An oilfield well site direct current microgrid capacity optimization method based on multi-objective column and constraint generation algorithm

Abstract

The application discloses an oilfield well site direct current micro-grid capacity optimization method based on a multi-target column and constraint generation algorithm, and mainly relates to the field of power system planning optimization. The method comprises the following steps: constructing a two-stage robust optimization model; the first stage is a capacity planning stage, and the optimal configuration capacity of a photovoltaic power generation system, an energy storage system and a diesel generator is determined; the second stage is a multi-target cost optimization stage; a column and constraint generation algorithm is used for decomposition and solving; the original problem is decomposed into a main problem and a sub-problem, the main problem is responsible for capacity configuration, and the sub-problem is responsible for finding the worst operating scenario under the given capacity configuration. The application has the beneficial effects that the problems of the prior art, such as the insufficient description of source and load bilateral uncertainty in the oilfield well site scene, the disconnection of capacity planning and operation scheduling, the difficulty in giving consideration to economy, reliability and environmental protection in single-target solving, and the difficulty in embedding a two-stage robust framework in multi-target optimization, are solved.

Description

Technical Field

[0001] This invention relates to the field of power system planning and optimization, specifically a method for optimizing the capacity of DC microgrids in oilfield well sites based on multi-objective columns and constraint generation algorithms. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] Oilfield well site energy systems are transforming from traditional single fossil fuel power supply to low-carbon, distributed, and intelligent systems. For oilfield well sites located in remote areas with weak power distribution conditions and complex load characteristics, introducing photovoltaic power generation systems, energy storage systems, and diesel generators to form independent or weakly grid-connected DC microgrids is an effective way to improve energy utilization efficiency, reduce carbon emissions, and enhance power supply reliability.

[0004] In the planning of well site DC microgrids, capacity optimization is a key factor determining the system's economy, reliability, and environmental friendliness. Capacity optimization refers to rationally determining the installed capacity of distributed power sources such as photovoltaic power generation systems, energy storage systems, and diesel generators, while meeting load power requirements and operational constraints, so as to achieve better overall performance throughout the system's life cycle.

[0005] However, unlike typical industrial parks or residential microgrids, oilfield well sites have significant industry-specific characteristics. On the one hand, the well site load is mainly driven by pumping unit loads, which are affected by factors such as formation pressure changes, pumping rates between pumping units, equipment start-up and shutdown processes, and seasonal temperature changes, exhibiting obvious periodic fluctuations and uncertainties. On the other hand, photovoltaic power generation output is affected by solar irradiance and ambient temperature, exhibiting typical randomness and intermittency.

[0006] Existing technical methods typically establish single-layer optimization models based on historical average daily load, typical daily photovoltaic curves, or predicted values, aiming to minimize system investment or operating costs to determine the capacity of photovoltaic, energy storage, and diesel generators. These methods are simple in structure and easy to solve, but they essentially treat source-load data as deterministic values, failing to effectively characterize uncertainty. Capacity optimization methods based on stochastic optimization usually presuppose that photovoltaic output and load demand follow a certain probability distribution, solving for the expected optimal solution of the system through a large number of random scenario samples. Their advantage lies in theoretically considering the impact of random fluctuations, but they rely on accurate probability distribution models and a large number of sample scenarios, resulting in a large computational scale, and in actual oilfield well sites, it is often difficult to obtain sufficient, accurate, and stable distribution information. Capacity optimization methods based on traditional robust optimization typically construct box-shaped uncertainty sets, polygonal uncertainty sets, etc., based on the upper and lower bounds of uncertain variables, improving the robustness of the configuration scheme by solving the optimization problem under the most unfavorable scenario. Many existing methods employ single-stage robust optimization or two-stage robust optimization frameworks. Two-stage robust optimization typically separates "capacity planning" and "operation scheduling" into layers. The first stage determines the capacity, and the second stage optimizes the operation decision under the worst-case scenario.

