Time sequence production model deduction method and device, terminal equipment and storage medium

Abstract

The invention discloses a time sequence production model deduction method and apparatus, a terminal device and a storage medium. The method comprises the steps of obtaining a first time sequence production model of a power system; four different thinking chains are constructed; wherein the thinking chain is solved by using time decomposition, space decomposition or parallel calculation; the first time sequence production model is solved according to the four thinking chains, and a fourth solving result is obtained; the result with the minimum Euclidean distance sum with all the solving results is calculated and serves as a target result; and planning and scheduling the power system according to the numerical value of the decision variable in the target result. According to the method, the simulation precision is remarkably improved through time decomposition and space decomposition. Through parallel calculation, the solving speed is improved, important details are reserved, and errors caused by simplification are reduced. In addition, the thinking chain integration effectively integrates the results of a plurality of reasoning paths, and the representativeness and accuracy of the simulation result are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power system analysis, and in particular to a method, device, terminal equipment and storage medium for deriving a time series production model. Background Art

[0002] With the rapid development of new energy and the advancement of the "carbon peak and carbon neutrality" goals, the production scheduling and operation management faced by the power system have become increasingly complex. The traditional power grid operation mode mainly relies on the management of conventional power generation units, and the supply and demand balance of the power system is relatively stable. However, with the continuous increase in the installed capacity of new energy, especially the high penetration rate of renewable energy such as photovoltaic and wind power, the power grid is facing unprecedented challenges. These new energy sources have strong randomness and volatility, making the time-series production simulation of the power system more complex and critical.

[0003] The current power system is facing challenges brought by the high penetration of renewable energy. Traditional time-series production simulation methods cannot fully cope with the randomness and volatility of renewable energy. Existing accelerated solution technologies often rely on model simplification, which may lose key details and lack accuracy, easily leading to error accumulation and unrepresentative final results. Summary of the invention

[0004] The present invention provides a time series production model deduction method, device, terminal equipment and storage medium to solve the problems of the prior art.

[0005] In order to solve the above technical problems, an embodiment of the present invention provides a time series production model deduction method, including:

[0006] Obtaining a first time-series production model of the power system;

[0007] Constructing a first solution thinking chain, a second solution thinking chain, a third solution thinking chain and a fourth solution thinking chain; wherein the first solution thinking chain adopts a method of time decomposition and non-parallel solution, the second solution thinking chain adopts a method of time decomposition and parallel solution; the third solution thinking chain adopts a method of time decomposition, space decomposition and non-parallel solution, and the fourth solution thinking chain adopts a method of time decomposition, space decomposition and parallel calculation;

[0008] Solving the first time series production model according to the first solution thinking chain to obtain a first solution result; solving the first time series production model according to the second solution thinking chain to obtain a second solution result; solving the first time series production model according to the third solution thinking chain to obtain a third solution result; solving the first time series production model according to the fourth solution thinking chain to obtain a fourth solution result;

[0009] Calculate the result with the smallest sum of Euclidean distances from the first solution, the second solution, the third solution, and the fourth solution as the target result;

[0010] The power system is planned and dispatched according to the values ​​of the decision variables in the target result.

[0011] As a preferred solution, solving the first sequential production model according to the first solution thinking chain to obtain a first solution result includes:

[0012] Get the time scale entered by the user;

[0013] According to the time scale, the first time series production model is time-decomposed to obtain a plurality of second time series production models of the power system in different short time periods;

[0014] Solve each of the second time series production models according to a preset short-time period solution order to obtain the first solution result;

[0015] Solving the first time series production model according to the second solution thinking chain to obtain a second solution result includes:

[0016] Get the time scale entered by the user;

[0017] According to the time scale, the first time series production model is time-decomposed to obtain a plurality of second time series production models of the power system in different short time periods;

[0018] Solve all second time series production models that meet the parallel solution condition in parallel, and solve all second time series production models that do not meet the parallel solution condition in the short time period solution order to obtain the second solution result;

[0019] The first time series production model is solved according to the third solution thinking chain to obtain a third solution result, including:

[0020] Get the time scale entered by the user;

[0021] According to the time scale, the first time series production model is time-decomposed to obtain a plurality of second time series production models of the power system in different short time periods;

[0022] For each of the short time periods, according to a preset regional division standard of the space where the power system is located, the second time series production model of the short time period is spatially decomposed to obtain a plurality of third time series production models of the power system in different regions within the short time period;

[0023] Select a short time according to the short time period solution sequence, and solve each of the third time series production models in the selected short time according to the preset regional solution sequence to obtain the third solution result;

[0024] Solving the first time series production model according to the fourth solution thinking chain to obtain a fourth solution result includes:

[0025] Get the time scale entered by the user;

[0026] According to the time scale, the first time series production model is time-decomposed to obtain a plurality of second time series production models of the power system in different short time periods;

[0027] For each of the short time periods, according to a preset regional division standard of the space where the power system is located, the second time series production model of the short time period is spatially decomposed to obtain a plurality of third time series production models of the power system in different regions within the short time period;

[0028] All the third time series production models within all short time periods that meet the parallel solution conditions are solved in parallel; for all short time periods that do not meet the parallel solution conditions, a short time period is selected according to the short time period solution order, and all the third time series production models within the selected short time period are solved in parallel to obtain the fourth solution result.

