Intelligent well completion multi-objective optimization method based on multi-objective genetic algorithm

Abstract

The invention provides an intelligent well completion multi-objective optimization method based on a multi-objective genetic algorithm, and the method comprises the steps: determining an optimization objective and a constraint condition according to a real environment of intelligent well completion; and constructing a multi-target genetic algorithm model, and executing optimization calculation of the multi-target genetic algorithm. And evaluating the influence of each ICV configuration on the optimization target by adopting a fitness function. And an initial population is generated for initialization, a current optimal individual is screened out through iterative operation, and the process is ended after data are output. The method has the beneficial effects that the calculation efficiency is improved by establishing the multi-target fitness model and optimizing the iteration strategy, so that the algorithm can quickly converge to the optimal solution under limited calculation resources. The fitness function of the scheme can accurately evaluate the influence of different parameter configurations on the optimization target, and the optimization result is more reliable and accurate. And an optimization method which is more efficient, more comprehensive and more suitable for actual production requirements is provided.

Description

Technical Field

[0001] The invention relates to the technical field of petroleum engineering, and more specifically to an intelligent well completion multi-objective optimization method based on a multi-objective genetic algorithm. Background Art

[0002] With the advancement of petroleum engineering technology, intelligent completion optimization has become the key to improving the efficiency of oil and gas field development. Although existing intelligent completion optimization technologies have made some progress, most of them still rely on traditional single-objective optimization strategies, especially the maximization of net present value (NPV). However, when multiple objectives such as maximizing oil production and minimizing water production need to be considered at the same time, these traditional methods are often unable to cope with the task. In addition, in the face of the changing conditions and complex constraints in the intelligent completion process, the existing optimization algorithms still have deficiencies in efficiency and adaptability.

[0003] In this context, genetic algorithms, as a powerful global search tool, show great potential in multi-objective optimization problems. Genetic algorithms have been widely studied and applied due to their outstanding performance in finding optimal solutions. For example, the Chinese patent document "A real-time reactive power optimization method for intelligent power systems based on genetic algorithms" is an innovative application example. The patent uses genetic algorithms to construct a real-time reactive power optimization model for intelligent power systems, highlighting the utility of genetic algorithms in real-time determination of the optimal operating settings of multi-stage capacitors and optimization of reactive power. The method provides an accurate solution within the computational time.

[0004] In addition, in the paper "Multi-objective Optimization of Robotic Laser Cladding Process Parameters Based on Genetic Algorithm", by combining the genetic algorithm with the objectives of specific industrial parameters, a multi-objective optimization mathematical model was established that can simultaneously meet the requirements of optimal performance and sub-objective minimization, providing a new perspective for the optimization of complex process parameters. In the research field of multi-agent systems, the solution to dynamic multi-objective optimization problems is also critical. The paper "Probability-driven Evolutionary Optimization of Dynamic Multi-objective Multi-agent Collaborative Scheduling" by Liu Xiaofang et al. further demonstrated the applicability of multi-objective optimization in dynamic environments and emphasized the necessity of achieving optimal collaborative scheduling among multiple objectives under changing conditions.

[0005] The key parameters in the intelligent completion industry have complex relationships. Not only do they need to maximize and minimize key outputs, but they also need to consider the impact of other factors such as economy, environment, and operational safety.

[0006] Based on this, the present invention proposes a new multi-objective genetic algorithm optimization model, which can more comprehensively solve the multi-dimensional and complex problems in the intelligent completion process, thereby improving the overall development efficiency and benefits of oil and gas fields. By combining the powerful search capability of genetic algorithms and the flexibility of multi-objective optimization, the present invention not only broadens the boundaries of intelligent completion optimization, but also promotes the innovation and progress of oil and gas field development technology. Summary of the invention

[0007] The present invention overcomes the deficiencies in the prior art and provides an intelligent well completion multi-objective optimization method based on a multi-objective genetic algorithm.

[0008] The purpose of the present invention is achieved through the following technical solutions.

