A method and system for determining the optimal drilling plan in oil extraction

Abstract

This invention discloses a method and system for determining the optimal drilling plan in oil extraction, including: constructing a 3D tensor of typical well production profiles; summing the 3D tensor multiple times according to the weights corresponding to the current drilling plan to obtain the total production profile corresponding to the current drilling plan tensor; calculating a loss function based on the target production sequence of the oil and gas field within the development period and the total production profile; differentiating the loss function and performing gradient descent calculations on the learnable parameters and the initial drilling plan tensor to be solved, obtaining the drilling plan tensor updated by gradient descent in this round; processing the updated drilling plan tensor to obtain a standardized drilling plan tensor; calculating the difference between the standardized drilling plan tensor at the end of this iteration and the initial drilling plan tensor to be solved; repeating steps S2 to S6, iteratively optimizing until convergence, to obtain the optimal drilling plan. This invention, based on tensor and gradient descent algorithms, can quickly solve for the optimal drilling plan.

Description

Technical Field

[0001] This invention relates to the field of oil extraction technology, and specifically to a method and system for determining the optimal drilling plan for oil extraction. Background Technology

[0002] When developing or adjusting oil and gas field development plans, it is necessary to plan future drilling workloads based on the future total production profile of the target oil and gas field. Currently, the commonly used method is to manually calculate future drilling workloads based on the production decline curve profile of typical wells until the target production profile is approximated. In traditional oil and gas reservoir planning, due to frequent changes in the target total production profile and decline curve of the oil and gas field, drilling plans also need to be frequently revised. This requires manual design of drilling plans through repeated trial calculations, which is extremely labor-intensive and relies heavily on engineers' experience, making it difficult to quickly respond to frequent changes in the plan's content. Summary of the Invention

[0003] To address the above problems, the present invention aims to provide a method and system for determining the optimal drilling plan in oil extraction. This method and system are designed to quickly solve for the optimal drilling plan, and the output drilling plan can achieve the goal of approximating the total production plan of the oil and gas field with the minimum number of wells of various types. This method can replace the traditional method of manually designing drilling plans through repeated trial calculations, does not rely on the experience of engineers, and can stably and quickly approach the optimal solution.

[0004] The first objective of this invention is to provide a method for determining the optimal drilling plan for oil extraction, comprising:

[0005] S1: Construct a 3D tensor for a typical well production profile;

[0006] S2: Summate the 3D tensor in S1 multiple times according to the weights corresponding to the current drilling plan to obtain the total production profile corresponding to the current drilling plan tensor;

[0007] S3: Calculate the loss function based on the target production sequence of the oil and gas field within the development period and the total production profile in S2;

[0008] S4: Take the derivative of the loss function and perform gradient descent calculations on the learnable parameters and the initial drilling plan tensor to be solved, to obtain the drilling plan tensor updated by gradient descent in this round;

[0009] S5: Process the updated drilling plan tensor in S4 to obtain a standardized drilling plan tensor;

[0010] S6: Calculate the difference between the standardized drilling plan tensor at the end of this iteration and the initial drilling plan tensor to be solved in S4;

[0011] S7: Repeat steps S2 to S6, iteratively optimizing until convergence, to obtain the optimal drilling plan.

[0012] This invention, based on tensor and gradient descent algorithms, rapidly solves for the optimal drilling plan given a total number of wells drilled and a planned total production profile for an oil and gas field. The output drilling plan approximates the planned total production of the oil and gas field with the minimum number of wells of each type. This invention integrates and summarizes the drilling plan preparation process into a mathematical problem, designed based on optimization principles. It can replace the traditional manual method of repeatedly calculating and designing drilling plans, without relying on engineers' experience, and can stably and quickly approximate the optimal solution. It can be directly applied to the design of oil and gas reservoir development schemes, and can quickly respond to frequent changes in scheme content.

[0013] In an optional embodiment, the process of constructing a typical well production profile 3D tensor in S1 is as follows: obtaining the drilling years T dl Development years T pl Typical wells of type N, from commissioning to the first Monthly production profile Oil and gas field target production sequence within the development period ;

[0014] Drilling will be completed and in the [missing information] The production output of wells put into production each month is divided into two periods: from month 1 to month 2. Monthly output was 0, the first Month to the first Monthly output was The first item to the second item item;

[0015] Based on the aforementioned classification, the production profiles of all drilling categories are organized into groups of size (N, T). dl T pl 3D tensor .

