Multi-target collaborative optimal scheduling method and device, electronic equipment and storage medium

Abstract

The invention relates to a multi-target collaborative optimal scheduling method and device, electronic equipment and a storage medium. The multi-objective collaborative optimal scheduling method comprises the steps that S1, an objective function is generated, an electric power system objective function comprising a thermal power generating unit, a wind turbine generator, a photovoltaic unit, an electrochemical energy storage system and a carbon trading market mechanism is generated, and a new energy base in the Sagomean region takes long-distance direct current delivery as a main consumption form; s2, constraint conditions are added, and constraint conditions related to power system operation are added; and S3, generating an optimal scheduling scheme, and generating the optimal scheduling scheme by adopting a sea image algorithm. According to the multi-target collaborative optimal scheduling method, the convergence speed of the algorithm can be ensured aiming at the problems that the model is difficult to establish due to multiple targets and the advantages and disadvantages of various solving methods participate in half, and a multi-target collaborative low-carbon optimal scheduling model is established; the purposes of improving the economical efficiency of system operation and reducing the carbon emission of the power system are achieved while new energy consumption is promoted.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic low-carbon technology in the desert region, and more specifically, to a method, apparatus, electronic device, and storage medium for multi-objective collaborative low-carbon optimal scheduling in the desert region. Background Technology

[0003] Despite a significant increase in installed capacity of new energy sources compared to the past, the contradiction between renewable energy consumption and grid integration remains prominent in the desert and Gobi areas due to the unique geographical location of Northwest China and the imperfect power transmission network architecture. In 2022, the amount of wind and solar power curtailed accounted for 90% of the national total. Technically, there are difficulties in model building due to multiple objectives and the mixed advantages and disadvantages of various solution methods.

[0004] Therefore, it is of great significance to broadly and effectively coordinate various peak-shaving resources in the power system to improve the grid's capacity to absorb new energy sources in desert and Gobi areas. There is an urgent need to develop a multi-objective, collaborative, low-carbon optimal dispatching technology for desert and Gobi areas to achieve the goals of promoting new energy absorption while improving the economic efficiency of system operation and reducing carbon emissions from the power system. Summary of the Invention

[0005] The technical problem this invention aims to solve is the difficulty in establishing models due to multiple objectives and the mixed advantages and disadvantages of various solution methods in the field of photovoltaic low-carbon technology in the desert region.

[0006] To address the aforementioned technical problems, according to one aspect of the present invention, a multi-objective collaborative optimal scheduling method is provided. This method is applied to a power system in the Gobi Desert region, comprising thermal power units, wind power units, photovoltaic units, electrochemical energy storage systems, and a carbon trading market mechanism, to generate an optimal scheduling scheme. The multi-objective collaborative optimal scheduling method includes the following steps: S1, generating an objective function for the power system, including thermal power units, wind power units, photovoltaic units, electrochemical energy storage systems, and a carbon trading market mechanism, wherein the new energy base in the Gobi Desert region primarily utilizes long-distance DC transmission for energy consumption; S2, adding constraints, incorporating constraints related to power system operation; S3, generating an optimal scheduling scheme, using the Walrus Algorithm to generate the optimal scheduling scheme.

[0007] According to an embodiment of the present invention, step S1 may include the following steps: S11, calculating the minimum coal consumption cost and start-up / shutdown cost of the thermal power unit using the following formula:

[0008] (1)

[0009] In the formula, C1 represents the operating cost of the thermal power unit, in yuan; C mh C qt- Coal consumption cost and start-up / shutdown cost of thermal power units, in yuan; T - Time of day, T=24h; a j b j c j - Coal consumption coefficient for thermal power units, t / MW; P G,j,t - Output value of thermal power unit j at time t, in MW; S jt - Start-up and shutdown cost of thermal power unit j, in ten thousand yuan; u j,t - The start-up and shutdown status of thermal power unit j at time t;

[0010] S12. Minimizing carbon emission costs: Both photovoltaic and wind power are clean energy sources that do not produce CO2. Therefore, the system's carbon emissions are calculated by considering thermal power units. The carbon allocation coefficient represents the proportion or weight of the carbon emission allowance allocated to different units, used to determine the allocation of carbon emission responsibilities or the allocation rules in carbon trading. The total carbon emission allowance for the entire system and the total carbon emissions of the system within a cycle are:

[0011] (2)

[0012] In the formula, M L - Total carbon emission allowance for the system, t; M P — Total carbon emissions of the system within one cycle, t; ϖ j - Carbon emission allocation coefficient per unit electricity of thermal power unit, t / MW; , - Carbon emission allocation coefficients for photovoltaic and wind turbine units, t / MW; P G,j,t ,P v,t ,P w,t - Output of thermal power units, photovoltaic power units, and wind power units at time t, in MW; λ j - Carbon emission intensity, t / MW; A tiered carbon trading cost model is adopted, dividing the purchase of carbon emission rights into multiple tiers. The more carbon emission rights required to purchase, the higher the price in the corresponding tier. The specific calculation formula of the model is as follows:

[0013] (3)

[0014] In the formula, C2 is the total carbon trading cost in yuan; ω is the carbon trading price in yuan / t; d is the length of the carbon emission range in t; and τ is the increase in carbon trading price for each step up in carbon emissions in yuan.

[0015] S13: Maximum new energy absorption capacity and lowest cost:

[0016] The wind and solar power absorption capacity is represented by the amount of wind and solar power curtailment during the dispatch cycle. The higher the amount of curtailment, the weaker the wind and solar power absorption capacity. Since wind power and photovoltaic power generation do not consume fuel, only operation and maintenance costs are considered here. The specific calculation formula is as follows:

[0017]

[0018] In the formula, C w,t C cw,t - Wind farm operation and maintenance costs, wind curtailment penalty costs, in yuan; C v,t C cv,t - Photovoltaic power plant operation and maintenance costs, curtailment penalty costs, in yuan / MW; k w k cw -Wind farm operating cost coefficient, wind curtailment penalty cost coefficient, yuan / MW; k v k cv - Photovoltaic power plant operating cost coefficient and curtailment penalty cost coefficient, in yuan; , -Predicted wind power and photovoltaic power generation at time t, in MW;

[0019] S14. Maximizing the operating benefits and minimizing the costs of energy storage systems:

[0020] (5)

[0021] In the formula, p price -Grid electricity price, yuan; C sy C cb - The discharge revenue and charging cost of an energy storage power station, in yuan; ν d ν c - The discharge and charging efficiency of the energy storage power station, %; P sd,t ,P sc,t - The discharge and charge power of the electrochemical energy storage power station at time t, in MW; c sc - Charging cost coefficient for energy storage power stations, in yuan / MW; c sd - Discharge cost coefficient, RMB / MW;

[0022] S15. Take the sum of the above four objective functions C1~C4 as the total objective function to be scheduled.

[0023] According to an embodiment of the present invention, step S2 may include the following steps: S21, establishing system power balance constraints:

[0024] (6)

[0025] In the formula, -Active power loss of the system at time t, in MW;

[0026] S22. Operating constraints of thermal power units, including:

[0027] Unit output constraints

[0028] (7)

[0029] In the formula, , -Minimum and maximum output of the j-th thermal power unit, in MW;

[0030] Unit ramp-up constraints,

[0031] (8)

[0032] In the formula, - The gradeability of thermal power unit j, in MW / h;

[0033] Unit start-up and shutdown status constraints.

[0034] (9)

[0035] In the formula, , - The start-up and shutdown actions of the jth thermal power unit at time t;

[0036] S23, Wind power output constraints

[0037] (10)

[0038] S24. Energy storage constraints, including:

[0039] Energy storage constraints

[0040] (11)

[0041] Energy storage charging and discharging constraints,

[0042] (12)

[0043] In the formula, S t - Energy storage capacity, MW; θ i - Self-loss rate; φ sc,t - Charging efficiency, % φ sd,t -Discharge efficiency, % u sc ,u sd,t - Charging state, discharging state; S t,min S t,max - Capacity upper limit, capacity lower limit, MW; P sc,max - Maximum charging power, MW; P sd,max - Maximum discharge power, MW.

