Joint scheduling optimization method for multi-energy complementary power grid considering wind and light load uncertainty

Abstract

The application discloses a kind of considering wind light load uncertainty multi-energy complementary grid joint scheduling optimization method, belong to grid scheduling optimization technical field, the method constructs a kind of wind light fire storage multi-energy complementary day joint scheduling optimization model, introduces nonlinear fuzzy opportunity constraint to quantify this uncertainty to represent the wind light load uncertainty in scheduling scene, proposes a kind of dynamic multi-objective genetic algorithm, by introducing dynamic control and disturbance correction mechanism, solve complex multi-objective optimization model, to realize joint scheduling day optimal plan.Example results show that, compared with the scheme without representing uncertainty, the optimization target values of low carbon, economy and stability are increased by 7%, 21% and 25% respectively, which can provide reliable protection for grid operation, meet the diversified needs of complex scenarios and improve the solution efficiency of multi-objective optimization.

Description

Technical Field

[0001] This invention belongs to the field of power grid dispatch optimization technology, specifically relating to a multi-energy complementary power grid joint dispatch optimization method that considers the uncertainty of wind, solar and load. Background Technology

[0002] The high proportion of renewable energy integrated into the grid poses challenges to the power system's power balance. While constructing a multi-energy complementary grid can improve the operational reliability of the power system, the uncertainty of wind and solar loads still affects grid stability and increases operating costs. Therefore, grid dispatch optimization techniques are needed to achieve multi-objective optimization within the grid and make reasonable and optimal grid dispatch decisions.

[0003] In the early stages of new energy technology development, wind and solar installed capacity accounted for a low proportion and had limited grid connection impact. Grid dispatch was centered on traditional energy sources, employing deterministic optimization methods with economic dispatch and stable operation constraints as core objectives. However, traditional models did not fully consider the intermittency and uncertainty of new energy power generation, resulting in insufficient adaptability of their dispatch strategies to fluctuating power sources. As the penetration rate of new energy increases, the impact of its output fluctuations on grid security strengthens, highlighting the limitations of traditional models. Single-energy dispatch is insufficient to mitigate wind and solar fluctuations, making multi-energy complementary grid joint dispatch a solution that utilizes flexible power sources and energy storage to address new energy fluctuations. Research methods include stochastic programming and robust optimization, with some methods using data-driven approaches to improve prediction accuracy and optimize multi-timescale dispatch.

[0004] Current methods typically employ multi-objective linear optimization, essentially transforming multiple objectives into a single objective, thus reducing the decision-making dimensionality and making it difficult to balance conflicting objectives. Furthermore, uncertainties significantly impact scheduling optimization, yet traditional methods often overlook them. While methods like stochastic programming and fuzzy theory exist, they suffer from inaccurate feature fitting and overly conservative optimization. In multi-energy complementary power grid scenarios, scheduling models need to consider conflicting optimization objectives such as environmental friendliness, economic efficiency, and stability. Existing hierarchical solution strategies and intelligent algorithms both suffer from unsatisfactory results. Summary of the Invention

[0005] Purpose of the invention: To address the shortcomings of the existing technology, this invention provides a multi-energy complementary power grid joint dispatch optimization method that considers the uncertainty of wind and solar loads.

[0006] Technical Solution: To achieve the above objectives, this invention provides a multi-energy complementary power grid joint dispatch optimization method considering the uncertainties of wind and solar loads, specifically including the following steps:

[0007] Step 1: Collect parameters of each energy supply device in the multi-energy complementary power grid, as well as the day-ahead forecast results of wind power, photovoltaic power generation and electricity load demand in the regional power grid, as input data for subsequent models;

[0008] Step 2: Utilizing the complementary and coordinated characteristics among wind, solar, thermal, and energy storage, and combining the power source composition and dispatching needs of the multi-energy complementary power system, a multi-objective and multi-constraint multi-energy complementary day-ahead joint dispatching optimization model is constructed from the day-ahead dispatching level, taking into account the optimization objectives of system operation economy, stability, and maximizing renewable energy consumption.

[0009] Step 3: Based on the scheduling optimization model, nonlinear fuzzy chance constraints are used to characterize the wind and solar load uncertainty of the system power balance constraint in the wind-solar-thermal-storage multi-energy complementary scheduling model.

[0010] Step 4: Use a dynamic multi-objective genetic algorithm to solve the proposed optimization model for joint dispatch of multi-energy complementary power grids under wind-solar load uncertainty;

[0011] Step 5: Obtain the optimal day-ahead dispatch plan, which is the joint day-ahead dispatch plan of the power grid under the multi-energy complementarity scenario, including the output plan of thermal power units, the working plan of energy storage batteries, and the grid interaction purchase and sale plan.

[0012] Preferably, step 2 includes the following sub-steps:

[0013] Step 2.1: Set the objective function of the scheduling optimization model. The multiple optimization objectives of the wind-solar-thermal-storage multi-energy complementary power system include wind and solar energy consumption, grid operating costs, and net load volatility.

[0014] Step 2.2: Set the constraints of the scheduling optimization model. The multiple constraints of the wind-solar-thermal-storage multi-energy complementary power system include system power balance constraints, thermal power unit operation constraints, energy storage battery operation constraints, and grid interaction constraints, which are represented as constraint 1 to constraint 4, respectively.

[0015] Preferably, in step 2.1, considering the low-carbon optimization objective, the maximum wind and solar energy consumption is taken as optimization objective 1. Since the overall operation economy of the system involves fuel consumption costs and energy market transactions, the minimum grid operation cost is taken as optimization objective 2. In order to ensure the stability of the system output, the minimum standard deviation of the net load fluctuation of thermal power is taken as optimization objective 3.

