System for predicting pore pressure and rock mechanical parameters based on real-time drilling parameters
By designing a system containing multiple pre-stored models and real-time data processing units, the problem of real-time prediction of formation pore pressure and rock mechanical parameters in the prior art is solved, and the flexibility and adjustment capabilities of the system are realized, and the demand for real-time continuous and accurate prediction is met.
Patent Information
- Application Number
- CN202311810242.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-12-26
- Publication Date
- 2025-06-27
AI Technical Summary
The existing logging parameter analysis system is difficult to achieve real-time prediction of formation pore pressure and rock mechanical parameters, and lacks flexibility and adjustment capabilities, which cannot meet the continuous accurate prediction requirements during real-time drilling.
A system is designed to predict pore pressure and rock mechanical parameters based on real-time drilling parameters, including a data transmission unit, a real-time drilling unit, a data preprocessing unit, a key information extraction unit, an empirical formula processing unit, a model training unit and a data prediction unit. Through the training and fine-tuning of multiple pre-stored models, adaptive prediction is performed in combination with real-time data.
Real-time and offline prediction of formation pore pressure and rock mechanical parameters is achieved, the flexibility and adjustment ability of the system are improved, and the geological characteristics and new data of different blocks are adapted to meet the needs of real-time continuous and accurate prediction.
Smart Images
Figure CN120211745A_ABST
Abstract
Description
Background Art
[0002] As the oil and gas industry gradually enters the development stage of digitalization and intelligentization, a large amount of geological engineering data has been accumulated. These big data resources provide valuable support for the integration of geology, engineering, and information technology in aspects such as drilling, directional wells, mud, and cementing. However, despite having rich data assets, the industry still faces the lack of professional guidance technology, the in-depth analysis and utilization of big data have not been fully realized, and the true potential of digitalization and intelligentization has not been fully explored.
[0003] In the specific application of predicting formation pore pressure and rock mechanics parameters, the current prediction methods mainly rely on seismic data interpretation, calculation of the drilling DC index while drilling, and well logging interpretation after drilling. The acquisition of rock mechanics parameters mainly depends on laboratory tests of cores after drilling and well logging interpretation after drilling. These traditional methods not only have limited application ranges but also have difficulties in achieving ideal accuracy. In addition, the calculation process is complex, the cost is high, the operation is cumbersome, and most of them can only be applied to post-drilling analysis, unable to meet the requirements of real-time continuous and accurate prediction.
[0004] In addition, existing systems often have difficulty adapting to the geological characteristics of different blocks and the continuous increase of new data, lack the flexibility and adjustment ability of the model, and cannot respond in a timely manner and meet the real-time prediction requirements of formation pore pressure and rock mechanics parameters in real-time drilling parameters. Summary of the Invention
[0001] To solve the above problems in the prior art, that is, the problem that the existing well logging parameter analysis system cannot respond in a timely manner and meet the real-time prediction requirements of formation pore pressure and rock mechanics parameters during real-time drilling, the present invention provides a system for predicting pore pressure and rock mechanics parameters based on real-time drilling parameters. The system includes:
[0002] A data transmission unit configured to receive historical well logging data uploaded by a user;
[0003] A real-time drilling unit configured to perform real-time drilling and obtain real-time drilling parameters;
[0004] A data preprocessing unit configured to unify the formats, standardize, and preprocess the well logging data and real-time drilling parameters to obtain effective historical data and effective real-time parameters;
[0005] A key information extraction unit configured to obtain key information of a set type based on the effective historical data;
[0006] An empirical formula processing unit configured to calculate based on the key information according to a set empirical formula to obtain an empirical value of pore pressure and an empirical value of rock mechanics parameters, and use them as standard values;
[0007] A model training unit, configured to pre-store a plurality of pore pressure prediction models and a plurality of rock mechanics parameter prediction models; the pore pressure prediction models and the plurality of rock mechanics parameter prediction models are trained based on the standard values, and support adding new prediction models, and fine-tuning the parameters of the pre-stored models according to the pore pressure prediction values and the rock mechanics parameter prediction values;
[0008] A data prediction unit, configured to select the pre-stored pore pressure prediction models and rock mechanics parameter prediction models based on the real-time drilling parameters for prediction, and obtain the pore pressure prediction values and the rock mechanics parameter prediction values.
