Security constraint unit commitment optimization acceleration method and system based on large language model

Abstract

The invention belongs to the technical field of power systems, and relates to a security constraint unit commitment optimization acceleration method and system based on a large language model. The method comprises the steps of obtaining economic operation basic data of a power system; constructing an integer relaxation security constraint unit commitment model and solving; generating a code corresponding to the seed algorithm; evolving a neighborhood search algorithm based on a large language model to obtain an optimal neighborhood search algorithm; constructing a security constraint unit commitment model of a limited neighborhood according to an optimal neighborhood search algorithm; solving a security constraint unit commitment model of a limited neighborhood; constructing a security constraint unit commitment model; and solving the security constraint unit commitment model. According to the method, high-quality neighborhood constraints can be generated, and efficient feasible solution optimization is realized; the solving efficiency of the security constraint unit combination problem is improved, the optimization solving efficiency of the power system based on the security constraint unit combination is improved, efficient optimization operation of a high-proportion power system is supported, and safe and stable operation of the power system is guaranteed.

Description

Technical Field

[0001] The present invention belongs to the technical field of power systems, and more specifically, relates to a method and system for accelerating the optimization of security-constrained unit commitment based on large language models. Background Art

[0002] Security-constrained unit commitment is the mathematical basis for power system dispatching operation, market clearing, system planning, etc. Security-constrained unit commitment is usually modeled as a mixed-integer programming problem. With the continuous increase in the proportion of new energy in the power system and the expansion of the scale of interconnected large power grids, the power system will face an expansion of the decision-making space in terms of time and space, and the number of constraints and the degree of constraint coupling of the unit commitment problem will increase greatly, thus bringing challenges to the solution of the unit commitment problem. There is an urgent need to study model dimension reduction and constraint reduction methods.

[0003] Data-driven methods and heuristic methods are commonly used methods to accelerate the solution of unit commitment. The former predicts by constructing a model to learn historical data, fixing integer variables or predicting ineffective constraints to achieve solution acceleration; the latter reduces the problem scale through special designs for the problem structure to accelerate the solution of security-constrained unit commitment. Existing research on data-driven methods and heuristic methods for accelerating the solution of unit commitment still has deficiencies. Data-driven methods require a large amount of data for training, and the computational cost of their offline training is large. In addition, most of the models trained by data-driven methods are black-box models and are not interpretable. Heuristic methods rely on historical experience and the design of human experts, and have certain limitations in various aspects. In addition, most heuristic methods cannot achieve good generalization acceleration for test cases, and usually can only achieve efficient acceleration for some test cases. Summary of the Invention

[0004] In order to solve the above technical problems, the present invention aims to improve the computational efficiency of security-constrained unit commitment from the perspective of the evolution of large language model-driven heuristic algorithms, and provides a method and system for accelerating the optimization of security-constrained unit commitment based on large language models.

[0005] In a first aspect, the present invention provides a method for accelerating the optimization of security-constrained unit commitment based on a large language model, including generating a large language model-driven neighborhood search algorithm based on the large language model and solving the security-constrained unit commitment; Generating a large language model-driven neighborhood search algorithm based on the large language model includes: obtaining basic data on the economic operation of the power system; constructing and solving an integer relaxation security-constrained unit commitment model; generating corresponding codes for the seed algorithm; and evolving the neighborhood search algorithm based on the large language model to obtain an optimal neighborhood search algorithm; Solving security-constrained unit commitment includes: constructing a security-constrained unit commitment model with a limited neighborhood based on the optimal neighborhood search algorithm; solving the security-constrained unit commitment model with a limited neighborhood; constructing a security-constrained unit commitment model; and solving the security-constrained unit commitment model.

