputs the predicted value of the wave impedance while drilling at the spatial position. The predicted values ​​of the wave impedance while drilling at multiple positions can form a prediction curve. Every time a prediction curve is generated, the two norms of the prediction curve and the wave impedance while drilling curve are calculated once. When it is lower than the threshold, it is proved that the wave impedance while drilling curve at the current depth can be accurately predicted according to the values ​​of the wave impedance data body corresponding to 8 different frequency bands. That is to say, the random forest model has learned the relationship between the seismic data body and the lithology at the current depth, and can obtain an updated lithology model.

[0088] Through the above embodiment, by using the second norm as the loss function of the random forest model, the random forest model can predict the wave impedance curve while drilling through the seismic data measured in advance. Through this method, the second wave impedance data body can be determined and the predicted lithology model can be generated, which is equivalent to updating the initial lithology model confirmed when designing the initial drilling trajectory, so that the predicted lithology model can match the actual measured wave impedance curve while drilling more closely than the predetermined initial lithology model. That is, within the depth after the current depth, the predicted lithology model can be closer to the true value.

[0089] It is understandable that determining the LWD wave impedance curve of the complete depth segment at one time through the random forest model will introduce interference due to the inconsistent relationship between the seismic data bodies and the wave impedance curve at different depth segments. Therefore, in order to improve the prediction accuracy, the training stage of the embodiment of the present disclosure does not use the method of inputting all seismic data bodies into the random forest model at one time to generate the prediction curve. The fifth embodiment of the present disclosure further explains how the present disclosure inputs frequency domain wave impedance curves based on multiple different frequency bands to generate a high-precision prediction curve.

[0090] In the fifth embodiment, the method of generating a high-precision prediction curve based on frequency-domain wave impedance curves of multiple different frequency bands includes the following steps.

[0091] The frequency domain wave impedance curves of the multiple different frequency bands are divided into multiple batches by means of a sliding window, and the multiple batches are input into the random forest model in sequence.

[0092] The prediction curve of the wave impedance curve is determined by the random forest model so that the prediction curve matches the wave impedance curve while drilling, and the random forest model corresponding to the current depth is obtained.

[0093] The step of sequentially inputting a plurality of batches into the random forest model includes inputting a plurality of batches into the random forest model in an order from shallow to deep.

[0094] It is understandable that in actual application scenarios, the trained random forest model can make a prediction at each depth segment, or at every other depth segment. Therefore, dividing the earthquake zone data into multiple batches through a sliding window and inputting the model from shallow to deep for training can enable the model to generate a smoother prediction curve and retain the characteristics of the corresponding depth.

[0095] Although the various steps in the above embodiment are described in the above-mentioned order, those skilled in the art can understand that in order to achieve the effect of this embodiment, different steps do not have to be executed in such an order. They can be executed simultaneously (in parallel) or in a reverse order. These simple changes are within the scope of protection of the present invention.

[0096] A system for real-time updating of lithology modeling in drilling location guidance according to a second embodiment of the present invention comprises:

[0097] A preparation module is used to obtain an initial lithology model with added lithology interpretation based on the seismic data volume, and obtain an initial well trajectory based on the initial lithology model; obtain multiple seismic data volumes of different frequency bands by wavelet transform method from the seismic data volume, and obtain multiple first wave impedance data volumes of different frequency bands by wave impedance inversion algorithm; based on the multiple first wave impedance data volumes of different frequency bands, determine frequency domain wave impedance curves of multiple different frequency bands by matching with the position of the initial well trajectory;

[0098] A collection module, used to control the drilling equipment to drill based on the initial well trajectory; collect the wave impedance curve while drilling during the drilling process;

[0099] The updating module is used to determine the prediction curve of the wave impedance curve through the random forest model based on the frequency domain wave impedance curves of multiple different frequency bands, so that the prediction curve matches the wave impedance curve while drilling, and obtain the random forest model corresponding to the current depth; based on the first wave impedance data body of the multiple different frequency bands, determine the wave impedance prediction data body through the random forest model corresponding to the current depth, and determine the predicted lithology model.

