Power grid dispatching method, electronic equipment, readable medium and program product

Abstract

The invention relates to the technical field of power grid dispatching, and discloses a power grid dispatching method, electronic equipment, a readable medium and a program product. The power grid dispatching method comprises the following steps: acquiring unit information of a unit, and modeling based on the unit information to obtain a power grid dispatching model; dividing the power grid dispatching model into K sub-models, and deleting at least part of operation data in the (i + 1) th sub-model to the Kth sub-model according to initial operation information and set constraints of an ith sub-model unit in the K sub-models to obtain an ith simplified dispatching model which is in time decoupling with other sub-models. Solving the ith simplified scheduling model to obtain an initial solution of the ith time period; and according to the initial solutions of the first to Kth time periods, solving the power grid dispatching model to obtain a final solution. Thus, time decoupling is carried out on the sub-model in each time period and the sub-models in other time periods, so that the scale of the sub-model in each time period is smaller, the sub-model can be solved independently, and the solving speed of the power grid dispatching model is increased.

Description

Technical Field

[0001] This application relates to the field of power grid dispatching technology, and in particular to a power grid dispatching method, electronic device, readable medium, and program product. Background Technology

[0002] To achieve structural optimization of power generation resources during power system operation and reserve sufficient optimization space for power generation resource scheduling, unit commitment (UC) can be determined based on the power grid information, enabling the power system to generate electricity based on UC. The power grid is a crucial component of the power system, and the operation of the power system depends on the physical connectivity and transmission capacity of the power grid. UC refers to the combination of generating units that minimizes the operating costs of the power system in each time period within a given planning period, while satisfying power system demand, spinning reserve constraints, and production process constraints, based on their on / off status and power generation.

[0003] Currently, for security-constrained unit commitments (SCUCs), determining the SCUC first requires obtaining grid information and then modeling the SCUC problem into a standardized mathematical model based on this information. The standardized mathematical model is then solved. The SCUC problem refers to the combination of generating units that minimizes the operating cost of the power system in each time period within a given planning period, while satisfying the power system's load demand and network security requirements. However, the standardized mathematical models are very large, often containing hundreds of thousands of variables, and the solution process is time-consuming, making it difficult to find a feasible solution in a short period. Summary of the Invention

[0004] To improve the solution speed of standardized mathematical models, this application provides a power grid dispatching method, electronic device, readable medium, and program product.

[0005] In a first aspect, embodiments of this application provide a power grid dispatching method. The power grid includes multiple generating units. The method includes: establishing a power grid dispatching model based on the generating unit information of the multiple generating units, the power grid dispatching model being used to determine the operating information of the multiple generating units within a set time period under set constraints; dividing the power grid dispatching model into K sub-models according to time, wherein the i-th sub-model corresponds to the i-th sub-time period in the set time period, K is a positive integer greater than or equal to 2, and i is an integer between 1 and K; based on the initial operating information of the generating units in the i-th sub-model and the set constraints, at least some of the operating data in the (i+1)-th to K-th sub-models are reduced to obtain the i-th simplified dispatching model; solving the i-th simplified dispatching model to obtain the initial solution of the power grid dispatching model in the i-th time period; and solving the power grid dispatching model based on the initial solution of the power grid dispatching model in the 1st to Kth time periods to obtain the final solution of the power grid dispatching model.

[0006] It is understandable that runtime data includes variables and constraints, and the variables include the switch variable u. g (t), start-stop variable x g (t), power variable p g (t) and power generation cost variables Startup cost variable S g,up (t) and shutdown cost variable S g,down Constraints include upper and lower bounds of generator power, generator generation cost constraints, unit ramp-up constraints, minimum start-up and shutdown time, system-level power balance constraints, and spinning reserve constraints R(t), etc.

[0007] In this embodiment of the application, by decoupling the sub-models in each time period from the sub-models in other time periods, a series of smaller sub-models can be obtained. The obtained sub-models are easier to solve. Furthermore, in the process of solving the power grid scheduling model based on the initial solutions of the power grid scheduling model in the first to the Kth time periods, the branch and bound tree corresponding to the power grid scheduling model can be pruned quickly based on the initial solutions of the first to the Kth time periods, or an exploratory solution can be performed on this basis, thereby achieving the effect of accelerating the solution.

[0008] In one possible implementation, at least a portion of the running data includes running data that has no effect on the solution of the initial solution for the i-th time period.

[0009] It is understandable that the operating data that has no impact on the solution of the initial solution in the i-th time period includes variables and constraints that are unrelated to the switching variables. Among them, variables unrelated to the switching variables include power variables and power generation cost variables, and constraints unrelated to the switching variables include the upper and lower bounds of generator power.

[0010] In this embodiment of the application, by deleting the running data that has no effect on the solution result of the initial solution in the i-th time period, the number of variables that need to be solved can be reduced, and the solution efficiency of the sub-model can be improved.

[0011] In one possible implementation, based on the initial operating information of the units in the i-th sub-model and the set constraints, at least some of the operating data in the (i+1)-th to K-th sub-models are reduced to obtain the i-th simplified scheduling model. This includes: obtaining the time coupling constraints of each unit in the i-th sub-model in the set constraints; determining the influence range of each unit in the i-th sub-model based on the initial operating information of the units and the time coupling constraints; deleting the operating data outside the influence range of each unit in the i-th sub-model from the (i+1)-th to K-th sub-models; and deleting the operating data within the influence range of each unit in the i-th sub-model from the (i+1)-th to K-th sub-models that has no impact on the solution results of the initial solution of the units in the i-th sub-model, thereby obtaining the i-th simplified scheduling model.

[0012] In this embodiment of the application, by deleting the running data that has no effect on the solution result of the initial solution of the i-th time period, the number of variables that need to be solved can be reduced and the solution efficiency of the sub-model can be improved.

[0013] In one possible implementation, the time coupling constraint includes at least one of minimum power-on / off time and ramp constraint.

[0014] One possible implementation includes: corresponding to the fact that the influence range of each unit in the i-th sub-model is not completely the same, obtaining the power demand value of the power grid and the minimum start-up time of each unit in the i-th sub-model; based on the power demand value and the minimum start-up time of each unit in the i-th sub-model, recovering at least part of the operating data outside the influence range of some units in the (i+1)-K-th sub-models, to obtain the i-th simplified scheduling model.

[0015] It is understood that the influence range of each unit in the i-th sub-model is not exactly the same, and the output of each unit in the i-th sub-model should meet the overall demand of the power grid system. In this embodiment, in order to ensure the normal operation of the power grid system, the power demand value and minimum start-up time of the power grid can be obtained; based on the power demand value and minimum start-up time, at least part of the operating data outside the influence range of some units in the (i+1)-K-th sub-models can be recovered to obtain the i-th simplified scheduling model. In this way, the stability of the power grid and the continuity of power supply can be guaranteed.

[0016] In one possible implementation, based on the power demand value and the minimum start-up and shutdown time of each unit in the i-th sub-model, at least some of the operating data outside the influence range of some units in the (i+1)-K-th sub-models is recovered to obtain the i-th simplified scheduling model. This includes: corresponding to the influence range of the n-th unit in the i-th sub-model at time t being smaller than the influence range of at least one unit in the i-th sub-model, obtaining the maximum power of the n-th unit; based on the power demand value and the minimum start-up and shutdown time of each unit in the i-th sub-model, at least some of the operating data outside the influence range of some units in the (i+1)-K-th sub-models is recovered; based on the power demand value and the maximum power of the n-th unit, at least some of the operating information of some units at time t is deleted to obtain the i-th simplified scheduling model.

[0017] It is understandable that recovering at least part of the operating data outside the influence range of some units in the (i+1)th to Kth sub-models based on spinning reserve constraints may be too much. In this embodiment, the influence range of the nth unit in the i-th sub-model at time t is smaller than the influence range of at least one unit in the i-th sub-model, so the maximum power of the nth unit can be obtained. Based on the power demand value and the maximum power of the nth unit, at least part of the operating information of some units at time t is deleted, thereby reducing the variables that need to be solved.

[0018] In one possible implementation, based on the initial operating information of the units in the i-th sub-model and the set constraints, at least some of the operating data in the (i+1)-th to K-th sub-models are reduced to obtain the i-th simplified scheduling model. This includes: determining the values ​​of some variables in the i-th sub-model based on the initial operating information of the units in the i-th sub-model and the set constraints, and reducing at least some of the operating data in the (i+1)-th to K-th sub-models to obtain the i-th simplified scheduling model.

[0019] In this embodiment, the values ​​of some variables in the i-th sub-model are determined based on the initial operating information of the unit in the i-th sub-model and the set constraints, thereby reducing the number of variables that need to be solved.

[0020] In one possible implementation, based on the initial operating information of the unit in the i-th sub-model and the set constraints, the values ​​of some variables in the i-th sub-model are determined, including: based on the initial solutions of the first to i-1 time periods and the constraints that affect the solution results of the initial solutions of the unit in the i-th sub-model, the values ​​of some variables in the i-th sub-model are determined.

[0021] In one possible implementation, solving the i-th simplified scheduling model to obtain the initial solution of the power grid scheduling model in the i-th time period includes: obtaining m sets of initial solutions for the first to (i-1)-th time periods; based on the initial solutions for the first to (i-1)-th time periods, solving the i-th simplified scheduling model to obtain p sets of initial solutions of the power grid scheduling model in the i-th time period; and determining the m sets of initial solutions of the power grid scheduling model in the i-th time period based on the superiority and difference of the p sets of initial solutions.

[0022] It is understood that, in the embodiments of this application, m sets of solutions in the p sets of solutions for the i-th time period can be determined based on the goodness and difference, and the m sets of solutions can be used as the initial solutions for the i-th time period, thereby reducing the input of the initial solutions for the (i+1)-th time period and reducing the possibility of needing to explore.