[0016] In this invention, based on the drilling plan, drilling will be completed and on the [number]th [day / month]... The production output of wells put into production each month is divided into two segments, and the production profiles of all types of wells are denoted as (N, T). dl T pl The 3D tensor of the well plan can be used to transform nonlinear calculations into linear calculations, enabling subsequent gradient descent algorithms and integrating the well planning process into a mathematical problem.

[0017] In an optional embodiment, the process of obtaining the total production profile corresponding to the current drilling plan tensor in S2 is as follows:

[0018] Using Einstein's summation convention, calculate all the th... Class of wells in The total monthly output was , It is a dimensionality reduction function;

[0019] Get in the first The total production of all wells in the month is ;

[0020] Get the first tensor corresponding to the current drilling plan tensor Annual total production profile Y pl The drilling and development period is measured in years.

[0021] The Einstein summation convention notation used in this invention can greatly simplify the formula algorithm, so as to obtain the total production profile corresponding to the current drilling plan tensor in a simple and efficient manner.

[0022] In an optional embodiment, the loss function in S3 .

[0023] In an optional embodiment, the gradient descent calculation process in S4 is as follows: To obtain the drilling plan tensor updated in this round of gradient descent. T dl This refers to the drilling lifespan.

[0024] In an optional embodiment, the normalized drilling plan tensor in S5 is a positive integer and less than or equal to the upper limit of the number of wells drilled. The standardized drilling plan tensor is used as the final drilling plan tensor for this iteration.

[0025] In an optional embodiment, the updated drilling plan tensor is used in S5. The total number of wells in each category is obtained by summing the data along the months of well commissioning. ;

[0026] calculate And use it as the final drilling plan tensor for this iteration;

[0027] To use the rounding function, This is a truncation function that assigns 0 to values ​​less than 0.

[0028] In an optional embodiment, the difference in step S6 ;

[0029] T dl For drilling years, Let be the initial drilling plan tensor to be solved.

[0030] In an optional embodiment, steps S2 to S6 are repeated in S7, iteratively optimizing until... Less than or equal to 0.5, the final result is This is the optimal drilling plan.

[0031] A second objective of this invention is to provide an optimal drilling plan determination system for oil extraction, comprising:

[0032] The production profile 3D tensor construction module is used to construct a typical well production profile 3D tensor.

[0033] The summation module is used to sum the 3D tensor multiple times according to the corresponding weights in the current drilling plan to obtain the total production profile corresponding to the current drilling plan tensor.

[0034] The loss function acquisition module is used to calculate the loss function based on the target production sequence of the oil and gas field within the development period and the total production profile corresponding to the current drilling plan tensor.

[0035] The derivative module is used to calculate the derivative of the loss function;

[0036] The gradient descent calculation module is used to perform gradient descent calculations on the learning parameters and the initial drilling plan tensor to be solved, so as to obtain the drilling plan tensor updated in this round of gradient descent.

[0037] The standardized data processing module is used to standardize the updated drilling plan tensor to obtain a standardized drilling plan tensor, wherein the standardized drilling plan tensor is a positive integer and is less than or equal to the upper limit of the number of wells drilled.

[0038] The difference calculation module is used to calculate the difference between the standardized drilling plan tensor at the end of this iteration and the initial drilling plan tensor to be solved.

[0039] The iteration module is used to perform iterative processing until convergence to obtain the optimal drilling plan.

[0040] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0041] This invention, based on tensor and gradient descent algorithms, rapidly solves for the optimal drilling plan given a total number of wells drilled and a planned total production profile for an oil and gas field. The output drilling plan approximates the planned total production of the oil and gas field with the minimum number of wells of each type. This invention integrates and summarizes the drilling plan preparation process into a mathematical problem, designed based on optimization principles. It can replace the traditional manual method of repeatedly calculating and designing drilling plans, without relying on engineers' experience, and can stably and quickly approximate the optimal solution. It can be directly applied to the design of oil and gas reservoir development schemes, and can quickly respond to frequent changes in scheme content. Attached Figure Description

[0042] To more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of the present invention and should not be considered as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort. In the drawings:

[0043] Figure 1 A flowchart of a method for determining the optimal drilling plan for oil extraction provided in an embodiment of the present invention;

[0044] Figure 2 For the given Typical wells from commissioning to the first Monthly production profile.