[0006] In a second aspect, the present invention provides a security-constrained unit commitment optimization acceleration system based on a large language model, including an algorithm generation unit and a solution unit; The algorithm generation unit is used to generate a large language model-driven neighborhood search algorithm based on the large language model; the algorithm generation unit includes an acquisition unit, a model construction and solution unit, a code generation unit, and an evolutionary update unit; The acquisition unit is used to acquire the basic data of the economic operation of the power system; The model construction and solution unit is used to construct and solve an integer relaxation security-constrained unit commitment model; The code generation unit is used to generate the corresponding code of the seed algorithm; The evolutionary update unit is used to evolve the neighborhood search algorithm based on the large language model to obtain the optimal neighborhood search algorithm; The solution unit is used to solve the security-constrained unit commitment; the solution unit includes a first combination model construction unit, a combination model solution unit, a second combination model construction unit, and a solution subunit; The first combination model construction unit is used to construct a security-constrained unit commitment model with a limited neighborhood based on the optimal neighborhood search algorithm; The combination model solution unit is used to solve the security-constrained unit commitment model with a limited neighborhood; The second combination model construction unit is used to construct a security-constrained unit commitment model; The solution subunit is used to solve the security-constrained unit commitment model.

[0007] Based on the above technical solutions, the present invention can be further improved as follows.

[0008] Further, the basic data of the economic operation of the power system includes the upper and lower limits of unit output, the upper limit of the unit's ramp rate for increasing and decreasing, the minimum continuous on-off time of the unit, the unit operation cost function, the unit start-stop cost function, the power flow transfer distribution factor, the line transmission capacity, and the system positive and negative reserve rates.

[0009] Further, constructing and solving an integer relaxation security-constrained unit commitment model includes: Determining the decision variables of the integer relaxation security-constrained unit commitment, where the decision variables include the output variables of the generating units in the set time period and the start-stop relaxation variables in the set time period; Determining the objective function corresponding to the integer relaxation security-constrained unit commitment model; the objective function is to minimize the total operation cost and start-stop cost in each time period; Establish the constraint conditions based on security-constrained unit commitment; the constraint conditions include system power balance constraint, system positive and negative reserve constraints, system positive and negative ramping constraints, minimum continuous on-off time constraints of generating units, and line security constraints; Use a mixed-integer linear programming solver to solve the objective function corresponding to the defined integer-relaxed security-constrained unit commitment model, and obtain the integer-relaxed security-constrained unit commitment result.

[0010] Furthermore, generate the corresponding code for the seed algorithm, including determining the seed neighborhood search algorithm and generating code text based on the seed neighborhood search algorithm: Determine the seed neighborhood search algorithm, including: defining the unit start-stop variables of the security-constrained unit commitment and the neighborhood distance of the optimal solution of the integer-relaxed security-constrained unit commitment, and obtaining the neighborhood constraint by limiting the neighborhood distance of the optimal solution to be no greater than the system reserve rate.

[0011] Furthermore, based on the algorithm evolution of the large language model, including: Construct an evaluation function for evaluating the neighborhood search algorithm, which is used to evaluate the performance of the neighborhood search algorithm; set a group of security-constrained unit commitment examples for evaluation, solve them based on the neighborhood search algorithm and the mixed-integer linear programming solver, and compare the solved results and time consumption with the results and time consumption of directly using the mixed-integer linear programming solver to solve the security-constrained unit commitment, and calculate the evaluation index for evaluating the neighborhood search algorithm through the evaluation function; Determine the code generation large model and the prompt generation large model; the code generation large model is used to generate the code of the neighborhood search algorithm; the prompt generation large model is used to summarize the characteristics of the generated code corresponding to the neighborhood search algorithm and input them into the code generation large model; Generate the neighborhood search algorithm based on the evolutionary algorithm; Based on the preset prompts and the seed algorithm code, use the code generation large model to generate a batch of codes corresponding to the neighborhood search algorithm; evaluate the performance of the neighborhood search algorithm based on the evaluation function for evaluating the neighborhood search algorithm, and sort all the neighborhood search algorithms; Based on the sorted neighborhood search algorithms, randomly select algorithm pairs from them, and use the prompt generation large model to process each pair of algorithms based on each pair of algorithms in the algorithm pair and the preset prompts, and generate algorithm improvement suggestions according to the algorithm performance to update the neighborhood search algorithm; Evaluate the performance of the generated neighborhood search algorithm based on the evaluation function for evaluating the neighborhood search algorithm, and sort the neighborhood search algorithms; Repeat the process of generating the neighborhood search algorithm until the evaluation score of the neighborhood search algorithm remains unchanged or reaches the maximum number of algorithm generations after evaluating the set number of neighborhood search algorithms, and obtain the optimal neighborhood search algorithm.