[0023] In one possible implementation, the power grid dispatch model is solved based on the initial solutions of the power grid dispatch model in the first to the Kth time periods to obtain the final solution of the power grid dispatch model. This includes: determining the final solution of the Kth time period based on the goodness and difference of the initial solution in the Kth time period; determining the final solution of the power grid dispatch model in the first to the (K-1)th time periods based on the final solution of the Kth time period; and solving the power grid dispatch model based on the final solution of the power grid dispatch model in the first to the Kth time periods to obtain the final solution of the power grid dispatch model.

[0024] In a second aspect, embodiments of this application provide an electronic device, including: a memory for storing instructions executed by one or more processors of the electronic device, and a processor, which is one of the one or more processors of the electronic device, for implementing any of the power grid dispatching methods provided by the first aspect and various possible implementations of the first aspect.

[0025] Thirdly, embodiments of this application provide a readable medium storing instructions that, when executed on an electronic device, cause the electronic device to implement any of the power grid dispatching methods provided by the first aspect and various possible implementations of the first aspect.

[0026] Fourthly, embodiments of this application provide a computer program product, which includes computer instructions. When executed by an electronic device, the electronic device performs any of the power grid dispatching methods provided by the first aspect and various possible implementations of the first aspect. Attached Figure Description

[0027] Figure 1 A flowchart of a power grid dispatching method is shown;

[0028] Figure 2 According to an embodiment of this application, a schematic diagram of a power grid dispatching method deployment platform is shown;

[0029] Figure 3According to an embodiment of this application, a flowchart of a power grid dispatching method is shown;

[0030] Figure 4 According to an embodiment of this application, a flowchart of a time decoupling method is shown;

[0031] Figure 5 According to an embodiment of this application, a timing diagram is shown;

[0032] Figure 6 According to an embodiment of this application, another timing diagram is shown;

[0033] Figure 7 According to an embodiment of this application, yet another timing diagram is shown;

[0034] Figure 8 According to an embodiment of this application, a flowchart of an iterative solution method is shown;

[0035] Figure 9 According to an embodiment of this application, a schematic diagram of parallel solution of subproblems is shown. Detailed Implementation

[0036] For a long time, achieving safe and economical operation has been the primary goal of power systems. Determining the core function (UC) of a power system is a crucial step in developing power generation plans. UC provides a foundation for economic dispatch and safety verification of the power system, and can address the uncertainties of future loads and the volatility of renewable energy generation. Simultaneously, UC can improve the operating efficiency of high-efficiency large generating units, maximizing their operation at their optimal operating points and enhancing the economic efficiency and energy conservation of the power system. From a safety perspective, UC research can be divided into three stages: the first stage is UC without considering cybersecurity constraints; the second stage is UC considering cybersecurity constraints (or SCUC); and the third stage is UC that comprehensively considers safety constraints such as transient stability and voltage stability.

[0037] The SCUC problem is an important problem in power systems. In some implementations, the SCUC problem can be solved directly using a mathematical programming solver. For example... Figure 1 As shown, the SCUC problem is solved directly using a mathematical programming solver, including: modeling a scenario based on power grid information to obtain a standardized mathematical model, and then using a mathematical programming solver to solve the standardized mathematical model to obtain the solution. Here, the power grid is the physical network in the power system used for transmitting and distributing electrical energy, including multiple generating units. Power grid information includes the operating data of each generating unit, such as power generation limits, minimum start-up and shutdown times, generation costs, line connections, dynamic demand, and security constraints.

[0038] It is understandable that power generation limits include an upper and lower limit. The upper limit is the maximum power output of the generating unit, and the lower limit is the minimum power output of the generating unit. Minimum start-up and shutdown time refers to the shortest operating time required for a generating unit to operate from startup to complete shutdown. This minimum start-up and shutdown time is set to ensure stable operation of the unit and avoid damage to equipment caused by frequent start-ups and shutdowns. Line connections include the electrical connections between various components of the power network, such as the connection between generator sets and electrical loads, and the connections within the power grid, such as different transmission lines and distribution networks. Dynamic demand refers to the information that the power grid needs to acquire and process in real-time or near real-time during operation, including the operating status of power generation resources, power generation, and power grid operating parameters, such as power output. Safety constraints include the range of values ​​for dynamic demand, such as power generation limits and minimum start-up and shutdown times.

[0039] Among them, scenario modeling based on power grid information yields a standardized mathematical model, including: based on power grid information, writing the SCUC problem in the form of variables, constraints, and objective functions, and then modeling the SCUC problem into a standardized mathematical model through linearization; that is, transforming the SCUC problem into a mixed integer linear programming (MILP) problem.

[0040] In some embodiments, variables include binary variables and continuous variables, wherein binary variables (or 0-1 variables) include switch variables u. g (t) and start / stop variable x g (t), continuous variables include power variable p g (t) and power generation cost variables Startup cost variable S g,up (t) and shutdown cost variable S g,down (t). Constraints include upper and lower limits of generator power, generator generation cost constraints, unit ramp-up constraints, minimum start-up and shutdown times, system-level power balance constraints, and spinning reserve constraints R(t). Among these, the spinning reserve constraint R(t) refers to the reserve capacity of generator units that must be kept in operation but not at full load to ensure stable operation of the power system in the face of unexpected situations (such as sudden generator shutdowns or load surges). The spinning reserve constraint R(t) is a constraint to ensure the continuity and reliability of power supply and can be determined based on factors such as the size of the power system, load characteristics, and possible types of faults. For example, the spinning reserve constraint R(t) = 60% × the capacity of generator units at full load.

[0041] Wherein, the switch variable u g(t) represents the operating state of unit g at time t. For example, u1(t1) = 1 indicates that unit 1 is in the on state at time t1, and u1(t2) = 0 indicates that unit 1 is in the off state at time t2. Start-up / shutdown variable x g (t) represents the action of unit g to open or close at time t. For example, x1(t1) = 1 indicates that the switching state of unit 1 at time t1 is different from the switching state at time t0, such as unit 1 being opened or closed at time t1. x1(t2) = 0 indicates that the switching state of unit 1 at time t2 is the same as the switching state at time t1. It can be understood that when the switching variable u g (t) changes from the on state to the off state, the start / stop variable x g (t) also changes from the on state to the off state, and the start / stop variable x g (t) represents the switching variable u. g (t) is an auxiliary variable. In other embodiments, the switching variable u is... g (t) is the start / stop variable x g Auxiliary variables of (t).

[0042] Power variable p g (t) represents the operating power of unit g at time t. For example, p1(t1) = 500 means that the power of unit 1 at time t1 is 150kW. Generation cost variable. This represents the power generation cost of unit g at time t with a power output of p, for example, This indicates that the cost of Unit 1 at time t1, with a generating capacity of 150kW, is 3500 yuan. Start-up cost variable S g,up (t) represents the start-up cost of unit g at time t, for example, S 1,up (t1) = 4000 indicates that the start-up cost of unit 1 at time t1 is 4000 yuan; the shutdown cost variable S g,down (t) represents the shutdown cost of unit g at time t, for example, S 1,up (t2) = 4000 means that the start-up cost of unit 1 at time t2 is 2000 yuan.

[0043] In some embodiments, the objective function can be determined based on the operating cost of the power system (or the power generation operating cost), for example, objective function = operating cost of the power system.

[0044] In other embodiments, the operating costs of the power system include the generation costs of the generator sets and the start-up and shutdown costs of the generator sets, for example...

[0045]

[0046] in, This indicates the cost of generating electricity from the generator set. The generator set's start-up and shutdown cost is represented by t; t represents the current time; T represents the scheduling time corresponding to the objective function; g represents the generator set; and N represents the total number of generator sets. U represents the power generation cost of unit g at time t with a power output of p. g,t U represents the on / off state of unit g at time t. g,t-1 S represents the on / off state of unit g at time (t-1). g,up (t) represents the startup cost of unit g at time t, S g,down (t) represents the shutdown cost of unit g at time t.

[0047] It is understandable that solving for large-scale power systems can be very time-consuming. In some embodiments, the generation cost of generator sets in the objective function can be considered. Linearization is performed to reduce the power generation cost of the generator set. Linearization to the power variable p g (t) related functions.

[0048] In some embodiments, a mathematical programming solver is used to solve a standardized mathematical model to obtain a solution result, including: using a mathematical programming solver to solve the standardized mathematical model based on the branch-and-bound (B&B) method, and using a decomposition algorithm to assist in the solution process, and finally obtaining a solution result.

[0049] Branch and bound is a search and iterative method that obtains feasible solutions by branching, bounding, and pruning different branch variables and subproblems. Branching involves repeatedly dividing the entire feasible solution space into increasingly smaller subsets; bounding involves calculating a target bound for the solution set within each subset (e.g., for a minimum problem, calculating a target lower bound for the solution set within each subset); and pruning involves preventing further branching of subsets whose bounds exceed the target value of the known feasible solution set after each branch, thus avoiding ineffective searches.

[0050] Decomposition algorithms include Benders decomposition algorithm, column generation algorithm, etc. Decomposition algorithms can break down complex problems into several smaller and easier-to-solve subproblems, and can distribute the subproblems to different processors or computing nodes for parallel solving. The results of the parallel solutions are then combined to obtain the solution to the original problem. Decomposition algorithms can significantly accelerate the solution process.

[0051] For example, the Benders decomposition algorithm can be used to decompose the SCUC problem into several interrelated subproblems, and then the branch and bound method can be applied to solve each subproblem.

[0052] It's understandable that decomposition algorithms are designed for general mixed-integer programming problems. However, the variables and constraints of the standardized mathematical model obtained from modeling the SCUC problem have real physical meanings. Decomposition algorithms cannot accelerate the decomposition based on the relevant knowledge of general mixed-integer programming problems. Furthermore, the standardized mathematical model obtained from modeling the SCUC problem is very large, often containing hundreds of thousands of integer variables. This causes mathematical programming solvers to spend a long time solving the standardized mathematical model, making it difficult to find a feasible solution in a short time. Integer variables are variables that can only take integer values.