[0007] In summary, these methods generally remain at the level of single-objective solution, general scenario modeling, or simple uncertainty description. They have not yet formed a capacity optimization method for column and constraint generation algorithms that addresses the dual uncertainties of oilfield well site sources and loads and can achieve multi-objective collaborative optimization within a two-stage robust framework. Summary of the Invention

[0008] The purpose of this invention is to provide a capacity optimization method for DC microgrids in oilfield well sites based on multi-objective column and constraint generation algorithms. It solves the problems of existing methods in oilfield well site scenarios, such as insufficient characterization of uncertainties on both the source and load sides, disconnect between capacity planning and operation scheduling, difficulty in balancing economy, reliability and environmental protection in single-objective solutions, and difficulty in embedding multi-objective optimization into a two-stage robust framework.

[0009] To achieve the above objectives, the present invention employs the following technical solution: A method for optimizing the capacity of a DC microgrid in an oilfield well site based on a multi-objective column and constraint generation algorithm includes the following steps: A two-stage robust optimization model is constructed: the first stage is the capacity planning stage, which determines the optimal configuration capacity of the photovoltaic power generation system, energy storage system and diesel generator based on historical data, equipment parameters and operating conditions analysis; The second stage is the multi-objective cost optimization stage. After the uncertain variables are revealed, the operation strategy is optimized with the goal of minimizing the total operating cost, the load shortage penalty cost, and the carbon emission penalty cost. The problem is solved by decomposing it using a column and constraint generation algorithm: the original problem is decomposed into a main problem and subproblems. The main problem is responsible for capacity configuration, and the subproblems are responsible for finding the worst operating scenario under a given capacity configuration. New columns and constraints are added to the main problem iteratively until convergence.

[0010] The objective function of the two-stage robust optimization model is constructed as follows: Objective function form: ; A preference coefficient is introduced, which is calculated based on actual system operating data and equipment parameters; The objective function contains , , , These are preference coefficients, used to adjust investment costs. Operating costs Penalty cost for load shortage and carbon emission penalty costs By assigning weights to each objective, multi-objective collaborative optimization can be achieved.

[0011] The main problem is constructed and solved as follows: Main Problem Model: The main problem is responsible for the system's capacity planning. Its model is expressed as minimizing the investment cost in the first stage and the total cost under the worst-case scenario in the second stage, given upper and lower bounds for uncertain variables. ,in As an auxiliary variable, it represents the total cost of the worst-case scenario in the second stage under a given capacity.

[0012] The subproblems are constructed and solved as follows: Sub-problem model: The sub-problem is responsible for finding the worst operating scenario under a given capacity configuration and optimizing the output of each distributed power source to minimize operating costs, load shortage penalty costs and carbon emission penalty costs; KKT condition transformation: By introducing KKT conditions, the inner optimization problem in the subproblem is transformed into a system of equations without a target, which is then combined with the outer problem for solution, thus improving the solution efficiency; Multi-objective optimization: Use linear weighting or constraint methods to handle multi-objective optimization problems to ensure a reasonable trade-off between different objectives.

[0013] Oilfield well site feature modeling: Pumping unit load modeling: Based on the characteristics of the pumping unit's electrical power diagram, periodicity, and backflow generation phenomenon, an accurate pumping unit load model is established; Energy storage system modeling: Based on the impact of low temperature environment on energy storage capacity, an effective capacity decay model of energy storage system at low temperature is established; Diesel generator modeling: Based on factors such as the backup support requirements and fuel consumption characteristics of diesel generators, an economic operation model for diesel generators is established.

[0014] The implementation method of the multi-objective optimization is as follows: Linear weighted method: The linear weighted method is used to achieve multi-objective optimization in sub-problems, and the trade-offs between different objectives are achieved by adjusting the preference coefficients; Constraint method: Transform some objectives into constraints, and achieve trade-offs between different objectives by adjusting the constraint boundaries; Interactive optimization: Introducing decision-maker interaction allows for flexible adjustment of optimization objectives and constraints based on actual needs, improving the practicality and satisfaction of optimization results.