[0029] As a preferred solution, the first time series production model is time-decomposed according to the time scale to obtain a plurality of second time series production models of the power system in different short time periods, including:

[0030] Acquire a total simulation period of the first time series production model; wherein the total simulation period refers to a total production period of the power system simulated by the first time series production model;

[0031] According to the time scale, the total simulation period is divided into a number of consecutive short time periods;

[0032] For each of the short time periods, constructing the time constraints and state constraints of the short time period for the first time series production model to obtain a second time series production model of the short time period;

[0033] Among them, the time constraints for building a short period of time include:

[0034] Constructing a time constraint condition according to the start time and the end time of the short time period;

[0035] Construct state constraints for short time periods, including:

[0036] If the short time period is the first short time period, the state variable value at the initial moment of the total simulation period is used as the state variable value at the initial moment of the first short time period to construct the state constraint of the first short time period;

[0037] If the short time period is a short time period other than the first short time period, the state variable value of the previous short time period at the end time is used as the state variable value at the initial time of the short time period to construct the state constraint of the short time period.

[0038] As a preferred solution, the second time series production model of the short time period is spatially decomposed according to the preset regional division standard of the space where the power system is located, and a plurality of third time series production models of the power system in different regions within the short time period are obtained, including:

[0039] According to the area division standard, the space where the power system is located is divided to obtain a plurality of areas;

[0040] For each of the regions, determining all adjacent regions of the region, constructing inter-regional power transmission constraints of the region for the second time series production model according to the adjacent regions, and obtaining a third time series production model of the region;

[0041] The inter-regional transmission constraints include:

[0042] The power transmission between the area and each of the adjacent areas does not exceed a preset power threshold;

[0043] For each adjacent region, the value of the power transmitted from the adjacent region to the region is equal to the value of the power transmitted from the region to the adjacent region.

[0044] As a preferred solution, before obtaining the first time-series production model of the power system, the method further includes:

[0045] Obtaining the optimization target of the time series production model to be constructed;

[0046] According to the optimization goal, construct a corresponding optimization objective function;

[0047] Determining the decision variables of the time series production model to be constructed according to the optimization objective function;

[0048] Determining the constraint conditions of the time series generation model to be constructed according to the decision variables;

[0049] The first time series production model is constructed according to the optimization objective function, decision variables and constraints.

[0050] As a preferred solution, the step of obtaining the optimization target of the time series production model to be constructed includes:

[0051] Obtaining an optimization target selected by a user from a preset optimization target set;

[0052] Among them, the optimization target set includes: the first optimization target: minimizing the operating cost of the power system, the second optimization target: maximizing the utilization rate of new energy, the third optimization target: taking into account both the operating cost of the power system and the utilization rate of new energy, and the fourth optimization target: minimizing the investment in the power system while taking into account the operating cost of the power system.

[0053] As a preferred solution, determining the constraint conditions of the time series generation model to be constructed according to the decision variables includes:

[0054] According to the decision variables, select corresponding constraints from a preset set of constraints;

[0055] Among them, the constraint condition set includes: power balance constraint, system reserve constraint, unit output constraint, unit climbing constraint, start and stop time constraint, carbon emission constraint, quota constraint, new energy installed capacity endowment constraint, new energy power generation constraint and investment restriction constraint.

[0056] Based on the above embodiment, another embodiment of the present invention provides a time series production model deduction device, including: a data acquisition module, a thinking chain construction module, a thinking chain solution module, a result integration module and a scheduling module;

[0057] The data acquisition module is used to acquire a first time-series production model of the power system;

[0058] The thought chain construction module is used to construct a first solution thought chain, a second solution thought chain, a third solution thought chain and a fourth solution thought chain; wherein the first solution thought chain adopts a method of time decomposition and non-parallel solution, the second solution thought chain adopts a method of time decomposition and parallel solution; the third solution thought chain adopts a method of time decomposition, space decomposition and non-parallel solution, and the fourth solution thought chain adopts a method of time decomposition, space decomposition and parallel calculation;

[0059] The thought chain solving module is used to solve the first time series production model according to the first solution thought chain to obtain a first solution result; solve the first time series production model according to the second solution thought chain to obtain a second solution result; solve the first time series production model according to the third solution thought chain to obtain a third solution result; solve the first time series production model according to the fourth solution thought chain to obtain a fourth solution result;

[0060] The result integration module is used to calculate the result with the smallest sum of Euclidean distances from the first solution result, the second solution result, the third solution result and the fourth solution result as the target result;

[0061] The scheduling module is used to plan and schedule the power system according to the values ​​of the decision variables in the target result.

[0062] Based on the above embodiments, another embodiment of the present invention provides a terminal device, which includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and when the processor executes the computer program, the timing production model deduction method described in the above invention embodiment is implemented.

[0063] Based on the above embodiments, another embodiment of the present invention provides a storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the storage medium is located is controlled to execute the time series production model deduction method described in the above invention embodiment.

[0064] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0065] The present invention obtains a first time-series production model of the power system; constructs a first solution thinking chain, a second solution thinking chain, a third solution thinking chain and a fourth solution thinking chain; solves the first time-series production model according to the first solution thinking chain to obtain a first solution result; solves the first time-series production model according to the second solution thinking chain to obtain a second solution result; solves the first time-series production model according to the third solution thinking chain to obtain a third solution result; solves the first time-series production model according to the fourth solution thinking chain to obtain a fourth solution result; calculates the result with the smallest sum of Euclidean distances with the first solution result, the second solution result, the third solution result and the fourth solution result as the target result; plans and dispatches the power system according to the numerical values ​​of the decision variables in the target result. The present invention significantly improves the simulation accuracy through time decomposition and space decomposition. In particular, time decomposition and space decomposition techniques significantly improve the solution speed through parallel computing, while retaining important details and reducing errors caused by simplification. In addition, the thought chain integration and self-consistency optimization methods effectively integrate the results of multiple reasoning paths and improve the representativeness and accuracy of the simulation results. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 It is a flow chart of a method for deducing a time series production model provided by an embodiment of the present invention;

[0067] Figure 2It is a structural schematic diagram of a time series production model deduction device provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0068] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0069] Embodiment 1

[0070] Please refer to Figure 1 , a method for deducing a time series production model provided by an embodiment of the present invention, comprising:

[0071] S1. Obtain a first time-series production model of the power system.