[0016] Among them, as a clean and renewable energy source, the use of wind and solar energy in the new energy power grid to replace thermal power reduces fossil fuel consumption and improves the system's economic and environmental benefits. Maximizing the absorption of wind and solar energy is taken as the optimization objective 1, as shown in the following formula:

[0017] in, This represents the total amount of renewable energy absorbed in the regional new energy power grid system; , These represent the amount of electricity consumed by wind power and solar power in the system, respectively, as shown in the following formulas: in, , express The system's wind and solar power consumption is measured in real time. This represents the entire time period for optimized scheduling of the regional new energy power grid system.

[0018] In a multi-energy complementary power grid integrating wind, solar, thermal, and energy storage, the economic efficiency of the power system mainly depends on the conventional operating costs of the power grid, while other additional costs must also be considered. Minimizing the power grid operating cost is taken as optimization objective 2, expressed by the formula: in, , , , , , These represent the fuel cost, operation and maintenance costs, grid interaction costs, depreciation costs of energy storage systems (charging and discharging), pollution gas treatment costs from thermal power units, and deviation penalty costs, respectively. The specific expressions are as follows: in, This indicates the number of thermal power generation units within the system. The price of fuel consumed per unit of thermal power generation. express time The power generation capacity of the station To consume fuel with low calorific value, Expresses the unit power generation efficiency of thermal power generation; in, , , , These represent the unit power operation and maintenance costs of wind turbines, photovoltaic power plants, thermal power plants, and energy storage batteries, respectively. , , , express The output power of each power source, including wind turbines, photovoltaic power, thermal power, and energy storage batteries, at all times;

[0019] in, , They are respectively Real-time electricity price and power consumption, , These are the electricity sales price and the electricity sales capacity, respectively.

[0020] in, This is the depreciation cost factor for energy storage batteries. This indicates the adjustable charging and discharging power of the energy storage battery;

[0021] in, Indicates the first Unit treatment cost for Class II emissions express Time of the first Emissions of Class II pollutants;

[0022] in, , This indicates that the effort was overestimated and the penalty price was underestimated. , express Always overestimate or underestimate the penalties for exertion;

[0023] Among them, thermal power units, as base load power sources, undertake the basic power support function in the power system; frequent adjustments to the output of thermal power units will directly affect the economic efficiency and environmental benefits of the power system; it is necessary to focus on smoothing the fluctuation amplitude of net load so that thermal power units can maintain a quasi-steady-state operation of power output. Minimizing the standard deviation of the net load borne by thermal power units is taken as the optimization objective 3, and the formula is as follows: in, This represents the standard deviation of the net load borne by thermal power plants. express The net load borne by thermal power plants within the power system at any given time.

[0024] Preferably, four constraints are set in step 2.2; among them, constraint 1 considers the balance between system power supply and total load power: in, , These represent the charging and discharging power of the energy storage battery, respectively. This represents the total load in the system;

[0025] Among them, constraint 2 considers the upper and lower power limits and ramping constraints that the thermal power unit needs to meet during operation: in, , for Upper and lower limits of the power generation capacity of thermal power units at the site. express Time and The difference in output at any moment , For thermal power units, this refers to the uphill and downhill climbing techniques.

[0026] Among them, constraint 3 considers the upper and lower limits of the energy storage battery power, capacitance, and operating state transition constraints: in, This indicates the maximum output power of the energy storage battery. , These represent the initial and final states of the energy storage battery capacitor, respectively. , For the minimum and maximum capacitance of the energy storage battery, express Constant state of the energy storage battery capacitor;

[0027] Among them, constraint condition 4 considers grid interaction power constraints: in, , This indicates the maximum power available for interaction with the power grid, including the purchase and sale of electricity.

[0028] Preferably, step 3 includes the following steps:

[0029] Step 3.1: Construct a model based on fuzzy chance-constrained programming, with the following formula: in, It is a parameter between 0 and 1. For decision variables in the scheduling model, For fuzzy variables in the scheduling model, To determine the probability of an event occurring;

[0030] Step 3.2: In intraday optimization scheduling, the uncertainty of wind and solar load is characterized using normally distributed fuzzy parameters: in, The center value of the fuzzy number corresponds to the predicted mean. This is a fuzziness parameter, related to the prediction error;

[0031] Step 3.3: Define fuzzy chance constraints based on system power balance, requiring that the probability of output satisfying the load is not less than the fuzzy confidence level pre-specified by the chance constraints. : ; Let be the fuzzy random variables corresponding to wind and solar loads, which follow a normal distribution membership function, and have fuzzy numbers for mean and variance: in, Indicates load The upper and lower bounds of the fuzzy interval of the mean. express Similarly, the upper and lower bounds of the fuzzy interval of variance are defined. Related variables;

[0032] For easy distinction, Set as , Set as Define combination variables Its mean and variance are: ; ;

[0033] Using the properties of the fuzzy normal distribution, the constraint is equivalent to: in For fuzzy confidence levels, The cumulative distribution function of the standard normal distribution;

[0034] Step 3.4: Using the clear equivalence class theorem of confidence measure, the fuzzy constraints are transformed into a deterministic form. When the confidence level... When the chance constraint is ≥0.5, the clear equivalence class is: in, It is the inverse function of the cumulative distribution function; its final result is the deterministic form of the fuzzy chance constraint of system power balance involving wind, solar and load uncertainties;

[0035] Step 3.5: Introduce the fuzzy chance constraint into the scheduling optimization model above to construct a joint scheduling optimization model for multi-energy complementary power grids under the uncertainty of wind, solar and load.