[0015] The verification and evaluation methods for the optimization results are as follows: By constructing a simulation platform, the system operation under different working conditions is simulated to verify the effectiveness and robustness of the capacity optimization scheme; field tests are conducted in actual oilfield well sites to collect actual operating data, which is then compared and analyzed with the optimization results to evaluate the performance of the optimization scheme; based on the simulation experiments and field test results, the optimization model and methods are continuously improved to enhance the optimization accuracy and practicality.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention provides a capacity optimization method for DC microgrids in oilfield well sites based on multi-objective column and constraint generation algorithms, in order to solve the problems of existing methods in oilfield well site scenarios, such as insufficient characterization of uncertainties on both the source and load sides, disconnect between capacity planning and operation scheduling, difficulty in balancing economy, reliability and environmental protection in single-objective solutions, and difficulty in embedding multi-objective optimization into a two-stage robust framework.

[0017] Specifically, a two-stage capacity optimization model applicable to DC microgrids in oilfield well sites is provided, which integrates system capacity planning and operation scheduling under uncertain scenarios into the same optimization framework; Specifically, a modeling method for the uncertainty of both sides of the source load in oilfield well sites is provided, so that the capacity optimization results are more in line with the actual operating requirements under the coupling of photovoltaic fluctuations and pumping unit periodic loads; Specifically, a multi-objective operation optimization method is provided that simultaneously considers operating costs, load shortage penalty costs, and carbon emission penalty costs in the second stage, achieving synergistic optimization of economy, reliability, and environmental protection; Specifically, a method is provided to embed multi-objective optimization problems into the solution framework of column and constraint generation algorithms, so that two-stage robust optimization models can still be solved efficiently through iterative methods of main problem and sub-problem; Specifically, it improved the engineering applicability, power supply reliability, and low-carbon operation level of the DC microgrid capacity configuration results for oilfield well sites. Attached Figure Description

[0018] Appendix Figure 1 This is the electrical diagram of the well site pumping unit of the present invention.

[0019] Appendix Figure 2 This is a flowchart of the operation of the low-carbon well site DC microgrid in this invention.

[0020] Appendix Figure 3 This is a diagram showing the capacity configuration of the low-carbon well site DC microgrid TSRO in this invention.

[0021] Appendix Figure 4 This is the power balance state diagram in this invention. Detailed Implementation

[0022] The present invention will be further illustrated below with reference to specific embodiments. It should be understood that these embodiments are for illustrative purposes only and are not intended to limit the scope of the invention. Furthermore, it should be understood that after reading the teachings of this invention, those skilled in the art can make various alterations or modifications to the invention, and these equivalent forms also fall within the scope defined in this application.

[0023] This invention describes a capacity optimization method for DC microgrids in oilfield well sites based on a multi-objective column and constraint generation algorithm. In DC microgrids in oilfield well sites, considering the uncertainties on both the source and load sides caused by the random fluctuations of photovoltaic power generation and the periodic changes in pumping unit load, this invention aims to construct a solvable, implementable, and engineering-applicable multi-objective two-stage robust capacity optimization method while simultaneously considering investment costs, operating costs, power supply reliability, and carbon emissions. The method achieves efficient solution through a column and constraint generation algorithm.

[0024] The low-carbon well site DC microgrid system comprises core equipment including: a photovoltaic power generation system, an energy storage system, a DC / DC converter, a DC / AC inverter, an AC / DC bidirectional converter, AC loads, and DC loads. The low-carbon well site DC microgrid adopts a three-level coordinated operation strategy of "photovoltaic priority - energy storage regulation - diesel backup," achieving dynamic matching of source, storage, and load through hierarchical control.

[0025] Based on the established low-carbon well site DC microgrid system model, to improve the accuracy of its uncertainty description, a TSRO capacity optimization configuration method oriented towards uncertainties on both the source and load sides is proposed. The first stage optimizes the distributed power supply configuration capacity of the well site microgrid, and the optimization result serves as the known quantity for the second stage. The second stage is a multi-objective cost optimization. The operating strategy is determined after the uncertain variables are revealed. The two stages are jointly optimized with the goal of minimizing the total operating cost, load shortage penalty cost, and carbon emission penalty cost under the worst-case distribution of random variables, aiming to minimize the total cost, load shortage rate, and carbon emissions over the entire life cycle.