[0012] Furthermore, when constructing a security-constrained unit commitment model with a limited neighborhood, the unit start-stop decision variables of the integer-relaxed security-constrained unit commitment model are adjusted to discrete variables, while the unit output decision variables remain continuous variables; based on the optimal neighborhood search algorithm, neighborhood constraints are generated to obtain a security-constrained unit commitment model with a limited neighborhood. Furthermore, solving the security-constrained unit commitment model with a limited neighborhood includes: Using a mixed-integer linear programming solver to solve the security-constrained unit commitment model with a limited neighborhood to obtain the result of the security-constrained unit commitment model with a limited neighborhood.

[0013] Furthermore, constructing a security-constrained unit commitment model includes: According to the security-constrained unit commitment model with a limited neighborhood, the neighborhood constraints are deleted to obtain a security-constrained unit commitment model.

[0014] Furthermore, solving the security-constrained unit commitment model includes: using the result of the security-constrained unit commitment model with a limited neighborhood obtained as the warm start solution of the mixed-integer linear programming solver, and using the mixed-integer linear programming solver to solve the security-constrained unit commitment model to obtain the result of the security-constrained unit commitment model.

[0015] The beneficial effects of the present invention are as follows: Based on the large language model and within the framework of the evolutionary algorithm, the present invention uses the neighborhood search algorithm to achieve directed evolution, generates a large language model-driven neighborhood search algorithm, which can generate high-quality neighborhood constraints and achieve efficient feasible solution optimization; without sacrificing the accuracy of the solution, based on the high-quality feasible solutions found, it improves the solution efficiency of the security-constrained unit commitment problem, enhances the optimization solution efficiency of the power system based on security-constrained unit commitment, supports the efficient optimization operation of a high-proportion power system, and ensures the safe and stable operation of the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a schematic diagram of the security-constrained unit commitment optimization acceleration method based on a large language model provided in Embodiment 1 of the present invention; Figure 2 It is a schematic block diagram of the security-constrained unit commitment optimization acceleration system based on a large language model provided in Embodiment 2 of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0017] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Usually, the components of the embodiments of the present invention described and illustrated herein can be arranged and designed in various different configurations.

[0018] Embodiment 1 As an embodiment, as shown in the appendix Figure 1 To solve the above technical problems, this embodiment provides a security-constrained unit commitment optimization acceleration method based on a large language model, including generating a neighborhood search algorithm driven by the large language model and solving the security-constrained unit commitment based on the large language model; Generating a neighborhood search algorithm driven by the large language model based on the large language model, including: obtaining the basic data of the economic operation of the power system; constructing and solving an integer-relaxed security-constrained unit commitment model; generating the corresponding code of the seed algorithm; evolving the neighborhood search algorithm based on the large language model to obtain the optimal neighborhood search algorithm; Solving the security-constrained unit commitment, including: constructing a security-constrained unit commitment model with a limited neighborhood according to the optimal neighborhood search algorithm; solving the security-constrained unit commitment model with a limited neighborhood; constructing a security-constrained unit commitment model; solving the security-constrained unit commitment model.

[0019] Based on the large language model, under the framework of the evolutionary algorithm, this invention uses the neighborhood search algorithm to achieve directed evolution, generates a neighborhood search algorithm driven by the large language model, can generate high-quality neighborhood constraints, and realizes efficient feasible solution optimization. Without sacrificing the accuracy of the solution, based on the obtained high-quality feasible solutions, it improves the solution efficiency of the security-constrained unit commitment problem, enhances the optimization solution efficiency of the power system based on the security-constrained unit commitment, supports the efficient optimization operation of the high-proportion power system, and ensures the safe and stable operation of the power system.

[0020] Optionally, the basic data of the economic operation of the power system includes the upper and lower limits of unit output, the upper limit of the unit's ramp rate of increase and decrease, the minimum continuous on-off time of the unit, the unit operation cost function, the unit start-stop cost function, the power flow transfer distribution factor, the line transmission capacity, and the system positive and negative reserve rates.