[0072] In a preferred embodiment, before obtaining the first time-series production model of the power system, the method further includes:

[0073] Obtaining the optimization target of the time series production model to be constructed;

[0074] According to the optimization goal, construct a corresponding optimization objective function;

[0075] Determining the decision variables of the time series production model to be constructed according to the optimization objective function;

[0076] Determining the constraint conditions of the time series generation model to be constructed according to the decision variables;

[0077] The first time series production model is constructed according to the optimization objective function, decision variables and constraints.

[0078] It should be noted that the first time series production model is used to simulate the production process of the entire power system during the entire simulation period.

[0079] In this embodiment, when obtaining the optimization target of the time series production model to be constructed, the input variable type is also obtained, and the preset initial input variable type is the predicted new energy output, load data and power grid topology. The optimization objective function is constructed according to the optimization target, and all decision variables are determined according to the optimization objective function. The common decision variables include group output, carbon emissions, fuel consumption and line transmission power. The corresponding constraints are determined according to the decision variables. Finally, the first time series production model is constructed according to the input variable type, the optimization objective function, the decision variables and the constraints.

[0080] It should be noted that the data that the present invention needs to collect include historical operating data and forecast data, among which the historical operating data is used to verify the accuracy and reliability of the method, including: new energy, load, unit output data and power grid topology data for the past five years; the forecast data is used to formulate a dispatch plan for the power system, including: new energy and load forecast data for the next year.

[0081] In a preferred embodiment, the step of obtaining the optimization target of the time series production model to be constructed includes:

[0082] Obtaining an optimization target selected by a user from a preset optimization target set;

[0083] Among them, the optimization target set includes: the first optimization target: minimizing the operating cost of the power system, the second optimization target: maximizing the utilization rate of new energy, the third optimization target: taking into account both the operating cost of the power system and the utilization rate of new energy, and the fourth optimization target: minimizing the investment in the power system while taking into account the operating cost of the power system.

[0084] It should be noted that, according to the optimization goal, constructing a corresponding optimization objective function includes:

[0085] If the obtained optimization target is the first optimization target, the expression of the constructed optimization objective function is:

[0086]

[0087] In the formula, T represents the total length of time scheduling; t represents the simulation time step; M represents the total number of units included in the power system; m represents the unit number; L(m,t) represents the coal consumption cost of the mth unit in the tth period; represents the startup cost of the mth unit in the tth period; represents the shutdown cost of the mth unit in the tth period;

[0088] If the obtained optimization target is the second optimization target, the expression of the constructed optimization target function is:

[0089]

[0090] Where P w (t) represents the wind power consumption of the power system in the tth period; P pv (t) represents the photovoltaic consumption of the power system in the tth period;

[0091] If the obtained optimization target is the third optimization target, the expression of the constructed optimization objective function is:

[0092]

[0093] Where λ represents the power abandonment penalty coefficient; P ea(t) represents the amount of power abandoned by the power system in the tth period;

[0094] If the obtained optimization target is the fourth optimization target, the expression of the constructed optimization target function is:

[0095] min(L f +L v );

[0096]

[0097]

[0098] Where, L f Represents the investment cost of the power system; L v represents the operating cost of the power system; J represents the total number of conventional units; j represents the number of conventional units; c j represents the annual unit capacity investment cost of conventional units; R j represents the investment decision of conventional units; c w Represents the annual unit capacity investment cost of wind turbines; R w represents the investment decision of wind turbines; c pv represents the annualized unit capacity investment cost of the photovoltaic unit; R pv Represents the investment decision of a photovoltaic unit.

[0099] The present invention provides the above different optimization objective functions for selection. Of course, it is not limited to the above optimization objective functions, and related optimization objective functions can be added later according to more complex requirements.

[0100] In a preferred embodiment, determining the constraint conditions of the time series generation model to be constructed according to the decision variables includes:

[0101] According to the decision variables, select corresponding constraints from a preset set of constraints;

[0102] Among them, the constraint condition set includes: power balance constraint, system reserve constraint, unit output constraint, unit climbing constraint, start and stop time constraint, carbon emission constraint, quota constraint, new energy installed capacity endowment constraint, new energy power generation constraint and investment restriction constraint.

[0103] In this embodiment, after the optimization objective function is determined, the corresponding constraint conditions can be automatically determined from a preset constraint condition set according to the decision variables.