[0053] To address the problem that solving the standardized mathematical model obtained from modeling the SCUC problem using a mathematical programming solver is time-consuming and difficult to find a feasible solution in a short time, this application provides a power grid dispatching method. The method includes: acquiring unit information of multiple generating units within the power grid, and performing scenario modeling based on the unit information to obtain a power grid dispatching model. The power grid dispatching model is divided into K sub-models according to time, where the i-th sub-model corresponds to the i-th sub-period in a set time period, K is a positive integer greater than 2, and i is an integer between 1 and K. Based on the initial operating information and set constraints of the generating units in the i-th sub-model, at least some operating data in the (i+1)-th to K-th sub-models are removed to obtain an i-th simplified dispatching model decoupled from other sub-models in time. The i-th simplified dispatching model is solved to obtain the initial solution of the power grid dispatching model in the i-th time period; based on the initial solutions of the power grid dispatching model in the 1st to Kth time periods, the power grid dispatching model is solved again to obtain the final solution of the power grid dispatching model. Temporal decoupling refers to decomposing the system's behavior along a time dimension, allowing it to be developed, deployed, and maintained independently at different points in time. The runtime data includes variables and constraints, with variables including switch variables u. g (t), start-stop variable x g (t), power variable p g (t) and power generation cost variables Startup cost variable S g,up (t) and shutdown cost variable S g,down The constraints include upper and lower bounds of generator power, generator generation cost constraints, unit ramp-up constraints, minimum start-up and shutdown times, system-level power balance constraints, and spinning reserve constraints R(t). Thus, by decoupling the sub-models within each time period from those within other time periods, a series of smaller sub-models can be obtained. These smaller sub-models are easier to solve, improving the speed of solving the power grid dispatch model.

[0054] In this embodiment of the application, the power grid dispatching model is divided into K sub-models according to time, including: dividing the dispatching time corresponding to the power grid dispatching model into K sub-models according to time periods, where K is a positive integer greater than 2, and the length of the time periods corresponding to each sub-model can be the same or different.

[0055] For example, the scheduling time corresponding to the power grid scheduling model can be divided into 24 sub-models with a time period of 1 hour each. Each sub-model corresponds to a time window, resulting in 24 time windows. Specifically, time window 1 has a time interval of [00:00, 01:00], time window 2 has a time interval of [01:00, 02:00], time window 3 has a time interval of [02:00, 03:00], time window 4 has a time interval of [03:00, 04:00], and so on. The time intervals for window 5 are [04:00, 05:00), time intervals for window 6 are [05:00, 06:00), time intervals for window 7 are [06:00, 07:00), time intervals for window 8 are [07:00, 08:00), time intervals for window 9 are [08:00, 09:00), time intervals for window 10 are [09:00, 10:00), and time intervals for window 11 are [10:00, 11:00]. Time window 12's time interval is [11:00, 12:00); Time window 13's time interval is [12:00, 13:00); Time window 14's time interval is [13:00, 14:00); Time window 15's time interval is [14:00, 15:00); Time window 16's time interval is [15:00, 16:00); Time window 17's time interval is [16:00, 17:00); Time window 18's time interval is [11:00, 12:00]. The time intervals for time windows are [17:00, 18:00), [18:00, 19:00], [19:00, 20:00], [20:00, 21:00], [21:00, 22:00], [22:00, 23:00], and [23:00, 0:00] (17:00, 18:00), [19:00, 19:00], [20:00, 20:00], [20:00, 20:00], [21:00, 21:00], [22:00, 22:00], [23:00, 23:00], and [23:00, 0:00]).

[0056] For example, the scheduling time corresponding to the power grid scheduling model is divided into 5-hour and 7-hour time periods, resulting in four sub-models. Each sub-model corresponds to a time window, i.e., four time windows. The time interval of time window 1 is [00:00, 05:00), the time interval of time window 2 is [05:00, 12:00), the time interval of time window 3 is [12:00, 17:00], and the time interval of time window 4 is [17:00, 00:00].

[0057] In this embodiment of the application, based on the initial operating information and set constraints of the units in the i-th sub-model, at least some of the operating data in the (i+1)-th to K-th sub-models are reduced to obtain the i-th simplified scheduling model that is decoupled from other sub-models in time. This includes: determining the influence range of the i-th sub-model, deleting the operating data outside the influence range of each unit in the i-th sub-model from the (i+1)-th to K-th sub-models, and deleting the operating data within the influence range of each unit in the i-th sub-model from the (i+1)-th to K-th sub-models that has no impact on the solution results of the initial solution of the units in the i-th sub-model, thereby obtaining the i-th simplified scheduling model.

[0058] Among them, the operating data within the influence range of each unit in the i-th sub-model from the (i+1)-th to the K-th sub-models, which has no impact on the solution results of the initial solution of the units in the i-th sub-model, includes the operating data within the influence range of each unit in the i-th sub-model from the (i+1)-th to the K-th sub-models, and the switching variable u. g (t) variables and constraints unrelated to the switch variable u, where u is related to the switch variable u. g (t) Unrelated variables include the power variable p g (t), power generation cost variables With switch variable u g (t) Unrelated constraints include the upper and lower bounds of generator power.

[0059] It is understandable that the SCUC problem can also be called the unit scheduling problem. The core of the SCUC problem is to formulate a multi-period unit start-up and shutdown plan, and the main variables involved include the switching variable u. g (t) and start / stop variable x g (t). In this embodiment of the application, by deleting the switch variable u g (t) Irrelevant variables and constraints can reduce the number of variables that need to be solved and improve the efficiency of solving subproblems.

[0060] Determining the influence range of the i-th sub-model includes: determining the time coupling constraints of the units in the i-th sub-model within the set constraints; and determining the influence range of the i-th sub-model based on the initial operating information of the units in the i-th sub-model and the time coupling constraints. The time coupling constraints include at least one of minimum start-up time and ramp-up constraints.

[0061] The following section introduces the method for determining the influence range of unit 1 in the i-th sub-model, taking time coupling constraints including minimum start-up time as an example.

[0062] For example, the minimum start-up and shutdown time of Unit 1 is 3 hours. The current period is time window 2, and the time interval of time window 2 is [01:00, 02:00). The previous period is time window 1, and the time interval of time window 1 is [00:00, 01:00]. The future period includes time window 3, time window 4, time window 5, and time window 6. The time interval of time window 3 is [02:00, 03:00]. The time interval of time window 4 is [03:00, 04:00]. The time interval of time window 5 is [04:00, 05:00]. The time interval of time window 6 is [05:00, 06:00].

[0063] If the initial operating information of Unit 1 determines that Unit 1 should be started at 01:00, it will affect the decisions at 02:00 and 03:00. That is, the decision in time window 2 will affect the decisions in time windows 3 and 4. In other words, the influence of Unit 1 on future time windows within the current time period is 2 hours. If the initial operating information of Unit 1 determines that Unit 1 should be started at 02:00, it will affect the decisions at 03:00 and 04:00. That is, the decision in time window 3 will affect the decisions in time windows 4 and 5. In other words, the influence of the decision on Unit 1 within time window 3 on future time windows is 2 hours.

[0064] For example, the minimum start-up and shutdown time of Unit 1 is 3 hours. The current period is time window 2, and the time interval of time window 2 is [04:00, 08:00). The previous period is time window 1, and the time interval of time window 1 is [00:00, 04:00). The future period includes time window 3, and the time interval of time window 3 is [08:00, 12:00].

[0065] If the initial operating information of Unit 1 determines that Unit 1 should be started at 05:00, it will affect the decisions at 06:00 and 07:00. That is, the decision in time window 2 will affect the decisions before 08:00 within time window 3. In other words, the influence range of Unit 1 on future time periods within the current time period is 0 hours. If the initial operating information of Unit 1 determines that Unit 1 should be started at 06:00, it will affect the decisions at 07:00 and 08:00. That is, the decision in time window 2 will affect the decision at 08:00 within time window 3. In other words, the influence range of Unit 1 on future time windows within the current time period is 1 hour. If Unit 1 is started at 07:00, it will affect the decisions at 08:00 and 09:00. That is, the decision in time window 2 will affect the decisions at 08:00 and 09:00 within time window 3. In other words, the influence range of Unit 1 on future time periods within the current time period is 2 hours.

[0066] It is understood that, regardless of when unit 1 in the i-th sub-model is turned on at any time in the current period, the range of influence of unit 1 on the future period is less than the time determined based on the time coupling constraint. Therefore, in some embodiments, the range of influence of each unit in the i-th sub-model can be determined according to the time coupling constraint of each unit in the i-th sub-model.

[0067] It is understood that the influence range of each unit in the i-th sub-model is not exactly the same, and the output of each unit in the i-th sub-model should meet the overall demand of the power grid system. In this embodiment, in order to ensure the normal operation of the power grid system, the spinning reserve constraint and minimum start-up time of the power grid can be obtained; based on the spinning reserve constraint and minimum start-up time, at least part of the operating data outside the influence range of some units in the (i+1)-K-th sub-models can be recovered to obtain the i-th simplified scheduling model. In this way, the stability of the power grid and the continuity of power supply can be guaranteed.

[0068] For example, based on spinning reserve constraints (spatial constraints between units). The minimum power-on / off time restores the switching variable u outside the scope of the current time window's influence on future time windows. g (t) and start / stop variable x g (t). Where R(t) represents the spinning reserve demand of the power grid system, and R(t) is a fixed value. This indicates the upper limit of the unit's power. This indicates that the total power of the generating units exceeds the spinning reserve requirement of the power grid system.

[0069] It is understandable that recovering at least part of the operating data outside the influence range of some units in the (i+1)th to Kth sub-models based on spinning reserve constraints may be too much. In this embodiment, the influence range of the nth unit in the i-th sub-model at time t is smaller than the influence range of at least one unit in the i-th sub-model, so the maximum power of the nth unit can be obtained. Based on the power demand value and the maximum power of the nth unit, at least part of the operating information of some units at time t is deleted, thereby reducing the variables that need to be solved.