[0026] This paper introduces a multi-objective collaborative optimization mechanism into the sub-problems. Since investment cost optimization has been completed in the first stage, the second stage only adjusts the preference coefficients for the operational objective. This approach ensures the separability of the modeling structure and the flexibility of multi-objective trade-offs, achieving a comprehensive balance between system economy, reliability, and environmental friendliness while guaranteeing robustness. The first stage mainly addresses investment and construction costs; the second stage addresses operating costs, load shortage penalty costs, and carbon emission penalty costs. A two-stage robust optimization objective function is constructed as follows:

[0027]

[0028] In the formula: This is the preference coefficient; , , , These are investment costs, operating costs, power shortage penalty costs, and carbon emission penalty costs, respectively.

[0029] Its compact form is:

[0030] In the formula: x represents the decision variable of the first stage, that is, the configuration capacity of each distributed power source; y represents the decision variable of the second stage, that is, the output of each distributed power source. Represents an uncertain variable; Let y represent the feasible region; A, B, C, D, E represent coefficient vectors; a, b, c, d represent coefficient matrices; and T represents a constant column vector.

[0031] The specific expression is as follows:

[0032] Based on the idea of ​​C&CG algorithm, the original problem is decomposed into a TSRO model by master problem (MP) and sub-problem (SP). The linear weighting method is used to consider multi-objective optimization of operating cost, load shortage and carbon emission penalty cost in SP.

[0033] In this paper, MP corresponds to the system's capacity planning process. Based on the two-stage robust optimization model established in 3.3, the MP model can be expressed as:

[0034] In the formula: As an auxiliary variable, it represents the worst-case scenario in the second stage under a given capacity; Indicates the number of iterations; This represents the solution to the subproblem after the j-th iteration; This represents the value of the uncertain variable under the worst-case scenario generated by SP after the j-th iteration.

[0035] Within the TSRO framework, SP (Power Supply Optimization) refers to the process of optimizing power supply scheduling under a given capacity configuration scheme. Its model can be represented as:

[0036] After the j-th iteration of SP is solved, a new set of second-stage variables is added to MP, and new constraints are added to MP:

[0037] By applying the KKT conditions, the inner problem is transformed into a system of equations without a target, which is then combined with the outer problem to obtain the following form:

[0038] In the formula: represents the auxiliary variable introduced by the inequality constraint; represents the auxiliary variable introduced by the equality constraint.

[0039] The main steps for solving TSRO using the C&CG algorithm are as follows:

[0040] This invention employs a two-stage robust optimization framework, unifying the first-stage capacity configuration with the second-stage operational optimization. This overcomes the shortcomings of traditional single-stage methods that focus solely on planning while neglecting operational considerations, resulting in capacity configuration outcomes that better align with actual operational needs. In the second stage, this invention simultaneously considers operating costs, load shortage penalty costs, and carbon emission penalty costs, using a unified weighting through preference coefficients to achieve multi-objective coordinated optimization and avoid the strong bias issues of single-objective methods. Through a unified multi-objective transformation mechanism, this invention transforms the second-stage multi-objective operational optimization problem into a form that can be handled by column and constraint generation algorithms, enabling multi-objective optimization to be embedded within the two-stage robust optimization master-sub-problem iterative framework. This invention models typical well site characteristics such as pumping unit cyclic load, reverse power generation, and the impact of low-temperature environments on energy storage capacity, making it more consistent with the realities of oilfield well site engineering compared to general microgrid optimization models. This invention identifies the worst-case operating conditions under uncertain source-load scenarios and optimizes capacity configuration accordingly, ensuring that the resulting scheme maintains strong adaptability and power supply guarantee capabilities under conditions of photovoltaic fluctuations and load changes. This invention solves the problem by decomposing the main problem into subproblems, with a clear computational path, making it suitable for deployment in actual well site microgrid planning and exhibiting good engineering feasibility.