[0009] A multi-objective optimization method for intelligent well completion based on a multi-objective genetic algorithm, the specific steps include:

[0010] S1. Determine the optimization objectives and constraints according to the actual environment of intelligent completion;

[0011] S2, construct a multi-objective genetic algorithm model;

[0012] S3, generate the initial population and evaluate the fitness of individuals;

[0013] S4, the initial population generated by S3 is combined with the hybrid elite strategy to iterate the genetic algorithm and evaluate whether each round of iteration meets the end condition. If the condition is met, it enters S5, and if the condition is not met, it returns to S4 to continue iteration;

[0014] S5. Output the current optimal individual and end the process.

[0015] The optimization objectives in S1 include: maximizing NPV, maximizing oil production, and minimizing water production.

[0016] The constraints in S1 include: the water production does not exceed the specified threshold of total production.

[0017] The construction process of the target genetic algorithm model in S2 specifically includes:

[0018] S2-1. Define the decision variables that affect optimization in intelligent completion.

[0019]

[0020] in, To define the decision variable vector, x i (i=1…n) represents the ICV opening of the i-th well section;

[0021] S2-2. Create a fitness function to comprehensively weigh multiple optimization objectives and constraints to evaluate the effect of each ICV configuration. The fitness function is as follows:

[0022]

[0023] Among them, ω1 and ω2 are weight coefficients used to adjust the contribution of each objective function to the total fitness. Represents the maximization of the net present value NPV, Represents the maximization of oil production, Based on water production Whether the penalty function exceeds the specified threshold, the specific formula is as follows:

[0024]

[0025] Among them, θ is the specified threshold, and k is the penalty coefficient, which is used to adjust the fitness reduction degree of the part exceeding the threshold.

[0026] The specific steps of S3 include:

[0027] S3-1. Determine the size N of the initial population based on the complexity of the optimization problem and available computing resources;

[0028] S3-2, randomly generate N initial individuals, each individual represents a parameter configuration of an intelligent completion section;

[0029] S3-3, using the fitness function defined in S2, evaluate the fitness of each initial individual;

[0030] S3-4. Store the generated initial population and the fitness value of the initial population.

[0031] The genetic algorithm iteration combined with the hybrid elite strategy in S4 includes the following specific steps:

[0032] S4-1, set α as the elite retention rate, select individuals with the top α% of fitness from the population generated in S3 as elite individuals, and copy them to the new population;

[0033] S4-2, perform crossover operation on the remaining (1-α)% of individuals to generate new offspring;

[0034] S4-3, performing mutation operation on the newly generated offspring;

[0035] S4-4, evaluate the fitness of individuals in the new generation population;

[0036] S4-5, check whether the end condition is met, terminate the iteration if it is met, and continue to execute the genetic algorithm iteration of the hybrid elite strategy if it is not met.

[0037] The end conditions in step S4-5 include:

[0038] The preset maximum number of iterations is reached;

[0039] The highest fitness value of the population has converged;

[0040] Find M Pareto optimal solutions that meet the multi-objective optimization conditions, where M is the required number preset in the specific implementation plan.

[0041] The beneficial effects of the present invention are:

[0042] (1) The present invention can simultaneously consider multiple optimization objectives, such as maximizing net present value (NPV), oil production, and minimizing water production, overcoming the technical problem that the prior art may only focus on a single objective. The multi-objective genetic algorithm has a powerful global search capability, which can effectively avoid falling into a local optimal solution and increase the probability of finding a global optimal solution. The algorithm of the present invention can be flexibly adjusted according to actual production conditions and technical requirements, and is applicable to various complex intelligent completion scenarios.

[0043] (2) By establishing a multi-objective fitness model and an optimization iteration strategy, the computational efficiency is improved, so that the algorithm can quickly converge to the optimal solution under limited computing resources. The fitness function of this solution can accurately evaluate the impact of different parameter configurations on the optimization target, and the optimization results are more reliable and accurate. It provides an optimization method that is more efficient, more comprehensive, and more adaptable to actual production needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 It is the principle diagram of the present invention;

[0045] Figure 2 Calculate analysis graphs for multi-objective intelligent optimization algorithms;

[0046] Figure 3 Calculate and analyze the graph for the single-objective intelligent optimization algorithm; DETAILED DESCRIPTION

[0047] Example

[0048] like Figure 1 As shown in Figure 2, the specific process of multi-objective intelligent optimization is as follows:

[0049] S1. Determine the optimization objectives and constraints based on the actual environment of intelligent completion.