[0036] Preferably, step 4 includes the following steps:

[0037] Step 4.1: Set initial parameters: population size N, maximum number of iterations iter max Crossover probability p c Mutation probability p m Switching coefficient ct, threshold coefficient Generate an initial population P0 of size N, with each individual randomly generated, and evaluate the objective function value F(x) for each individual x in the initial population P0.

[0038] Step 4.2: Perform selection, crossover, and mutation operations according to the dynamic control mechanism to generate the offspring population Q of generation t. t ;

[0039] Step 4.3: Merge the parent population P of generation t. t-1 and offspring population Q t Obtain the combined population R t ;

[0040] Step 4.4: For the feasible solution set R t Perform non-dominated sorting to generate the front set of solutions;

[0041] Step 4.5: Generate a new parent population P t ;

[0042] Step 4.6: Apply the perturbation correction strategy to redistribute candidate solutions to the potential solution space in later iterations, and repeat steps 4.4 and 4.5.

[0043] Step 4.7: Output the solution set. The Pareto front solution set obtained by the algorithm is P. final = {x ∈ Piter max}

[0044] Preferably, in step 4.2, the dynamic control mechanism is as follows: The dynamic control mechanism adopts a phased adjustment of weight coefficients to control the population selection behavior under the roulette wheel strategy. The iterative process is divided into two phases, and the switching between the two phases is determined by variables. Sure: in, This represents the maximum value of the iteration. It is the switching coefficient;

[0045] In the first stage, the fitness weight coefficient will be reduced during early iterations, and the population search space will be expanded in conjunction with the crowding distance penalty, gradually approaching the optimal solution; the cumulative probability function of the first stage is shown in the following equation: in, Represents an individual fitness For individuals Crowded distance, This represents the sum of the fitness of all individuals in the population. It is the weight coefficient of the first iteration stage. In the early stage, it controls the weight of reducing the fitness of candidate individuals, increases population diversity, and avoids getting trapped in local optima.

[0046] In the second stage, population selection is not affected by the parameter settings of earlier iterations, and the selection pressure is adjusted according to the current individual fitness distribution; the cumulative probability function of the second stage is shown in the following equation: in, The weighting coefficients in the second stage set a higher parental selection pressure. By adjusting the probability formula, the selection opportunities for individuals with higher fitness are enhanced, ensuring better convergence of the algorithm.

[0047] Preferably, the perturbation correction strategy in step 4.6 is as follows:

[0048] In later iterations of NSGA-II, the perturbation correction strategy is triggered under the following conditions: in, This is the threshold coefficient; when the convergence speed slows down, it is based on the current non-dominated solution set. Calculate the new boundary of the individual: ; in, , These represent the upper and lower bounds of the potential solution space. For the solution set Inner individual fitness These are expansion coefficients used to control the extent of expansion in the search space; for non-dominated solution sets... Each individual in Perform Gaussian perturbation operation: ; ;in, The perturbation intensity is used; after the operation, a projection is performed to correct the boundary and generate a new solution set. ,Will They merge with the original population and undergo non-dominant ranking and environmental selection.

[0049] Compared with the prior art, the present invention, employing the above technical solution, has the following technical effects:

[0050] First, this invention constructs a multi-energy complementary day-ahead joint dispatch optimization model for wind, solar, thermal and energy storage to coordinate the dispatch of various power generation units in the power grid, providing reliable guarantees for power grid operation and meeting the diverse needs of complex scenarios;

[0051] Second, this invention introduces nonlinear fuzzy chance constraints to quantify this uncertainty, balancing the risks and benefits during operation, in order to cope with the uncertainty of wind and solar load in scheduling scenarios;

[0052] Third, this invention proposes a dynamic multi-objective genetic algorithm, which introduces dynamic control and perturbation correction mechanisms to accelerate the convergence speed of the optimization model solution, improve the quality of the solution set, and enhance the solution efficiency of multi-objective optimization. Attached Figure Description

[0053] Figure 1 This is a flowchart of the method of the present invention.

[0054] Figure 2 This is a schematic diagram of the overall framework of the method of the present invention.

[0055] Figure 3 This is a schematic diagram of a multi-energy complementary power grid scenario for the method of the present invention.

[0056] Figure 4 The figure shows the day-ahead scheduling results in the embodiment of the method of the present invention; in the figure, (a) is the optimized day-ahead scheduling plan, including energy storage, thermal power output, wind power output, and load power; (b) is the output plan curve of thermal power units with natural gas and coal as the main fuel sources; (c) is the output plan curve of energy storage battery and the corresponding state of charge value; (d) is the day-ahead predicted load and wind and solar absorption curve.

[0057] Figure 5 This is a diagram showing the actual scheduling results in an embodiment of the method of the present invention.

[0058] Figure 6This is a comparison diagram of the optimization algorithm iterations of the method of the present invention. Detailed Implementation

[0059] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and should not be used to limit the scope of protection of the present invention.