[0009] In some preferred embodiments, the preprocessing includes filtering, denoising, outlier detection and data repair.
[0010] In some preferred embodiments, the system is designed based on the streamlit framework and includes an application presentation layer and a logic control layer.
[0011] In some preferred embodiments, the application presentation layer is used to upload data, download data, select prediction models, adjust prediction models, display real-time prediction results and browse historical prediction results.
[0012] In some preferred embodiments, the logic control layer is used to collect, clean and store data, and build a pore pressure prediction model and a rock mechanics parameter prediction model; the pore pressure prediction model and the rock mechanics parameter prediction model include a random forest model, an artificial neural network and a support vector machine.
[0013] In some preferred embodiments, the key information includes the real-time drilling parameters such as the hook load, the weight on bit, the torque, the rotary speed, the rate of penetration, the pump pressure, the displacement, and the density and viscosity of the drilling fluid.
[0014] In some preferred embodiments, the system further includes a data visualization unit, configured to fit and display trend lines for the key information, the pore pressure prediction values and the rock mechanics parameter prediction values.
[0015] In some preferred embodiments, the empirical formula includes the Eaton formula.
[0016] In some preferred embodiments, the selection of the pre-stored pore pressure prediction models and rock mechanics parameter prediction models specifically includes:
[0017] Dividing the real-time drilling parameters into continuous windows according to the well depth, and each window covers a set drilling length;
[0018] Calculating the empirical cumulative distribution function (ECDF) of the selected real-time drilling parameters within each window;
[0019] Compare the empirical cumulative distribution functions (ECDFs) of two adjacent windows to obtain the cumulative distributions of the adjacent windows:
[0020]
[0021]
[0022] where I(x i ≤ x) represents the indicator function, and the indicator function is 1 when the data point x i in window 1 is less than or equal to the evaluation point x of the ECDF, and F 1,n (x) represents the ECDF cumulative distribution of the adjacent window 1, F 2,m (x) represents the ECDF cumulative distribution of the adjacent window 2, n represents the total number of parameters of the adjacent window 1, and m represents the total number of parameters of the adjacent window 2;
[0023] Calculate the maximum difference between the ECDFs, i.e., the Kolmogorov-Smirnov (KS) statistic D n,m :
[0024]
[0025] represents taking the maximum value over all possible x values;
[0026] Compare the KS statistic with a pre-set critical value; if it is greater than the critical value, it indicates that a concept drift has occurred in the drilling parameters;
[0027] When a concept drift is detected, by comparing the similarity between the data distribution of the real-time drilling parameters at the current depth and the optimal data distributions of the pre-stored pore pressure prediction model and the rock mechanics parameter prediction model, select the pore pressure prediction model and the rock mechanics parameter prediction model with the highest similarity for prediction.
[0029] Advantages of the present invention:
[0030] (1) The prediction model of the present invention does not need to rely on a separate logging tool to obtain logging data, and can directly use real-time conventional drilling data to quickly complete the prediction of rock mechanics parameters and pore pressure.
[0031] (2) By pre-storing multiple prediction models, the present invention can adaptively select the best prediction model according to real-time data and support fine-tuning of the model, thereby improving the adaptability to geological characteristics of different blocks and enhancing the flexibility and adjustment ability of the model.
[0032] (3) The system proposed by the present invention not only integrates the functions of well logging data processing and interpretation, but also can flexibly fine-tune and create new models. At the same time, it has the dual capabilities of real-time and offline prediction of rock mechanical parameters and pore pressure based on drilling parameters, and is equipped with a database query function, thus providing users with a comprehensive integrated solution. Description of the Drawings
[0033] Other features, objects, and advantages of the present application will become more apparent by reading the detailed description of the non-limiting embodiments with reference to the following drawings:
[0034] Figure 1 It is a structural block diagram of the system for predicting well logging parameters based on real-time drilling parameters in an embodiment of the present invention;
[0035] Figure 2 It is a schematic diagram of the architecture of the system for predicting well logging parameters based on real-time drilling parameters in an embodiment of the present invention. Detailed Embodiments
[0036] The following further details the present application with reference to the drawings and embodiments. It can be understood that the specific embodiments described herein are only used to explain the related invention and are not intended to limit the invention. Additionally, it should be noted that for ease of description, only parts related to the relevant invention are shown in the drawings.