[0104] It should be noted that the expression of the power balance constraint is:

[0105]

[0106] Where P w (t) represents the wind power consumption of the power system in the tth period; P pv (t) represents the photovoltaic consumption of the power system in the tth period; P(m,t) represents the active power output of the mth unit in the tth period; P t represents the total load of the power system in the tth period;

[0107] The expression of the system backup constraint is:

[0108]

[0109] In the formula, represents the maximum output that the mth unit can provide in the tth period; O t represents the reserve capacity demand of the power system in the tth period;

[0110] The expression of the unit output constraint is:

[0111]

[0112] In the formula, K(m,t) represents the startup status of the mth unit in the tth period; Indicates the minimum active output of the mth unit; Indicates the maximum active output of the mth unit;

[0113] The expression of the unit climbing constraint is:

[0114] P(m,t)-P(m,t-1)≤W u,m +a[1-K(m,t-1)]+a[1-K(m,t)];

[0115] P(m,t-1)-P(m,t)≤W d,m +a[1-K(m,t-1)]+a[1-K(m,t)];

[0116]

[0117] Where P(m,t-1) represents the active power output of the mth unit in the t-1th period; K(m,t-1) represents the startup status of the mth unit in the t-1th period; K(m,t+1) represents the startup status of the mth unit in the t+1th period; W u,m Indicates the climbing ability of the mth unit; W d,m represents the down-scaling capability of the mth unit; a represents the weight coefficient;

[0118] The expression of the start-stop time constraint is:

[0119]

[0120] T u (m,t)=min{M u (m),T-t+1};

[0121] T d (m,t)=min{M d (m),T-t+1};

[0122] Where, T u (m,t) indicates the time that the mth unit needs to continue to operate in the tth period; u m (T) represents the operating status of the mth unit in the tth period. When its value is 1, it indicates operation, and when it is 0, it indicates shutdown; T d (m,t) indicates the time that the mth unit needs to remain shut down in the tth period; M u (m) represents the minimum start-up time of the mth unit; M d (m) represents the minimum downtime of the mth unit;

[0123] The carbon emission constraint is expressed as:

[0124]

[0125] In the formula, e m represents the carbon emission coefficient of the mth unit; Q represents the carbon emission limit;

[0126] The quota constraint expression is:

[0127]

[0128] In the formula, Q rps Indicates the new energy quota requirement;

[0129] The expression of the new energy installed capacity endowment constraint is:

[0130]

[0131] In the formula, B s Indicates the photovoltaic installations that have been connected to the grid; B w Indicates wind power installed capacity that has been connected to the grid; Indicates the upper limit of photovoltaic installed capacity; Indicates the upper limit of wind power installed capacity;

[0132] The expression of the new energy generation constraint is:

[0133] P pv (t)+Ppv,cur (t)=(B s + B s )β pv (t);

[0134] P w (t)+P w,cur (t)=(B w + B w )β w (t);

[0135] Where P pv,cur (t) represents the photovoltaic power curtailment of the power system in the tth period; β pv (t represents the photovoltaic power output per unit of the power system in the tth period; P w,cur (t) represents the amount of wind power abandoned in the power system in the tth period; β w (t) represents the wind power output per unit of the power system in the tth period;

[0136] The expression of the investment restriction constraint is:

[0137] C j =I w +I s ;

[0138] In the formula, C j represents the cost of the jth investment decision; I w I represents the annualized unit capacity investment cost of wind power in the power system; s Represents the annualized unit capacity investment cost of photovoltaic power system.

[0139] The present invention provides the above constraints, but is not limited to the above constraints. Related constraints can be added later according to more complex requirements.

[0140] So far, a variety of sequential production optimization models have been built in the sequential production simulation modeling tool module, and the corresponding optimization objective function selection and supporting constraints can be performed according to user needs. It should be noted that the optimization objective function and constraint conditions in this tool module are not limited to this and can be expanded according to needs.

[0141] S2. Construct a first solution thinking chain, a second solution thinking chain, a third solution thinking chain and a fourth solution thinking chain; wherein, the first solution thinking chain adopts a method of time decomposition and non-parallel solution, and the second solution thinking chain adopts a method of time decomposition and parallel solution; the third solution thinking chain adopts a method of time decomposition, space decomposition and non-parallel solution, and the fourth solution thinking chain adopts a method of time decomposition, space decomposition and parallel calculation.

[0142] In step S2, the present invention constructs multiple thought chains for solving the time series production model. Different thought chains need to determine the following questions in sequence during the construction process:

[0143] 1. Should we further adopt spatial decomposition based on time decomposition?

[0144] 2. Is the parallel solution method adopted?

[0145] By answering these two questions, four chains of thought for solving problems can be generated.

[0146] S3. Solve the first time series production model according to the first solution thinking chain to obtain a first solution result; solve the first time series production model according to the second solution thinking chain to obtain a second solution result; solve the first time series production model according to the third solution thinking chain to obtain a third solution result; solve the first time series production model according to the fourth solution thinking chain to obtain a fourth solution result.

[0147] In step S3, the first sequential production model is solved through different chains of thought, which may involve the following techniques:

[0148] (1) Time dimension decomposition: Decompose the constructed time-series production model in the time dimension to analyze the production activities in different time periods and extract their changing patterns in the time dimension;

[0149] (2) Spatial dimension decomposition: Further decompose the time series production model into spatial dimensions to analyze the dynamic behaviors of different locations or links in the production process in detail;

[0150] (3) Accelerated solution of time-series production simulation: Through parallel computing, the solution process of the time-series production model is accelerated to improve computing efficiency and quickly obtain production simulation results.