[0050] The optimization goal was established based on the specific conditions of the oil field: to maximize NPV while ensuring maximum oil production and minimum water production.

[0051] Considering equipment performance limitations, environmental regulations and economic budget, corresponding technical and economic constraints were formulated. Constraints include that water production must not exceed the specified threshold of total production.

[0052] S2. Construct a multi-objective genetic algorithm model to perform the optimization calculation of the multi-objective genetic algorithm. The model can accurately simulate the multi-parameter operating environment of intelligent completion and prepare for the subsequent optimization calculation.

[0053] The decision variables that affect optimization in the model are:

[0054]

[0055] in, To define the decision variable vector, x i (i=1…n) represents the ICV opening of the i-th well section;

[0056] The fitness function is as follows:

[0057]

[0058] Among them, ω1 and ω2 are weight coefficients used to adjust the contribution of each objective function to the total fitness. In this example, the relative importance of NPV and oil production is close, so both are set to 0.5. Objective function Represents the maximization of the net present value NPV, Represents the maximization of oil production, Based on water production The penalty function for whether the specified threshold is exceeded is as follows:

[0059]

[0060] Among them, θ is taken as 20%, if the water production If the output exceeds 20%, the item will reduce the fitness value. The size of the reduction is determined by the penalty coefficient k, which is used to adjust the fitness reduction of the part exceeding the threshold.

[0061] The k value can be obtained through experiments by observing its influence on the optimization process. For example, it can be set starting from 10 and gradually increased until a value is found that can effectively constrain the excess part without affecting the overall convergence performance. In this embodiment, the k value is 50.

[0062] S3. Generate an initial population with a population size of 100. Each population represents a different well section parameter configuration of the intelligent completion. The ICV opening sequence contained in each population individual is randomly initialized within a reasonable operating range. Then, the fitness function is used to evaluate the fitness of each individual in the population and store it.

[0063] S4, iteratively execute the genetic algorithm on the subpopulation initialized in S3, and perform subsequent selection, crossover and mutation operations according to the score value of the fitness function. The specific process is as follows:

[0064] Select subpopulation: select 20% of the individuals with the highest fitness according to the fitness value. These individuals are directly retained to the next generation and do not participate in crossover mutation.

[0065] Crossover operation: After the hybrid elite strategy is executed, the top 50% of individuals in terms of fitness are selected from the remaining non-elite individuals (i.e., the 80% of individuals not directly retained) for random pairing crossover operation, with the crossover rate set to 50%. This operation aims to generate offspring with new characteristics by mixing the genetic configurations between different individuals, thereby promoting population diversity and the algorithm's exploration ability.

[0066] Mutation operation: 5% of the above non-elite individuals are mutated to introduce genetic mutations and explore new areas of the solution space. This step ensures that all solutions are within the technical constraints, especially maintaining the water production constraint to avoid exceeding the acceptable technical limits.

[0067] Afterwards, fitness evaluation is performed to assess the fitness of individuals in the new generation population.

[0068] Finally, determine whether the iteration termination condition is met. If the termination condition is met, the iteration is terminated. If the iteration termination condition is not met, the iteration is re-executed.

[0069] The iteration termination condition is:

[0070] (1) The maximum number of iterations is 10,000;

[0071] (2) The fitness convergence standard is that there is no significant change in 10 consecutive iterations, and the improvement of the fitness value is less than the preset threshold of 0.01, thus terminating the iteration process;

[0072] (3) Find at least 10 Pareto optimal solutions.

[0073] S5. Output the current optimal individual and end the process.

[0074] The experimental results of this scheme are as follows:

[0075] like Figure 2 As shown in the figure, in the initial stage of the multi-objective intelligent optimization algorithm, NPV shows a clear growth trend. With the increase of the number of iterations, when the algorithm causes excessive water production during the optimization process, the growth rate of NPV slows down, and may even decrease, which reflects the impact of the penalty mechanism. At the same time, the percentage curves of oil production and water production show that the algorithm is trying its best to adjust the production ratio to optimize NPV, while avoiding additional consumption caused by excessive water production.