[0045] Figure 3 A graph showing the relationship between the total number of wells in the drilling plan tensor and the number of iterations.

[0046] Figure 4 The graph shows the relationship between the loss function MSE and the number of iterations.

[0047] Figure 5 This is a comparison chart of the initial drilling plan tensor and the production curves corresponding to the obtained optimal drilling plan.

[0048] Figure 6 This is a diagram showing the optimal drilling plan results obtained using the method provided in this embodiment of the invention.

[0049] Figure 7 The system schematic diagram provided for an embodiment of the present invention.

[0050] The components and markings in the attached diagram are as follows:

[0051] 101 - Production Profile 3D Tensor Construction Module, 102 - Summation Processing Module, 103 - Loss Function Acquisition Module, 104 - Differentiation Module, 105 - Gradient Descent Calculation Module, 106 - Standardized Data Processing Module, 107 - Difference Calculation Module, 108 - Iteration Module. Detailed Implementation

[0052] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the embodiments and accompanying drawings. The illustrative embodiments and descriptions of the present invention are only used to explain the present invention and are not intended to limit the present invention.

[0053] Example 1:

[0054] Taking the mid-term adjustment of a certain gas field plan as an example, the gas field has entered a stable production period.

[0055] A method for determining the optimal drilling plan for oil extraction includes the following steps:

[0056] 1. Establish production profile tensors for typical wells

[0057] Given drilling years ( Development period () ), and convert them to monthly units, respectively denoted as , Based on the preliminary analysis of declining production, given Typical wells from commissioning to the first Monthly production profile, , indicating the first Type 1 well in its production profile Monthly production profile, such as Figure 1 As shown; given the upper limit of the number of wells to be drilled for each type of well, denoted as... In this case, the maximum number of wells for categories I, II, and III is 40, 80, and 80 respectively, totaling 200 wells. Given the target production sequence of the oil and gas field within the development period. ,like Figure 5 The total planned output curve is shown in the figure.

[0058] For the first Based on the specific month in which it was drilled and put into production, there are a total of Actual production profile of a well: Suppose the time when a well is completed drilling and put into production is the nth day. In the month, the actual output is divided into two segments, from month 1 to month 2. Monthly output was 0, the first Month to the first Monthly output was The first item to the second item Based on the data structure defined above, the production profile of all well categories can be denoted as a sequence of size (N, T). dl T pl 3D tensor , , , , indicating that it belongs to the first Class and in the first The number of wells that were drilled and put into production in the month was [number missing]. Monthly output, such as Figure 5 .

[0059] 2. Calculate the total production profile corresponding to the tensor of the current drilling scheme.

[0060] The 3D tensor in step (1) is summed multiple times according to the weights corresponding to the current drilling plan to obtain the total production profile corresponding to the current drilling plan tensor, specifically:

[0061] Using Einstein's summation convention, then according to the current drilling plan, calculate all the... Class of wells in The total monthly output was . The dimension reduction function can remove one and only one dimension; here, a 2-dimensional vector can be obtained using this function. Clearly, by summing the values ​​for each well category, we can obtain the vector at the 1st... The total monthly production of all wells (including undrilled wells with a production of 0) was: Considering 1D tensors It is the first The total output for the month, therefore summing it once every 12 months yields the result for the month. Total annual output, i.e. This is the total production profile corresponding to the tensor of the current drilling scheme.

[0062] 3. Calculate the loss function

[0063] loss function .

[0064] 4. Gradient Descent Optimization Algorithm

[0065] Taking the derivative of the loss function, i.e.

[0066]

[0067] And the drilling plan tensor to be solved Defined as learnable parameters, This is the initial drilling plan tensor to be solved. Through the tensor calculation design process described above, and with a reasonable data structure, G is calculated in step (2). y The process involves only linear computation, and the calculation is differentiable step-by-step; therefore, gradient descent can be used for optimization. Perform one gradient descent algorithm, and let the drilling plan tensor obtained after this round of gradient descent update be . .

[0068] The specific gradient descent algorithm process is as follows:

[0069]

[0070] 5. Standardized and updated drilling plan tensor

[0071] After step (4) update The total number of wells in each category is obtained by summing the data along the months of well commissioning. .

[0072] calculate And use it as the final drilling plan tensor of this iteration. This is a rounding function that uses rounding to the nearest integer. This is a truncation function that assigns 0 to values ​​less than 0. This refers to the standardized drilling plan tensor, which is a positive integer and does not exceed the upper limit of the number of wells drilled for each type of well. Positive integers.