[0037] It should be noted that, without conflict, the embodiments in the present application and the features in the embodiments can be combined with each other. The present application will be described in detail below with reference to the drawings and embodiments.
[0038] For a clearer description of the system of the present invention for predicting well logging parameters based on real-time drilling parameters, the following combines Figure 1 and Figure 2 to detail each functional module in the embodiments of the present invention.
[0039] The system for predicting well logging parameters based on real-time drilling parameters in the first embodiment of the present invention includes a step data transmission unit, a data preprocessing unit, a key information extraction unit, an empirical formula processing unit, a data prediction unit, and a model adjustment unit. The detailed description of each functional module is as follows:
[0040] The system is designed based on the streamlit framework. As Figure 2 shown, it includes an application presentation layer and a logic control layer.
[0041] The application presentation layer is used to upload data, download data, select a prediction model, adjust the prediction model, display real-time prediction results, and view historical prediction results.
[0042] At the application presentation layer, as the user interface of the system, it brings a convenient and friendly operation experience to users. With the powerful data visualization function of Streamlit, this presentation layer can display drilling data and prediction results in real time, thus transmitting important real-time information to the drilling site personnel.
[0043] The logic control layer is used to collect, clean, and store data, and build a pore pressure prediction model and a rock mechanics parameter prediction model; the pore pressure prediction model and the rock mechanics parameter prediction model include a random forest model, an artificial neural network, and a support vector machine.
[0044] The logic control layer is the core part of the system, undertaking all background operations, closely integrating with the application presentation layer to ensure the unobstructed transmission of data and instructions between the two layers. This control layer also has the ability to access external databases, facilitating the acquisition and storage of a large amount of drilling data and prediction models. To ensure the prediction quality and speed, many modern machine learning and statistical techniques have been introduced into this layer, and it has undergone multiple optimization iterations.
[0045] The system includes:
[0046] A data transmission unit configured to receive historical logging data uploaded by users;
[0047] In this embodiment, as Figure 1 shown, the database module can regularly read drilling parameters through the data center and the API interface module, and can also read historical data in EXCEL through the offline data interface.
[0048] A real-time drilling unit configured to perform real-time drilling and obtain real-time drilling parameters;
[0049] The database module supports data storage, transfer, and query.
[0050] A data preprocessing unit configured to unify the formats, standardize, and preprocess the logging data and real-time drilling parameters to obtain effective historical data and effective real-time parameters;
[0051] In this embodiment, the preprocessing includes filtering, denoising, outlier detection, and data repair.
[0052] A key information extraction unit configured to obtain key information of a set type based on the effective historical data;
[0053] In this embodiment, the key information includes real-time drilling parameters such as the hook load, weight on bit, torque, rotary speed, rate of penetration, pump pressure, displacement, and drilling fluid density and viscosity.
[0054] An empirical formula processing unit, configured to calculate based on the key information according to a set empirical formula to obtain an empirical value of pore pressure and an empirical value of rock mechanical parameters;
[0055] The database module provides data support for the prediction module, and the model fine-tuning module provides a prediction model for the prediction module; the data prediction unit performs rock physical parameters and pore pressure based on the collected data and stores them in the database module
[0056] A model training unit, configured to pre-store multiple pore pressure prediction models and multiple rock mechanical parameter prediction models; the pore pressure prediction models and multiple rock mechanical parameter prediction models are trained based on the standard values, and support adding new prediction models, and fine-tuning the parameters of the pre-stored models according to the pore pressure prediction values and rock mechanical parameter prediction values;
[0057] A data prediction unit, configured to select the pre-stored pore pressure prediction model and rock mechanical parameter prediction model based on the real-time drilling parameters for prediction to obtain a pore pressure prediction value and a rock mechanical parameter prediction value.
[0058] During the prediction process, it also supports the function of adaptively selecting the best prediction model for real-time data, and applies the Kolmogorov-Smirnov Test (KS Test) algorithm in the statistical method for detecting concept drift.
[0059] The real-time drilling parameters are divided into continuous windows according to the well depth, and each window covers a set drilling length; within each window, the empirical cumulative distribution function (ECDF) of the selected drilling parameters will be calculated to reflect its distribution characteristics within this drilling section.