[0044] According to an embodiment of the present invention, step S3 may include the following steps:

[0045] S31. Initialize the population, determining the population size as N and the dimension m of each individual. Each individual is a vector composed of a set of parameters to be optimized, representing a candidate solution to an optimization problem, i.e., a single objective function value; the initial value of each individual in the search space is a vector x. i,j , that is, x i,j =Lb i,j +rand(x i,j (Ub) i,j -Lb i,j ), that is, the decision variables within a single objective function, and the total objective function value matrix X obtained from all candidate solutions:

[0046] (13)

[0047] Among them, Ub i,j Lb i,j The upper and lower boundary values ​​of the candidate solution represent the constraints of each decision variable; rand(x) i,j ) is a random uniform function, and rand(x) i,j X ∈ [0,1]; X is the matrix of all individual candidate positions, i.e., the total objective function value; X i Xi,j is the vector of the i-th individual, representing the objective function value of a single individual; Xi,j are the decision variables within a single objective function; N is the value of the overall objective function; i∈{1,2,⋯,N}, j is the j-th dimension of the i-th individual, and m is the individual dimension, i.e., the number of single objective functions; where Xi,j are the decision variables within a single objective function, including: T - time of day, T=24h; a j b j c j - Coal consumption coefficient for thermal power units, t / MW; P G,j,t - Output value of thermal power unit j at time t, in MW; S jt - Start-up and shutdown cost of thermal power unit j, in ten thousand yuan; u j,t - The start-up and shutdown status of thermal power unit j at time t; M L - Total carbon emission allowance for the system, t; M P — Total carbon emissions of the system within one cycle, t; ϖ j - Carbon emission allocation coefficient per unit electricity of thermal power unit, t / MW; , - Carbon emission allocation coefficients for photovoltaic and wind turbine units, t / MW; P G,j,t ,P v,t ,P w,t - Output of thermal power units, photovoltaic power units, and wind power units at time t, in MW; λ j- Carbon emission intensity, t / MW; ω - Carbon trading price, yuan / t; d - Length of carbon emission range, t; τ - Increase in carbon trading price for each carbon emission increment, yuan; k w k cw -Wind farm operating cost coefficient, wind curtailment penalty cost coefficient, yuan / MW; k v k cv - Photovoltaic power plant operating cost coefficient and curtailment penalty cost coefficient, in yuan; , -Predicted wind power and photovoltaic power generation at time t, in MW; p price -Grid electricity price, yuan; ν d ν c - The discharge and charging efficiency of the energy storage power station, %; P sd,t ,P sc,t - The discharge and charge power of the electrochemical energy storage power station at time t, in MW; c sc —Charging cost coefficient for energy storage power stations, yuan / MW; c sd - Discharge cost coefficient, yuan / MW; X i Let X be the vector of the i-th individual, i.e., the objective function value of a single individual, namely: C1 - operating cost of thermal power unit, yuan; C2 - total carbon trading cost, yuan; C3 - new energy absorption capacity and cost, yuan; C4 - operating revenue and cost of energy storage system, yuan; X is the matrix of candidate positions of all individuals, i.e. the total objective function value, i.e., minC1+minC2+minC3+maxC4, which means minimizing the coal consumption cost and start-up and shutdown cost of thermal power unit + minimizing carbon emission cost + maximizing the new energy absorption capacity and minimizing the cost + maximizing the operating revenue and minimizing the cost of energy storage system. Achieving these four aspects simultaneously is considered as the realization of multi-objective coordinated low-carbon optimal scheduling in the desert region.

[0048] S32. Fitness Calculation: Each individual is a candidate solution to the problem. Based on its fitness value (in this invention, the smaller value of individual objective function values ​​C1~C3 and the larger value of C4 indicate better fitness), the individual objective function can be evaluated. The fitness value is calculated for each candidate position, resulting in:

[0049] (14)

[0050] Where F is the fitness vector, and F i This is the fitness vector of the i-th candidate position; we explore here to find the leader individual; the leader individual is the individual with the best fitness in the population, with its fitness value being either the minimum or the maximum, to guide other individuals in optimization; we update the individual positions to generate a new leader individual, i.e.:

[0051] (15)

[0052] (16)

[0053] (17)

[0054] (18)

[0055] Among them, X i It is the new position of the i-th generated individual; It is the j-th dimension of the new position; It is the objective function value at the new position; rand i,j It is a random number in the interval [0,1]; SW j It is the best candidate solution for the strongest individual; , Indicates the safety and danger signals at the new location; T is the total number of iterations, and t is the current iteration number;

[0056] S33. Different update strategies are used to adjust the position of individuals in the population based on the different value ranges of safety signals and danger signals.

[0057] Migration behavior is judged when the risk factor is too high (Dange_signal≧1), the walrus population will migrate to an area more suitable for the population's survival; assuming that each individual migrates from one region of the search space to another random location, the new location is generated by equation (19), that is:

[0058] (19)

[0059] (20)

[0060] According to Equation 19, if the new position changes the value of the objective function, then the new position replaces the previous position. Wherein, It is the position generated by the migration of the i-th individual; It is the objective function value at the new position; X k， k∈{1,2,⋯,N}, and k≠i is the location chosen for the migration of the i-th individual, F k It is the objective function value of the migration location;

[0061] During migration, the walrus herd expands its search area, engaging in a global search. However, with increasing iterations, the ordinary walrus algorithm tends towards local search development, especially after iteration T / 2, when the herd fully enters the development phase. In the early stages of the algorithm, the herd not only performs global search exploration but also engages in significant development. Initially, the distribution of global and local searches is not well balanced, potentially leading to premature convergence and hindering the herd's exploration of the solution space, thus impacting the final optimization. To enhance the global search capability of the walrus algorithm, a spiral search strategy inspired by natural spiral motion is adopted. This strategy strengthens the algorithm's global optimization ability, ensuring convergence speed and increasing individual diversity. The spiral search formula is as follows:

[0062] (twenty one)

[0063] Where X(t+1) represents the position of the i-th walrus in the (k+1)-th iteration; X*(t) represents the position of the i-th walrus in the k-th iteration; X(t) represents the position of a randomly selected walrus individual; D' represents the position difference between the randomly selected walrus individual X(t) and the i-th walrus individual X*(t) in the k-th iteration; here, c is a constant with a value of 1; l is a random number in the range [-1, 1].

[0064] In the spiral search formula, e cl A random nonlinear expansion of the search radius was achieved, and cos(2πl) can generate a circular search trajectory at any angle, ensuring the diversity of search directions and effectively covering the corner regions of the solution space. In the early stage, there are significant differences between walrus individuals. By randomly selecting individuals, population difference information is introduced to perturb the search direction and maintain population diversity. In the later stage, the differences between individuals decrease, and walruses begin to explore locally. Through the synergistic effect of dynamic radius, omnidirectional search, and group perturbation, the exploration of the search space by walruses is expanded, achieving a balance between exploration and development.

[0065] Compared to migration, walrus groups tend to breed in the present when the risk factor is low (Dange_signal<1). Breeding behavior mainly consists of two types of behaviors: resting and foraging. When the safety factor is high (Safety_signal>0.5), they engage in resting behavior, and vice versa.

[0066] In habitat behavior, walrus population members are divided into three categories: males, females, and juveniles, each defined as X. male X female X Juvenile They update their positions in different ways; Xmale Position updates are performed using Halton sequences, X female and X Juvenile The update method is shown in the following formula:

[0067] (twenty two)

[0068] (twenty three)

[0069] (twenty four)

[0070] Where α is the iterative convergence factor, P represents the distress coefficient of juvenile walruses, which is a random number between 0 and 1; O represents the reference safe position, LF is a random number vector based on the Levy distribution, representing Levy movement; and foraging behavior includes two behaviors: escape and gathering.

[0071] Escape behavior: Walruses, even when foraging underwater, are vulnerable to predators. They will flee their current area based on danger signals from their companions (Dange_signal > 0.5). The update method is shown in the following formula:

[0072] (25)

[0073] (26)

[0074] Where R is as shown in Formula 26, and r1 and r4 are random numbers ranging from 0 to 1;

[0075] In terms of gregarious behavior, walruses can cooperate in foraging and move based on the location of other walruses in the population, as shown in the following formula:

[0076] (27)

[0077] (28)

[0078] (29)

[0079] (30)

[0080] Among them, X1 and X z These are two weighted factors influencing walrus aggregation behavior, X best (t) is the optimal solution, X second (t) is the suboptimal solution, a and b are the aggregation coefficients, r5 is a random number between 0 and 1, and θ is a random number between 0 and π.

[0081] S34. Recalculate the fitness value and update the walrus's position. The number of iterations in the walrus optimization algorithm is determined based on the actual situation. In each iteration, calculate the fitness of each individual, determine the position of the leader individual and the migrating individuals. When the maximum number of iterations is reached, the algorithm terminates and outputs the final optimal solution.

[0082] (31)

[0083] (32)

[0084] (33)

[0085] Among them, lls j and uls j These are the upper and lower bound values ​​of the j-th variable in the new generation during iterative updates; and These are the local upper and lower boundary values ​​of the j-th variable in the new generation during iterative updates.

[0086] According to a second aspect of the present invention, a multi-objective collaborative optimal scheduling apparatus is provided, wherein the apparatus is applied to a power system in the desert region, including thermal power units, wind power units, photovoltaic units, electrochemical energy storage systems, and a carbon trading market mechanism, to generate an optimal scheduling scheme. The multi-objective collaborative optimal scheduling apparatus includes the following modules: an objective function generation module, used to generate an objective function for the power system including thermal power units, wind power units, photovoltaic units, electrochemical energy storage systems, and a carbon trading market mechanism, wherein the new energy base in the desert region mainly consumes energy through long-distance DC transmission; a constraint condition addition module, used to add constraints related to the operation of the power system; and an optimal scheduling scheme generation module, used to generate an optimal scheduling scheme using the walrus algorithm.