[0070] In this embodiment of the application, the values ​​of some variables in the i-th sub-model can also be determined based on the initial operating information and set constraints of the unit in the i-th sub-model. For example, the values ​​of some variables in the i-th sub-model can be determined based on the initial solutions of the first to (i-1)th time periods and the constraints that affect the solution results of the initial solutions of the unit in the i-th sub-model.

[0071] In this embodiment of the application, solving the i-th simplified scheduling model to obtain the initial solution of the power grid scheduling model in the i-th time period includes: obtaining m sets of initial solutions for the first to (i-1)-th time periods; based on the m sets of initial solutions for the first to (i-1)-th time periods, solving the i-th simplified scheduling model to obtain p sets of initial solutions of the power grid scheduling model in the i-th time period, where p = m × q, and q is the solution obtained by solving the subproblem created based on each set of solutions; and determining the m sets of initial solutions of the power grid scheduling model in the i-th time period based on the superiority and difference of the p sets of initial solutions.

[0072] It is understandable that, in the process of solving the sub-model within each time period based on beam search, the p initial solutions of the sub-model within each time period can be sorted according to goodness and dissimilarity, and m initial solutions (or feasible solutions) from the p feasible solutions can be retained, where m ≤ p; then, the sub-model of the next time period can be solved based on the m initial solutions. Here, m and p can be set according to the parameters of the mathematical programming solver.

[0073] For example, if m=2, after obtaining the solutions A, B, and C for the current time window based on the beam search, if it is determined that the difference between A and B is small and the goodness of A is higher than that of B, then solutions A and C are retained, and the next time window sub-model is solved based on A and C.

[0074] In the process of solving the sub-model of the current time period, the initial solutions obtained in this embodiment can be screened again based on their goodness and difference, thereby reducing the number of initial solutions to be input into the next time window, reducing the scale that needs to be solved in the next time window, and ensuring exploration efficiency.

[0075] In this embodiment, for solutions that require traversing all possibilities (e.g., solutions that cannot be obtained through linear programming), a mathematical programming solver can be used to solve the power grid dispatch model based on the initial solutions of the power grid dispatch model in the first to Kth time periods, thus ensuring the accuracy of the final solution of the power grid dispatch model. Furthermore, during the process of the mathematical programming solver using the branch and bound method to solve the power grid dispatch model, the branch and bound tree corresponding to the power grid dispatch model can be quickly pruned based on the initial solutions in the first to Kth time periods, or an exploratory solution can be performed based on this, achieving the effect of accelerating the solution process.

[0076] The power grid dispatching method provided in this application can be applied to… Figure 2 The operations research optimization platform shown can be understood to provide scenario optimization solution services and time-decoupled iterative solution algorithms. It can also acquire anonymized data, solution results, and scenario models obtained from the time-decoupled iterative solution algorithms, and upload these data to the storage service.

[0077] The scenario optimization solution service includes data modeling service and mathematical programming solver service. The data modeling service includes scenario modeling based on power grid data to obtain a standardized mathematical model (or power grid dispatch model). The power grid data includes unit information of multiple units within the power grid. The mathematical programming solver service includes solving the standardized mathematical model through a mathematical programming solver to obtain the calculation results.

[0078] The time-decoupled iterative solution algorithm includes a data preprocessing module, a problem decomposition module, and an iterative solution module. The data preprocessing module is used to preprocess the data to obtain preprocessed data, such as data anonymization processing to obtain anonymized data. The problem decomposition module is used to break down the problem in the standardized mathematical model to obtain time-decoupled subproblems (or simplified scheduling models). The iterative solution module is used to solve the time-decoupled subproblems to obtain high-quality feasible solutions (or initial solutions).

[0079] In some embodiments, users can upload power grid information to the scenario optimization solution service. The mathematical modeling service within the scenario optimization solution service then performs scenario modeling based on this power grid information, obtaining a standardized mathematical model. The standardized mathematical model is then decoupled in time using a problem decomposition module within a time-decoupled iterative solution algorithm, resulting in a series of smaller sub-problems. These smaller sub-problems are then solved using an iterative solution module to obtain high-quality feasible solutions, which are then returned to the scenario optimization solution service. The mathematical programming solver service within the scenario optimization solution service can then solve the standardized mathematical model again based on the high-quality feasible solutions, obtaining computational results, which are then returned to the user.

[0080] In other embodiments, users may also request the operations research platform to invoke a time-decoupled iterative solution algorithm. The problem decomposition module in the time-decoupled iterative solution algorithm decouples the user-input model in time, resulting in a series of smaller sub-problems. The iterative solution module then solves the series of smaller sub-problems to obtain a high-quality feasible solution, which is then returned to the user.

[0081] In this embodiment of the application, the time-decoupled iterative solution algorithm can also preprocess the power grid data in the standardized mathematical model through the data preprocessing module, such as desensitization processing, and upload the preprocessed data to the operations research and optimization platform.

[0082] In this embodiment of the application, a time-decoupled iterative solution algorithm is used to decouple the standardized mathematical model or the user-input model in time, and the original problem is broken down into a series of smaller sub-problems, which can accelerate the solution process of the original problem and improve the solution speed of the original problem.

[0083] It is understood that the technical solution of this application can be applied not only to power grid scenarios but also to other time-series scheduling-related scenarios, such as factory production scheduling, traffic control, and medical equipment control. The problem decomposition module and iterative solution module provided in this application can decompose the problem according to the characteristics of the scenario during the problem-solving process in other scenarios, thereby reducing the scale of the problem to be solved.

[0084] The following example uses a power grid scenario, combined with... Figure 3 The power grid dispatching method provided in the embodiments of this application is described. Among them, Figure 3 The power grid dispatching method shown is similar to Figure 1 The difference in the power grid dispatching methods shown lies in the fact that, after obtaining the standardized mathematical model, it first performs time decoupling on the standardized mathematical model to obtain a high-quality feasible solution. Then, the high-quality feasible solution is input into a mathematical programming solver, which solves the standardized mathematical model again based on the high-quality feasible solution to obtain the target feasible solution. For example... Figure 3 As shown, the power grid dispatching method includes:

[0085] 101: Power Grid Information.

[0086] In this embodiment, users can upload power grid information to the scenario optimization solution service, and then use the mathematical modeling service within the scenario optimization solution service to model based on the power grid information. The power grid information includes unit information for multiple generating units within the power grid, such as power generation limits, minimum start-up and shutdown times, generation costs, line connections, dynamic demand, and safety constraints.

[0087] 102: Scene Modeling.

[0088] In this embodiment of the application, the mathematical modeling service in the scenario optimization solution service can be used to write the SCUC problem into the form of variables, constraints and objective function based on the input power grid information.

[0089] The variables include binary variables and continuous variables. Binary variables (or 0-1 variables) include the switch variable u. g (t) and start / stop variables x g (t), continuous variables include power variable p g (t) and electricity generation cost variables Startup cost variable S g,up (t), shutdown cost variable S g,down (t).

[0090] It is understandable that the switch variable u g(t) represents the operating state of unit g at time t. For example, u1(t1) = 1 indicates that unit 1 is in the on state at time t1, and u1(t2) = 0 indicates that unit 1 is in the off state at time t2. Start-up / shutdown variable x g (t) represents the action of unit g to open or close at time t. For example, x1(t1) = 1 indicates that the switching state of unit 1 at time t1 is different from the switching state at time t0, such as unit 1 being opened or closed at time t1. x1(t2) = 0 indicates that the switching state of unit 1 at time t2 is the same as the switching state at time t1. It can be understood that when the switching variable u g (t) changes from the on state to the off state, the start / stop variable x g (t) also changes from the on state to the off state, and the start / stop variable x g (t) represents the switching variable u. g (t) is an auxiliary variable. In other embodiments, the switching variable u is... g (t) is the start / stop variable x g Auxiliary variables of (t).

[0091] Power variable p g (t) represents the operating power of unit g at time t. For example, p1(t1) = 500 means that the power of unit 1 at time t1 is 150kW. Generation cost variable. This represents the power generation cost of unit g at time t with a power output of p, for example, This indicates that the cost of Unit 1 at time t1, with a generating capacity of 150kW, is 3500 yuan. Start-up cost variable S g,up (t) represents the start-up cost of unit g at time t, for example, S 1,up (t1) = 4000 indicates that the start-up cost of unit 1 at time t1 is 4000 yuan; the shutdown cost variable S g,down (t) represents the shutdown cost of unit g at time t, for example, S 1,up (t2) = 4000 means that the start-up cost of unit 1 at time t2 is 2000 yuan.

[0092] Constraints include upper and lower limits for generator power, generator generation cost constraints, unit ramp-up constraints, minimum start-up and shutdown times, system-level power balance constraints, and spinning reserve constraints R(t). Among these, the spinning reserve constraint R(t) is the reserve capacity of generators that must be maintained in operation but not at full load to ensure the power system can maintain stable operation in the face of unexpected situations (such as sudden generator shutdowns or load surges). The spinning reserve constraint R(t) is a constraint to ensure the continuity and reliability of power supply and can be determined based on factors such as the size of the power system, load characteristics, and possible types of faults. For example, the spinning reserve constraint R(t) = 60% × the capacity of generators at full load.

[0093] In this embodiment of the application, the objective function can be determined based on the operating cost of the power system (or the power generation operating cost). The operating cost of the power system includes the power generation cost of the generator sets and the start-up and shutdown cost of the generator sets, i.e.

[0094] in, This indicates the cost of generating electricity from the generator set. The generator set's start-up and shutdown cost is represented by t; t represents the current time; T represents the scheduling time corresponding to the objective function; g represents the generator set; and N represents the total number of generator sets. U represents the power generation cost of unit g at time t with a power output of p. g,t U represents the on / off state of unit g at time t. g,t -1 indicates the on / off state of unit g at time (t-1), S g,up (t) represents the startup cost of unit g at time t, S g,down (t) represents the shutdown cost of unit g at time t.

[0095] 103: Standardized mathematical model.