[0060] like Figure 1 As shown, this invention relates to a joint dispatch optimization method for a multi-energy complementary power grid considering the uncertainty of wind and solar loads. The overall process framework of this invention is as follows: Figure 2 As shown, the method flow includes input data, establishing a multi-energy complementary scheduling model, and a model based on a dynamic multi-objective genetic algorithm. Specifically, the method includes the following steps:

[0061] Step 1: Collect parameters of each energy supply device in the multi-energy complementary power grid, as well as the day-ahead forecast results of wind power, photovoltaic power generation and electricity load demand in the regional power grid, as input data for subsequent models;

[0062] The multi-energy complementary power grid mentioned in the method of this invention is a multi-energy complementary power grid with coordinated and complementary wind, solar, thermal, and energy storage, such as... Figure 3 As shown, its main energy supply methods include wind power generation, photovoltaic power generation, thermal power output, battery discharge, and external grid interaction power purchase. It can also store electricity through energy storage batteries. Its electricity demand mainly comes from user load, energy storage charging, and grid interaction power sales.

[0063] Step 2: Utilizing the complementary and coordinated characteristics among wind, solar, thermal, and energy storage, and considering the power source composition and dispatch needs of the multi-energy complementary power system, a multi-objective, multi-constraint multi-energy complementary day-ahead joint dispatch optimization model is constructed from the day-ahead dispatch level, taking into account the optimization objectives of system operation economy, stability, and maximizing renewable energy consumption. The specific model is as follows:

[0064] Step 2.1: Define the objective function of the scheduling optimization model. The multiple optimization objectives of the wind-solar-thermal-storage multi-energy complementary power system include wind and solar energy consumption, grid operating costs, and net load volatility. Considering the low-carbon optimization objective, maximizing wind and solar energy consumption is taken as optimization objective 1. Since the overall operating economy of the system involves fuel consumption costs and energy market transactions, minimizing grid operating costs is taken as optimization objective 2. To ensure the stability of system output, minimizing the standard deviation of net load volatility borne by thermal power is taken as optimization objective 3.

[0065] Among these, wind and solar energy, as clean and renewable energy sources, are used in new energy power grids to replace thermal power, reducing fossil fuel consumption and improving the system's economic and environmental benefits. Therefore, maximizing the absorption of wind and solar energy is taken as optimization objective 1, as shown in the following formula: in, This represents the total amount of renewable energy absorbed in the regional new energy power grid system. , These represent the amount of electricity consumed by wind power and solar power in the system, respectively, as shown in the following formulas: in, , express The system's wind and solar power consumption is measured in real time. This represents the entire time period for optimized scheduling of the regional new energy power grid system.

[0066] In a multi-energy complementary power grid integrating wind, solar, thermal, and energy storage, the economic efficiency of the power system mainly depends on the conventional operating costs of the grid, while other additional costs must also be considered. Therefore, minimizing the grid operating cost is taken as optimization objective 2, expressed by the formula: in, , , , , , These represent the fuel cost, operation and maintenance cost, grid interaction cost, depreciation cost of energy storage system charging and discharging, cost of treating pollutants emitted by thermal power units, and deviation penalty cost, respectively.

[0067] The specific expression is as follows: in, This indicates the number of thermal power generation units within the system. The price of fuel consumed per unit of thermal power generation. express time The power generation capacity of the station To consume fuel with low calorific value, Expresses the unit power generation efficiency of thermal power generation;

[0068] in, , , , These represent the unit power operation and maintenance costs of wind turbines, photovoltaic power plants, thermal power plants, and energy storage batteries, respectively. , , , express The output power of each power source, including wind turbines, photovoltaic power, thermal power, and energy storage batteries, at all times;

[0069] in, , They are respectively Real-time electricity price and power consumption, , These are the electricity sales price and the electricity sales capacity, respectively.

[0070] in, This is the depreciation cost factor for energy storage batteries. This indicates the adjustable charging and discharging power of the energy storage battery;

[0071] in, Indicates the first Unit treatment cost for Class II emissions express Time of the first Emissions of Class II pollutants;

[0072] in, , This indicates that the effort was overestimated and the penalty price was underestimated. , express Always overestimate or underestimate the penalties for exertion;

[0073] Thermal power units, as baseload power sources, play a crucial role in supporting basic power output in the power system. However, thermal power units have high start-up and shutdown costs and low ramp rates, resulting in relatively poor operational flexibility. Frequent adjustments to the output of thermal power units directly impact the economic efficiency and environmental benefits of the power system. Therefore, it is necessary to focus on smoothing out net load fluctuations to enable thermal power units to maintain a quasi-steady-state operating state in terms of power output. Minimizing the standard deviation of the net load borne by thermal power units is set as optimization objective 3, as shown in the following formula:

[0074] in, This represents the standard deviation of the net load borne by thermal power plants. express The net load borne by thermal power plants within the power system at any given time.

[0075] Step 2.2: Set the constraints of the scheduling optimization model. The multiple constraints of the wind-solar-thermal-storage multi-energy complementary power system mainly include system power balance constraints, thermal power unit operation constraints, energy storage battery operation constraints, and grid interaction constraints, which are represented as constraint 1 to constraint 4 respectively.

[0076] Among them, constraint 1 considers the balance between system power supply and total load power: in, , These represent the charging and discharging power of the energy storage battery, respectively. This represents the total load in the system;

[0077] Among them, constraint 2 considers the upper and lower power limits and ramping constraints that the thermal power unit needs to meet during operation: in, , for Upper and lower limits of the power generation capacity of thermal power units at the site. express Time and The difference in output at any moment , For thermal power units, this refers to the uphill and downhill climbing techniques.

[0078] Among them, constraint 3 considers the upper and lower limits of the energy storage battery power, capacitance, and operating state transition constraints: in, This indicates the maximum output power of the energy storage battery. , These represent the initial and final states of the energy storage battery capacitor, respectively. , For the minimum and maximum capacitance of the energy storage battery, express Constant state of the energy storage battery capacitor;

[0079] Among them, constraint condition 4 considers grid interaction power constraints: in, , This indicates the maximum power available for interaction with the power grid, including the purchase and sale of electricity.