[0060] Each window covers data with a drilling length of 100 meters, so as to ensure the continuity and representativeness of the data;
[0062] Calculate the cumulative distribution ECDF of the selected real-time drilling parameters within each window;
[0063] Compare the cumulative distribution ECDF of two adjacent windows to obtain the cumulative distribution of adjacent windows:
[0064]
[0065]
[0066] where, I(x i ≤x) represents the indicator function, and the indicator function represents that the data point x i in window 1 is 1 when it is less than or equal to the evaluation point x of the ECDF, F 1,n(x) represents the ECDF cumulative distribution of the adjacent window 1, F 2,m (x) represents the ECDF cumulative distribution of the adjacent window 2, n represents the total number of parameters of the adjacent window 1, and m represents the total number of parameters of the adjacent window 2;
[0067] Calculate the maximum difference between the ECDFs based on the cumulative distributions of the adjacent windows, i.e., the KS statistic D n,m :
[0068]
[0069] denotes taking the maximum value over all possible x values;
[0070] Compare the KS statistic with a pre-set critical value; if it is greater than the critical value, it indicates that a concept drift has occurred in the drilling parameters; in this embodiment, the critical value is selected as 0.05, and being greater than the critical value indicates that there are significant differences in the distributions of the drilling parameters in the two windows;
[0071] When a concept drift is detected, by comparing the similarity between the data distribution of the real-time drilling parameters at the current depth and the optimal data distributions of the pre-stored pore pressure prediction model and the rock mechanics parameter prediction model, select the pore pressure prediction model and the rock mechanics parameter prediction model with the highest similarity for prediction.
[0072] When a concept drift is detected, the system will adaptively select or adjust the prediction model to better adapt to the new data distribution and ensure the accuracy and timeliness of the prediction.
[0073] Divide the drilling parameter data into continuous windows according to the well depth sequence;
[0074] The data prediction in this embodiment includes a real-time prediction link and a historical prediction link. The real-time prediction link is to make predictions based on the real-time acquired data, and can also make predictions through the historically collected data to verify the prediction results of the model.
[0075] In this embodiment, the user can select a suitable model or make appropriate fine-tuning of the existing model according to the operating block where they are located and the actual requirements to further ensure the accuracy of the prediction results.
[0076] Furthermore, as the real-time prediction link progresses and new drilling data is continuously input while drilling, the system will instantaneously refresh the prediction results. This real-time feedback is very helpful to the staff on the drilling site, assisting them to make accurate decisions quickly.
[0077] In this embodiment, the system further includes a data visualization unit for fitting and displaying trend lines for key information, pore pressure prediction values, and rock mechanics parameter prediction values.
[0078] The system proposed by the present invention not only integrates the functions of well logging data processing and interpretation, but also can flexibly fine-tune and create new models. More importantly, it can achieve dual real-time and offline prediction of rock mechanics parameters and pore pressure, and is equipped with a database query function, thus providing users with a comprehensive integrated solution. The fully functional database query module of this embodiment enables users to conveniently query, compare and deeply analyze various drilling data and prediction results, thereby providing solid data support for the subsequent plans and decisions of drilling.
[0079] Those skilled in the art can clearly understand that for the convenience and conciseness of description, the specific working processes and related descriptions of the above-described storage device and processing device can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.
[0080] Those skilled in the art should be able to realize that the modules and method steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. The programs corresponding to the software modules and method steps can be placed in a random access memory (RAM), memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art. To clearly illustrate the interchangeability of electronic hardware and software, the components and steps of each example have been generally described according to their functions in the above description. Whether these functions are executed in the form of 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.
[0081] Terms such as "first", "second", etc. are used to distinguish similar objects, rather than to describe or represent a specific order or sequence.
[0082] The term "comprising" or any other similar term is intended to cover non-exclusive inclusion, so that a process, method, article, or device / equipment including a series of elements not only includes those elements, but also includes other elements not explicitly listed, or also includes elements inherent in these processes, methods, articles, or devices / equipment.
[0083] So far, the technical solution of the present invention has been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easily understood by those skilled in the art 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 all fall within the protection scope of the present invention.