[0087] According to an embodiment of the present invention, the objective function generation module generates the objective function through the following steps: calculating the minimum coal consumption cost and start-up / shutdown cost of the thermal power unit, using the following formula:

[0088] (1)

[0089] In the formula, C1 represents the operating cost of the thermal power unit, in yuan; C mh C qt - Coal consumption cost and start-up / shutdown cost of thermal power units, in yuan; T - Time of day, T=24h; a j b j c j - Coal consumption coefficient for thermal power units, t / MW; P G,j,t - Output value of thermal power unit j at time t, in MW; Sjt - Start-up and shutdown cost of thermal power unit j, in ten thousand yuan; u j,t - The start-up and shutdown status of thermal power unit j at time t;

[0090] The calculation of carbon emissions minimizes costs. Both solar and wind power are clean energy sources that do not produce CO2; therefore, the system's carbon emissions are considered in relation to thermal power units. The carbon allocation coefficient represents the proportion or weight of the allowable carbon emissions allocated to different units, used to determine the allocation of carbon emission responsibilities or the allocation rules in carbon trading. The total carbon emission allowance for the entire system and the total carbon emissions within a cycle are:

[0091] (2)

[0092] In the formula, M L - Total carbon emission allowance for the system, t; M P — Total carbon emissions of the system within one cycle, t; ϖ j - Carbon emission allocation coefficient per unit electricity of thermal power unit, t / MW; , - Carbon emission allocation coefficients for photovoltaic and wind turbine units, t / MW; P G,j,t ,P v,t ,P w,t - Output of thermal power units, photovoltaic power units, and wind power units at time t, in MW; λ j - Carbon emission intensity, t / MW;

[0093] A tiered carbon trading cost model is adopted, dividing the purchase of carbon emission rights into multiple tiers. The more carbon emission rights required to purchase, the higher the price in the corresponding tier. The specific calculation formula of the model is as follows:

[0094] (3)

[0095] In the formula, C2 is the total carbon trading cost in yuan; ω is the carbon trading price in yuan / t; d is the length of the carbon emission range in t; and τ is the increase in carbon trading price for each step up in carbon emissions in yuan.

[0096] The maximum renewable energy absorption capacity and minimum cost are achieved through wind and solar power integration. Wind and solar power integration capacity is represented by the amount of wind and solar power curtailment during the dispatch cycle; the higher the curtailment, the weaker the integration capacity. Since wind and solar power generation do not consume fuel, only operation and maintenance costs are considered here. The specific calculation formula is as follows:

[0097]

[0098] In the formula, C w,t C cw,t - Wind farm operation and maintenance costs, wind curtailment penalty costs, in yuan; C v,t Ccv,t - Photovoltaic power plant operation and maintenance costs, curtailment penalty costs, in yuan / MW; k w k cw -Wind farm operating cost coefficient, wind curtailment penalty cost coefficient, yuan / MW; k v k cv - Photovoltaic power plant operating cost coefficient and curtailment penalty cost coefficient, in yuan; , -Predicted wind power and photovoltaic power generation at time t, in MW;

[0099] The system aims to maximize its benefits and minimize its costs.

[0100] (5)

[0101] In the formula, p price -Grid electricity price, yuan; C sy C cb - The discharge revenue and charging cost of an energy storage power station, in yuan; ν d ν c - The discharge and charging efficiency of the energy storage power station, %; P sd,t ,P sc,t - The discharge and charge power of the electrochemical energy storage power station at time t, in MW; c sc - Charging cost coefficient for energy storage power stations, in yuan / MW; c sd - Discharge cost coefficient, RMB / MW;

[0102] The sum of the four objective functions C1 to C4 is taken as the overall objective function to be scheduled.

[0103] According to an embodiment of the present invention, the constraint addition module can add the following constraints:

[0104] Establish system power balance constraints:

[0105] (6)

[0106] In the formula, -Active power loss of the system at time t, in MW;

[0107] Operating constraints for thermal power units include:

[0108] Unit output constraints

[0109] (7)

[0110] In the formula, , -Minimum and maximum output of the j-th thermal power unit, in MW;

[0111] Unit ramp-up constraints,

[0112] (8)

[0113] In the formula, - The gradeability of thermal power unit j, in MW / h;

[0114] Unit start-up and shutdown status constraints.

[0115] (9)

[0116] In the formula, , - The start-up and shutdown actions of the jth thermal power unit at time t;

[0117] Wind power output constraints

[0118] (10)

[0119] Energy storage constraints include:

[0120] Energy storage constraints

[0121] (11)

[0122] Energy storage charging and discharging constraints,

[0123] (12)

[0124] In the formula, S t - Energy storage capacity, MW; θ i - Self-loss rate; φ sc,t - Charging efficiency, % φ sd,t -Discharge efficiency, % u sc ,u sd,t - Charging state, discharging state; S t,min S t,max - Capacity upper limit, capacity lower limit, MW; P sc,max - Maximum charging power, MW; P sd,max - Maximum discharge power, MW.

[0125] According to an embodiment of the present invention, the module for generating the optimal scheduling scheme can generate the optimal scheduling scheme according to the following steps:

[0126] The first step is to initialize the population, determining the population size as N and the dimension m of each individual. Each individual is a vector composed of a set of parameters to be optimized, representing a candidate solution to an optimization problem, i.e., a single objective function value; the initial value of each individual in the search space is a vector x. i,j , that is, x i,j =Lb i,j +rand(xi,j (Ub) i,j -Lb i,j ), that is, the decision variables within a single objective function, and the total objective function value matrix X obtained from all candidate solutions:

[0127] (13)

[0128] Among them, Ub i,j Lb i,j The upper and lower boundary values ​​of the candidate solution represent the constraints of each decision variable; rand(x) i,j ) is a random uniform function, and rand(x) i,j X ∈ [0,1]; X is the matrix of all individual candidate positions, i.e., the total objective function value; X i Xi,j is the vector of the i-th individual, representing the objective function value of a single individual; Xi,j are the decision variables within a single objective function; N is the value of the overall objective function; i∈{1,2,⋯,N}, j is the j-th dimension of the i-th individual, and m is the individual dimension, i.e., the number of single objective functions; where Xi,j are the decision variables within a single objective function, including: T - time of day, T=24h; a j b j c j - Coal consumption coefficient for thermal power units, t / MW; P G,j,t - Output value of thermal power unit j at time t, in MW; S jt - Start-up and shutdown cost of thermal power unit j, in ten thousand yuan; u j,t - The start-up and shutdown status of thermal power unit j at time t; M L - Total carbon emission allowance for the system, t; M P — Total carbon emissions of the system within one cycle, t; ϖ j - Carbon emission allocation coefficient per unit electricity of thermal power unit, t / MW; , - Carbon emission allocation coefficients for photovoltaic and wind turbine units, t / MW; P G,j,t ,P v,t ,P w,t - Output of thermal power units, photovoltaic power units, and wind power units at time t, in MW; λ j - Carbon emission intensity, t / MW; ω - Carbon trading price, yuan / t; d - Length of carbon emission range, t; τ - Increase in carbon trading price for each carbon emission increment, yuan; k w k cw -Wind farm operating cost coefficient, wind curtailment penalty cost coefficient, yuan / MW; k v k cv - Photovoltaic power plant operating cost coefficient and curtailment penalty cost coefficient, in yuan; , -Predicted wind power and photovoltaic power generation at time t, in MW; p price -Grid electricity price, yuan; ν d ν c - The discharge and charging efficiency of the energy storage power station, %; P sd,t ,P sc,t - The discharge and charge power of the electrochemical energy storage power station at time t, in MW; c sc —Charging cost coefficient for energy storage power stations, yuan / MW; c sd - Discharge cost coefficient, RMB / MW;

[0129] X i Let be the vector of the i-th individual, i.e., the objective function value of a single individual, namely: C1 - operating cost of thermal power unit, yuan; C2 - total carbon trading cost, yuan; C3 - new energy absorption capacity and cost, yuan; C4 - operating revenue and cost of energy storage system, yuan;

[0130] X is the matrix of all individual candidate positions, which is the total objective function value, namely minC1+minC2+minC3+maxC4. This means that the coal consumption cost and start-up and shutdown cost of thermal power units are minimized, carbon emission cost is minimized, new energy absorption capacity is maximized and cost is minimized, and the operating benefits of energy storage system are maximized and cost is minimized. Achieving these four aspects simultaneously is considered as the realization of multi-objective coordinated low-carbon optimal scheduling in the desert area.