[0096] In the embodiments of this application, the variables, constraints and objective function can be modeled into a standardized mathematical model by linearization, that is, the SCUC problem is transformed into a mixed integer programming problem.

[0097] In some embodiments, for a polynomial objective function, the objective function can be linearized by introducing new variables and constraints; for a nonlinear objective function, the objective function can be linearized by introducing auxiliary variables; for a nonlinear objective function, and where the nonlinear constraints can be represented as implicit functions, the implicit function theorem can be used to transform the constraints, thereby linearizing the objective function.

[0098] It is understandable that linear programming has mature theories and algorithms, and most optimization software and solvers currently support linear problems, while nonlinear problems require specific algorithms; that is, linear problems are easier to solve than nonlinear problems. This application improves the speed and accuracy of solving nonlinear or complex problems by transforming them into linear problems.

[0099] It is understood that the standardized mathematical model obtained from the embodiments of this application includes hundreds of thousands of variables and constraints.

[0100] 104: Determine whether it is an iterative solution.

[0101] In this embodiment of the application, the user can determine whether to perform iterative solution on the standardized mathematical model. If the determination result is yes, then execute 105: time window division; if the determination result is no, then execute 109: mathematical programming solver.

[0102] In some other embodiments, 104: whether to solve iteratively can be deleted, and 105: time window division can be executed after 103 is completed.

[0103] 105: Time window division.

[0104] As can be understood, time window partitioning refers to dividing a standardized mathematical model into K sub-problems according to time. The i-th sub-problem corresponds to the i-th time window (or sub-segment) within a given time period, where K is a positive integer greater than or equal to 2, and i is an integer between 1 and K. For example, the scheduling time corresponding to the standardized mathematical model can be divided into K sub-problems according to time periods. The length of the time periods corresponding to each sub-problem can be the same or different.

[0105] For example, if the scheduling time corresponding to the standardized mathematical model is 24 hours, the standardized mathematical model can be divided into 4 sub-problems in 6-hour intervals or into 23 sub-problems in 1-hour intervals. It can be understood that the granularity of the time window division determines the granularity of the problem subdivision.

[0106] The following section uses a standardized mathematical model with a scheduling time of 24 hours as an example to introduce the method of dividing the time window.

[0107] For example, the scheduling time corresponding to the standardized mathematical model can be divided into 1-hour time intervals, resulting in 24 sub-problems. Each sub-problem corresponds to a time window, resulting in 23 time windows. Specifically, time window 1 has a time interval of [00:00, 01:00], time window 2 has a time interval of [01:00, 02:00], time window 3 has a time interval of [02:00, 03:00], time window 4 has a time interval of [03:00, 04:00], and so on. The time intervals for window 5 are [04:00, 05:00), time intervals for window 6 are [05:00, 06:00), time intervals for window 7 are [06:00, 07:00), time intervals for window 8 are [07:00, 08:00), time intervals for window 9 are [08:00, 09:00), time intervals for window 10 are [09:00, 10:00), and time intervals for window 11 are [10:00, 11:00]. Time window 12's time interval is [11:00, 12:00); Time window 13's time interval is [12:00, 13:00); Time window 14's time interval is [13:00, 14:00); Time window 15's time interval is [14:00, 15:00); Time window 16's time interval is [15:00, 16:00); Time window 17's time interval is [16:00, 17:00); Time window 18's time interval is [11:00, 12:00]. The time intervals for time windows are [17:00, 18:00), [18:00, 19:00], [19:00, 20:00], [20:00, 21:00], [21:00, 22:00], [22:00, 23:00], and [23:00, 00:00] (17:00, 18:00), [19:00, 19:00], [20:00, 20:00], [21:00, 21:00], [22:00, 22:00], [23:00, 23:00], and [23:00, 00:00]).

[0108] For example, the scheduling time corresponding to the standardized mathematical model can be divided into six sub-problems, each corresponding to a time window, resulting in six time windows. The time intervals of time window 1 are [00:00, 04:00), time window 2 are [04:00, 08:00), time window 3 are [8:00, 12:00), time window 4 are [12:00, 16:00), time window 5 are [16:00, 20:00], and time window 6 are [20:00, 00:00].

[0109] For example, the scheduling time corresponding to the standardized mathematical model can be divided into four sub-problems, each with a time window, in 6-hour intervals. This results in four time windows: time window 1 has a time interval of [00:00, 6:00], time window 2 has a time interval of [06:00, 12:00], time window 3 has a time interval of [12:00, 18:00], and time window 4 has a time interval of [18:00, 00:00].

[0110] For example, by dividing the scheduling time corresponding to the standardized mathematical model into 5-hour and 7-hour time periods, four sub-problems are obtained. Each sub-problem corresponds to a time window, resulting in four time windows. The time intervals of time window 1 are [00:00, 05:00), time window 2 are [05:00, 12:00], time window 3 are [12:00, 17:00], and time window 4 are [17:00, 00:00].

[0111] 106: Problem decomposition of time decoupling.

[0112] It is understandable that after dividing the scheduling time corresponding to the standardized mathematical model into time windows, the solution of the current time window sub-problem may affect the solution of the next time window sub-problem, the constraints of the next time window may affect the solution of the current time window sub-problem, and the solution and constraints of the previous time window may affect the solution of the current time window sub-problem.

[0113] The following example uses a standardized mathematical model with a scheduling time of 24 hours, a power grid consisting of generating units 1 and 2, a minimum start-up time of 3 hours for generating unit 1 and 5 hours for generating unit 2, a current time window (or the i-th time period corresponding to the i-th sub-model) as time window 2 with a time interval of [04:00, 08:00), a previous time window (or the i-1th time period corresponding to the (i-1)-th sub-model) as time window 1 with a time interval of [00:00, 04:00), and a next time window (or the i+1th time period corresponding to the (i+1)-th sub-model) as time window 3 with a time interval of [08:00, 12:00). This example illustrates the temporal coupling between the time window sub-problems of generating units 1 and 2.

[0114] For example, if Unit 1 is started at 05:00 and operates based on the minimum start-up and shutdown time, it will affect the decisions made at 06:00 and 07:00. If Unit 2 is started at 05:00 and operates based on the minimum start-up and shutdown time, it will affect the decisions made from 06:00 to 09:00. Unit 2 cannot be shut down until 10:00 at the earliest. That is, the decision made about Unit 2 at 05:00 within time window 2 will affect the decisions made about Unit 2 at 08:00 and 09:00 within time window 3. In other words, the solution to the problem in the current time window may affect the solution to the problem in the next time window.

[0115] If Unit 1 and Unit 2 are started at 05:00, and Unit 1 operates based on the minimum start-up time, then Unit 1 may be shut down after 08:00. If the sum of the output power of Unit 2 and other started units within [08:00, 10:00) does not reach the spinning reserve constraint, the stability of the power grid and the continuity of power supply cannot be guaranteed. Therefore, Unit 1 still needs to be started within [08:00, 10:00), that is, the operating time of Unit 1 should be greater than the minimum start-up time. In other words, the constraint of the next time window may affect the solution of the current time window problem. Units need to adjust their power output according to the overall demand of the power grid and the output of other units to maintain supply and demand balance.

[0116] If Unit 2 is started at 02:00 and operates based on the minimum start-up time, then Unit 2 can not be shut down until 07:00 at the earliest; if Unit 2 is started at 03:00 and operates based on the minimum start-up time, then Unit 2 can not be shut down until 08:00 at the earliest; that is, the solution and constraints of the previous time window may affect the solution of the problem in the current time window.

[0117] That is, the feasible solutions (or initial solutions) of Unit 1 and Unit 2 in time window 2 are temporally coupled with the feasible solutions of future time windows, the constraints of Unit 1 and Unit 2 in future time windows, and the feasible solutions and constraints of Unit 1 and Unit 2 in time window 1. When solving for the feasible solutions of Unit 1 and Unit 2 in time window 2, if only the feasible solutions of Unit 1 and Unit 2 in time window 2 are considered, without considering the impact of the feasible solutions of Unit 1 and Unit 2 in time window 2 on future time windows, the constraints of Unit 1 and Unit 2 in future time windows, and the feasible solutions and constraints of Unit 1 and Unit 2 in time window 1, the accuracy of obtaining the feasible solutions of time window 2 will be reduced.

[0118] In this embodiment of the application, a time-decoupled problem decomposition framework can be used to decouple a series of smaller sub-problems obtained based on "105: time window partitioning" in time. The method of decoupling a series of smaller sub-problems based on "105: time window partitioning" using this time-decoupled problem decomposition framework is... Figure 4 The corresponding implementation examples are described below.

[0119] 107: Solving iterative subproblems.

[0120] In the embodiments of this application, an iterative solution framework based on bundle search technology can be used to solve the sub-problems within each time window obtained by the time decoupling problem decomposition framework in the order of time windows, and finally form a high-quality feasible solution to the SCUC problem (or the initial solution of the power grid dispatch model in the first to the Kth time periods).

[0121] For example, the divided time windows include six time windows: time window 1 (00:00, 04:00), time window 2 (04:00, 08:00), time window 3 (8:00, 12:00), time window 4 (12:00, 16:00), time window 5 (16:00, 20:00), and time window 6 (20:00, 00:00). First, an iterative solution framework based on bundle search can be used to solve the subproblems within time window 1, obtaining p sets of solutions for the subproblems within time window 1. Then, based on m sets of solutions from the p sets of solutions for the subproblems within time window 1, the subproblems of time window 2 can be solved, and so on, until all subproblems within time windows 3 to 6 are solved.

[0122] 108: High-quality feasible solution.

[0123] In this embodiment of the application, the solution with the highest quality and difference among the feasible solutions of the Kth time period can be determined as the final solution of the Kth time period based on the quality and difference of the solutions to the subproblems of the Kth time window; the final solution of the power grid dispatch model in the 1st to K-1th time periods can be determined based on the final solution of the Kth time period; and the final solution of the power grid dispatch model in the 1st to Kth time periods can be taken as a high-quality feasible solution.

