Power system planning typical day selection method and system based on operation simulation

Abstract

The invention relates to the technical field of electric power systems, in particular to an electric power system planning typical day selection method and system based on operation simulation, and the method mainly comprises the steps: constructing an operation day residual load curve, obtaining the classification label and solving difficulty of each operation day based on a Mama model, and carrying out the calculation of the classification label. The method comprises the following steps: selecting an operation day with the highest solving difficulty in various operation days as a typical day, solving unit combination starting and stopping results of various typical days, and solving unit output of the operation days based on the unit starting and stopping results of the typical days. And aggregation of operation days with similar load characteristics but different unit starting and stopping or large load difference but the same unit starting and stopping is avoided, and the defect that the clustering error is large or the clustering result is rough is overcome. Historical operation simulation data is autonomously learned through a Mama model, and a typical daily label vector is directly output. Manual selection of important information such as load curve characteristics and the number of typical days is avoided, and the result is more objective.

Description

Technical Field

[0001] The present invention relates to the technical field of power systems, and in particular to a method and system for selecting typical days for power system planning based on operation simulation. Background Art

[0002] The power planning model is a basic tool for power system operation analysis. On the premise of meeting the system safety and stability constraints, it simulates different operation scenarios, predicts possible system failures and overloads, and rationally allocates power resources, thereby ensuring the safe and stable operation of the power system while achieving economic maximization.

[0003] However, power planning models are often simulated over a time span of months or even years, and contain a large number of integer variables and complex constraints, resulting in extremely long computational times for direct solutions. With the advancement of global energy transformation, the proportion of renewable energy in the power system has increased significantly. Renewable energy has strong volatility and uncertainty, which makes the constraints of power planning models more complex and the computational difficulty further increased. In order to improve computational efficiency, power planning models often select typical days that can represent the operating scenarios throughout the year for solution. How to select an appropriate number of typical days with valid information is crucial to the efficiency of running simulations.

[0004] Existing research at home and abroad mainly selects typical days through three methods: clustering method, sampling method and heuristic scenario reduction method. The clustering method classifies the operating day load curve through an algorithm and obtains the load curve at the center of the class as the typical day load curve. The sampling method randomly selects sample points to capture the load curve with a higher probability of occurrence in different scenarios as the typical day load curve. The heuristic scenario reduction method selects representative load curves as the typical day load curve through experience and simple approximate calculations.

[0005] The clustering method ignores the different contributions of different points in the daily load curve to the load curve characteristics, and may not be able to accurately aggregate load curves with similar key characteristics. At the same time, the number of clusters depends on human selection, which is highly subjective and may lead to large errors.

[0006] When the sample size is small or the sampling process is not sophisticated enough, the selected typical daily load curve may not be representative enough. At the same time, the sampling method may ignore some extreme load conditions, which may have a greater impact on the operation and planning of the power system.

[0007] The heuristic scenario reduction method mainly relies on experience and simple approximate calculations, which may not accurately evaluate the high-proportion renewable energy load curve with strong volatility, but lacks certain generalization. At the same time, when the system scale becomes larger or the situation is complex, the stability of the heuristic algorithm may be poor, resulting in large errors. Summary of the invention

[0008] The purpose of the present invention is to provide a method and system for selecting typical days for power system planning based on operation simulation to solve the above-mentioned problems in the prior art.

[0009] The present invention is achieved through the following technical solutions: In a first aspect, a typical day selection method for power system planning based on operation simulation includes: Obtain historical operating day data and construct the historical operating day residual load curve; Solve the unit combination according to the residual load curve of the historical operation day, obtain the unit start and shutdown results of each historical operation day, and classify and mark the unit start and shutdown results of the historical operation day to obtain the classification label of each historical operation day; Based on the Mamba model and the classification labels of each historical operating day, the difficulty of solving the classification labels of each historical operating day is output, and the operating day with the greatest difficulty in solving among various types of operating days is obtained as a typical day; Solve the start-up and shutdown results of various typical day unit combinations, establish a unit combination model based on the full consumption of new energy output, and obtain the output results of each period of each operating day based on the start-up and shutdown results of various typical day unit combinations and the unit combination model, and output the output results.