[0131] The second step is fitness calculation. Each individual is a candidate solution to the problem. Based on its fitness value (in this invention, the smaller value of individual objective function values ​​C1~C3 and the larger value of C4 indicate better fitness), the individual objective function can be evaluated. The fitness value is calculated for each candidate position, resulting in:

[0132] (14)

[0133] Where F is the fitness vector, and F i It is the fitness vector of the i-th candidate position;

[0134] Here, we explore and identify the leader individual; the leader individual is the one with the best fitness in the population, with either the minimum or maximum fitness value, to guide other individuals in optimization. We then update the individual positions and generate a new leader individual.

[0135] (15)

[0136] (16)

[0137] (17)

[0138] (18)

[0139] Among them, X i It is the new position of the i-th generated individual; It is the j-th dimension of the new position; It is the objective function value at the new position; rand i,j It is a random number in the interval [0,1]; SW j It is the best candidate solution for the strongest individual; , Indicates the safety and danger signals at the new location; T is the total number of iterations, and t is the current iteration number;

[0140] The third step involves using different update strategies to adjust the positions of individuals in the population based on the different value ranges of the safety and danger signals.

[0141] Migration behavior is judged when the risk factor is too high (Dange_signal≧1), the walrus population will migrate to an area more suitable for the population's survival; assuming that each individual migrates from one region of the search space to another random location, the new location is generated by equation (19), that is:

[0142] (19)

[0143] (20)

[0144] According to Equation 19, if the new position changes the value of the objective function, then the new position replaces the previous position. Wherein, It is the position generated by the migration of the i-th individual; It is the objective function value at the new position; X k， k∈{1,2,⋯,N}, and k≠i is the location chosen for the migration of the i-th individual, F k It is the objective function value of the migration location;

[0145] During migration, the walrus herd expands its search area, engaging in a global search. However, with increasing iterations, the ordinary walrus algorithm tends towards local search development, especially after iteration T / 2, when the herd fully enters the development phase. In the early stages of the algorithm, the herd not only performs global search exploration but also engages in significant development. Initially, the distribution of global and local searches is not well balanced, potentially leading to premature convergence and hindering the herd's exploration of the solution space, thus impacting the final optimization. To enhance the global search capability of the walrus algorithm, a spiral search strategy inspired by natural spiral motion is adopted. This strategy strengthens the algorithm's global optimization ability, ensuring convergence speed and increasing individual diversity. The spiral search formula is as follows:

[0146] (twenty one)

[0147] Where X(t+1) represents the position of the i-th walrus in the (k+1)-th iteration; X*(t) represents the position of the i-th walrus in the k-th iteration; X(t) represents the position of a randomly selected walrus individual; D' represents the position difference between the randomly selected walrus individual X(t) and the i-th walrus individual X*(t) in the k-th iteration; here, c is a constant with a value of 1; l is a random number in the range [-1, 1].

[0148] In the spiral search formula, e cl A random nonlinear expansion of the search radius was achieved, and cos(2πl) can generate a circular search trajectory at any angle, ensuring the diversity of search directions and effectively covering the corner regions of the solution space. In the early stage, there are significant differences between walrus individuals. By randomly selecting individuals, population difference information is introduced to perturb the search direction and maintain population diversity. In the later stage, the differences between individuals decrease, and walruses begin to explore locally. Through the synergistic effect of dynamic radius, omnidirectional search, and group perturbation, the exploration of the search space by walruses is expanded, achieving a balance between exploration and development.

[0149] Compared to migration, walrus groups tend to breed in the present when the risk factor is low (Dange_signal<1). Breeding behavior mainly consists of two types of behaviors: resting and foraging. When the safety factor is high (Safety_signal>0.5), they engage in resting behavior, and vice versa.

[0150] In habitat behavior, walrus population members are divided into three categories: males, females, and juveniles, each defined as X. male X female X Juvenile They update their positions in different ways; Xmale Position updates are performed using Halton sequences, X female and X Juvenile The update method is shown in the following formula:

[0151] (twenty two)

[0152] (twenty three)

[0153] (twenty four)

[0154] Where α is the iterative convergence factor, P represents the distress coefficient of juvenile walruses, which is a random number between 0 and 1; O represents the reference safe position, LF is a random number vector based on the Levy distribution, representing Levy movement; and foraging behavior includes two behaviors: escape and gathering.

[0155] Escape behavior: Walruses, even when foraging underwater, are vulnerable to predators. They will flee their current area based on danger signals from their companions (Dange_signal > 0.5). The update method is shown in the following formula:

[0156] (25)

[0157] (26)

[0158] Where R is as shown in Formula 26, and r1 and r4 are random numbers ranging from 0 to 1;

[0159] In terms of gregarious behavior, walruses can cooperate in foraging and move based on the location of other walruses in the population, as shown in the following formula:

[0160] (27)

[0161] (28)

[0162] (29)

[0163] (30)

[0164] Among them, X1 and X z These are two weighted factors influencing walrus aggregation behavior, X best (t) is the optimal solution, X second (t) is the suboptimal solution, a and b are the aggregation coefficients, r5 is a random number between 0 and 1, and θ is a random number between 0 and π.

[0165] The fourth step is to recalculate the fitness values ​​and update the walrus positions. The number of iterations in the walrus optimization algorithm is determined based on the actual situation. In each iteration, the fitness of each individual is calculated, and the positions of the leader and migrating individuals are determined. When the maximum number of iterations is reached, the algorithm terminates and outputs the final optimal solution.

[0166] (31)

[0167] (32)

[0168] (33)

[0169] Among them, lls j and uls j These are the upper and lower bound values ​​of the j-th variable in the new generation during iterative updates; and These are the local upper and lower boundary values ​​of the j-th variable in the new generation during iterative updates.

[0170] According to a third aspect of the present invention, an electronic device is provided, comprising: a memory, a processor, and a multi-objective cooperative optimal scheduler stored in the memory and executable on the processor, wherein the multi-objective cooperative optimal scheduler, when executed by the processor, implements the steps of the multi-objective cooperative optimal scheduling method described above.

[0171] According to a fourth aspect of the present invention, a computer storage medium is provided, wherein a multi-objective cooperative optimal scheduler is stored on the computer storage medium, and when the multi-objective cooperative optimal scheduler is executed by a processor, it implements the steps of the multi-objective cooperative optimal scheduling method described above.

[0172] Compared with the prior art, the technical solution provided by the embodiments of the present invention can achieve at least the following beneficial effects:

[0173] The method, apparatus, electronic device, and storage medium for multi-objective cooperative optimal scheduling provided by this invention address the difficulties in model establishment caused by multiple objectives and the mixed advantages and disadvantages of various solution methods. Based on the walrus algorithm optimization scheme, a spiral search strategy is introduced to enhance the global search capability of the walrus algorithm, ensuring the convergence speed of the algorithm and increasing the diversity of individuals. In addition, a tiered carbon trading cost is introduced into the objective function to establish a multi-objective cooperative low-carbon optimal scheduling model, achieving the goals of promoting the consumption of new energy sources while improving the economic efficiency of system operation and reducing the carbon emissions of the power system. Attached Figure Description

[0174] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings of the embodiments will be briefly described below. Obviously, the drawings described below only relate to some embodiments of the present invention and are not intended to limit the present invention.

[0175] Figure 1 This is a flowchart illustrating the generation of the optimal scheduling scheme for multi-objective cooperative optimal scheduling according to an embodiment of the present invention. Detailed Implementation

[0176] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the described embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0177] Unless otherwise defined, the technical or scientific terms used herein shall have the ordinary meaning as understood by one of ordinary skill in the art to which this invention pertains. The terms “first,” “second,” and similar terms used in the specification and claims of this patent application do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Similarly, the terms “an” or “a” and similar terms do not indicate a limitation of quantity, but rather indicate the presence of at least one.

[0178] The multi-objective collaborative optimal scheduling method is applied to the power system in the desert region, which includes thermal power units, wind power units, photovoltaic units, electrochemical energy storage systems, and carbon trading market mechanisms, to generate optimal scheduling schemes.

[0179] The method for multi-objective cooperative optimal scheduling includes the following steps:

[0180] S1. Generate objective function: Generate the objective function of the power system, including thermal power units, wind power units, photovoltaic units, electrochemical energy storage systems and carbon trading market mechanisms. Among them, the new energy base in the Shagohuang area mainly uses long-distance DC transmission as the consumption form.

[0181] S2. Add constraints: Add constraints related to the operation of the power system.

[0182] S3. Generate the optimal scheduling scheme using the Walrus algorithm.