[0151] In a preferred embodiment, solving the first sequential production model according to the first solution thinking chain to obtain a first solution result includes:

[0152] Get the time scale entered by the user;

[0153] According to the time scale, the first time series production model is time-decomposed to obtain a plurality of second time series production models of the power system in different short time periods;

[0154] Solve each of the second time series production models according to a preset short-time period solution order to obtain the first solution result;

[0155] Solving the first time series production model according to the second solution thinking chain to obtain a second solution result includes:

[0156] Get the time scale entered by the user;

[0157] According to the time scale, the first time series production model is time-decomposed to obtain a plurality of second time series production models of the power system in different short time periods;

[0158] Solve all second time series production models that meet the parallel solution condition in parallel, and solve all second time series production models that do not meet the parallel solution condition in the short time period solution order to obtain the second solution result;

[0159] The first time series production model is solved according to the third solution thinking chain to obtain a third solution result, including:

[0160] Get the time scale entered by the user;

[0161] According to the time scale, the first time series production model is time-decomposed to obtain a plurality of second time series production models of the power system in different short time periods;

[0162] For each of the short time periods, according to a preset regional division standard of the space where the power system is located, the second time series production model of the short time period is spatially decomposed to obtain a plurality of third time series production models of the power system in different regions within the short time period;

[0163] Select a short time according to the short time period solution sequence, and solve each of the third time series production models in the selected short time according to the preset regional solution sequence to obtain the third solution result;

[0164] Solving the first time series production model according to the fourth solution thinking chain to obtain a fourth solution result includes:

[0165] Get the time scale entered by the user;

[0166] According to the time scale, the first time series production model is time-decomposed to obtain a plurality of second time series production models of the power system in different short time periods;

[0167] For each of the short time periods, according to a preset regional division standard of the space where the power system is located, the second time series production model of the short time period is spatially decomposed to obtain a plurality of third time series production models of the power system in different regions within the short time period;

[0168] All the third time series production models within all short time periods that meet the parallel solution conditions are solved in parallel; for all short time periods that do not meet the parallel solution conditions, a short time period is selected according to the short time period solution order, and all the third time series production models within the selected short time period are solved in parallel to obtain the fourth solution result.

[0169] In a preferred embodiment, the first time series production model is time-decomposed according to the time scale to obtain a plurality of second time series production models of the power system in different short time periods, including:

[0170] Acquire a total simulation period of the first time series production model; wherein the total simulation period refers to a total production period of the power system simulated by the first time series production model;

[0171] According to the time scale, the total simulation period is divided into a number of consecutive short time periods;

[0172] For each of the short time periods, constructing the time constraints and state constraints of the short time period for the first time series production model to obtain a second time series production model of the short time period;

[0173] Among them, the time constraints for building a short period of time include:

[0174] Constructing a time constraint condition according to the start time and the end time of the short time period;

[0175] Construct state constraints for short time periods, including:

[0176] If the short time period is the first short time period, the state variable value at the initial moment of the total simulation period is used as the state variable value at the initial moment of the first short time period to construct the state constraint of the first short time period;

[0177] If the short time period is a short time period other than the first short time period, the state variable value of the previous short time period at the end time is used as the state variable value at the initial time of the short time period to construct the state constraint of the short time period.

[0178] It should be noted that the main task of time decomposition is to decompose the long-term production simulation task into multiple shorter time periods so as to gradually solve it, thereby reducing the computational complexity, improving computational efficiency, and maintaining overall accuracy and consistency. The time decomposition process needs to consider the relationship between the unit status in different time periods, the load change trend, and the power generation fluctuation of new energy.

[0179] In this embodiment, time is decomposed according to the time scale input by the user, and the user can choose to decompose time in the form of day, week or month.

[0180] The specific time decomposition steps are as follows:

[0181] (1) Divide the long time period into the time scale selected by the user. Each short time period is independent of each other. Let the total time length be T total , decompose it into several short time periods, each of which has a length of Δt, and there are a total of F short time periods:

[0182]

[0183] Among them, the starting time of the i-th short time period is End time is but:

[0184]

[0185] (2) According to the start and end times of the short time period, add the following time constraints to each time period i:

[0186]

[0187] (3) Construct the state constraints of short time periods. For adjacent time periods i and i+1, the continuity of information such as the unit status needs to be ensured. Let x i represents the state change in the i-th time period (such as the start and stop state of the unit, the state of the energy storage device, etc.). In the time period i+1, the state variable x i+1 The initial value should be:

[0188]

[0189] Therefore, when the solution of time period i+1 begins, its initial state is the final state of the previous time period.

[0190] In a preferred embodiment, the second time series production model of the short time period is spatially decomposed according to a preset regional division standard of the space where the power system is located, to obtain a plurality of third time series production models of the power system in different regions within the short time period, including:

[0191] According to the area division standard, the space where the power system is located is divided to obtain a plurality of areas;

[0192] For each of the regions, determining all adjacent regions of the region, constructing inter-regional power transmission constraints of the region for the second time series production model according to the adjacent regions, and obtaining a third time series production model of the region;

[0193] The inter-regional transmission constraints include:

[0194] The power transmission between the area and each of the adjacent areas does not exceed a preset power threshold;

[0195] For each adjacent region, the value of the power transmitted from the adjacent region to the region is equal to the value of the power transmitted from the region to the adjacent region.

[0196] It should be noted that spatial decomposition needs to be divided into multiple smaller areas according to the preset regional division standards, and each area needs to be solved separately, and finally the solution results of each area need to be integrated through a certain coordination mechanism.