[0124] It is understandable that the final solution of the Kth time period is determined based on the feasible solutions of the power grid dispatch model in the 1st to K-1th time periods. Therefore, the final solution of the power grid dispatch model in the 1st to K-1th time periods can be determined based on the final solution of the Kth time period.

[0125] It is understandable that a high-quality feasible solution is a set of variables that satisfy the constraints (or setting conditions) of the power grid.

[0126] 109: Mathematical Programming Solver.

[0127] It is understandable that continuous variables in a standardized mathematical model (or SCUC problem) need to traverse all possibilities. To improve the accuracy of solutions for continuous variables, the continuous variables in this embodiment can be solved again based on the values ​​of the binary variables using a mathematical programming solver. For example, MATLAB, GAMS, CPLEX, Gurobi, etc., can be used to solve the problem again based on the values ​​of the binary variables to obtain the target feasible solution. The binary variables include the switch variable u. g (t) and start / stop variable x g (t).

[0128] In this embodiment, after obtaining a high-quality feasible solution, the high-quality feasible solution and the standardized mathematical model are input together into a mathematical programming solver. The mathematical programming solver then accelerates the solution based on the high-quality feasible solution. It can be understood that the standardized mathematical model contains more than 60% binary variables. With the high-quality feasible solution as input, the mathematical programming solver can quickly prune the branch and bound tree corresponding to the standardized mathematical model or explore solutions based on it, eliminating the need to traverse all possibilities, reducing the scope of exploration, and achieving a rapid solution to the original problem.

[0129] In other embodiments, a high-quality feasible solution can be directly output as the solution to the SCUC problem, or a high-quality feasible solution can be directly output as the solution to the standardized mathematical model.

[0130] 310: Solution results.

[0131] In this embodiment, power grid information is used as input, and a standardized mathematical model of the SCUC problem is obtained through power grid scenario modeling. After a given time window is divided, a series of smaller, time-decoupled subproblems are obtained through a time-decoupled problem decomposition framework. Then, the subproblems are solved through an iterative solution framework to obtain high-quality feasible solutions. The high-quality feasible solutions are then input into a mathematical programming solver for solving. In the process of solving the problem, it is no longer necessary to traverse all possibilities, thus achieving the effect of problem decomposition and solution acceleration, which can improve the solution time by up to 90%.

[0132] The following section introduces a method for dividing the SCUC problem into time-decoupled subproblems using a time-decoupled problem decomposition framework, such as... Figure 4 As shown, it includes:

[0133] 1061: Scene Model.

[0134] It is understood that the scenario model can be a standardized mathematical model obtained by modeling power grid information based on the embodiments of this application.

[0135] 1062: The previous step was a feasible solution.

[0136] It is understandable that if the current time window is the first time window, the feasible solution in the previous step is empty; if the current time window is a time window after the first time window, the feasible solution in the previous step is the feasible solution obtained in the previous time window.

[0137] 1063: Create a subproblem.

[0138] In this embodiment, the influence range of decisions made by the unit within the current time window on variables in future time windows can be determined first. Variables and constraints outside the influence range are deleted, and variables and constraints within the influence range are reduced to obtain preliminary sub-problems, thereby achieving time decoupling between the sub-problems of the current time window and the sub-problems of the future time window. The method for obtaining preliminary sub-problems is described below.

[0139] For example, the overall scheduling time interval for the SCUC problem is [0, T), the current time window is [t, t+x), and the length of the time window is x. Figure 5 The time series diagram shown is represented in white; the past time interval is [0, t). Figure 5 The time series diagram shown is represented in gray; the future time interval is [t+x, T). Figure 5 The timing diagram shown is represented in black. For example... Figure 5 As shown, the current time window is [04:00, 08:00), the past time interval is [00:00, 04:00), and the future time interval is [08:00, 00:00).

[0140] First, obtain the initial operating information and set constraints of each unit within the current time window, and determine the influence range of the sub-problems of each unit within the current time window on future time windows based on the initial operating information and set constraints of each unit within the current time window, and delete constraints and variables outside the influence range.

[0141] For example, determine Figure 5 If the influence range of the current time window A2 of Unit 1 on the future time window is the black area A3, then all constraints and variables in the future time window, excluding the black area A3, can be deleted. (Confirm) Figure 5 If the current time window B2 of Unit 2 affects the future time window within the black area B3, then all constraints and variables in the future time window, excluding the black area B3, can be deleted. (Confirm) Figure 5 If the current time window C2 of unit 3 has an influence on the future time window within the black area C3, then all constraints and variables in the future time window, excluding the black area C3, can be deleted.

[0142] In this embodiment of the application, by deleting constraints and variables outside the scope of influence in the future time window, the number of variables that need to be explored can be reduced, the amount of computation can be reduced, and the computational efficiency can be improved.

[0143] In this embodiment, the influence range of each unit's sub-problems on future time windows is determined based on the initial operating information and set constraints of each unit within the current time window (or the i-th sub-period), and constraints and variables outside the influence range are deleted. This includes: determining the influence range of each unit's sub-problems on future time windows based on the initial operating information and time coupling constraints of each unit within the current time window. The time coupling constraints include at least one of minimum start-up / start-up time and ramp-up constraints. Variables outside the influence range that are deleted include the switching variable u. g (t), power variable p g (t), power generation cost variables Constraints outside the scope of the deletion include the upper and lower limits of generator power, unit ramp-up constraints, minimum start-up and shutdown time, and spinning reserve constraints R(t), etc.

[0144] As can be understood, ramp-up constraints limit the rate of power change of a generator unit within adjacent time periods; that is, the unit's output power cannot change too rapidly within a short period. Ramp-up constraints include positive ramp-up rate and negative ramp-down rate, representing the maximum power that the unit can increase or decrease per unit time, respectively. While ramp-up constraints only control a single point in time, they may control power for a period of time, thus requiring consideration. The upper and lower bounds of generator power and the unit ramp-up constraints are related to the power variable p. g The constraints R(t) on the switching variable u include minimum start-up time and spin-up reserve constraints. g (t) and start / stop variable x g (t) constraint.

[0145] The following section uses the time coupling constraint, including the minimum start-up time, as an example to introduce the method for determining the influence range of unit 1 in the i-th sub-problem corresponding to the i-th time period.

[0146] For example, the minimum start-up and shutdown time of Unit 1 is 3 hours. The current period is time window 2, and the time interval of time window 2 is [01:00, 02:00). The previous period is time window 1, and the time interval of time window 1 is [00:00, 01:00]. The future period includes time window 3, time window 4, time window 5, and time window 6. The time interval of time window 3 is [02:00, 03:00]. The time interval of time window 4 is [03:00, 04:00]. The time interval of time window 5 is [04:00, 05:00]. The time interval of time window 6 is [05:00, 06:00].

[0147] If the initial operating information of Unit 1 determines that Unit 1 should be started at 01:00, it will affect the decisions at 02:00 and 03:00. That is, Unit 1 can be shut down at 04:00 at the earliest. In other words, the decision of time window 2 will affect the decisions of time windows 3 and 4. That is, the influence of Unit 1 on future time windows within the current time period is within 2 hours. If the initial operating information of Unit 1 determines that Unit 1 should be started at 02:00, it will affect the decisions at 03:00 and 04:00. That is, Unit 1 can be shut down at 05:00 at the earliest. That is, the decision of time window 3 will affect the decisions of time windows 4 and 5. That is, the influence of the decision of time window 3 on future time windows is within 2 hours.

[0148] For example, the minimum start-up and shutdown time of Unit 1 is 3 hours. The current period is time window 2, and the time interval of time window 2 is [04:00, 08:00). The previous period is time window 1, and the time interval of time window 1 is [00:00, 04:00). The future period includes time window 3, and the time interval of time window 3 is [08:00, 12:00].

[0149] If the initial operating information of Unit 1 determines that Unit 1 should be started at 05:00, it will affect the decisions at 06:00 and 07:00. Unit 1 can be shut down as early as 08:00. This means that the decision in time window 2 affects the decisions before time window 3 (08:00). Therefore, the influence of Unit 1 on future time periods within the current time period is 0 hours. If the initial operating information of Unit 1 determines that Unit 1 should be started at 06:00, it will affect the decisions at 07:00 and 08:00. Unit 1 can be shut down as early as 08:00. The unit can be shut down as early as 09:00, meaning that the decision in time window 2 will affect the decision at 08:00 within time window 3. In other words, the influence of unit 1 on future time windows within the current time period is 1 hour. If unit 1 is turned on at 07:00, it will affect the decisions at 08:00 and 09:00. Unit 1 can be shut down as early as 10:00, meaning that the decision in time window 2 will affect the decisions at 08:00 and 09:00 within time window 3. In other words, the influence of unit 1 on future time periods within the current time period is 2 hours.

[0150] It is understood that, regardless of when unit 1 in the i-th sub-model is turned on at any time in the current period, the range of influence of unit 1 on the future period is less than the time determined based on the time coupling constraint. Therefore, in some embodiments, the range of influence of each unit in the i-th sub-model can be determined according to the time coupling constraint of each unit in the i-th sub-model.

[0151] In this embodiment of the application, in order to further reduce the number of variables that need to be solved, constraints and variables that have no effect on the solution results of feasible solutions for units in the current time window can be deleted within the scope of the influence of the current time window on future time windows.

[0152] In some embodiments, constraints and variables that have no effect on the solution results of the feasible solution for the unit in the current time window include constraints and variables that are independent of switching variables. For example, variables independent of switching variables include the power variable p. g (t), power generation cost variables Variables related to the switching variable include the switching variable u. g (t) and start / stop variable x g (t), and the switching variable u g (t) Unrelated constraints include the upper and lower bounds of generator power.

[0153] For example, such as Figure 5 As shown, delete the variables u in black areas A3, B3, and C3. g (t) Irrelevant constraints and variables, for example, removing the upper and lower bounds of generator power.