[0183] According to one or more embodiments of the present invention, step S1 includes the following steps:

[0184] S11. The following formula minimizes the coal consumption cost and start-up / shutdown cost of thermal power units:

[0185] (1)

[0186] In the formula, C1 represents the operating cost of the thermal power unit, in yuan; C mh C qt - Coal consumption cost and start-up / shutdown cost of thermal power units, in yuan; T - Time of day, T=24h; a j b j c j - Coal consumption coefficient for thermal power units, t / MW; P G,j,t - Output value of thermal power unit j at time t, in MW; S jt - Start-up and shutdown cost of thermal power unit j, in ten thousand yuan; u j,t - The start-up and shutdown status of thermal power unit j at time t;

[0187] S12, Minimizes the cost of calculating carbon emissions:

[0188] Both photovoltaic and wind power are clean energy sources that do not produce CO2; therefore, the system's carbon emissions are considered in relation to thermal power units. The carbon allocation factor represents the proportion or weight of the carbon emission allowance allocated to different units, used to determine the allocation of carbon emission responsibilities or the allocation rules in carbon trading. The total carbon emission allowance for the entire system and the total carbon emissions of the system within a cycle are:

[0189] (2)

[0190] In the formula, M L - Total carbon emission allowance for the system, t; M P — Total carbon emissions of the system within one cycle, t; ϖ j - Carbon emission allocation coefficient per unit electricity of thermal power unit, t / MW; , - Carbon emission allocation coefficients for photovoltaic and wind turbine units, t / MW; P G,j,t ,P v,t ,P w,t - Output of thermal power units, photovoltaic power units, and wind power units at time t, in MW; λ j - Carbon emission intensity, t / MW;

[0191] A tiered carbon trading cost model is adopted, dividing the purchase of carbon emission rights into multiple tiers. The more carbon emission rights required to purchase, the higher the price in the corresponding tier. The specific calculation formula of the model is as follows:

[0192] (3)

[0193] In the formula, C2 is the total carbon trading cost in yuan; ω is the carbon trading price in yuan / t; d is the length of the carbon emission range in t; and τ is the increase in carbon trading price for each step up in carbon emissions in yuan.

[0194] S13: Maximum new energy absorption capacity and lowest cost:

[0195] The wind and solar power absorption capacity is represented by the amount of wind and solar power curtailment during the dispatch cycle. The higher the amount of curtailment, the weaker the wind and solar power absorption capacity. Since wind power and photovoltaic power generation do not consume fuel, only operation and maintenance costs are considered here. The specific calculation formula is as follows:

[0196]

[0197] In the formula, C w,t C cw,t - Wind farm operation and maintenance costs, wind curtailment penalty costs, in yuan; C v,t C cv,t - Photovoltaic power plant operation and maintenance costs, curtailment penalty costs, in yuan / MW; k w k cw -Wind farm operating cost coefficient, wind curtailment penalty cost coefficient, yuan / MW; k v k cv - Photovoltaic power plant operating cost coefficient and curtailment penalty cost coefficient, in yuan; , -Predicted wind power and photovoltaic power generation at time t, in MW;

[0198] S14. Maximizing the operating benefits and minimizing the costs of energy storage systems:

[0199] (5)

[0200] In the formula, p price -Grid electricity price, yuan; C sy C cb - The discharge revenue and charging cost of an energy storage power station, in yuan; ν d ν c - The discharge and charging efficiency of the energy storage power station, %; P sd,t ,P sc,t - The discharge and charge power of the electrochemical energy storage power station at time t, in MW; c sc - Charging cost coefficient for energy storage power stations, in yuan / MW; c sd - Discharge cost coefficient, RMB / MW;

[0201] S15. Take the sum of the above four objective functions C1~C4 as the total objective function to be scheduled.

[0202] According to one or more embodiments of the present invention, step S2 includes the following steps:

[0203] S21. Establish system power balance constraints:

[0204] (6)

[0205] In the formula, -Active power loss of the system at time t, in MW;

[0206] S22. Operating constraints of thermal power units, including:

[0207] Unit output constraints

[0208] (7)

[0209] In the formula, , -Minimum and maximum output of the j-th thermal power unit, in MW;

[0210] Unit ramp-up constraints,

[0211] (8)

[0212] In the formula, - The gradeability of thermal power unit j, in MW / h;

[0213] Unit start-up and shutdown status constraints.

[0214] (9)

[0215] In the formula, , - The start-up and shutdown actions of the jth thermal power unit at time t;

[0216] S23, Wind power output constraints

[0217] (10)

[0218] S24. Energy storage constraints, including:

[0219] Energy storage constraints

[0220] (11)

[0221] Energy storage charging and discharging constraints,

[0222] (12)

[0223] In the formula, S t - Energy storage capacity, MW; θ i - Self-loss rate; φ sc,t - Charging efficiency, % φsd,t -Discharge efficiency, % u sc ,u sd,t - Charging state, discharging state; S t,min S t,max - Capacity upper limit, capacity lower limit, MW; P sc,max - Maximum charging power, MW; P sd,max - Maximum discharge power, MW.

[0224] Figure 1 This is a flowchart illustrating a method for multi-objective cooperative optimal scheduling according to an embodiment of the present invention.

[0225] The Walrus Optimization Algorithm is an optimization algorithm inspired by the natural behavior of walrus populations. Its design draws on behaviors such as migration, reproduction, habitat roosting, foraging, gathering, and escape, where the population determines its behavioral strategies based on danger and safety signals. In the Walrus Optimization Algorithm, these population behaviors are abstracted as different processes for solving the optimization problem. At the start of the algorithm, an initial population is randomly generated, with each individual representing a walrus in the simulated environment, corresponding to an initial solution in the solution space. The algorithm then simulates the walrus's behavior adjustment based on its perception of danger and safety signals. When the algorithm determines that the current environmental risk is high, the simulated walrus population migrates, expanding the search area and thus conducting a global exploration to discover new, potentially better solutions. Conversely, when the environmental risk is low, the algorithm simulates more localized activities of the walrus population, such as reproduction, habitat roosting, or gathering, allowing for a more refined local search around the current solution, thereby improving the accuracy of the solution.

[0226] like Figure 1 As shown, step S3 includes the following steps:

[0227] S31. Initialize the population, determining the population size as N and the dimension m of each individual. Each individual is a vector composed of a set of parameters to be optimized, representing a candidate solution to an optimization problem, i.e., a single objective function value; the initial value of each individual in the search space is a vector x. i,j , that is, x i,j =Lb i,j +rand(x i,j (Ub) i,j -Lb i,j ), that is, the decision variables within a single objective function, and the total objective function value matrix X obtained from all candidate solutions:

[0228] (13)

[0229] Among them, Ub i,j Lb i,j The upper and lower boundary values ​​of the candidate solution represent the constraints of each decision variable; rand(x)i,j ) is a random uniform function, and rand(x) i,j X ∈ [0,1]; X is the matrix of all individual candidate positions, i.e., the total objective function value; X i Xi,j is the vector of the i-th individual, i.e., the objective function value of a single individual; Xi,j is the decision variable within a single objective function; N is the value of the overall objective function; i∈{1,2,⋯,N}, j is the j-th dimension of the i-th individual, and m is the individual dimension, i.e., the number of single objective functions;

[0230] Where Xi,j are decision variables within a single objective function, including: T - time of day, T=24h; a j b j c j - Coal consumption coefficient for thermal power units, t / MW; P G,j,t - Output value of thermal power unit j at time t, in MW; S jt - Start-up and shutdown cost of thermal power unit j, in ten thousand yuan; u j,t - The start-up and shutdown status of thermal power unit j at time t; M L - Total carbon emission allowance for the system, t; M P — Total carbon emissions of the system within one cycle, t; ϖ j - Carbon emission allocation coefficient per unit electricity of thermal power unit, t / MW; , - Carbon emission allocation coefficients for photovoltaic and wind turbine units, t / MW; P G,j,t ,P v,t ,P w,t - Output of thermal power units, photovoltaic power units, and wind power units at time t, in MW; λ j - Carbon emission intensity, t / MW; ω - Carbon trading price, yuan / t; d - Length of carbon emission range, t; τ - Increase in carbon trading price for each carbon emission increment, yuan; k w k cw -Wind farm operating cost coefficient, wind curtailment penalty cost coefficient, yuan / MW; k v k cv - Photovoltaic power plant operating cost coefficient and curtailment penalty cost coefficient, in yuan; , -Predicted wind power and photovoltaic power generation at time t, in MW; p price -Grid electricity price, yuan; ν d ν c - The discharge and charging efficiency of the energy storage power station, %; P sd,t ,P sc,t - The discharge and charge power of the electrochemical energy storage power station at time t, in MW; c sc —Charging cost coefficient for energy storage power stations, yuan / MW; c sd- Discharge cost coefficient, RMB / MW;

[0231] X i Let be the vector of the i-th individual, i.e., the objective function value of a single individual, namely: C1 - operating cost of thermal power unit, yuan; C2 - total carbon trading cost, yuan; C3 - new energy absorption capacity and cost, yuan; C4 - operating revenue and cost of energy storage system, yuan;

[0232] X is the matrix of all individual candidate positions, which is the total objective function value, namely minC1+minC2+minC3+maxC4. This means that the coal consumption cost and start-up and shutdown cost of thermal power units are minimized, carbon emission cost is minimized, new energy absorption capacity is maximized and cost is minimized, and the operating benefits of energy storage system are maximized and cost is minimized. Achieving these four aspects simultaneously is considered as the realization of multi-objective coordinated low-carbon optimal scheduling in the desert area.