[0080] Step 3: Based on the scheduling optimization model, nonlinear fuzzy chance constraints are used to characterize the wind and solar load uncertainty of the system power balance constraint in the wind-solar-thermal-storage multi-energy complementary scheduling model. Specifically, this includes:

[0081] Step 3.1: Construct a model based on fuzzy chance-constrained programming, with the following formula: in, It is a parameter between 0 and 1. For decision variables in the scheduling model, For fuzzy variables in the scheduling model, To determine the probability of an event occurring;

[0082] Step 3.2: In intraday optimization scheduling, the uncertainty of wind and solar load is characterized using normally distributed fuzzy parameters: in, The center value of the fuzzy number corresponds to the predicted mean. This is the ambiguity parameter, which is related to the prediction error. Let be the fuzzy random variables corresponding to wind and solar loads, which follow a normal distribution membership function, and have fuzzy numbers for mean and variance: in, Indicates load The upper and lower bounds of the fuzzy interval of the mean. express Similarly, the upper and lower bounds of the fuzzy interval of variance are defined. Related variables;

[0083] Step 3.3: Define fuzzy chance constraints based on system power balance, requiring that the probability of output satisfying the load is not less than the fuzzy confidence level pre-specified by the chance constraints. : For easy distinction, Set as , Set as Define combined variables. Its mean and variance are: ; ;

[0084] Using the properties of the fuzzy normal distribution, the constraint is equivalent to: in For fuzzy confidence levels, The cumulative distribution function of the standard normal distribution;

[0085] Step 3.4: Using the clear equivalence class theorem of confidence measure, the fuzzy constraints are transformed into a deterministic form. When the confidence level... When the chance constraint is ≥0.5, the clear equivalence class is: in, It is the inverse function of the cumulative distribution function. The final result is the deterministic form of the fuzzy chance constraint for system power balance involving uncertainties in wind, solar, and load.

[0086] Step 3.5: Introduce the fuzzy chance constraint into the scheduling optimization model above to construct a joint scheduling optimization model for multi-energy complementary power grids under the uncertainty of wind, solar and load.

[0087] Step 4: The dynamic multi-objective genetic algorithm is used to solve the proposed multi-energy complementary power grid joint dispatch optimization model under wind-solar load uncertainty, as detailed below:

[0088] Step 4.1: Set initial parameters: population size N, maximum number of iterations iter max Crossover probability p c Mutation probability p m Switching coefficient ct, threshold coefficient Generate an initial population P0 of size N, with each individual randomly generated, and evaluate the objective function value F(x) for each individual x in the initial population P0.

[0089] Step 4.2: Perform selection, crossover, and mutation operations according to the dynamic control mechanism to generate the offspring population Q of generation t. t The dynamic control mechanism is as follows:

[0090] The dynamic control mechanism employs a phased adjustment of weight coefficients to control population selection behavior under the roulette wheel strategy. The iterative process consists of two phases, and the timing of the switch between the two phases is determined by variables. Sure: in, This represents the maximum value of the iteration. It is the switching coefficient.

[0091] In the first stage, the fitness weight coefficient is reduced during early iterations, and a crowding distance penalty is used to expand the search space of the population, gradually approaching the optimal solution. The cumulative probability function for the first stage is shown in the following equation: in, Represents an individual fitness For individuals Crowded distance, This represents the sum of the fitness of all individuals in the population. It is the weight coefficient of the first iteration stage. In the early stage, it controls the weight of reducing the fitness of candidate individuals, increases population diversity, and avoids getting trapped in local optima.

[0092] In the second stage, population selection is not affected by the parameter settings of earlier iterations; instead, the selection pressure is adjusted based on the current individual fitness distribution. The cumulative probability function for the second stage is shown in the following equation: in, The weighting coefficients in the second stage set a higher parental selection pressure. By adjusting the probability formula, the selection opportunities for individuals with higher fitness are enhanced, ensuring better convergence of the algorithm.

[0093] Step 4.3: Merge the parent population P of generation t. t-1 and offspring population Q t Obtain the combined population R t ;

[0094] Step 4.4: For the feasible solution set R t Perform non-dominated sorting to generate the front set of solutions;

[0095] Step 4.5: Generate a new parent population P t The specific steps are as follows:

[0096] Let P t Let P be an empty set, i = 1; tMerge with the i-th non-dominated front (Fronts[i]), increment i by 1, and repeat this operation until P. t The size of the set with Fronts[i] is greater than or equal to N; if |P t If |< N, then assign congestion level to Fronts[i] and sort Fronts[i] in descending order of congestion level, and then assign P t = P t ∪ Fronts[i] [1 : (N - |P t |)];

[0097] Step 4.6: Apply the perturbation correction strategy to reallocate candidate solutions to the potential solution space in later iterations, and repeat steps 4.4 and 4.5. The perturbation correction strategy is as follows:

[0098] In later iterations of NSGA-II, the perturbation correction strategy is triggered under the following conditions: in, This is the threshold coefficient. When the number of iterations reaches a certain number or the convergence speed slows down, it is based on the current non-dominated solution set. Calculate the new boundary of the individual: ; in, , These represent the upper and lower bounds of the potential solution space. For the solution set Inner individual fitness For the expansion coefficient, it controls the extent of the search space expansion. For the non-dominated solution set... Each individual in Perform Gaussian perturbation operation: ; ; The disturbance intensity is used. After the operation, a projection is performed to correct the boundary and generate a new solution set. ,Will They merge with the original population and undergo non-dominant ranking and environmental selection.