[0100] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process and related instructions of the system described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0101] It should be noted that the system for real-time updating of lithology modeling in the drilling address guidance provided by the above embodiment is only illustrated by the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the modules or steps in the embodiments of the present invention can be decomposed or combined. For example, the modules in the above embodiment can be combined into one module, or further divided into multiple sub-modules to complete all or part of the functions described above. The names of the modules and steps involved in the embodiments of the present invention are only for distinguishing the modules or steps, and are not regarded as improper limitations of the present invention.

[0102] An electronic device according to a third embodiment of the present invention includes:

[0103] at least one processor; and

[0104] a memory communicatively connected to at least one of the processors; wherein,

[0105] The memory stores instructions executable by the processor, and the instructions are used to be executed by the processor to implement the above-mentioned method for real-time updating of lithology modeling in geosteering while drilling.

[0106] A fourth embodiment of the present invention is a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to be executed by the computer to implement the above-mentioned method for real-time updating of lithology modeling in geosteering while drilling.

[0107] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process and related instructions of the storage device and processing device described above can refer to the corresponding process in the aforementioned method embodiment and will not be repeated here.

[0108] Those skilled in the art should be able to appreciate that the modules and method steps of each example described in conjunction with the embodiments disclosed herein can be implemented with electronic hardware, computer software, or a combination of the two, and the programs corresponding to the software modules and method steps can be placed in random access memory (RAM), memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disks, removable disks, CD-ROMs, or any other form of storage medium known in the technical field. In order to clearly illustrate the interchangeability of electronic hardware and software, the composition and steps of each example have been generally described in the above description according to the function. Whether these functions are performed in electronic hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0109] The terms "first", "second", etc. are used to distinguish similar objects rather than to describe or indicate a particular order or sequence.

[0110] The term "comprise" or any other similar term is intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus / device that includes a list of elements includes not only those elements but also other elements not expressly listed, or also includes elements inherent to such process, method, article, or apparatus / device.

[0111] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

Claims

1. A method for real-time updating of lithology modeling in geosteering while drilling, characterized in that: The method comprises: Acquire an initial lithology model with added lithology interpretation based on the seismic data volume, and acquire an initial well trajectory based on the initial lithology model; The seismic data volume is used to obtain multiple seismic data volumes of different frequency bands by wavelet transform method, and multiple first wave impedance data volumes of different frequency bands are obtained by wave impedance inversion algorithm, wherein the first wave impedance data volume represents the wave impedance data volume obtained by inverting the seismic data volume of a single frequency band; Based on the first wave impedance data volumes of the plurality of different frequency bands, a plurality of first wave impedance curves are determined by matching with the positions of the initial well trajectory; Controlling the drilling equipment to perform drilling based on the initial well trajectory; Collect the wave impedance curve while drilling during the drilling process; Based on the first wave impedance curves of multiple different frequency bands, a prediction curve of the frequency domain wave impedance curves of different frequency bands is determined by a random forest model, so that the prediction curve matches the wave impedance curve while drilling, and a random forest model corresponding to the current depth is obtained; Based on the first wave impedance data volumes of the multiple different frequency bands, a second wave impedance data volume is determined by the random forest model corresponding to the current depth, and then the predicted lithology model is determined, wherein the second wave impedance data volume is a wave impedance data volume after the current depth obtained by nonlinearly weighting the multiple first wave impedance data volumes through the random forest model.