[0233] S32. Fitness Calculation: Each individual is a candidate solution to the problem. Based on its fitness value (in this invention, the smaller value of individual objective function values ​​C1~C3 and the larger value of C4 indicate better fitness), the individual objective function can be evaluated. The fitness value is calculated for each candidate position, resulting in:

[0234] (14)

[0235] Where F is the fitness vector, and F i It is the fitness vector of the i-th candidate position;

[0236] Here, we explore and identify the leader individual; the leader individual is the one with the best fitness in the population, with either the minimum or maximum fitness value, to guide other individuals in optimization. We then update the individual positions and generate a new leader individual.

[0237] (15)

[0238] (16)

[0239] (17)

[0240] (18)

[0241] Among them, X i It is the new position of the i-th generated individual; It is the j-th dimension of the new position; It is the objective function value at the new position; rand i,j It is a random number in the interval [0,1]; SW j It is the best candidate solution for the strongest individual; , Indicates the safety and danger signals at the new location; T is the total number of iterations, and t is the current iteration number;

[0242] S33. Different update strategies are used to adjust the position of individuals in the population based on the different value ranges of safety signals and danger signals.

[0243] Migration behavior is judged when the risk factor is too high (Dange_signal≧1), the walrus population will migrate to an area more suitable for the population's survival; assuming that each individual migrates from one region of the search space to another random location, the new location is generated by equation (19), that is:

[0244] (19)

[0245] (20)

[0246] According to Equation 19, if the new position changes the value of the objective function, then the new position replaces the previous position. Wherein, It is the position generated by the migration of the i-th individual; It is the objective function value at the new position; X k， k∈{1,2,⋯,N}, and k≠i is the location chosen for the migration of the i-th individual, F k It is the objective function value of the migration location;

[0247] During migration, the walrus herd expands its search area, engaging in a global search. However, with increasing iterations, the ordinary walrus algorithm tends towards local search development, especially after iteration T / 2, when the herd fully enters the development phase. In the early stages of the algorithm, the herd not only performs global search exploration but also engages in significant development. Initially, the distribution of global and local searches is not well balanced, potentially leading to premature convergence and hindering the herd's exploration of the solution space, thus impacting the final optimization. To enhance the global search capability of the walrus algorithm, a spiral search strategy inspired by natural spiral motion is adopted. This strategy strengthens the algorithm's global optimization ability, ensuring convergence speed and increasing individual diversity. The spiral search formula is as follows:

[0248] (twenty one)

[0249] Where X(t+1) represents the position of the i-th walrus in the (k+1)-th iteration; X*(t) represents the position of the i-th walrus in the k-th iteration; X(t) represents the position of a randomly selected walrus individual; D' represents the position difference between the randomly selected walrus individual X(t) and the i-th walrus individual X*(t) in the k-th iteration; here, c is a constant with a value of 1; l is a random number in the range [-1, 1].

[0250] In the spiral search formula, e cl A random nonlinear expansion of the search radius was achieved, and cos(2πl) can generate a circular search trajectory at any angle, ensuring the diversity of search directions and effectively covering the corner regions of the solution space. In the early stage, there are significant differences between walrus individuals. By randomly selecting individuals, population difference information is introduced to perturb the search direction and maintain population diversity. In the later stage, the differences between individuals decrease, and walruses begin to explore locally. Through the synergistic effect of dynamic radius, omnidirectional search, and group perturbation, the exploration of the search space by walruses is expanded, achieving a balance between exploration and development.

[0251] Compared to migration, walrus groups tend to breed in the present when the risk factor is low (Dange_signal<1). Breeding behavior mainly consists of two types of behaviors: resting and foraging. When the safety factor is high (Safety_signal>0.5), they engage in resting behavior, and vice versa.

[0252] In habitat behavior, walrus population members are divided into three categories: males, females, and juveniles, each defined as X. male X female X Juvenile They update their positions in different ways; X male Position updates are performed using Halton sequences, X female and X Juvenile The update method is shown in the following formula:

[0253] (twenty two)

[0254] (twenty three)

[0255] (twenty four)

[0256] Where α is the iterative convergence factor, P represents the distress coefficient of juvenile walruses, which is a random number between 0 and 1; O represents the reference safe position, LF is a random number vector based on the Levy distribution, representing Levy movement; and foraging behavior includes two behaviors: escape and gathering.

[0257] Escape behavior: Walruses, even when foraging underwater, are vulnerable to predators. They will flee their current area based on danger signals from their companions (Dange_signal > 0.5). The update method is shown in the following formula:

[0258] (25)

[0259] (26)

[0260] Where R is as shown in Formula 26, and r1 and r4 are random numbers ranging from 0 to 1;

[0261] In terms of gregarious behavior, walruses can cooperate in foraging and move based on the location of other walruses in the population, as shown in the following formula:

[0262] (27)

[0263] (28)

[0264] (29)

[0265] (30)

[0266] Among them, X1 and X z These are two weighted factors influencing walrus aggregation behavior, X best (t) is the optimal solution, X second (t) is the suboptimal solution, a and b are the aggregation coefficients, r5 is a random number between 0 and 1, and θ is a random number between 0 and π.

[0267] S34. Recalculate the fitness value and update the walrus's position. The number of iterations in the walrus optimization algorithm is determined based on the actual situation. In each iteration, calculate the fitness of each individual, determine the position of the leader individual and the migrating individuals. When the maximum number of iterations is reached, the algorithm terminates and outputs the final optimal solution.

[0268] (31)

[0269] (32)

[0270] (33)

[0271] Among them, lls j and uls j These are the upper and lower bound values ​​of the j-th variable in the new generation during iterative updates; and These are the local upper and lower boundary values ​​of the j-th variable in the new generation during iterative updates.

[0272] According to a second aspect of the present invention, a multi-objective cooperative optimal scheduling apparatus is provided, which is implemented using the aforementioned multi-objective cooperative optimal scheduling method. The multi-objective cooperative optimal scheduling apparatus includes the following modules: an objective function generation module, used to generate an objective function for a power system including thermal power units, wind power units, photovoltaic units, electrochemical energy storage systems, and a carbon trading market mechanism, wherein the new energy base in the desert region mainly consumes power through long-distance DC transmission; a constraint condition addition module, used to add constraints related to power system operation; and an optimal scheduling scheme generation module, used to generate an optimal scheduling scheme using the walrus algorithm.

[0273] The execution pseudocode of this invention is as follows:

[0274] Input: Initialized population, population size N, maximum number of iterations T

[0275] Output: The optimal solution to the overall objective function

[0276] WaOA_FOS(int N, int T)

[0277] {

[0278] 1) Initialize the population and calculate fitness values:

[0279] 2) Follow step 1) to create the initial population → X;

[0280] 3) Calculate the fitness value of the initial population → F;

[0281] 4) for t = 1 to T do

[0282] 5) Calculate the control parameters Dsign and Ssign;

[0283] 6) Select different behavior patterns based on the values ​​of Dsign and Ssign:

[0284] 7) ifDsign>= 1 then

[0285] 8) For each individual, update the strongest individual using equations (15) to (18);

[0286] 9) else if Ssign >= 0.5 then

[0287] 10) For each individual, calculate the new position using equation (19) and update the position using equation (21);

[0288] 11) else if Dsign >= 0.5 then

[0289] 12) For each individual, use equations (31) and (32) to calculate the new location of the individual neighbors;

[0290] 13) else

[0291] 14) For each individual, update the position using equation (33);

[0292] 15) Update the location of each individual;

[0293] 16) Calculate the fitness value of the population;

[0294] 17) Update the current optimal position and save the optimal candidate solution;

[0295] 18) endfor

[0296] 19) Output the location of the optimal solution.

[0297] }

[0298] The method, apparatus, electronic device, and storage medium for multi-objective cooperative optimal scheduling provided by this invention address the difficulties in model establishment caused by multiple objectives and the mixed advantages and disadvantages of various solution methods. Based on the walrus algorithm optimization scheme, a spiral search strategy is introduced to enhance the global search capability of the walrus algorithm, ensuring the convergence speed of the algorithm and increasing the diversity of individuals. In addition, a tiered carbon trading cost is introduced into the objective function to establish a multi-objective cooperative low-carbon optimal scheduling model, achieving the goals of promoting the consumption of new energy sources while improving the economic efficiency of system operation and reducing the carbon emissions of the power system.

[0299] According to another aspect of the present invention, a device for multi-objective cooperative optimal scheduling is provided, comprising: a memory, a processor, and a multi-objective cooperative optimal scheduling program stored in the memory and executable on the processor, wherein the multi-objective cooperative optimal scheduling program implements the steps of the multi-objective cooperative optimal scheduling method described above when executed by the processor.