[0154] For example, the minimum start-up and shutdown time of Unit 1 is 3 hours. The current time window is Time Window 2, and the time interval of Time Window 2 is [04:00, 08:00). The previous time window is Time Window 1, and the time interval of Time Window 1 is [00:00, 04:00). The time interval of the future time window is [08:00, 00:00].

[0155] If the initial operating information of Unit 1 determines that Unit 1 will be started at 05:00, and Unit 1 operates based on the minimum start-up and shutdown time, then Unit 1 can be shut down at the earliest at 08:00, which means that it will not affect time window 3.

[0156] If the initial operating information of Unit 1 determines that Unit 1 will be started at 06:00, and Unit 1 operates based on the minimum start-up and shutdown time, then Unit 1 can be shut down as early as 09:00. This affects the decision at 08:00 within time window 3. Therefore, the influence range of Unit 1 on future time windows within the current time window is 1 hour, or the influence range of Unit 1 on future time windows within the current time window is [08:00, 09:00). Thus, constraints and variables unrelated to the start-up and shutdown variables within [08:00, 09:00) can be deleted. For example, the power variable p can be deleted. g (t), power generation cost variables The upper and lower limits of generator power.

[0157] If the initial operating information of Unit 1 determines that Unit 1 will be started at 07:00, and Unit 1 operates based on the minimum start-up and shutdown time, then Unit 1 can be shut down as early as 10:00. This affects the decisions made at 08:00 and 09:00 within time window 3. Therefore, the influence range of Unit 1 on future time windows within the current time window is 2 hours, or the influence range of Unit 1 on future time windows within the current time window is [08:00, 10:00). Thus, constraints and variables unrelated to the start-up and shutdown variables within [08:00, 10:00) can be deleted. For example, the power variable p can be deleted. g (t), power generation cost variables The upper and lower limits of generator power.

[0158] It is understandable that the SCUC problem can also be called the unit scheduling problem. The core of the SCUC problem is to formulate a multi-period unit start-up and shutdown plan, and the main variables involved include the switching variable u. g (t) and start / stop variable x g (t). In this embodiment of the application, by deleting the switch variable u g (t) Irrelevant variables and constraints can reduce the number of variables that need to be solved and improve the efficiency of solving subproblems.

[0159] It is understood that the start-up and shutdown times of various generating units in a power grid system are not entirely the same, and the output of the power grid system should meet the overall demand of the power grid system. In this embodiment of the application, in order to ensure the normal operation of the power grid system, the spinning reserve constraint and minimum start-up time of the power grid can be obtained; based on the spinning reserve constraint and minimum start-up time, at least part of the operating data outside the influence range of some generating units in the (i+1)th to Kth sub-models can be recovered to obtain the i-th simplified scheduling model. In this way, the stability of the power grid and the continuity of power supply can be guaranteed.

[0160] For example, the spinning reserve constraint (the spatial constraint between units) can be restored first. And minimum start-up and shutdown times, then based on spinning reserve constraints (space constraints between units). The minimum power-on / off time restores the switching variable u outside the scope of the current time window's influence on future time windows. g (t) and start / stop variable x g (t).

[0161] Where R(t) represents the spinning reserve demand of the power grid system, and R(t) is a fixed value. This indicates the upper limit of the unit's power. This indicates that the total power of the generating units exceeds the spinning reserve requirement of the power grid system.

[0162] For example, based on spinning reserve constraints (spatial constraints between units). and minimum power-on / off time, restore Figure 5 The switching variable u outside the black area A3 in the future time window of Unit 1 g (t) and start / stop variable x g (t), recovery Figure 5 The switching variable u outside the black area B3 in the future time window of Unit 2 g (t) and start / stop variable x g (t), recovery Figure 5 The switching variable u outside the black area C3 in the future time window of Unit 3. g (t) and start / stop variable x g (t).

[0163] In summary, variables and constraints in future time windows that are affected by the current time window of the unit can be largely removed, leaving only the switching and start-stop operations to be solved together, thus creating a subproblem of moderate size.

[0164] 1064: Add a cutting plane.

[0165] It is understandable that the start-up and shutdown times of each unit are inconsistent. It may be too much to recover at least some of the operating data outside the influence range of some units within the future time window (or the time period corresponding to the i+1 to Kth sub-models) based on the spinning reserve constraint. In the embodiments of this application, the variables and constraints of the sub-problems of the units within the current time window can be scaled down by adding a cutting plane and utilizing the power characteristics of the units, thereby reducing the variables and constraints that need to be solved.

[0166] For example, if the influence range of the nth unit at time t in the i-th time window (or i-th sub-model) is smaller than the influence range of at least one unit in the i-th time window, the maximum power of the nth unit can be obtained; based on the cutting plane (or power demand value and the maximum power of the nth unit), the constraints and variables of some units at time t are deleted, thereby reducing the variables that need to be solved.

[0167] The cutting plane can be represented as R(t) represents the sum of power output from the power grid system, and R(t) represents the spinning reserve constraint. This represents the sum of the maximum output power of units whose influence range at time t is less than that of at least one other unit.

[0168] For example, such as Figure 6 As shown, first, determine the unit G' whose influence range at time t is smaller than the influence range of at least one unit, and then based on... Remove the switch variable (t) that is outside the scope of the current time window's influence on future time windows.

[0169] It is understandable that, since the added cutting plane excludes the space of infeasible solutions, or variables that do not need to be solved again, it achieves lossless reduction of the subproblems of the unit within the current time window without destroying the original feasible solutions, significantly reducing the combination variables (or switching variables) and further reducing the size of the subproblems.

[0170] 1065: Decompression of space.

[0171] It is understandable that the solutions of the unit in the past time window will affect the solutions of the current time window. Therefore, in order to reduce the variables that need to be solved again, the embodiments of this application can obtain the solutions and constraints in the past time window that will affect the current time window, assign relevant variables to the past solutions, and further reduce the variables of the subproblems in the current time window based on the constraints in the past time window that will affect the current time window and the solutions of the past time window of the unit, so as to obtain the subproblems of the current time window.

[0172] For example, such as Figure 7 As shown, firstly, all solutions from past time window A1 for unit 1, all solutions from past time window B1 for unit 2, and all solutions from past time window C1 for unit 3 can be obtained. Then, all solutions from past time window A1 are assigned to variables related to the current time window A2, all solutions from past time window B1 are assigned to variables related to the current time window B2, and all solutions from past time window C1 are assigned to variables related to the current time window C2. The solutions from past time window A1 include the switching variable u. g The solution to (t), the start and stop variables x g The solution to (t), the power variable p g Solution of (t), cost variables The solutions are equal. For example, u1(0)=0, x1(0)=0, p1(0)=150, Then i1(0) = 0, x1(0) = 0, p1(0) = 150, Assign values ​​to related variables.

[0173] Secondly, constraints that only apply within past time windows A1, B1, and C1 can be removed, retaining only those that affect the current time windows A2, B2, and C2, such as minimum power-on / off times. Finally, based on all solutions within past time windows A1, B1, and C1, and the constraints that only affect the current time windows A2, B2, and C2, the variables in the current time windows A2, B2, and C2 can be assigned values ​​and reduced, thereby reducing the number of variables requiring decision-making within the current time window.

[0174] In summary, by working through steps 1063: Creating Subproblems to 1065: Solving Space Compression, the temporal coupling between the subproblems of the current time window and the subproblems of the past and future time windows can be removed, thus allowing the subproblems of the current time window to be solved independently.

[0175] 1066: Subproblems of time decoupling.

[0176] In this embodiment, the massive SCUC problem is first divided into multiple subproblems based on time windows, and the subproblems in each time window are decoupled from the subproblems in other time windows in time, so that the subproblems in each time window are independent of each other and can be solved independently. This not only reduces the size of the subproblems in each time window, but also reduces the difficulty of solving the subproblems in each time window and improves the solution speed of the subproblems in each time window.

[0177] In some embodiments, after obtaining a series of time-decoupled subproblems, a For loop can be used to solve each subproblem sequentially until a feasible solution is obtained. However, during the process of solving each subproblem using a For loop, if a solution is found to be infeasible, it is necessary to solve the problem again based on all future variables and constraints between the current solutions. Furthermore, the For loop can sometimes get stuck in local optima.

[0178] In this embodiment of the application, sub-problems within a time window can be solved using an iterative solution framework to obtain high-quality feasible solutions. For example... Figure 8 As shown, it includes:

[0179] 1071: m feasible solutions.

[0180] It can be understood that m feasible solutions are solutions to subproblems of the previous time window of the current time window, where m is a positive integer greater than or equal to 1.

[0181] In some embodiments, m can be set according to the parameters of the mathematical programming solver.

[0182] 1072: Parallel creation of m subgroups problem.

[0183] In this embodiment of the application, the solution path of the subproblem within the current time window can be explored in parallel based on beam search, thereby improving the quality of the solution.

[0184] It is understandable that the beam search-based solution algorithm can collect m feasible solutions obtained from the previous solution process, and explore multiple solution paths for the subproblem within the current time window in parallel based on the m feasible solutions from the previous solution process, which can significantly improve the quality of the solution.

[0185] 1073: Mathematical Programming Solver.

[0186] In this embodiment of the application, the solution for the current time window can be obtained by solving the subproblems of the current time window based on the solution of the previous time window and the bundle search.

[0187] For example, by solving the m subproblems of the current time window, we can obtain p solutions for the current time window, where p = m × q, and q is the solution obtained by solving the subproblems created based on each solution.

[0188] For example, such as Figure 9 As shown, the subproblems within time window 2 are solved based on solutions A and C from time window 1, resulting in three sets of solutions X, Y, and Z corresponding to A, and three sets of solutions X, Y, and Z corresponding to B.

[0189] 1074: Priority queue ordering and filtering.

[0190] In this embodiment, if the subproblem corresponding to the current time window includes at least one set of solutions, m sets of solutions among the p sets of solutions for the current time window can be determined based on goodness and dissimilarity, and these m sets of solutions can be taken as feasible solutions for the current time window. If the subproblem corresponding to the time window has only one set of solutions, this set of solutions can be taken as feasible solutions for the current time window. Where m is less than p.