[0197] The specific steps of space decomposition are as follows:

[0198] (1) Assume that the area of ​​the entire power system is R total , we divide it into K sub-regions, each sub-region R k The boundaries of the region are determined by geography, load, power generation resources, etc. The total region satisfies:

[0199]

[0200] And each sub-region R k Disjoint:

[0201]

[0202] (2) Add inter-regional transmission constraints:

[0203]

[0204] Among them, T kl (t) represents the adjacent area R k and R l The power transmission between is the maximum transmission capacity. And the power exchange between regions must meet:

[0205] T kl (t) = -T lk (t),

[0206] After each region is optimized independently, the results still need to be integrated through the inter-regional coordination mechanism. Let the inter-regional coordination variable be z kl (t), represents the region R k and R l The amount of power exchanged between them at time t. k The optimization problem contains coordination constraints:

[0207] T kl (t) = z kl (t),

[0208] Among them, N k Represents the region R k The set of all adjacent regions. Through the inter-region coordination mechanism, it can ensure that the solution results of each region are spatially consistent.

[0209] When solving the first sequential production model, by using a large model to accelerate the solution of sequential production simulation, the Gurobi optimization solver can be effectively called to improve the efficiency of solving large-scale problems. Specifically, optimization can be performed through the following two parallel solution strategies:

[0210] 1. Parallel solving and result integration in different time periods:

[0211] The large model can identify the coupling strength between multiple time periods. When the coupling between time periods is weak, multiple time periods can be solved in parallel. The optimization problem of each time period is solved by the large model calling the Gurobi solver.

[0212] When the coupling between time periods is weak, multiple time periods can be solved in parallel. Let the solution result of the time period be i, then the parallel solution can be expressed as:

[0213] r1,r2,...,r N =SolveParallel([t1,t2,...,t N ]);

[0214] If the decision variables between time periods are relatively independent, or the influence of the variables that affect each other in subsequent time periods is small or negligible, then it can be determined that the coupling between time periods is weak.

[0215] The decision variables xi and xi+1 corresponding to time periods ti and ti+1 satisfy:

[0216]

[0217] This means that the optimal solution of the current time period has a negligible impact on the decision-making of the next time period.

[0218] Integrate the solution results of all time periods r1, r2, ..., r N To obtain the final result of the entire timing production simulation total .

[0219] In this way, the model can fully utilize the advantages of parallel computing, which can significantly improve the solution speed, especially when dealing with large-scale time series problems.

[0220] 2. Parallel solution and result integration between regions:

[0221] The large model can not only handle parallel processing of time periods, but also perform parallel solutions for different regions. Assuming that the entire system is divided into multiple regions, the optimization problem s of each region is solved by the Gurobi solver scheduled by the large model.

[0222] Assume that the optimization problem P for each region is k It is expressed as:

[0223] P k : subject to constraints;

[0224] Then for all sub-regions R k Perform parallel solving:

[0225] r1,r2,...,r K =SolveParallel([P1,P2,...,P K ]);

[0226] The solution results of each region are integrated through the coordination mechanism, and finally the optimal solution of the entire provincial region is obtained. total :

[0227] r total =IntegrateResults(r1,r2,...,r K ,z kl ).

[0228] In summary, when calling the Gurobi optimization solver, the large model can achieve seamless integration of parallel solving and result integration by automatically scheduling each sub-problem. The large model can not only adaptively allocate computing resources according to the complexity and scale of the problem, but also optimize the parallel solving strategy based on the specific structure of the problem, thereby maximizing the use of hardware resources and improving overall computing efficiency. Through the combination of the large model and the Gurobi solver, the solution of time-series production simulation and regional optimization problems has been significantly accelerated and optimized, achieving efficient solution of complex problems and rapid acquisition of global optimal solutions.

[0229] S4. Calculate a result with the smallest sum of Euclidean distances from the first solution, the second solution, the third solution, and the fourth solution as the target result.

[0230] In step S4, multiple reasoning paths are constructed using thought chain integration and self-consistency. By integrating the reasoning results of multiple time-series production simulations, the self-consistency method is applied to optimize different reasoning results, thereby selecting the optimal solution. Specifically, by calculating the differences between the results generated by each reasoning path, the formulated self-consistency criterion is used to evaluate and minimize the differences between the reasoning results, and finally the time-series production simulation results with the strongest consistency and the smallest error are obtained, ensuring the stability of the simulation process and the accuracy of the results.

[0231] Suppose the inference results of multiple thinking chains are r1′, r2′, ..., r g ′, each result is the output generated by a different reasoning path of the sequential production simulation process. By integrating these results, we can get the candidate result set r1′, r2′, ..., r g ′. Through the self-consistency method, the candidate result set r1′, r2′, ..., r g ′ is optimized, the goal is to find the best result r that best matches the majority of inference results * . Let each inference result r i ′ and r j ′ are all output vectors of the sequential production simulation, denoted by r i ′=(r i ′1,r i ′2,...,r i ' h ) and r j ′=(r j ′1,r j ′2,...,r j ' h ), where h represents the dimension of the result vector (i.e., the multiple features or values ​​of the time series). Then the inference result r i and r j The Euclidean distance d(r i ′,r j ′) is defined as:

[0232]

[0233] Based on this, the optimal inference result r * This can be determined by minimizing the sum of the Euclidean distances between it and all other results:

[0234]

[0235] Where n is the number of integrated inference results and m is the dimension of the result vector. By minimizing the sum of this distance, the optimal solution r that is closest to and most consistent with other inference results can be found. *This result not only satisfies the reasoning logic of most thought chains, but also can optimally reflect the actual situation of sequential production simulation.