[0300] The present invention also provides a computer storage medium.

[0301] The computer storage medium stores a multi-objective cooperative optimal scheduler, which, when executed by the processor, implements the steps of the multi-objective cooperative optimal scheduling method described above.

[0302] The method implemented when the multi-objective cooperative optimal scheduler running on the processor is executed can be referred to in various embodiments of the multi-objective cooperative optimal scheduling method of the present invention, and will not be repeated here.

[0303] The present invention also provides a computer program product.

[0304] The computer program product of the present invention includes a multi-objective cooperative optimal scheduler, which, when executed by a processor, implements the steps of the multi-objective cooperative optimal scheduling method as described above.

[0305] The method implemented when the multi-objective cooperative optimal scheduler running on the processor is executed can be referred to in various embodiments of the multi-objective cooperative optimal scheduling method of the present invention, and will not be repeated here.

[0306] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0307] The above description is merely an exemplary embodiment of the present invention and is not intended to limit the scope of protection of the present invention, which is determined by the appended claims.

Claims

1. A method for multi-objective cooperative optimal scheduling, wherein, The multi-objective collaborative optimal scheduling method is applied to the power system in the Sakhalin region, which includes thermal power units, wind power units, photovoltaic units, electrochemical energy storage systems, and carbon trading market mechanisms, to generate optimal scheduling schemes. The multi-objective cooperative optimal scheduling method includes the following steps: S1. Generate objective function: Generate the objective function of the power system, including thermal power units, wind power units, photovoltaic units, electrochemical energy storage systems and carbon trading market mechanisms. Among them, the new energy base in the Shagohuang area mainly uses long-distance DC transmission as the consumption form. S2. Add constraints: Add constraints related to the operation of the power system. S3. Generate the optimal scheduling scheme using the Walrus algorithm.

2. The multi-objective cooperative optimal scheduling method as described in claim 1, wherein, Step S1 includes the following steps: S11. The following formula minimizes the coal consumption cost and start-up / shutdown cost of thermal power units: （1） In the formula, C1 represents the operating cost of the thermal power unit, in yuan; C mh C qt - Coal consumption cost and start-up / shutdown cost of thermal power units, in yuan; T - Time of day, T=24h; a j b j c j - Coal consumption coefficient for thermal power units, t / MW; P G,j,t - Output value of thermal power unit j at time t, in MW; S jt - Start-up and shutdown cost of thermal power unit j, in ten thousand yuan; u j,t - The start-up and shutdown status of thermal power unit j at time t; S12, Minimizes the cost of calculating carbon emissions: Both photovoltaic and wind power are clean energy sources that do not produce CO2; therefore, the system's carbon emissions are considered in relation to thermal power units. The carbon allocation factor represents the proportion or weight of the carbon emission allowance allocated to different units, used to determine the allocation of carbon emission responsibilities or the allocation rules in carbon trading. The total carbon emission allowance for the entire system and the total carbon emissions of the system within a cycle are: （2） In the formula, M L - Total carbon emission allowance for the system, t; M P — Total carbon emissions of the system within one cycle, t; ϖ j - Carbon emission allocation coefficient per unit electricity of thermal power unit, t / MW; , - Carbon emission allocation coefficients for photovoltaic and wind turbine units, t / MW; P G,j,t ,P v,t ,P w,t - Output of thermal power units, photovoltaic power units, and wind power units at time t, in MW; λ j - Carbon emission intensity, t / MW; A tiered carbon trading cost model is adopted, dividing the purchase of carbon emission rights into multiple tiers. The more carbon emission rights required to purchase, the higher the price in the corresponding tier. The specific calculation formula of the model is as follows: （3） In the formula, C2 is the total carbon trading cost in yuan; ω is the carbon trading price in yuan / t; d is the length of the carbon emission range in t; and τ is the increase in carbon trading price for each step up in carbon emissions in yuan. S13: Maximum new energy absorption capacity and lowest cost: The wind and solar power absorption capacity is represented by the amount of wind and solar power curtailment during the dispatch cycle. The higher the amount of curtailment, the weaker the wind and solar power absorption capacity. Since wind power and photovoltaic power generation do not consume fuel, only operation and maintenance costs are considered here. The specific calculation formula is as follows: ， Among them, C w,t C cw,t - Wind farm operation and maintenance costs, wind curtailment penalty costs, in yuan; C v,t C cv,t - Photovoltaic power plant operation and maintenance costs, curtailment penalty costs, in yuan / MW; k w k cw -Wind farm operating cost coefficient, wind curtailment penalty cost coefficient, yuan / MW; k v k cv - Photovoltaic power plant operating cost coefficient and curtailment penalty cost coefficient, in yuan; , -Predicted wind power and photovoltaic power generation at time t, in MW; S14. Maximizing the operating benefits and minimizing the costs of energy storage systems: （5） In the formula, p price -Grid electricity price, yuan; C sy C cb - The discharge revenue and charging cost of an energy storage power station, in yuan; ν d ν c - The discharge and charging efficiency of the energy storage power station, %; P sd,t ,P sc,t - The discharge and charge power of the electrochemical energy storage power station at time t, in MW; c sc - Charging cost coefficient for energy storage power stations, in yuan / MW; c sd - Discharge cost coefficient, RMB / MW; S15. Take the sum of the above four objective functions C1~C4 as the total objective function to be scheduled.

3. The multi-objective cooperative optimal scheduling method as described in claim 2, wherein, Step S2 includes the following steps: S21. Establish system power balance constraints: （6） In the formula, -Active power loss of the system at time t, in MW; S22. Operating constraints of thermal power units, including: Unit output constraints （7） In the formula, , -Minimum and maximum output of the j-th thermal power unit, in MW; Unit ramp-up constraints, （8） In the formula, - The gradeability of thermal power unit j, in MW / h; Unit start-up and shutdown status constraints. （9） In the formula, , - The start-up and shutdown actions of the jth thermal power unit at time t; S23, Wind power output constraint. （10） S24. Energy storage constraints, including: Energy storage constraints (11) Energy storage charging and discharging constraints, (12) In the formula, S t - Energy storage capacity, MW; θ i - Self-loss rate; φ sc,t - Charging efficiency, % φ sd,t -Discharge efficiency, % u sc ,u sd,t - Charging state, discharging state; S t,min S t,max - Capacity upper limit, capacity lower limit, MW; P sc,max - Maximum charging power, MW; P sd,max - Maximum discharge power, MW.