[0021] Optionally, constructing and solving an integer-relaxed security-constrained unit commitment model includes: Determining the decision variables of the integer-relaxed security-constrained unit commitment, where the decision variables include the output variables of the generating units in the set time period and the start-stop relaxation variables in the set time period; Determining the objective function corresponding to the integer-relaxed security-constrained unit commitment model; the objective function is to minimize the total operation cost and start-stop cost in each time period; Establishing the constraint conditions based on the security-constrained unit commitment; the constraint conditions include system power balance constraints, system positive and negative reserve constraints, system positive and negative ramp constraints, minimum continuous on-off time constraints of generating units, and line security constraints; Solve the objective function corresponding to the defined integer-relaxed security-constrained unit commitment model using a mixed-integer linear programming solver to obtain the integer-relaxed security-constrained unit commitment result.

[0022] Generating unit At time period The output variable of is At time period The start-stop relaxation variable of generating unit is which takes continuous values between 0 and 1. The start-stop relaxation variable of generating unit At time period is is is the start-stop cost function of unit , is the output cost function of unit . In the formula, is defined as the initial start-stop state of generating unit , is the total time, is the total number of units. The expression of the objective function corresponding to the integer-relaxed security-constrained unit commitment model is: .

[0023] Establish the constraints based on the security-constrained unit commitment. Let be the system net load within time period , be the system positive reserve rate, be the system negative reserve rate, be the maximum value of the system net load within time period , be the maximum value of the system net load within time period , be the output upper limit of generating unit , be the output lower limit of generating unit , be the upper limit of the increased output of generating unit , be the upper limit of the decreased output of generating unit , be the output upper limit when generating unit starts, be the output upper limit before generating unit shuts down, be the minimum continuous on-time of generating unit , be the minimum continuous off-time of generating unit The minimum continuous outage time, in hours, is the number of time periods for unit commitment, denotes the th line, is the line 's maximum power flow, is the node 's power flow transfer distribution factor for line , is the node where unit is located, is the node where unit is located, then: The system power balance constraint is expressed as: ; The system positive reserve constraint is expressed as: ; The system negative reserve constraint is expressed as: ; The system positive ramp reserve constraint is expressed as: ; The system negative ramp reserve constraint is expressed as: ; The generating unit output range constraint is expressed as: ; The generating unit positive ramp constraint is expressed as: ; The generating unit negative ramp constraint is expressed as: ; The generating unit minimum continuous on - off time constraint is expressed as: ; The generating unit minimum continuous on - off time constraint is expressed as: ; The line security constraint is expressed as: .

[0024] Optionally, generate the corresponding code for the seed algorithm, including determining the seed neighborhood search algorithm and generating code text based on the seed neighborhood search algorithm: Determine the seed neighborhood search algorithm, including: defining the start - stop variables of the security - constrained unit commitment and the neighborhood distance of the integer - relaxed optimal solution of the security - constrained unit commitment, and obtaining the neighborhood constraint by limiting the neighborhood distance of the optimal solution to be no greater than the system reserve rate.

[0025] An example of the algorithm is: define the start - stop variables of the security - constrained unit commitment as , and the integer - relaxed optimal solution of the security - constrained unit commitment as , the neighborhood distance is , and then by restricting this distance to be no greater than to obtain neighborhood constraints, is the weighting coefficient, and the weighting coefficient is used to describe the strength of the actual solution approaching the relaxed solution, as follows: ; ; .