Claims
1. A system for predicting pore pressure and rock mechanical parameters based on real-time drilling parameters, characterized in that, The system includes: A data transmission unit configured to receive historical logging data uploaded by a user; A real-time drilling unit configured to drill in real time and obtain real-time drilling parameters; A data preprocessing unit configured to unify the formats, standardize, and preprocess the logging data and real-time drilling parameters to obtain valid historical data and valid real-time parameters; A key information extraction unit configured to obtain key information of a set type based on the valid historical data; An empirical formula processing unit configured to calculate based on the key information according to a set empirical formula to obtain an empirical value of pore pressure and an empirical value of rock mechanics parameters, and use them as standard values; A model training unit configured to pre-store multiple pore pressure prediction models and multiple rock mechanics parameter prediction models; the pore pressure prediction models and multiple rock mechanics parameter prediction models are trained based on the standard values, support adding new prediction models, and fine-tune the parameters of the pre-stored models according to the pore pressure prediction values and rock mechanics parameter prediction values; A data prediction unit configured to select the pre-stored pore pressure prediction models and rock mechanics parameter prediction models based on the real-time drilling parameters for prediction to obtain pore pressure prediction values and rock mechanics parameter prediction values.
2. The system for predicting pore pressure and rock mechanical parameters based on real-time drilling parameters according to claim 1, wherein The preprocessing includes filtering, denoising, outlier detection, and data repair.
3. The system for predicting pore pressure and rock mechanical parameters based on real-time drilling parameters according to claim 1, wherein The system is designed based on the streamlit framework and includes an application presentation layer and a logic control layer.
4. The system for predicting pore pressure and rock mechanical parameters based on real-time drilling parameters according to claim 1, wherein The application presentation layer is used to upload data, download data, select a prediction model, adjust the prediction model, display real-time prediction results, and browse historical prediction results.
5. The system for predicting pore pressure and rock mechanical parameters based on real-time drilling parameters according to claim 1, characterized in that, The logic control layer is used to collect, clean, and store data, and build a pore pressure prediction model and a rock mechanics parameter prediction model; the pore pressure prediction model and the rock mechanics parameter prediction model include a random forest model, an artificial neural network, and a support vector machine.
6. The system for predicting pore pressure and rock mechanical parameters based on real-time drilling parameters according to claim 1, wherein The key information includes real-time drilling parameters such as hook load, weight on bit, torque, rotary speed, rate of penetration, pump pressure, displacement, and drilling fluid density and viscosity.
7. The system for predicting pore pressure and rock mechanical parameters based on real-time drilling parameters according to claim 1, wherein, The system further includes a data visualization unit configured to fit and display trend lines for the key information, pore pressure prediction values, and rock mechanics parameter prediction values.
8. The system for predicting pore pressure and rock mechanical parameters based on real-time drilling parameters according to claim 1, wherein The empirical formula includes the Eaton formula.
9. The system for predicting pore pressure and rock mechanical parameters based on real-time drilling parameters according to claim 1, wherein The selection of the pre-stored pore pressure prediction models and rock mechanics parameter prediction models specifically includes: Dividing the real-time drilling parameters into continuous windows according to well depth, with each window covering a set drilling length; Calculating the empirical cumulative distribution function (ECDF) of the selected real-time drilling parameters within each window; Comparing the ECDFs of two adjacent windows to obtain the cumulative distributions of the adjacent windows: Among them, I(x i ≤ x) represents an indicator function, and the indicator function represents that the data point x i in window 1 is 1 when it is less than or equal to the evaluation point x of the ECDF, and 0 otherwise. F 1,n (x) represents the ECDF cumulative distribution of the adjacent window 1, and F 2,m (x) represents the ECDF cumulative distribution of the adjacent window 2. n represents the total number of parameters of the adjacent window 1, and m represents the total number of parameters of the adjacent window 2; Calculate the maximum difference between the ECDFs based on the cumulative distributions of the adjacent windows, i.e., the KS statistic D n,m : represents taking the maximum value over all possible x values; Comparing the Kolmogorov-Smirnov (KS) statistic with a pre-set critical value; if it is greater than the critical value, it indicates that a concept drift has occurred in the drilling parameters; When a concept drift is detected, by comparing the similarity between the data distribution of the real-time drilling parameters at the current depth and the optimal data distributions of the pre-stored pore pressure prediction models and rock mechanics parameter prediction models, select the pore pressure prediction model and rock mechanics parameter prediction model with the highest similarity for prediction.