[0236] Finally, the historical time series production data is applied to the constructed model for verification. The model's simulation output is compared with historical data to evaluate the accuracy and robustness of the model. Specifically, the performance of the model is verified by calculating the error between the model simulation results and the historical actual data, using the mean square error (MSE) or mean absolute error (MAE). At the same time, the verification results are also used to further adjust and optimize the model to ensure that it can accurately reflect the actual time series production behavior and improve the ability to predict load user demand.

[0237] S5. Plan and dispatch the power system according to the values ​​of the decision variables in the target result.

[0238] Embodiment 2

[0239] Please refer to Figure 2 , is a schematic diagram of the structure of a time series production model deduction device provided by an embodiment of the present invention, comprising: a data acquisition module, a thinking chain construction module, a thinking chain solution module, a result integration module and a scheduling module;

[0240] The data acquisition module is used to acquire a first time-series production model of the power system;

[0241] The thought chain construction module is used to construct a first solution thought chain, a second solution thought chain, a third solution thought chain and a fourth solution thought chain; wherein the first solution thought chain adopts a method of time decomposition and non-parallel solution, the second solution thought chain adopts a method of time decomposition and parallel solution; the third solution thought chain adopts a method of time decomposition, space decomposition and non-parallel solution, and the fourth solution thought chain adopts a method of time decomposition, space decomposition and parallel calculation;

[0242] The thought chain solving module is used to solve the first time series production model according to the first solution thought chain to obtain a first solution result; solve the first time series production model according to the second solution thought chain to obtain a second solution result; solve the first time series production model according to the third solution thought chain to obtain a third solution result; solve the first time series production model according to the fourth solution thought chain to obtain a fourth solution result;

[0243] The result integration module is used to calculate the result with the smallest sum of Euclidean distances from the first solution result, the second solution result, the third solution result and the fourth solution result as the target result;

[0244] The scheduling module is used to plan and schedule the power system according to the values ​​of the decision variables in the target result.

[0245] Embodiment 3

[0246] Accordingly, an embodiment of the present invention provides a terminal device, which includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and when the processor executes the computer program, it implements the timing production model deduction method described in the above-mentioned embodiment of the invention.

[0247] Embodiment 4

[0248] Accordingly, an embodiment of the present invention provides a storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the storage medium is located is controlled to execute the time series production model deduction method described in the above-mentioned embodiment of the invention.

[0249] It should be noted that the device embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, in the accompanying drawings of the device embodiments provided by the present invention, the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines. A person of ordinary skill in the art may understand and implement it without paying any creative effort.

[0250] Those skilled in the art can clearly understand that for the sake of convenience and brevity, the specific working process of the device described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0251] The terminal device may be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory.

[0252] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the device, and various interfaces and lines are used to connect various parts of the entire device.

[0253] The memory can be used to store the computer program, and the processor realizes various functions of the device by running or executing the computer program stored in the memory and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application required for a function, etc.; the data storage area can store data created according to the use of the mobile phone, etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (FlashCard), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0254] The storage medium is a storage medium, and the computer program is stored in the storage medium. When the computer program is executed by the processor, the steps of each method embodiment described above can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form, etc. The computer readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content contained in the computer readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer readable media do not include electric carrier signals and telecommunication signals.

[0255] The above is a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the principle of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for deducing a time series production model, characterized in that: include: Obtaining a first time-series production model of the power system; Constructing a first solution thinking chain, a second solution thinking chain, a third solution thinking chain and a fourth solution thinking chain; wherein the first solution thinking chain adopts a method of time decomposition and non-parallel solution, the second solution thinking chain adopts a method of time decomposition and parallel solution; the third solution thinking chain adopts a method of time decomposition, space decomposition and non-parallel solution, and the fourth solution thinking chain adopts a method of time decomposition, space decomposition and parallel calculation; Solving the first time series production model according to the first solution thinking chain to obtain a first solution result; solving the first time series production model according to the second solution thinking chain to obtain a second solution result; solving the first time series production model according to the third solution thinking chain to obtain a third solution result; solving the first time series production model according to the fourth solution thinking chain to obtain a fourth solution result; Calculate the result with the smallest sum of Euclidean distances from the first solution, the second solution, the third solution, and the fourth solution as the target result; The power system is planned and dispatched according to the values ​​of the decision variables in the target result.

2. The method for deducing a time series production model according to claim 1, characterized in that: The first sequential production model is solved according to the first solution thinking chain to obtain a first solution result, including: Get the time scale entered by the user; According to the time scale, the first time series production model is time-decomposed to obtain a plurality of second time series production models of the power system in different short time periods; Solve each of the second time series production models according to a preset short-time period solution order to obtain the first solution result; Solving the first time series production model according to the second solution thinking chain to obtain a second solution result includes: Get the time scale entered by the user; According to the time scale, the first time series production model is time-decomposed to obtain a plurality of second time series production models of the power system in different short time periods; Solve all second time series production models that meet the parallel solution condition in parallel, and solve all second time series production models that do not meet the parallel solution condition in the short time period solution order to obtain the second solution result; The first time series production model is solved according to the third solution thinking chain to obtain a third solution result, including: Get the time scale entered by the user; According to the time scale, the first time series production model is time-decomposed to obtain a plurality of second time series production models of the power system in different short time periods; For each of the short time periods, according to a preset regional division standard of the space where the power system is located, the second time series production model of the short time period is spatially decomposed to obtain a plurality of third time series production models of the power system in different regions within the short time period; Select a short time according to the short time period solution sequence, and solve each of the third time series production models in the selected short time according to the preset regional solution sequence to obtain the third solution result; Solving the first time series production model according to the fourth solution thinking chain to obtain a fourth solution result includes: Get the time scale entered by the user; According to the time scale, the first time series production model is time-decomposed to obtain a plurality of second time series production models of the power system in different short time periods; For each of the short time periods, according to a preset regional division standard of the space where the power system is located, the second time series production model of the short time period is spatially decomposed to obtain a plurality of third time series production models of the power system in different regions within the short time period; All the third time series production models within all short time periods that meet the parallel solution conditions are solved in parallel; for all short time periods that do not meet the parallel solution conditions, a short time period is selected according to the short time period solution order, and all the third time series production models within the selected short time period are solved in parallel to obtain the fourth solution result.