[0191] For example, when m = 2, such as Figure 9 As shown, time window 1 includes three sets of solutions A, B, and C, which can be sorted based on their superiority and differences. When it is determined that the superiority of A and C is higher than that of B, A and C are used as inputs for time window 2 to simplify the variables that need to be solved in the subproblems of time window 2.

[0192] Time window 2 includes six sets of solutions AX, AY, AZ, CX, CY, and CZ. These solutions can be sorted based on their quality and differences. For example, if among the solutions AX, AY, and AZ corresponding to A, AX has the highest quality, and among the solutions CX, CY, and CZ corresponding to C, CZ has the highest quality, then AY, AZ, CX, and CY can be deleted, and AX and CZ can be input into the current time window to solve the subproblems of the current time window based on AX and CZ.

[0193] It is understood that the method provided by the embodiments of this application can significantly reduce the solution time of the optimal solution.

[0194] In some embodiments according to this application Figure 3 The time required to obtain the objective feasible solution using the provided power grid dispatching method is faster than that based on... Figure 1 The power grid dispatching method shown reduces the time required to obtain the target feasible solution by more than 90%.

[0195] The various embodiments of the mechanisms disclosed in this application can be implemented in hardware, software, firmware, or a combination of these implementation methods. Embodiments of this application can be implemented as computer programs or program code executable on a programmable system, the programmable system including at least one processor, a storage system (including volatile and non-volatile memory and / or storage elements), at least one input device, and at least one output device.

[0196] Program code can be applied to input instructions to execute the functions described in this application and generate output information. The output information can be applied to one or more output devices in a known manner. For the purposes of this application, the processing system includes any system having a processor such as, for example, a digital signal processor (DSP), a microcontroller, an application-specific integrated circuit (ASIC), or a microprocessor.

[0197] The program code can be implemented using a high-level procedural language or an object-oriented programming language to communicate with the processing system. Assembly language or machine language can also be used when needed. In fact, the mechanisms described in this application are not limited to any particular programming language. In either case, the language can be a compiled language or an interpreted language.

[0198] In some cases, the disclosed embodiments may be implemented in hardware, firmware, software, or any combination thereof. The disclosed embodiments may also be implemented as instructions carried on or stored thereon on one or more transient or non-transitory machine-readable (e.g., computer-readable) storage media, which may be read and executed by one or more processors. For example, the instructions may be distributed via a network or through other computer-readable media. Therefore, machine-readable media can include any mechanism for storing or transmitting information in a machine-readable (e.g., computer-readable) form, including but not limited to floppy disks, optical disks, CD-ROMs, compact disc-read-only memory (CD-ROMs), magneto-optical disks, read-only memory (ROM), random access memory (RAM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), magnetic cards or optical cards, flash memory, or tangible machine-readable storage for transmitting information (e.g., carrier waves, infrared signals, digital signals, etc.) using the Internet in the form of electrical, optical, acoustic, or other forms of propagation signals. Therefore, machine-readable media includes any type of machine-readable medium suitable for storing or transmitting electronic instructions or information in a machine-readable (e.g., computer-readable) form.

[0199] In the accompanying drawings, some structural or methodological features may be shown in a specific arrangement and / or order. However, it should be understood that such a specific arrangement and / or order may not be necessary. Rather, in some embodiments, these features may be arranged in a manner and / or order different from that shown in the illustrative drawings. Furthermore, the inclusion of structural or methodological features in a particular figure does not imply that such features are required in all embodiments, and in some embodiments, these features may be omitted or may be combined with other features.

[0200] It should be noted that all units / modules mentioned in the device embodiments of this application are logical units / modules. Physically, a logical unit / module can be a physical unit / module, a part of a physical unit / module, or a combination of multiple physical units / modules. The physical implementation of these logical units / modules themselves is not the most important factor; the combination of functions implemented by these logical units / modules is the key to solving the technical problems proposed in this application. Furthermore, to highlight the innovative aspects of this application, the above-described device embodiments of this application have not introduced units / modules that are not closely related to solving the technical problems proposed in this application. This does not mean that the above-described device embodiments do not contain other units / modules.

[0201] It should be noted that, in the examples and description of this patent, the terms "comprising," "including," or any other variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one" does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0202] Although this application has been illustrated and described with reference to certain preferred embodiments thereof, those skilled in the art should understand that various changes in form and detail may be made thereto without departing from the spirit and scope of this application.

Claims

1. A power grid dispatching method, wherein the power grid comprises multiple generating units, characterized in that, The method includes: Based on the unit information of the multiple generating units, a power grid dispatching model is established. The power grid dispatching model is used to determine the operating information of the multiple generating units within a set time period under set constraints. The power grid dispatching model is divided into K sub-models according to time, where the i-th sub-model corresponds to the i-th sub-time period in the set time period, K is a positive integer greater than 2, and i is an integer between 1 and K; Based on the initial operating information of the unit in the i-th sub-model and the set constraints, at least some of the operating data in the (i+1)-K-th sub-models are deleted to obtain the i-th simplified scheduling model. Solving the i-th simplified scheduling model yields the initial solution of the power grid scheduling model in the i-th time period; Based on the initial solutions of the power grid dispatch model in the first to the Kth time periods, the power grid dispatch model is solved to obtain the final solution of the power grid dispatch model.

2. The method according to claim 1, characterized in that, The at least part of the operating data includes operating data that has no impact on the solution results of the initial solution for the i-th time period.

3. The method according to claim 1, characterized in that, The step of reducing at least a portion of the operational data in the (i+1)th to (K)th sub-models based on the initial operating information of the unit in the i-th sub-model and the set constraints to obtain the i-th simplified scheduling model includes: Obtain the time coupling constraints of each unit in the i-th sub-model of the set constraints; The influence range of each unit in the i-th sub-model is determined based on the initial operating information of the unit and the time coupling constraint. Delete the operating data outside the influence range of each unit in the i-th sub-model from the (i+1)-th to the K-th sub-models, and delete the operating data within the influence range of each unit in the i-th sub-model from the (i+1)-th to the K-th sub-models that has no impact on the solution results of the initial solution of the units in the i-th sub-model, to obtain the i-th simplified scheduling model.

4. The method according to claim 3, characterized in that, The time coupling constraint includes at least one of minimum power-on / off time and ramp constraint.

5. The method according to claim 3, characterized in that, include: Since the influence range of each unit in the i-th sub-model is not exactly the same, the power demand value of the power grid and the minimum start-up and shutdown time of each unit in the i-th sub-model are obtained. Based on the power demand value and the minimum start-up and shutdown time of each unit in the i-th sub-model, at least some of the operating data outside the influence range of some units in the (i+1)-K-th sub-models are recovered to obtain the i-th simplified scheduling model.

6. The method according to claim 5, characterized in that, The simplified scheduling model is obtained by restoring at least a portion of the operating data outside the influence range of some units in the (i+1)th to Kth sub-models based on the power demand value and the minimum start-up and shutdown time of each unit in the i-th sub-model, including: If the influence range of the nth unit in the i-th sub-model at time t is smaller than the influence range of at least one unit in the i-th sub-model, then the maximum power of the nth unit is obtained. Based on the power demand value and the minimum start-up and shutdown time of each unit in the i-th sub-model, at least some of the operating data outside the influence range of some units in the (i+1)-K-th sub-models are recovered. Based on the power demand value and the maximum power of the n-th unit, at least some of the operating information of some units at time t is deleted to obtain the i-th simplified scheduling model.

7. The method according to claim 1, characterized in that, The step of reducing at least a portion of the operational data in the (i+1)th to (K)th sub-models based on the initial operating information of the unit in the i-th sub-model and the set constraints to obtain the i-th simplified scheduling model includes: Based on the initial operating information of the unit in the i-th sub-model and the set constraints, the values ​​of some variables in the i-th sub-model are determined, and at least some operating data in the (i+1)-th to K-th sub-models are deleted to obtain the i-th simplified scheduling model.

8. The method according to claim 7, characterized in that, The step of determining the values ​​of some variables in the i-th sub-model based on the initial operating information of the unit in the i-th sub-model and the set constraints includes: Based on the initial solutions for the first to (i-1)th time periods and the constraints in the set constraints that affect the solution results of the initial solution of the unit in the i-th sub-model, the values ​​of some variables in the i-th sub-model are determined.

9. The method according to claim 1, characterized in that, Solving the i-th simplified scheduling model to obtain the initial solution of the power grid scheduling model in the i-th time period includes: Obtain m initial solutions for time periods 1 to i-1; Based on the initial solutions for the first to (i-1)th time periods, the i-th simplified scheduling model is solved to obtain p sets of initial solutions for the power grid scheduling model in the i-th time period; Based on the superiority and difference of the p initial solutions, the m initial solutions of the power grid dispatching model in the i-th time period are determined.

10. The method according to claim 1, characterized in that, The step of solving the power grid dispatch model based on the initial solutions of the power grid dispatch model in the first to Kth time periods to obtain the final solution of the power grid dispatch model includes: Based on the goodness and difference of the initial solution in the Kth time period, determine the final solution for the Kth time period; The final solution of the power grid dispatching model in the first to the (K-1)th time periods is determined based on the final solution in the Kth time period. Based on the final solution of the power grid dispatch model in the first to the Kth time periods, the power grid dispatch model is solved to obtain the final solution of the power grid dispatch model.

11. An electronic device, characterized in that, It includes: a memory for storing instructions executed by one or more processors of the electronic device, and the processor being one of the one or more processors of the electronic device for executing the power grid dispatching method according to any one of claims 1 to 10.

12. A readable medium, characterized in that, The readable medium stores instructions that, when executed on an electronic device, cause the electronic device to perform the power grid dispatching method according to any one of claims 1 to 10.

13. A computer program product, characterized in that, The computer program product includes computer instructions, which, when executed by an electronic device, enable the electronic device to perform the power grid dispatching method as described in any one of claims 1 to 10.

Images