[0010] Preferably, the difficulty of solving the classification labels of each historical operation day based on the Mamba model and the classification labels of each historical operation day output includes: Constructing a time series characteristic matrix of all historical operation days, wherein the first dimension of the time series characteristic matrix is ​​a load curve, the second dimension of the time series characteristic matrix is ​​a wind power output curve, and the third dimension of the time series characteristic matrix is ​​a photovoltaic output curve; Constructing a label vector for all historical running days, wherein the first dimension of the label vector is the classification label, and the second dimension is the solution time; The Mamba model is trained with the time series feature matrix of historical operating days as input and the label vector of historical operating days as output. The solution time in the label vector of each historical operating day is obtained through the trained Mamba model, and the operating day with the longest solution time is taken as the operating day with the greatest solution difficulty.

[0011] Preferably, the establishing of the unit combination model comprises: Establish the objective function of the unit commitment model and set the constraints of the objective function; The constraints include load balance constraints, system positive reserve constraints, system negative reserve constraints, unit output upper and lower limit constraints, unit ramp rate constraints and unit minimum continuous start and stop time constraints.

[0012] Preferably, the objective function for establishing the unit commitment model includes: A functional formula is constructed with the goal of minimizing the total cost of system operation, and the functional formula includes:

[0013] In the formula, For a typical day The thermal power output cost of unit i in period t, N is the total number of units, For a typical day The startup cost of unit i in time period t, where T is the number of time periods in a day.

[0014] Preferably, the load balancing constraints include:

[0015] Where: For a typical day Medium thermal power unit In the period of efforts, For a typical day The total system load, For a typical day China New Energy contributes.

[0016] Preferably, the system normal and standby constraints include:

[0017] The system negative reserve constraints include:

[0018] In the formula, For a typical day Medium thermal power unit In the period The start and stop status of the typical daily unit combination, For a typical day Medium thermal power unit In the period The maximum output, For a typical day Medium thermal power unit In the period The minimum output, For a typical day Mid-session of the system's positive spare capacity, For a typical day Mid-session The system negative spare capacity.

[0019] Preferably, the upper and lower limit constraints of the unit output include: .

[0020] Preferably, the unit climbing rate constraint includes:

[0021] ;

[0022] ; In the formula, For thermal power units Maximum climbing rate; For thermal power units Maximum downhill climbing rate, For a typical day Medium thermal power unit In the period The maximum output, For a typical day Medium thermal power unit In the period The start and stop status.

[0023] Preferably, the minimum continuous start and stop time constraint of the unit includes:

[0024] Where: For a typical day Medium thermal power unit In the period Continuous power-on time; For a typical day Medium thermal power unit In the period Continuous downtime. is the minimum continuous start time of the unit, It is the minimum continuous downtime of the unit.

[0025] In a second aspect, the present invention further provides a typical day selection system for power system planning based on operation simulation, comprising: The data processing module is configured to obtain historical operating day data and construct a residual load curve for the historical operating day; solve the unit combination according to the residual load curve for the historical operating day to obtain the unit start and stop results for each operating day, and classify and label the historical unit start and stop results to obtain the classification labels for each operating day; The model operation model is configured to output the classification labels and solution difficulties of each operating day based on the Mamba model and the classification labels of each operating day, and obtain the operating day with the greatest solution difficulty among various operating days as the typical day, solve the start and stop results of the unit combination of various typical days, establish a unit combination model based on the full consumption of new energy output, and obtain the output results of each period of each operating day based on the start and stop results of the unit combination of various typical days and the unit combination model, and output the output results; The main control module is connected to the data processing module and the model operation model, and is used to execute the above-mentioned method for selecting typical days for power system planning based on operation simulation.

[0026] The technical solution of the present invention has at least the following advantages and beneficial effects: 1. The present invention classifies the operating days based on the historical unit start and stop results, avoiding aggregating the operating days with similar load characteristics but different unit start and stop or with large load differences but the same unit start and stop, and preventing the disadvantages of large clustering errors or rough clustering results.

[0027] 2. The present invention uses the Mamba model to autonomously learn historical operation simulation data and directly output typical day label vectors, thus avoiding the artificial selection of important information such as load curve characteristics and the number of typical days, and the results are more objective and more accurate.