[0076] Figure 3 It is a single-objective optimization algorithm. Compared with the multi-objective intelligent optimization algorithm, the single-objective optimization is more focused on directly increasing revenue, even if it means accepting higher water production. Increased oil production usually directly increases sales revenue, which may have a greater impact on NPV than the increased cost caused by increased water production. In single-objective optimization, more aggressive production strategies are adopted, such as increasing water injection pressure or using chemical methods to improve recovery, which can increase oil production and NPV in the short term, but may lead to higher water production and faster oil field decline in the long term. Therefore, the intelligent completion process should be comprehensively optimized and designed with multiple objectives in mind.

[0077] The embodiments of the present invention are described in detail above, but the contents described are only preferred embodiments of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A multi-objective optimization method for intelligent well completion based on a multi-objective genetic algorithm, characterized in that: The specific steps include: S1. Determine the optimization objectives and constraints according to the actual environment of intelligent completion; S2, construct a multi-objective genetic algorithm model; S3, generate the initial population and evaluate the fitness of individuals; S4, the initial population generated by S3 is combined with the hybrid elite strategy to iterate the genetic algorithm and evaluate whether each round of iteration meets the end condition. If the condition is met, enter S5, and if the condition is not met, return to S4 to continue iteration; S5. Output the current optimal individual and end the process.

2. The multi-objective optimization method for intelligent well completion based on a multi-objective genetic algorithm according to claim 1, characterized in that: The optimization objectives in S1 include: maximizing NPV, maximizing oil production, and minimizing water production.

3. The multi-objective optimization method for intelligent well completion based on a multi-objective genetic algorithm according to claim 1, characterized in that: The constraints in S1 include: the water production does not exceed the specified threshold of total production.

4. The multi-objective optimization method for intelligent well completion based on a multi-objective genetic algorithm according to claim 1, characterized in that: The construction process of the target genetic algorithm model in S2 specifically includes: S2-1. Define the decision variables that affect optimization in intelligent completion. in, To define the decision variable vector, x i (i=1…n) represents the ICV opening of the i-th well section; S2-2. Create a fitness function to comprehensively weigh multiple optimization objectives and constraints to evaluate the effect of each ICV configuration. The fitness function is as follows: Among them, ω1 and ω2 are weight coefficients used to adjust the contribution of each objective function to the total fitness. Represents the maximization of the net present value NPV, Represents the maximization of oil production, Based on water production Whether the penalty function exceeds the specified threshold, the specific formula is as follows: Among them, θ is the specified threshold, and k is the penalty coefficient, which is used to adjust the fitness reduction degree of the part exceeding the threshold.

5. The multi-objective optimization method for intelligent well completion based on a multi-objective genetic algorithm according to any one of claims 1 to 4, characterized in that: The specific steps of S3 include: S3-1. Determine the size N of the initial population based on the complexity of the optimization problem and available computing resources; S3-2, randomly generate N initial individuals, each individual represents a parameter configuration of an intelligent completion section; S3-3, using the fitness function defined in S2, evaluate the fitness of each initial individual; S3-4. Store the generated initial population and the fitness value of the initial population.

6. The multi-objective optimization method for intelligent well completion based on a multi-objective genetic algorithm according to claim 1, characterized in that: The genetic algorithm iteration combined with the hybrid elite strategy in S4 includes the following specific steps: S4-1, set α as the elite retention rate, select individuals with the top α% of fitness from the population generated in S3 as elite individuals, and copy them to the new population; S4-2, perform crossover operation on the remaining (1-α)% of individuals to generate new offspring; S4-3, performing mutation operation on the newly generated offspring; S4-4, evaluate the fitness of individuals in the new generation population; S4-5, check whether the end condition is met, terminate the iteration if it is met, and continue to execute the genetic algorithm iteration of the hybrid elite strategy if it is not met.

7. The intelligent well completion multi-objective optimization method based on a multi-objective genetic algorithm according to claim 6, characterized in that: The end conditions in step S4-5 include: The preset maximum number of iterations is reached; The highest fitness value of the population has converged; Find M Pareto optimal solutions that meet the multi-objective optimization conditions, where M is the required number preset in the specific implementation plan.

Images