[0073] 6. Calculate the difference between the standardized drilling plan tensor at the end of this iteration and the initial drilling plan tensor.

[0074] difference .

[0075] Will Set as .

[0076] 7. Iterate and optimize until convergence.

[0077] Repeat steps 2-6 until... This serves as the termination condition for iterative optimization.

[0078] Final optimization This refers to the optimal drilling plan. Specifically, as follows: Figure 6 As shown.

[0079] For this case, a total of 454 iterations were performed. The total number of wells decreased from the planned 200 to 126 (e.g., Figure 3 Not only did the number of wells decrease, but the MSE (Mean Sequence Equivalent) also became closer to the target production curve (e.g., ...). Figure 4 ).

[0080] This invention provides a method for accurately generating an optimized drilling plan that most closely approximates the target total production profile of an oil and gas field. Based on tensor computation, a systematic mathematical model of the oil and gas field drilling plan development process is constructed, and relevant data structures are designed. Under the premise of automatic differentiation at each step of the calculation process, the entire process is reduced to a mathematical optimization problem. Based on this, a solution process for this mathematical optimization problem is designed. Since the entire process involves step-differentiable linear tensor computation, the optimal solution can theoretically be approximated using the gradient descent algorithm. The resulting optimized drilling plan tensor satisfies the requirement of approximating the given target total production profile of the oil and gas field with the fewest total wells. This invention can replace the traditional method of manually calculating and recursively developing drilling plans. It overcomes the disadvantages of manual calculations, such as reliance on experience, high workload, and inability to determine whether the solution is optimal. The drilling plan obtained using this invention can be directly used for the development and adjustment of oil and gas field development plans.

[0081] Example 2:

[0082] Corresponding to the above method embodiments, this invention also provides an optimal drilling plan determination system for oil extraction, which can be referred to in conjunction with the optimal drilling plan determination method for oil extraction described in the above embodiments.

[0083] like Figure 7 As shown, an optimal drilling plan determination system for oil extraction includes:

[0084] Production profile 3D tensor construction module 101 is used to construct a typical well production profile 3D tensor;

[0085] The summation processing module 102 is used to sum the 3D tensor multiple times according to the corresponding weight in the current drilling plan to obtain the total production profile corresponding to the current drilling plan tensor.

[0086] The loss function acquisition module 103 is used to calculate the loss function based on the target production sequence of the oil and gas field within the development period and the total production profile corresponding to the current drilling plan tensor.

[0087] The derivative module 104 is used to calculate the derivative of the loss function;

[0088] The gradient descent calculation module 105 is used to perform gradient descent calculation on the learnable parameters and the initial drilling plan tensor to be solved based on the derivative of the loss function, so as to obtain the drilling plan tensor updated in this round of gradient descent.

[0089] The standardized data processing module 106 is used to standardize the updated drilling plan tensor to obtain a standardized drilling plan tensor, wherein the standardized drilling plan tensor is a positive integer and is less than or equal to the upper limit of the number of wells drilled.

[0090] The difference calculation module 107 is used to calculate the difference between the standardized drilling plan tensor at the end of this iteration and the initial drilling plan tensor to be solved.

[0091] Iteration module 108 is used to perform iterative processing until convergence to obtain the optimal drilling plan.

[0092] Methods, processes, and equipment not mentioned in this patent application are all obtained or performed using existing methods or methods that can be used.

[0093] Those skilled in the art will further recognize that the modules and algorithmic steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in 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 implementations should not be considered beyond the scope of this invention.

[0094] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for determining the optimal drilling plan for oil extraction, characterized in that, include: S1: Construct a 3D tensor for a typical well production profile; S2: Summate the 3D tensor in S1 multiple times according to the weights corresponding to the current drilling plan to obtain the total production profile corresponding to the current drilling plan tensor; S3: Calculate the loss function based on the target production sequence of the oil and gas field within the development period and the total production profile in S2; S4: Take the derivative of the loss function and perform gradient descent calculations on the learnable parameters and the initial drilling plan tensor to be solved, to obtain the drilling plan tensor updated by gradient descent in this round; S5: Process the updated drilling plan tensor in S4 to obtain a standardized drilling plan tensor; S6: Calculate the difference between the standardized drilling plan tensor at the end of this iteration and the initial drilling plan tensor to be solved in S4; S7: Repeat steps S2 to S6, iteratively optimize until convergence, and obtain the optimal drilling plan; The 3D tensor is of size (N, T) dl T pl 3D tensor , , , , indicating that it belongs to the first Class and in the first The number of wells that were drilled and put into production in the month was [number missing]. Monthly output; The process of obtaining the total production profile corresponding to the current drilling plan tensor in S2 is as follows: Using Einstein's summation convention, calculate all the th... Class of wells in The total monthly output was , It is a dimensionality reduction function; Get in the first The total production of all wells in the month is ; Get the first tensor corresponding to the current drilling plan. Annual total production profile Y pl The drilling and development period is measured in years. Loss function in S3 ; The drilling plan tensor to be solved , , Defined as learnable parameters.