[0099] Step 4.7: Output the solution set. The Pareto front solution set obtained by the algorithm is P. final = {x ∈ Piter max}

[0100] Step 5: Obtain the optimal day-ahead dispatch plan, which is the joint day-ahead dispatch plan of the power grid under the multi-energy complementarity scenario, including the output plan of thermal power units, the working plan of energy storage batteries, and the grid interaction purchase and sale plan.

[0101] In this specific embodiment of the invention, a multi-energy complementary power grid in the study area is taken as the research object. The power production scenario on December 1, 2023, is used as a simulation example. The dispatching operation cycle is 24 hours, and the unit dispatching time is 15 minutes. The system includes a photovoltaic power generation capacity of 6563MW, a wind power generation capacity of 5212MW, a thermal power generation capacity of 10400MW, and a maximum energy storage battery system power of 3000MW. The power output equipment involved in the example is as mentioned above. The abbreviations and operating parameters of each piece of equipment are shown in Table 1.

[0102] Table 1

[0103]

[0104] The actual time-of-use electricity price data for interaction with the external power grid is shown in Table 2.

[0105] Table 2

[0106]

[0107] The proposed multi-energy complementary scheduling model for wind, solar, thermal, and energy storage under uncertain wind-solar conditions was solved using a dynamic multi-objective genetic algorithm. The current-day scheduling results are as follows: Figure 4 The day-ahead dispatch plan, as the result of the day-ahead joint dispatch optimization method for multi-energy complementary power grids that takes into account the uncertainties of wind and solar power, is optimized based on day-ahead forecasts. Its actual dispatch results under the actual day-ahead scenario are as follows: Figure 5 As shown.

[0108] To visually represent the various optimization target values ​​of the daily plan and the actual results, a comparison of the daily scheduling plan and the actual scheduling results is shown in Table 3.

[0109] Table 3

[0110]

[0111] To compare the differences between single-objective and multi-objective scenarios, three additional scenarios were set up outside the example scenario, each with a different objective function as the optimization objective, and the results were compared with those of the multi-objective optimization. Table 4 shows the results of each single-objective algorithm in each scenario and the results of the multi-objective algorithm in the standard scenario.

[0112] Table 4

[0113]

[0114] Under the same model conditions, NSGAII, SPEA2, and MOPSO were selected as comparative optimization algorithms to compare their performance in day-ahead scheduling optimization of multi-energy complementary power grids. The optimization results are shown in Table 5, and the optimization efficiency during the iteration process is also shown in Table 5. Figure 6 .

[0115] Table 5

[0116]

[0117] The results of experiments comparing day-ahead scheduling plans with actual scheduling results, single-objective and multi-objective scenarios, and optimization algorithms further validate the effectiveness and superiority of the proposed method in the scheduling optimization problem of a multi-energy complementary power grid under wind-solar uncertainty. The numerical examples show that the proposed day-ahead scheduling optimization model can coordinate the scheduling of various power sources, constructing a multi-objective, multi-constraint scheduling model that meets the needs of the target scenario. By using nonlinear opportunistic fuzzy constraints, the uncertainty of wind and solar power in the scheduling scenario can be reasonably quantified, optimizing decisions to reduce the adverse effects of errors, thereby obtaining better actual scheduling results. The function values ​​of each optimization objective are improved by 7%, 21%, and 25%, respectively. Furthermore, the dynamic multi-objective genetic algorithm outperforms other multi-objective optimization algorithms in both optimization results and efficiency.

[0118] This invention provides a multi-energy complementary power grid joint dispatch optimization method considering the uncertainties of wind and solar loads. It should also be noted that the specific technical features described in the above embodiments can be combined in any suitable manner, provided they do not contradict each other. To avoid unnecessary repetition, this invention will not further describe the various possible combinations.

Claims

1. A multi-energy complementary power grid joint dispatch optimization method considering the uncertainty of wind and solar loads, characterized in that, Includes the following steps: S1. Collect parameters of each energy supply device in the multi-energy complementary power grid, as well as the day-ahead forecast results of wind power, photovoltaic power generation and electricity load demand in the regional power grid, as input data for subsequent models; S2. Utilizing the complementary and coordinated characteristics among wind, solar, thermal, and energy storage, and combining the power source composition and dispatching needs of the multi-energy complementary power system, a multi-objective and multi-constraint multi-energy complementary day-ahead joint dispatching optimization model is constructed from the day-ahead dispatching level, taking into account the optimization objectives of system operation economy, stability, and maximization of renewable energy consumption. S3. Based on the scheduling optimization model, nonlinear fuzzy chance constraints are used to characterize the wind and solar load uncertainty of the system power balance constraint in the wind-solar-thermal-storage multi-energy complementary scheduling model. S4. The dynamic multi-objective genetic algorithm is used to solve the proposed multi-energy complementary power grid joint dispatch optimization model under the uncertainty of wind and solar load. S5. Obtain the optimal day-ahead dispatch plan, i.e. the joint day-ahead dispatch plan of the power grid under the multi-energy complementarity scenario, including the output plan of thermal power units, the working plan of energy storage batteries, and the grid interaction purchase and sale plan; Step S4 includes the following steps: Step 4.1: Set initial parameters: population size N, maximum number of iterations iter max Crossover probability p c Mutation probability p m Set the switching coefficient ct and the threshold coefficient α, generate an initial population P0 of size N, with each individual randomly generated, and evaluate the objective function value F(x) of each individual x in the initial population P0. Step 4.2: Perform selection, crossover, and mutation operations according to the dynamic control mechanism to generate the offspring population Q of generation t. t ; Step 4.3: Merge the parent population P of generation t. t-1 and offspring population Q t Obtain the combined population R t ; Step 4.4: For the feasible solution set R t Perform non-dominated sorting to generate the front set of solutions; Step 4.5: Generate a new parent population P t ; Step 4.6: Apply the perturbation correction strategy to redistribute candidate solutions to the potential solution space in later iterations, and repeat steps 4.4 and 4.