[0026] Optionally, based on the algorithm evolution of the large language model, including: Construct an evaluation function for evaluating the neighborhood search algorithm. The evaluation function is used to evaluate the performance of the neighborhood search algorithm; set a set of security-constrained unit commitment test cases for evaluation, solve based on the neighborhood search algorithm and the mixed-integer linear programming solver, and compare the solved results and the time consumption with the results and time consumption of directly using the mixed-integer linear programming solver to solve the security-constrained unit commitment, and calculate the evaluation index of the evaluation neighborhood search algorithm through the evaluation function; set is the index of the algorithm, is the number of test cases used to evaluate the algorithm performance, is the weight of the efficiency index and the accuracy index, is the time consumption for solving the test case based on the neighborhood search algorithm or directly using the solver , is the time consumption for solving the test case based on the neighborhood search algorithm or directly using the solver . The lower the evaluation index, the better the algorithm performance, then: ; Determine the code generation large model and the prompt generation large model; the code generation large model is used to generate the code of the neighborhood search algorithm; the prompt generation large model is used to summarize the characteristics of the generated code corresponding to the neighborhood search algorithm and input it into the code generation large model; Generate the neighborhood search algorithm based on the evolutionary algorithm; Based on the preset prompts and the seed algorithm code, use the code generation large model to generate a batch of codes corresponding to the neighborhood search algorithm; evaluate the performance of the neighborhood search algorithm based on the evaluation function of the evaluation neighborhood search algorithm, and sort all the neighborhood search algorithms; Based on the sorted neighborhood search algorithms, randomly select algorithm pairs from them. Based on each pair of algorithms in the algorithm pair and the preset prompts, use the prompt generation large model to process each pair of algorithm pairs, compare and summarize the characteristics of the algorithms with better performance, and generate algorithm improvement suggestions for updating the neighborhood search algorithm; Evaluate the performance of the generated neighborhood search algorithm based on the evaluation function of the neighborhood search algorithm, and sort the neighborhood search algorithms; Repeat the process of the neighborhood search algorithm until the evaluation score of the neighborhood search algorithm remains unchanged after evaluating a set number of neighborhood search algorithms or reaches the maximum number of algorithm generations, and obtain the optimal neighborhood search algorithm.

[0027] For the optimal unit output of the integer relaxation security-constrained unit commitment, For the optimal unit relaxation start-stop result of the integer relaxation security-constrained unit commitment, For the basic data of the economic operation of the power system, then: 。

[0028] Optionally, when constructing the security-constrained unit commitment model with a limited neighborhood, adjust the unit start-stop decision variables of the integer relaxation security-constrained unit commitment model to discrete variables, and keep the unit output decision variables as continuous variables; generate neighborhood constraints based on the optimal neighborhood search algorithm to obtain the security-constrained unit commitment model with a limited neighborhood. The neighborhood constraint is expressed as: 。

[0029] Optionally, solving the security-constrained unit commitment model with a limited neighborhood includes: Use a mixed-integer linear programming solver to solve the security-constrained unit commitment model with a limited neighborhood and obtain the result of the security-constrained unit commitment model with a limited neighborhood.

[0030] Optionally, constructing the security-constrained unit commitment model includes: According to the security-constrained unit commitment model with a limited neighborhood, delete the neighborhood constraints to obtain the security-constrained unit commitment model.

[0031] Optionally, solving the security-constrained unit commitment model includes: using the result of the security-constrained unit commitment model with a limited neighborhood obtained as the warm start solution of the mixed-integer linear programming solver, and using the mixed-integer linear programming solver to solve the security-constrained unit commitment model to obtain the result of the security-constrained unit commitment model.

[0032] Based on the large language model, within the framework of the evolutionary algorithm, this invention uses the neighborhood search algorithm to achieve directed evolution, generates a large language model-driven neighborhood search algorithm, can generate high-quality neighborhood constraints, and realizes efficient feasible solution optimization. Without sacrificing the accuracy of the solution, it improves the solution efficiency of the security-constrained unit commitment problem, enhances the optimization solution efficiency of the power system based on the security-constrained unit commitment, supports the efficient optimization operation of the high-proportion power system, and ensures the safe and stable operation of the power system.

[0033] Embodiment 2 Based on the same principle as the method shown in Embodiment 1 of the present invention, as shown in the appendixFigure 2 As shown in the figure, in the embodiments of the present invention, there is also provided an accelerated system for optimizing the security-constrained unit commitment based on a large language model, including an algorithm generation unit and a solution unit; The algorithm generation unit is used to generate a neighborhood search algorithm driven by a large language model based on the large language model; the algorithm generation unit includes an acquisition unit, a model construction and solution unit, a code generation unit, and an evolutionary update unit; The acquisition unit is used to acquire the basic data of the economic operation of the power system; The model construction and solution unit is used to construct and solve an integer-relaxed security-constrained unit commitment model; The code generation unit is used to generate the corresponding code of the seed algorithm; The evolutionary update unit is used to evolve the neighborhood search algorithm based on the large language model to obtain the optimal neighborhood search algorithm; The solution unit is used to solve the security-constrained unit commitment; the solution unit includes a first combined model construction unit, a combined model solution unit, a second combined model construction unit, and a solution subunit; The first combined model construction unit is used to construct a security-constrained unit commitment model with a limited neighborhood by using the optimal neighborhood search algorithm; The combined model solution unit is used to solve the security-constrained unit commitment model with a limited neighborhood; The second combined model construction unit is used to construct a security-constrained unit commitment model; The solution subunit is used to solve the security-constrained unit commitment model.