Claims

1. A method for optimizing the capacity of a DC microgrid in an oilfield well site based on a multi-objective column and constraint generation algorithm, characterized in that: Includes the following steps: A two-stage robust optimization model is constructed: the first stage is the capacity planning stage, which determines the optimal configuration capacity of the photovoltaic power generation system, energy storage system and diesel generator based on historical data, equipment parameters and operating conditions analysis; The second stage is the multi-objective cost optimization stage. After the uncertain variables are revealed, the operation strategy is optimized with the goal of minimizing the total operating cost, the load shortage penalty cost, and the carbon emission penalty cost. The problem is solved by decomposing it using a column and constraint generation algorithm: the original problem is decomposed into a main problem and subproblems. The main problem is responsible for capacity configuration, and the subproblems are responsible for finding the worst operating scenario under a given capacity configuration. New columns and constraints are added to the main problem iteratively until convergence.

2. The method for optimizing the capacity of a DC microgrid in an oilfield well site based on a multi-objective column and constraint generation algorithm as described in claim 1, characterized in that: The objective function of the two-stage robust optimization model is constructed as follows: Objective function form: ; A preference coefficient is introduced, which is calculated based on actual system operating data and equipment parameters; The objective function contains , , , These are preference coefficients, used to adjust investment costs. Operating costs Penalty cost for load shortage and carbon emission penalty costs By assigning weights to each objective, multi-objective collaborative optimization can be achieved.

3. The method for optimizing the capacity of a DC microgrid in an oilfield well site based on a multi-objective column and constraint generation algorithm as described in claim 2, characterized in that: The main problem is constructed and solved as follows: Main Problem Model: The main problem is responsible for the capacity planning of the system. Its model is expressed as minimizing the investment cost of the first stage and the total cost of the worst-case scenario in the second stage, given the upper and lower bounds of the uncertain variables. That is, where is the auxiliary variable, representing the total cost of the worst-case scenario in the second stage under the given capacity.

4. The method for optimizing the capacity of a DC microgrid in an oilfield well site based on a multi-objective column and constraint generation algorithm as described in claim 3, characterized in that: The subproblems are constructed and solved as follows: Sub-problem model: The sub-problem is responsible for finding the worst operating scenario under a given capacity configuration and optimizing the output of each distributed power source to minimize operating costs, load shortage penalty costs and carbon emission penalty costs; KKT condition transformation: By introducing KKT conditions, the inner optimization problem in the subproblem is transformed into a system of equations without a target, which is then combined with the outer problem for solution, thus improving the solution efficiency; Multi-objective optimization: Use linear weighting or constraint methods to handle multi-objective optimization problems to ensure a reasonable trade-off between different objectives.

5. The method for optimizing the capacity of a DC microgrid in an oilfield well site based on a multi-objective column and constraint generation algorithm as described in claim 4, characterized in that: The oilfield well site feature modeling: Pumping unit load modeling: Based on the characteristics of the pumping unit's electrical power diagram, periodicity, and backflow generation phenomenon, an accurate pumping unit load model is established; Energy storage system modeling: Based on the impact of low temperature environment on energy storage capacity, an effective capacity decay model of energy storage system at low temperature is established; Diesel generator modeling: Based on factors such as the backup support requirements and fuel consumption characteristics of diesel generators, an economic operation model for diesel generators is established.

6. The method for optimizing the capacity of a DC microgrid in an oilfield well site based on a multi-objective column and constraint generation algorithm as described in claim 5, characterized in that: The implementation method of the multi-objective optimization is as follows: Linear weighted method: The linear weighted method is used to achieve multi-objective optimization in sub-problems, and the trade-offs between different objectives are achieved by adjusting the preference coefficients; Constraint method: Transform some objectives into constraints, and achieve trade-offs between different objectives by adjusting the constraint boundaries; Interactive optimization: Introducing decision-maker interaction allows for flexible adjustment of optimization objectives and constraints based on actual needs, improving the practicality and satisfaction of optimization results.

7. The method for optimizing the capacity of a DC microgrid in an oilfield well site based on a multi-objective column and constraint generation algorithm as described in claim 6, characterized in that: The verification and evaluation methods for the optimization results are as follows: By constructing a simulation platform, the system operation under different working conditions is simulated to verify the effectiveness and robustness of the capacity optimization scheme; field tests are conducted in actual oilfield well sites to collect actual operating data, which is then compared and analyzed with the optimization results to evaluate the performance of the optimization scheme; based on the simulation experiments and field test results, the optimization model and methods are continuously improved to enhance the optimization accuracy and practicality.

Images