5. Step 4.7: Output the solution set. The Pareto front solution set obtained by the algorithm is P. final ={x∈Piter max } 2. The multi-energy complementary power grid joint dispatch optimization method considering the uncertainty of wind and solar loads according to claim 1, characterized in that, Step S2 includes the following sub-steps: S2.

1. Set the objective function of the scheduling optimization model. The multiple optimization objectives of the wind-solar-thermal-storage multi-energy complementary power system include wind and solar energy consumption, grid operating costs and net load volatility. S2.

2. Set the constraints of the scheduling optimization model. The multiple constraints of the wind-solar-thermal-storage multi-energy complementary power system include system power balance constraints, thermal power unit operation constraints, energy storage battery operation constraints, and grid interaction constraints, which are represented as constraint 1 to constraint 4 respectively.

3. The multi-energy complementary power grid joint dispatch optimization method considering the uncertainty of wind and solar loads according to claim 2, characterized in that, Step S2.1 Considering the low-carbon optimization goal, the maximum wind and solar energy consumption is taken as optimization goal 1. Since the overall operation economy of the system involves fuel consumption costs and energy market transactions, the minimum grid operation cost is taken as optimization goal 2. In order to ensure the stability of the system output, the minimum standard deviation of the net load fluctuation of thermal power is taken as optimization goal 3. Among them, as a clean and renewable energy source, the use of wind and solar energy in the new energy power grid to replace thermal power reduces fossil fuel consumption and improves the system's economic and environmental benefits. Maximizing the absorption of wind and solar energy is taken as the optimization objective 1, as shown in the following formula: in, This represents the total amount of renewable energy absorbed in the regional new energy power grid system; These represent the amount of electricity consumed by wind power and solar power in the system, respectively, as shown in the following formulas: in, This represents the wind and solar power consumption of the system at time t, where T represents the entire time period for the optimized scheduling of the regional new energy power grid system. In a multi-energy complementary power grid integrating wind, solar, thermal, and energy storage, the economic efficiency of the power system operation depends on the conventional operating costs of the power grid, while other additional costs must also be considered. Minimizing the power grid operating cost is taken as optimization objective 2, expressed by the following formula: minC=C fuel +C OM +C grid +C storage +C emission +C punish Among them, C fuel C OM C grid C storage C emission C punish These represent the fuel cost, operation and maintenance costs, grid interaction costs, depreciation costs of energy storage systems (charging and discharging), pollution gas treatment costs from thermal power units, and deviation penalty costs, respectively. The specific expressions are as follows: Where N represents the number of thermal power generation units in the system, c fuel P is the price of fuel consumed per unit of thermal power generation. i,t L represents the power generation of station i at time t. fuel To consume fuel with a low calorific value, η fuel Expresses the unit power generation efficiency of thermal power generation; Where, k wind k pv k fuel k storage P represents the unit power operation and maintenance cost of wind turbines, photovoltaic power, thermal power, and energy storage batteries, respectively. wind,t P pv,t P fuel,t P storage,t This represents the output power of each power source at time t: wind turbine, photovoltaic power, thermal power, and energy storage battery. Among them, w buy,t P buy,t The electricity purchase price and power consumption at time t are respectively, w sell,t P sell,t These are the electricity sales price and the electricity sales capacity, respectively. Among them, c storage ΔP is the depreciation cost factor for energy storage batteries. storage,t This indicates the adjustable charging and discharging power of the energy storage battery; Among them, c j v represents the unit treatment cost for emissions of type j. t,j This represents the emission amount of the j-th type of emission at time t; Among them, c + c - This indicates that the effort was overestimated and the penalty price was underestimated. This indicates that the penalty for overestimating or underestimating the output at time t is negligible. Among them, thermal power units, as base load power sources, undertake the basic power support function in the power system; frequent adjustments to the output of thermal power units will directly affect the economic efficiency and environmental benefits of the power system; it is necessary to focus on smoothing the fluctuation amplitude of net load so that thermal power units can maintain a quasi-steady-state operation of power output. Minimizing the standard deviation of the net load borne by thermal power units is taken as the optimization objective 3, and the formula is as follows: Where F represents the standard deviation of the net load borne by thermal power, and L net,t This represents the net load borne by thermal power plants in the power system at time t.