2. The method for real-time updating of lithology modeling in geosteering while drilling according to claim 1, characterized in that: Based on the first wave impedance curves of multiple different frequency bands, a prediction curve of the frequency domain wave impedance curves of different frequency bands is determined by a random forest model so that the prediction curve matches the wave impedance curve while drilling, and before obtaining the random forest model corresponding to the current depth, the method further includes: For the target area, the probability density function of the data points corresponding to different lithological interpretations is obtained through the collected frequency domain wave impedance curves of multiple different frequency bands and the corresponding lithological interpretations; The first wave impedance data volume based on the multiple different frequency bands is used to determine the second wave impedance prediction data volume through the random forest model corresponding to the current depth, and then determine the predicted lithology model, including: Based on the first wave impedance data bodies of the multiple different frequency bands, a second wave impedance prediction data body is determined by a random forest model corresponding to the current depth; for each lithology, a probability density function of the lithology is matched with data points in the second wave impedance data body to confirm a predicted lithology model for the lithology; The predicted lithology model of all lithologies is determined as the predicted lithology model.

3. The method for real-time updating of lithology modeling in geosteering while drilling according to claim 2, characterized in that: The method of acquiring the probability density functions of the data points corresponding to the different lithological interpretations through the collected frequency domain wave impedance curves of multiple different frequency bands and the corresponding lithological interpretations includes: The lithology interpretation is applied to the Gaussian distribution model to obtain probability density functions of data points corresponding to different lithology interpretations.

4. The method for real-time updating of lithology modeling in geosteering while drilling according to claim 1, characterized in that: The method of obtaining a plurality of seismic data volumes of different frequency bands by using a wavelet transform method for the seismic data volume, and obtaining a plurality of first wave impedance data volumes of different frequency bands by using a wave impedance inversion algorithm, comprises: The seismic data volume is subjected to wavelet transformation at the integration interval of a plurality of set frequency factors to obtain a plurality of first wave impedance data volumes of different frequency bands, wherein each frequency factor corresponds to a frequency band.

5. The method for real-time updating of lithology modeling in geosteering while drilling according to claim 1, characterized in that: The random forest model is trained using the bi-norm as the loss function.

6. The method for real-time updating of lithology modeling in geosteering while drilling according to claim 1, characterized in that: The method of determining a prediction curve of the frequency domain wave impedance curves of different frequency bands by a random forest model based on the frequency domain wave impedance curves of multiple different frequency bands, so that the prediction curve matches the wave impedance curve while drilling, and obtaining a random forest model corresponding to the current depth includes: The frequency domain wave impedance curves of the multiple different frequency bands are divided into multiple batches by means of a sliding window, and the multiple batches are input into the random forest model in sequence.

7. The method for real-time updating of lithology modeling in geosteering while drilling according to claim 6, characterized in that: The step of sequentially inputting a plurality of batches into the random forest model comprises inputting a plurality of batches into the random forest model in an order from shallow to deep.

8. A system for real-time updating of lithology modeling in drilling location guidance, characterized in that: The system comprises: A preparation module is used to obtain an initial lithology model with added lithology interpretation based on the seismic data volume, and obtain an initial well trajectory based on the initial lithology model; obtain multiple seismic data volumes of different frequency bands by wavelet transform method from the seismic data volume, and obtain multiple first wave impedance data volumes of different frequency bands by wave impedance inversion algorithm; based on the multiple first wave impedance data volumes of different frequency bands, determine frequency domain wave impedance curves of multiple different frequency bands by matching with the position of the initial well trajectory; A collection module, used to control the drilling equipment to drill based on the initial well trajectory; collect the wave impedance curve while drilling during the drilling process; The updating module is used to determine the prediction curve of the frequency domain wave impedance curves of different frequency bands through the random forest model based on the frequency domain wave impedance curves of multiple different frequency bands, so that the prediction curve matches the wave impedance curve while drilling, and obtain the random forest model corresponding to the current depth; based on the first wave impedance data body of the multiple different frequency bands, determine the wave impedance prediction data body through the random forest model corresponding to the current depth, and determine the predicted lithology model.

9. An electronic device, characterized in that: include: at least one processor; as well as a memory communicatively connected to at least one of the processors; wherein, The memory stores instructions executable by the processor, wherein the instructions are used to be executed by the processor to implement the method for real-time updating of lithology modeling in geosteering while drilling according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to be executed by the computer to implement the method for real-time updating of lithology modeling in geosteering while drilling according to any one of claims 1 to 7.

Images