[0034] Optionally, the basic data of the economic operation of the power system includes the upper and lower limits of unit output, the upper limit of the unit's ramp rate of increase and decrease, the minimum continuous on-off time of the unit, the unit's operating cost function, the unit's start-stop cost function, the power flow transfer distribution factor, the line transmission capacity, and the system's positive and negative reserve rates.

[0035] Optionally, constructing and solving an integer-relaxed security-constrained unit commitment model includes: Determining the decision variables of the integer-relaxed security-constrained unit commitment, where the decision variables include the output variables of the generating units in the set time period and the start-stop relaxation variables in the set time period; Determining the objective function corresponding to the integer-relaxed security-constrained unit commitment model; the objective function is to minimize the total operating cost and start-stop cost in each time period; Establishing the constraint conditions based on the security-constrained unit commitment; the constraint conditions include system power balance constraints, system positive and negative reserve constraints, system positive and negative ramp constraints, minimum continuous on-off time constraints of generating units, and line security constraints; Solve the objective function corresponding to the defined integer-relaxed security-constrained unit commitment model using a mixed-integer linear programming solver to obtain the integer-relaxed security-constrained unit commitment result.

[0036] Optionally, generate the corresponding code for the seed algorithm, including determining the seed neighborhood search algorithm and generating code text based on the seed neighborhood search algorithm: Determine the seed neighborhood search algorithm, including: defining the unit start-stop variables of the security-constrained unit commitment and the neighborhood distance of the optimal solution of the integer-relaxed security-constrained unit commitment, and obtaining the neighborhood constraint by limiting the neighborhood distance of the optimal solution to be no greater than the system reserve rate.

[0037] Optionally, the algorithm evolution based on the large language model includes: Construct an evaluation function for evaluating the neighborhood search algorithm. The evaluation function is used to evaluate the performance of the neighborhood search algorithm; set a group of security-constrained unit commitment examples for evaluation, solve them based on the neighborhood search algorithm and the mixed-integer linear programming solver, and compare the obtained results and the time consumption with the results and time consumption of directly using the mixed-integer linear programming solver to solve the security-constrained unit commitment, and calculate the evaluation index for evaluating the neighborhood search algorithm through the evaluation function; Determine the code generation large model and the prompt generation large model; the code generation large model is used to generate the code of the neighborhood search algorithm; the prompt generation large model is used to summarize the characteristics of the code corresponding to the generated neighborhood search algorithm and input them into the code generation large model; Generate the neighborhood search algorithm based on the evolutionary algorithm; Based on the preset prompts and the seed algorithm code, use the code generation large model to generate a batch of codes corresponding to the neighborhood search algorithm; evaluate the performance of the neighborhood search algorithm based on the evaluation function for evaluating the neighborhood search algorithm, and sort all the neighborhood search algorithms; Based on the sorted neighborhood search algorithms, randomly select algorithm pairs from them. Based on each pair of algorithms in the algorithm pair and the preset prompts, use the prompt generation large model to process each pair of algorithm pairs, compare and summarize the characteristics of the algorithms with better performance, and generate algorithm improvement suggestions for updating the neighborhood search algorithm; Evaluate the performance of the generated neighborhood search algorithm based on the evaluation function for evaluating the neighborhood search algorithm, and sort the neighborhood search algorithms; Repeat the process of generating the neighborhood search algorithm until the evaluation score of the neighborhood search algorithm remains unchanged after evaluating the set number of neighborhood search algorithms or reaches the maximum number of algorithm generations, and obtain the optimal neighborhood search algorithm.