[0028] 3. The present invention reflects the difficulty of solving the unit combination that is difficult to quantify through the easily available unit combination solution time, greatly reducing the difficulty of model training while ensuring data validity, thereby enhancing practicality. The start and stop of the unit on the operating day with the greatest difficulty in solving the typical day is selected as the typical day result, ensuring that the typical day can take into account the most extreme load scenario, thereby enhancing the robustness of the system. 4. The present invention selects the start and stop results of the computer unit on a typical day, greatly reducing the number of model variables and the calculation time. The start and stop results of the unit are then used to solve the unit output in each period of each operating day, reducing the complexity of the unit and obtaining more accurate model simulation results. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0030] Figure 1 It is a schematic diagram of the process of the present invention; Figure 2 It is a schematic diagram of the system structure of the present invention. DETAILED DESCRIPTION

[0031] In order to make the purpose, 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 in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0032] The division of modules in this application is a logical division. There may be other division methods when implemented in actual applications. For example, multiple modules can be combined or integrated into another system, or some features can be ignored or not executed.

[0033] The modules or submodules described independently may be physically separated or not: they may be implemented by software or hardware, and some modules or submodules may be implemented by software, and the processor may call the software to implement the functions of these modules or submodules, and other modules or submodules may be implemented by hardware, such as by hardware circuits. In addition, some or all of the modules may be selected according to actual needs to achieve the purpose of the present application.

[0034] Please refer to Figure 1-Figure 2 The present invention provides a typical day selection method for power system planning based on operation simulation, comprising: S101: Acquire historical operation day data and construct a historical operation day residual load curve; In this embodiment, each operating day has a load curve , a wind power output curve , a photovoltaic output curve Assumptions are made to define the residual load curve for each operating day. .

[0035] S102: Solving the unit combination according to the residual load curve of the historical operation day, obtaining the unit start and stop results of each historical operation day, and classifying and marking the historical unit start and stop results to obtain the classification label of each historical operation day; Classify the historical unit start and stop results. Set the same label for the operation days with the same unit start and stop results in each period. For the first type of unit start and stop result, set the label to 1, the second type of unit start and stop result to 2, the third type to 3, and so on, until all the unit start and stop results that appeared in the historical operation days are marked.

[0036] S103: outputting the difficulty of solving the classification labels of each operating day based on the Mamba model and the classification labels of each operating day, and obtaining the operating day with the greatest difficulty of solving among various types of operating days as a typical day; Among them, since the difficulty of solving the unit combination on each operating day is difficult to quantify accurately, and the solution time can better reflect the difficulty of solving, this patent uses the solution time of historical operating days to train the Mamba model to predict the difficulty of solving the operating day.

[0037] Secondly, since the operation simulation needs to ensure that the system can operate stably under the most challenging situations, this patent selects the operating day with the greatest difficulty in solving each type of operating day as the typical day of that type of operating day, and uses the residual load curve of the typical day for production simulation, thereby ensuring that the load conditions of the most extreme operating days can be taken into account, thereby enhancing the robustness of the planning scheme.

[0038] S104: Solve the start-up and shutdown results of various typical day unit combinations, establish a unit combination model based on the full consumption of new energy output, and obtain the output results of each period of each operating day based on the start-up and shutdown results of various typical day unit combinations and the unit combination model, and output the output results.

[0039] Among them, the unit commitment (UC) problem is usually a mixed integer programming problem (MIP), which can be solved using solvers such as CPLEX and Gurobi. The goal of UC is to determine the start and stop status of the unit in each period.

[0040] Economic Dispatch (ED) is based on the start and stop status of the units determined by UC, and ED further optimizes the output of each unit to minimize the power generation cost. ED is usually a linear programming problem (LP) that can be solved using the same solver.

[0041] Commonly used tools include MATLAB combined with Yalmip and CPLEX / Gurobi. For example, in MATLAB, define decision variables (such as unit status u and output p), set the objective function and constraints, and then call the solver to obtain the optimal solution. After the solution is completed, the unit output results for each operating day and each period (i.e., the unit output plan for each period) are obtained. These results can be used to guide actual operation to ensure that the system achieves optimal economic performance while meeting load demand.

[0042] In view of the requirement that the power planning model needs to select appropriate typical days to improve the efficiency of operation simulation calculations, the present invention establishes a method for selecting typical days for power system planning based on historical operation simulation data. The present invention regards the operating days with the same unit start and stop in each time period as the same type of operating days, and represents them with a classification label. The operating day with the greatest difficulty in solving this type of operating day is selected as the typical day for unit combination solution, so as to ensure the robustness of the typical day. The unit start and stop results of the typical day are used as the unit start and stop results of all operating days of this type. Based on the start and stop results, the residual load curve of each operating day is used to solve the safety constrained economic dispatch model, and the unit output in each time period of each operating day is obtained as the final result of the operation simulation.