2. The method for determining the optimal drilling plan for oil extraction according to claim 1, characterized in that, The process of constructing the 3D tensor of a typical well production profile in S1 is as follows: Obtain drilling years T dl Development years T pl Typical wells of type N, from commissioning to the first Monthly production profile Oil and gas field target production sequence within the development period ; Drilling will be completed and in the [missing information] The production output of wells put into production each month is divided into two periods: from month 1 to month 2. Monthly output was 0, the first Month to the first Monthly output was The first item to the second item item; Based on the aforementioned classification, the production profiles of all drilling categories are organized as (N, T) dl T pl 3D tensor .

3. The method for determining the optimal drilling plan for oil extraction according to claim 1, characterized in that, The gradient descent calculation process in S4 is as follows: To obtain the drilling plan tensor updated in this round of gradient descent. T dl This refers to the drilling lifespan.

4. The method for determining the optimal drilling plan for oil extraction according to claim 3, characterized in that, In S5, the standardized drilling plan tensor is a positive integer and less than or equal to the upper limit of the number of wells drilled. The standardized drilling plan tensor is used as the final drilling plan tensor for this iteration.

5. The method for determining the optimal drilling plan for oil extraction according to claim 4, characterized in that, The drilling plan tensor will be updated in S5. The total number of wells in each category is obtained by summing the data along the months of well commissioning. ; calculate And use it as the final drilling plan tensor for this iteration; To use the rounding function, This is a truncation function that assigns 0 to values ​​less than 0.

6. The method for determining the optimal drilling plan for oil extraction according to claim 5, characterized in that, Differences in step S6 ; T dl For drilling years, Let be the initial drilling plan tensor to be solved.

7. The method for determining the optimal drilling plan for oil extraction according to claim 6, characterized in that, Repeat steps S2 through S6 in S7, iteratively optimizing until... Less than or equal to 0.5, the final result is This is the optimal drilling plan.

8. A system for determining the optimal drilling plan for oil extraction, characterized in that, include: The production profile 3D tensor construction module is used to construct a typical well production profile 3D tensor. The summation module is used to sum the 3D tensor multiple times according to the corresponding weights in the current drilling plan to obtain the total production profile corresponding to the current drilling plan tensor. The loss function acquisition module is used to calculate the loss function based on the target production sequence of the oil and gas field within the development period and the total production profile corresponding to the current drilling plan tensor. The derivative module is used to calculate the derivative of the loss function; The gradient descent calculation module is used to perform gradient descent calculations on the learning parameters and the initial drilling plan tensor to be solved, so as to obtain the drilling plan tensor updated in this round of gradient descent. The standardized data processing module is used to standardize the updated drilling plan tensor to obtain a standardized drilling plan tensor, wherein the standardized drilling plan tensor is a positive integer and is less than or equal to the upper limit of the number of wells drilled. The difference calculation module is used to calculate the difference between the standardized drilling plan tensor at the end of this iteration and the initial drilling plan tensor to be solved. The iteration module is used to perform iterative processing until convergence to obtain the optimal drilling plan; The 3D tensor is of size (N, T) dl T pl 3D tensor , , , , indicating that it belongs to the first Class and in the first The number of wells that were drilled and put into production in the month was [number missing]. Monthly output; The process for obtaining the total production profile corresponding to the current drilling plan tensor is as follows: Using Einstein's summation convention, calculate all the th... Class of wells in The total monthly output was , It is a dimensionality reduction function; Get in the first The total production of all wells in the month is ; Get the first tensor corresponding to the current drilling plan. Annual total production profile Y pl The drilling and development period is measured in years. loss function ; The drilling plan tensor to be solved , , Defined as learnable parameters.

Images