[0038] Optionally, when constructing a security-constrained unit commitment model with a limited neighborhood, adjust the unit start-stop decision variables of the integer-relaxed security-constrained unit commitment model to discrete variables, and keep the unit power output decision variables as continuous variables; generate neighborhood constraints based on the optimal neighborhood search algorithm to obtain a security-constrained unit commitment model with a limited neighborhood. Optionally, solving the security-constrained unit commitment model with a limited neighborhood includes: Use a mixed-integer linear programming solver to solve the security-constrained unit commitment model with a limited neighborhood, and obtain the result of the security-constrained unit commitment model with a limited neighborhood.

[0039] Optionally, constructing a security-constrained unit commitment model includes: According to the security-constrained unit commitment model with a limited neighborhood, delete the neighborhood constraints to obtain a security-constrained unit commitment model.

[0040] Optionally, solving the security-constrained unit commitment model includes: using the result of the obtained security-constrained unit commitment model with a limited neighborhood as the warm start solution of the mixed-integer linear programming solver, and using the mixed-integer linear programming solver to solve the security-constrained unit commitment model to obtain the result of the security-constrained unit commitment model.

[0041] The above are only the preferred embodiments of the present invention and are not used to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for accelerating the security-constrained unit commitment optimization based on large language models, characterized in that, Including generating a large language model-driven neighborhood search algorithm based on a large language model and solving a security-constrained unit commitment; Generating a large language model-driven neighborhood search algorithm based on a large language model, including: obtaining basic data on the economic operation of the power system; constructing and solving an integer relaxation security-constrained unit commitment model; generating corresponding code for the seed algorithm; evolving the neighborhood search algorithm based on the large language model to obtain the optimal neighborhood search algorithm; Solving the security-constrained unit commitment, including: constructing a security-constrained unit commitment model with a limited neighborhood according to the optimal neighborhood search algorithm; solving the security-constrained unit commitment model with a limited neighborhood; constructing a security-constrained unit commitment model; solving the security-constrained unit commitment model.

2. The security-constrained unit commitment optimization acceleration method based on large language models according to claim 1, wherein The basic data on the economic operation of the power system includes the upper and lower limits of unit output, the upper limit of the unit's ramp-up and ramp-down rates, the minimum continuous on-off time of the unit, the unit operation cost function, the unit start-stop cost function, the power flow transfer distribution factor, the line transmission capacity, and the system positive and negative reserve rates.

3. The safety-constrained unit commitment optimization acceleration method based on large language models according to claim 1, wherein Constructing and solving an integer relaxation security-constrained unit commitment model, including: Determining the decision variables of the integer relaxation security-constrained unit commitment, where the decision variables include the output variables of the generating units in the set time period and the start-stop relaxation variables in the set time period; Determining the objective function corresponding to the integer relaxation security-constrained unit commitment model; the objective function is to minimize the total operation cost and start-stop cost in each time period; Establishing the constraint conditions based on the security-constrained unit commitment; the constraint conditions include system power balance constraints, system positive and negative reserve constraints, system positive and negative ramp constraints, minimum continuous on-off time constraints of generating units, and line security constraints; Using a mixed-integer linear programming solver to solve the objective function corresponding to the defined integer relaxation security-constrained unit commitment model to obtain the integer relaxation security-constrained unit commitment result.

4. The safety-constrained unit commitment optimization acceleration method based on a large language model according to claim 1, wherein Generating the corresponding code for the seed algorithm, including determining the seed neighborhood search algorithm and generating code text based on the seed neighborhood search algorithm: Determining the seed neighborhood search algorithm, including: defining the start-stop variables of the generating units in the security-constrained unit commitment and the neighborhood distance of the optimal solution of the integer relaxation security-constrained unit commitment, and obtaining the neighborhood constraint by limiting the neighborhood distance of the optimal solution to be no greater than the system reserve rate.