[0043] The classification labels and solution difficulty of the running day are output through the Mamba model. The Mamba model can predict the classification labels and solution difficulty of each running day based on the time series feature matrix of each running day by autonomously learning the running simulation data of historical running days.

[0044] In an exemplary embodiment of the present invention, the difficulty of solving the classification labels of each operating day based on the output of the Mamba model includes: Construct a time series feature matrix for all historical operating days, wherein the first dimension of the time series feature matrix is ​​the load curve, the second dimension of the time series feature matrix is ​​the wind power output curve, and the third dimension of the time series feature matrix is ​​the photovoltaic output curve; construct a label vector for all historical operating days, wherein the first dimension of the label vector is the classification label, and the second dimension is the solution time; train a Mamba model with the time series feature matrix of the historical operating days as input and the label vector of the historical operating days as output, and obtain the solution time in the label vector of each historical operating day through the trained Mamba model, and take the operating day with the longest solution time as the operating day with the greatest solution difficulty.

[0045] In an exemplary embodiment of the present invention, establishing a unit combination model includes: Establish the objective function of the unit commitment model and set the constraints of the objective function; The constraints include load balance constraints, system positive reserve constraints, system negative reserve constraints, unit output upper and lower limit constraints, unit ramp rate constraints and unit minimum continuous start and stop time constraints.

[0046] Specifically, the objective function of establishing the unit commitment model includes: A functional formula is constructed with the goal of minimizing the total cost of system operation, and the functional formula includes:

[0047] In the formula, For a typical day The thermal power output cost of unit i in period t, N is the total number of units, For a typical day The startup cost of unit i in time period t, where T is the number of time periods in a day.

[0048] Load balancing constraints include:

[0049] Where: For a typical day Medium thermal power unit In the period of efforts, For a typical day The total system load, For a typical day China New Energy contributes.

[0050] System positive and standby constraints include:

[0051] The system negative reserve constraints include:

[0052] In the formula, For a typical day Medium thermal power unit In the period The start and stop status of the typical daily unit combination, For a typical day Medium thermal power unit In the period The maximum output, For a typical day Medium thermal power unit In the period The minimum output, For a typical day Mid-session of the system's positive spare capacity, For a typical day Mid-session The system negative spare capacity.

[0053] The upper and lower limits of unit output include: .

[0054] The unit ramp rate constraints include:

[0055] ;

[0056] ; In the formula, For thermal power units Maximum climbing rate; For thermal power units Maximum downhill climbing rate, For a typical day Medium thermal power unit In the period The maximum output, For a typical day Medium thermal power unit In the period The start and stop status.

[0057] The minimum continuous start and stop time constraints of the unit include:

[0058] Where: For a typical day Medium thermal power unit In the period Continuous power-on time; For a typical day Medium thermal power unit In the period Continuous downtime. is the minimum continuous start time of the unit, It is the minimum continuous downtime of the unit.

[0059] In a second aspect, the present invention further provides a typical day selection system for power system planning based on operation simulation, comprising: The data processing module is configured to obtain historical operating day data and construct a residual load curve for the historical operating day; solve the unit combination according to the residual load curve for the historical operating day to obtain the unit start and stop results for each operating day, and classify and label the historical unit start and stop results to obtain the classification labels for each operating day; The model operation model is configured to output the classification labels and solution difficulties of each operating day based on the Mamba model and the classification labels of each operating day, and obtain the operating day with the greatest solution difficulty among various operating days as the typical day, solve the start and stop results of the unit combination of various typical days, establish a unit combination model based on the full consumption of new energy output, and obtain the output results of each period of each operating day based on the start and stop results of the unit combination of various typical days and the unit combination model, and output the output results; The main control module is connected to the data processing module and the model operation model, and is used to execute the above-mentioned method for selecting typical days for power system planning based on operation simulation.

[0060] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0061] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. The computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods of various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0062] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A typical day selection method for power system planning based on operation simulation, characterized in that: include: Obtain historical operating day data and construct the historical operating day residual load curve; Solve the unit combination according to the residual load curve of the historical operation day, obtain the unit start and shutdown results of each historical operation day, and classify and mark the unit start and shutdown results of the historical operation day to obtain the classification label of each historical operation day; Based on the Mamba model and the classification labels of each historical operating day, the difficulty of solving the classification labels of each historical operating day is output, and the operating day with the greatest difficulty in solving among various types of operating days is obtained as a typical day; Solve the start-up and shutdown results of various typical day unit combinations, establish a unit combination model based on the full consumption of new energy output, and obtain the output results of each period of each operating day based on the start-up and shutdown results of various typical day unit combinations and the unit combination model, and output the output results.