4. The multi-energy complementary power grid joint dispatch optimization method considering the uncertainty of wind and solar loads according to claim 3, characterized in that, Step S2.2 sets four constraints; Among them, constraint 1 considers the balance between system power supply and total load power: P wind,t +P pv,t +P fuel,t +P buy,t +P dis,t =L t +P sell,t +P ch,t Among them, P ch,t P dis,t These represent the charging and discharging power of the energy storage battery, L respectively. t This represents the total load in the system; Among them, constraint 2 considers the upper and lower power limits and ramping constraints that the thermal power unit needs to meet during operation: in, ΔP represents the upper and lower limits of the generating capacity of the thermal power units at site i. fuel,i,t This represents the difference in output between time t and time t-1. For thermal power units, this refers to the uphill and downhill climbing techniques. Among them, constraint 3 considers the upper and lower limits of the energy storage battery power, capacitance, and operating state transition constraints: in, V represents the maximum output power of the energy storage battery. ps,init V ps,end These represent the initial and final states of the energy storage battery capacitor, V. t min V t max V represents the minimum and maximum capacitance of the energy storage battery. t This indicates the state of the energy storage battery capacitor at time t; Among them, constraint condition 4 considers grid interaction power constraints: in, This indicates the maximum power available for interaction with the power grid, including the purchase and sale of electricity.

5. The multi-energy complementary power grid joint dispatch optimization method considering the uncertainty of wind and solar loads according to claim 4, characterized in that, Step S3 includes the following steps: Step 3.1: Construct a model based on fuzzy chance-constrained programming, with the following formula: s.t.Pos{ξ|g i (x,ξ)≤0}≥α,i=1,2,…,p Where α is a parameter between 0 and 1, x = {P} fuel,t ,P buy,t ,P sell,t ,P dis,t ,P ch,t Let ξ = {P} be the decision variable in the scheduling model. wind,t ,P pv,t ,L t } represents the fuzzy variable in the scheduling model, and Pos{} represents the probability of the event occurring; Step 3.2: In intraday optimization scheduling, the uncertainty of wind and solar load is characterized using normally distributed fuzzy parameters: Where m is the center value of the fuzzy number, corresponding to the prediction mean, and σ is the fuzziness parameter, which is related to the prediction error: Where, m L * (α),m L* (α) represents the load L t The upper and lower bounds of the fuzzy interval of the mean, σ 2 L * (α),σ 2 L* (α) represents L t Similarly, P is defined as the upper and lower bounds of the fuzzy interval of variance. wind,t ,P pv,t Related variables; Step 3.3: Define fuzzy chance constraints based on system power balance, requiring that the probability of output satisfying the load is not less than the fuzzy confidence level pre-specified by the chance constraints. For easy distinction, P wind,t ,P pv,t ,L t Let W, P, L, P fuel,t +P buy,t +P dis,t -P sell,t -P ch,t Let X be the variable; define the combination variable Z = LW - P + X, with its mean and variance as follows: Using the properties of the fuzzy normal distribution, the constraint is equivalent to: in Let F be the fuzzy confidence level, and F be the cumulative distribution function of the standard normal distribution. Step 3.4: Using the clear equivalence class theorem of confidence measure, the fuzzy constraint is transformed into a deterministic form. When the confidence level α ≥ 0.5, the clear equivalence class of the chance constraint is: Among them, F -1 It is the inverse function of the cumulative distribution function, and its final result is the deterministic form of the fuzzy chance constraint of system power balance involving wind-solar-load uncertainty; Step 3.5: Introduce the fuzzy chance constraint into the scheduling optimization model to construct a joint scheduling optimization model for multi-energy complementary power grids under wind-solar-load uncertainty.

6. The multi-energy complementary power grid joint dispatch optimization method considering the uncertainty of wind and solar loads according to claim 1, characterized in that, In step S4.2, the dynamic control mechanism is as follows: The dynamic control mechanism employs a phased adjustment of weight coefficients to control population selection behavior under the roulette wheel strategy. The iterative process consists of two phases, and the switching between the two phases is determined by variable T. Among them, iter max This represents the maximum value of the iteration, and ct is the switching coefficient; In the first stage, the fitness weight coefficient will be reduced during early iterations, and the population search space will be expanded in conjunction with the crowding distance penalty, gradually approaching the optimal solution; the cumulative probability function of the first stage is shown in the following equation: Where fitness represents individual x i fitness, i distance For individual x i Crowded distance, W1 represents the sum of the fitness of all individuals in the population. W1 is the weight coefficient in the first iteration stage. In the early stage, it controls the weight of reducing the fitness of candidate individuals, increases population diversity, and avoids getting trapped in local optima. In the second stage, population selection is not affected by the parameter settings of earlier iterations, and the selection pressure is adjusted according to the current individual fitness distribution; the cumulative probability function of the second stage is shown in the following equation: W2 is the weight coefficient for the second stage, which sets a higher parental selection pressure. By adjusting the probability formula, the selection opportunities of individuals with higher fitness are enhanced, ensuring better convergence of the algorithm.

7. The multi-energy complementary power grid joint dispatch optimization method considering the uncertainty of wind and solar loads according to claim 6, characterized in that, The perturbation correction strategy in step S4.6 is as follows: In later iterations of NSGA-II, the perturbation correction strategy is triggered under the following conditions: Where α is the threshold coefficient; based on the current non-dominated solution set P n Calculate the new boundary of the individual: ub (t) =max(f i ∈P n ))+δ·|max(f i )-min(f i )| lb (t) =max(f i ∈P n ))-δ·|max(f i )-min(f i )| Among them, ub (t) lb (t) Let f represent the upper and lower bounds of the potential solution space. i For the solution set P n Intra-individual x i The fitness of the solution set P is denoted by δ, which is the expansion coefficient used to control the expansion range of the search space. n Each individual x in i Perform Gaussian perturbation operation: f i ′=f i +Ν(0,η(lb (t) ,ub (t) )) Where η is the disturbance intensity; after the operation, a projection is performed to correct the boundary and generate a new solution set P. n,new , will P n,new They merge with the original population and undergo non-dominant ranking and environmental selection.

Images