3. The method for deducing a time series production model according to claim 2, characterized in that: The step of performing time decomposition on the first time series production model according to the time scale to obtain a plurality of second time series production models of the power system in different short time periods includes: Acquire a total simulation period of the first time series production model; wherein the total simulation period refers to a total production period of the power system simulated by the first time series production model; According to the time scale, the total simulation period is divided into a number of consecutive short time periods; For each of the short time periods, constructing the time constraints and state constraints of the short time period for the first time series production model to obtain a second time series production model of the short time period; Among them, the time constraints for building a short period of time include: Constructing a time constraint condition according to the start time and the end time of the short time period; Construct state constraints for short time periods, including: If the short time period is the first short time period, the state variable value at the initial moment of the total simulation period is used as the state variable value at the initial moment of the first short time period to construct the state constraint of the first short time period; If the short time period is a short time period other than the first short time period, the state variable value of the previous short time period at the end time is used as the state variable value at the initial time of the short time period to construct the state constraint of the short time period.

4. The method for deducing a time series production model according to claim 2, characterized in that: The second time series production model of the short time period is spatially decomposed according to a preset regional division standard of the space where the power system is located, to obtain a plurality of third time series production models of the power system in different regions within the short time period, including: According to the area division standard, the space where the power system is located is divided to obtain a plurality of areas; For each of the regions, determining all adjacent regions of the region, constructing inter-regional power transmission constraints of the region for the second time series production model according to the adjacent regions, and obtaining a third time series production model of the region; The inter-regional transmission constraints include: The power transmission between the area and each of the adjacent areas does not exceed a preset power threshold; For each adjacent region, the value of the power transmitted from the adjacent region to the region is equal to the value of the power transmitted from the region to the adjacent region.

5. The method for deducing a time series production model according to claim 1, wherein: Before obtaining the first time-series production model of the power system, it also includes: Obtaining the optimization target of the time series production model to be constructed; According to the optimization goal, construct a corresponding optimization objective function; Determining the decision variables of the time series production model to be constructed according to the optimization objective function; Determining the constraint conditions of the time series generation model to be constructed according to the decision variables; The first time series production model is constructed according to the optimization objective function, decision variables and constraints.

6. The method for deducing a time series production model according to claim 5, characterized in that: The step of obtaining the optimization target of the time series production model to be constructed includes: Obtaining an optimization target selected by a user from a preset optimization target set; Among them, the optimization target set includes: the first optimization target: minimizing the operating cost of the power system, the second optimization target: maximizing the utilization rate of new energy, the third optimization target: taking into account both the operating cost of the power system and the utilization rate of new energy, and the fourth optimization target: minimizing the investment in the power system while taking into account the operating cost of the power system.

7. The method for deducing a time series production model according to claim 5, characterized in that: Determining the constraint conditions of the time series production model to be constructed according to the decision variables includes: According to the decision variables, select corresponding constraints from a preset set of constraints; Among them, the constraint condition set includes: power balance constraint, system reserve constraint, unit output constraint, unit climbing constraint, start and stop time constraint, carbon emission constraint, quota constraint, new energy installed capacity endowment constraint, new energy power generation constraint and investment restriction constraint.

8. A time series production model deduction device, characterized in that: include: Data acquisition module, thinking chain building module, thinking chain solving module, result integration module and scheduling module; The data acquisition module is used to acquire a first time-series production model of the power system; The thought chain construction module is used to construct a first solution thought chain, a second solution thought chain, a third solution thought chain and a fourth solution thought chain; wherein the first solution thought chain adopts a method of time decomposition and non-parallel solution, the second solution thought chain adopts a method of time decomposition and parallel solution; the third solution thought chain adopts a method of time decomposition, space decomposition and non-parallel solution, and the fourth solution thought chain adopts a method of time decomposition, space decomposition and parallel calculation; The thought chain solving module is used to solve the first time series production model according to the first solution thought chain to obtain a first solution result; solve the first time series production model according to the second solution thought chain to obtain a second solution result; solve the first time series production model according to the third solution thought chain to obtain a third solution result; solve the first time series production model according to the fourth solution thought chain to obtain a fourth solution result; The result integration module is used to calculate the result with the smallest sum of Euclidean distances from the first solution result, the second solution result, the third solution result and the fourth solution result as the target result; The scheduling module is used to plan and schedule the power system according to the values ​​of the decision variables in the target result.

9. A terminal device, characterized in that: The method comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor implements the time series production model deduction method as described in any one of claims 1 to 7 when executing the computer program.

10. A storage medium, characterized in that: The storage medium includes a stored computer program, wherein when the computer program is executed, the device where the storage medium is located is controlled to execute the time series production model deduction method according to any one of claims 1 to 7.

Images