5. The safety-constrained unit commitment optimization acceleration method based on a large language model according to claim 1, wherein The algorithm evolution based on the large language model, including: Constructing an evaluation function for evaluating the neighborhood search algorithm, where the evaluation function is used to evaluate the performance of the neighborhood search algorithm; setting a set of security-constrained unit commitment examples for evaluation, solving based on the neighborhood search algorithm and a mixed-integer linear programming solver, and comparing the obtained results and time consumption with the results and time consumption of directly using the mixed-integer linear programming solver to solve the security-constrained unit commitment, and calculating the evaluation index of the neighborhood search algorithm through the evaluation function; Determining the code generation large model and the prompt generation large model; the code generation large model is used to generate the code of the neighborhood search algorithm; the prompt generation large model is used to summarize the characteristics of the code corresponding to the generated neighborhood search algorithm and input it into the code generation large model; Generating a neighborhood search algorithm based on the evolutionary algorithm; Based on the preset prompt words and seed algorithm code, use the code generation large model to generate a batch of codes corresponding to the neighborhood search algorithm; evaluate the performance of the neighborhood search algorithm based on the evaluation function for the neighborhood search algorithm, and sort all the neighborhood search algorithms; Based on the sorted neighborhood search algorithms, randomly select algorithm pairs from them. Based on each pair of algorithms in the algorithm pair and the preset prompt words, use the prompt word generation large model to process each pair of algorithm pairs, and generate algorithm improvement suggestions according to the algorithm performance to update the neighborhood search algorithm; Evaluate the performance of the generated neighborhood search algorithm based on the evaluation function of the neighborhood search algorithm, and sort the neighborhood search algorithms; Repeat the process of generating the neighborhood search algorithm until the evaluation score of the neighborhood search algorithm remains unchanged or reaches the maximum number of algorithm generations after evaluating the set number of neighborhood search algorithms, and obtain the optimal neighborhood search algorithm.

6. The safety-constrained unit commitment optimization acceleration method based on large language models according to claim 1, wherein When constructing the security-constrained unit commitment model with a limited neighborhood, adjust the unit start-stop decision variables of the integer-relaxed security-constrained unit commitment model to discrete variables, and keep the unit output decision variables as continuous variables; generate neighborhood constraints based on the optimal neighborhood search algorithm to obtain the security-constrained unit commitment model with a limited neighborhood.

7. The safety-constrained unit commitment optimization acceleration method based on large language models according to claim 1, wherein Solve the security-constrained unit commitment model with a limited neighborhood, including: Use a mixed-integer linear programming solver to solve the security-constrained unit commitment model with a limited neighborhood to obtain the result of the security-constrained unit commitment model with a limited neighborhood.

8. The safety-constrained unit commitment optimization acceleration method based on a large language model according to claim 1, wherein, Construct a security-constrained unit commitment model, including: According to the security-constrained unit commitment model with a limited neighborhood, delete the neighborhood constraints to obtain the security-constrained unit commitment model.

9. The safety-constrained unit commitment optimization acceleration method based on a large language model according to claim 1, wherein Solve the security-constrained unit commitment model, including: use the obtained result of the security-constrained unit commitment model with a limited neighborhood as the warm start solution of the mixed-integer linear programming solver, and use the mixed-integer linear programming solver to solve the security-constrained unit commitment model to obtain the result of the security-constrained unit commitment model.

10. A security-constrained unit commitment optimization acceleration system based on a large language model, characterized in that, It includes an algorithm generation unit and a solution unit; The algorithm generation unit is used to generate a large language model-driven neighborhood search algorithm based on the large language model; the algorithm generation unit includes an acquisition unit, a model construction and solution unit, a code generation unit, and an evolution and update unit; The acquisition unit is used to acquire the basic data of the economic operation of the power system; The model construction and solution unit is used to construct and solve the integer-relaxed security-constrained unit commitment model; The code generation unit is used to generate the code corresponding to the seed algorithm; The evolution and update unit is used to evolve the neighborhood search algorithm based on the large language model to obtain the optimal neighborhood search algorithm; The solution unit is used to solve the security-constrained unit commitment; the solution unit includes a first combined model construction unit, a combined model solution unit, a second combined model construction unit, and a solution subunit; The first combined model construction unit is used to construct the security-constrained unit commitment model with a limited neighborhood using the optimal neighborhood search algorithm; The combined model solution unit is used to solve the security-constrained unit commitment model with a limited neighborhood; The second combined model construction unit is used to construct the security-constrained unit commitment model; The solution subunit is used to solve the security-constrained unit commitment model.

Images