2. A typical day selection method for power system planning based on operation simulation according to claim 1, characterized in that: The output of the classification labels of each operating day and the difficulty of solving the classification labels based on the Mamba model and the classification labels of each historical operating day includes: Constructing a time series characteristic matrix of all historical operation days, wherein the first dimension of the time series characteristic matrix is ​​a load curve, the second dimension of the time series characteristic matrix is ​​a wind power output curve, and the third dimension of the time series characteristic matrix is ​​a photovoltaic output curve; Constructing a label vector for all historical running days, wherein the first dimension of the label vector is the classification label, and the second dimension is the solution time; The Mamba model is trained with the time series feature matrix of historical operating days as input and the label vector of historical operating days as output. The solution time in the label vector of each historical operating day is obtained through the trained Mamba model, and the historical operating day with the longest solution time is taken as the operating day with the greatest solution difficulty.

3. A typical day selection method for power system planning based on operation simulation according to claim 1, characterized in that: The establishing of the unit combination model comprises: Establish the objective function of the unit commitment model and set the constraints of the objective function; The constraints include load balance constraints, system positive reserve constraints, system negative reserve constraints, unit output upper and lower limit constraints, unit ramp rate constraints and unit minimum continuous start and stop time constraints.

4. A typical day selection method for power system planning based on operation simulation according to claim 3, characterized in that: The objective function of establishing the unit commitment model includes: A functional formula is constructed with the goal of minimizing the total cost of system operation, and the functional formula includes: In the formula, For a typical day The thermal power output cost of unit i in period t, N is the total number of units, For a typical day The startup cost of unit i in time period t, where T is the number of time periods in a day.

5. A typical day selection method for power system planning based on operation simulation according to claim 4, characterized in that: The load balancing constraints include: Where: For a typical day Medium thermal power unit In the period of efforts, For a typical day Total system load, For a typical day China New Energy contributes.

6. A typical day selection method for power system planning based on operation simulation according to claim 5, characterized in that: The system positive and standby constraints include: The system negative reserve constraints include: In the formula, For a typical day Medium thermal power unit In the period The start and stop status of the unit combination on a typical day, For a typical day Medium thermal power unit In the period The maximum output, For a typical day Medium thermal power unit In the period The minimum output, For a typical day Mid-session of the system's positive spare capacity, For a typical day Mid-session The system negative spare capacity.

7. A typical day selection method for power system planning based on operation simulation according to claim 6, characterized in that: The upper and lower limits of the unit output constraints include: 。 8. A typical day selection method for power system planning based on operation simulation according to claim 7, characterized in that: The unit climbing rate constraints include: ； ； In the formula, For thermal power units Maximum climbing rate; For thermal power units Maximum downhill climbing rate, For a typical day Medium thermal power unit In the period The maximum output, For a typical day China Thermal Power Plant In the period The start and stop status.

9. A typical day selection method for power system planning based on operation simulation according to claim 8, characterized in that: The minimum continuous start and stop time constraints of the unit include: Where: For a typical day Medium thermal power unit In the period Continuous power-on time; For a typical day Medium thermal power unit In the period Continuous downtime. is the minimum continuous start time of the unit, It is the minimum continuous downtime of the unit.

10. A typical day selection system for power system planning based on operation simulation, characterized in that: include: A data processing module is configured to obtain historical operation day data and construct a historical operation day residual load curve; Solve the unit combination according to the residual load curve of the historical operation day, obtain the unit start and stop results of each historical operation day, and classify and mark the unit start and stop results of the historical operation day to obtain the classification label of each historical operation day; The model operation model is configured to output the difficulty of solving the classification labels of each operating day based on the Mamba model and the classification labels of each operating day, and obtain the operating day with the greatest difficulty in solving among various operating days as the typical day, solve the start-up and shutdown results of the unit combination of various typical days, establish a unit combination model based on the full consumption of new energy output, and solve the output results of each period of each operating day based on the start-up and shutdown results of the unit combination of various typical days and the unit combination model, and output the output results; A main control module is connected to the data processing module and the model operation model, and is used to execute a typical day selection method for power system planning based on operation simulation as described in any one of claims 1-9.

Images