4. The multi-objective cooperative optimal scheduling method as described in claim 3, wherein, Step S3 includes the following steps: S31. Initialize the population, determining the population size as N and the dimension m of each individual. Each individual is a vector composed of a set of parameters to be optimized, representing a candidate solution to an optimization problem, i.e., a single objective function value; the initial value of each individual in the search space is a vector x. i,j , that is, x i,j =Lb i,j +rand(x i,j (Ub) i,j -Lb i,j ), that is, the decision variables within a single objective function, and the total objective function value matrix X obtained from all candidate solutions: (13) Among them, Ub i,j Lb i,j The upper and lower boundary values ​​of the candidate solution represent the constraints of each decision variable; rand(x) i,j ) is a random uniform function, and rand(x) i,j X ∈ [0,1]; X is the matrix of all individual candidate positions, i.e., the total objective function value; X i Xi,j is the vector of the i-th individual, i.e., the objective function value of a single individual; Xi,j is the decision variable within a single objective function; N is the value of the overall objective function; i∈{1,2,⋯,N}, j is the j-th dimension of the i-th individual, and m is the individual dimension, i.e., the number of single objective functions; Where Xi,j are decision variables within a single objective function, including: T - time of day, T=24h; a j b j c j - Coal consumption coefficient for thermal power units, t / MW; P G,j,t - Output value of thermal power unit j at time t, in MW; S jt - Start-up and shutdown cost of thermal power unit j, in ten thousand yuan; u j,t - The start-up and shutdown status of thermal power unit j at time t; M L - Total carbon emission allowance for the system, t; M P — Total carbon emissions of the system within one cycle, t; ϖ j - Carbon emission allocation coefficient per unit electricity of thermal power unit, t / MW; , - Carbon emission allocation coefficients for photovoltaic and wind turbine units, t / MW; P G,j,t ,P v,t ,P w,t - Output of thermal power units, photovoltaic power units, and wind power units at time t, in MW; λ j - Carbon emission intensity, t / MW; ω - Carbon trading price, yuan / t; d - Length of carbon emission range, t; τ - Increase in carbon trading price for each carbon emission increment, yuan; k w k cw -Wind farm operating cost coefficient, wind curtailment penalty cost coefficient, yuan / MW; k v k cv - Photovoltaic power plant operating cost coefficient and curtailment penalty cost coefficient, in yuan; , -Predicted wind power and photovoltaic power generation at time t, in MW; p price -Grid electricity price, yuan; ν d ν c - The discharge and charging efficiency of the energy storage power station, %; P sd,t ,P sc,t - The discharge and charge power of the electrochemical energy storage power station at time t, in MW; c sc —Charging cost coefficient for energy storage power stations, yuan / MW; c sd - Discharge cost coefficient, RMB / MW; X i Let be the vector of the i-th individual, i.e., the objective function value of a single individual, namely: C1 - operating cost of thermal power unit, yuan; C2 - total carbon trading cost, yuan; C3 - new energy absorption capacity and cost, yuan; C4 - operating revenue and cost of energy storage system, yuan; X is the matrix of all individual candidate positions, which is the total objective function value, namely minC1+minC2+minC3+maxC4. This means that the coal consumption cost and start-up and shutdown cost of thermal power units are minimized, carbon emission cost is minimized, new energy absorption capacity is maximized and cost is minimized, and the operating benefits of energy storage system are maximized and cost is minimized. Achieving these four aspects simultaneously is considered as the realization of multi-objective coordinated low-carbon optimal scheduling in the desert area. S32. Fitness calculation: Each individual is a candidate solution to the problem. Based on its fitness value (in this invention, the smaller value of individual objective function values ​​C1~C3 and the larger value of C4 indicate better fitness), the individual objective function can be evaluated. The fitness value is calculated for each candidate position, yielding: （14） Where F is the fitness vector, and F i It is the fitness vector of the i-th candidate position; Here, we explore and identify the leader individual; the leader individual is the one with the best fitness in the population, with either the minimum or maximum fitness value, to guide other individuals in optimization. We then update the individual positions to generate a new leader individual. （15） （16） （17） （18） Among them, X i It is the new position of the i-th generated individual; It is the j-th dimension of the new position; It is the objective function value at the new position; rand i,j It is a random number in the interval [0,1]; SW j It is the best candidate solution for the strongest individual; , Indicates the safety and danger signals at the new location; T is the total number of iterations, and t is the current iteration number; S33. Different update strategies are used to adjust the position of individuals in the population based on the different value ranges of safety and danger signals. Migration behavior is judged when the risk factor is too high (Dange_signal≧1), the walrus population will migrate to an area more suitable for the population's survival; assuming that each individual migrates from one region of the search space to another random location, the new location is generated by equation (19), that is: （19） （20） According to Equation 19, if the new position changes the value of the objective function, then the new position replaces the previous position. Wherein, It is the position generated by the migration of the i-th individual; It is the objective function value at the new position; X k， k∈{1,2,⋯,N}, and k≠i is the location chosen for the migration of the i-th individual, F k It is the objective function value of the migration location; During migration, the walrus herd expands its search area, engaging in a global search. However, with increasing iterations, the ordinary walrus algorithm tends towards local search development, especially after iteration T / 2, when the herd fully enters the development phase. In the early stages of the algorithm, the herd not only performs global search but also engages in significant development. Initially, the distribution of global and local searches is not well balanced, potentially leading to premature convergence and hindering the herd's exploration of the solution space, thus impacting the final optimization. To enhance the global search capability of the walrus algorithm, a spiral search strategy inspired by natural spiral motion is adopted. This strategy strengthens the algorithm's global optimization ability, ensuring convergence speed and increasing individual diversity. The spiral search formula is as follows: （21） Where X(t+1) represents the position of the i-th walrus in the (k+1)-th iteration; X*(t) represents the position of the i-th walrus in the k-th iteration; X(t) represents the position of a randomly selected walrus individual; D' represents the position difference between the randomly selected walrus individual X(t) and the i-th walrus individual X*(t) in the k-th iteration; here, c is a constant with a value of 1; l is a random number in the range [-1, 1]. In the spiral search formula, e cl A random nonlinear expansion of the search radius was achieved, and cos(2πl) can generate a circular search trajectory at any angle, ensuring the diversity of search directions and effectively covering the corner regions of the solution space. In the early stage, there are significant differences between walrus individuals. By randomly selecting individuals, population difference information is introduced to perturb the search direction and maintain population diversity. In the later stage, the differences between individuals decrease, and walruses begin to explore locally. Through the synergistic effect of dynamic radius, omnidirectional search, and group perturbation, the exploration of the search space by walruses is expanded, achieving a balance between exploration and development. Compared to migration, walrus groups tend to breed in the present when the risk factor is low (Dange_signal<1). Breeding behavior mainly consists of two types of behaviors: resting and foraging. When the safety factor is high (Safety_signal>0.5), they engage in resting behavior, and vice versa. In habitat behavior, walrus population members are divided into three categories: males, females, and juveniles, each defined as X. male X female X Juvenile They update their positions in different ways; X male Position updates are performed using Halton sequences, X female and X Juvenile The update method is shown in the following formula: （22） （23） （24） Where α is the iterative convergence factor, P represents the distress coefficient of juvenile walruses, which is a random number between 0 and 1; O represents the reference safe position, LF is a random number vector based on the Levy distribution, representing Levy movement; and foraging behavior includes two behaviors: escape and gathering. Escape behavior: Walruses, even when foraging underwater, are vulnerable to predators. They will flee their current area based on danger signals from their companions (Dange_signal > 0.5). The update method is shown in the following formula: （25） （26） Where R is as shown in Formula 26, and r1 and r4 are random numbers ranging from 0 to 1; In terms of gregarious behavior, walruses can cooperate in foraging and move based on the location of other walruses in the population, as shown in the following formula: （27） （28） （29） （30） Among them, X1 and X z These are two weighted factors influencing walrus aggregation behavior, X best (t) is the optimal solution, X second (t) is the suboptimal solution, a and b are the aggregation coefficients, r5 is a random number between 0 and 1, and θ is a random number between 0 and π. S34. Recalculate the fitness value and update the walrus's position. The number of iterations in the walrus optimization algorithm is determined based on the actual situation. In each iteration, calculate the fitness of each individual, determine the position of the leader individual and the migrating individuals. When the maximum number of iterations is reached, the algorithm terminates and outputs the final optimal solution. （31） （32） （33） Among them, lls j and uls j These are the upper and lower bound values ​​of the j-th variable in the new generation during iterative updates; and These are the local upper and lower boundary values ​​of the j-th variable in the new generation during iterative updates.

5. A device for multi-objective cooperative optimal scheduling, wherein, The multi-objective collaborative optimal scheduling device is applied to the power system in the Sakhalin region, which includes thermal power units, wind power units, photovoltaic units, electrochemical energy storage systems, and carbon trading market mechanisms, to generate optimal scheduling schemes. The multi-objective cooperative optimal scheduling device includes the following modules: The objective function generation module is used to generate objective functions for power systems, including thermal power units, wind power units, photovoltaic units, electrochemical energy storage systems, and carbon trading market mechanisms. Among them, the new energy base in the Shagohuang area mainly uses long-distance DC transmission for consumption. The constraint module is used to add constraints related to the operation of the power system; The module for generating the optimal scheduling scheme is used to generate the optimal scheduling scheme using the Walrus algorithm.

6. The multi-objective cooperative optimal scheduling method as described in claim 5, wherein, The objective function generation module generates the objective function through the following steps: The following formula is used to minimize the coal consumption cost and start-up / shutdown cost of thermal power units: （1） In the formula, C1 represents the operating cost of the thermal power unit, in yuan; C mh C qt - Coal consumption cost and start-up / shutdown cost of thermal power units, in yuan; T - Time of day, T=24h; a j b j c j - Coal consumption coefficient for thermal power units, t / MW; P G,j,t - Output value of thermal power unit j at time t, in MW; S jt - Start-up and shutdown cost of thermal power unit j, in ten thousand yuan; u j,t - The start-up and shutdown status of thermal power unit j at time t; The cost of calculating carbon emissions is minimized. Both photovoltaic and wind power are clean energy sources that do not produce CO2; therefore, the system's carbon emissions are considered in relation to thermal power units. The carbon allocation factor represents the proportion or weight of the carbon emission allowance allocated to different units, used to determine the allocation of carbon emission responsibilities or the allocation rules in carbon trading. The total carbon emission allowance for the entire system and the total carbon emissions of the system within a cycle are: （2） In the formula, M L - Total carbon emission allowance for the system, t; M P — Total carbon emissions of the system within one cycle, t; ϖ j - Carbon emission allocation coefficient per unit electricity of thermal power unit, t / MW; , - Carbon emission allocation coefficients for photovoltaic and wind turbine units, t / MW; P G,j,t ,P v,t ,P w,t - Output of thermal power units, photovoltaic power units, and wind power units at time t, in MW; λ j - Carbon emission intensity, t / MW; A tiered carbon trading cost model is adopted, dividing the purchase of carbon emission rights into multiple tiers. The more carbon emission rights required to purchase, the higher the price in the corresponding tier. The specific calculation formula of the model is as follows: （3） In the formula, C2 is the total carbon trading cost in yuan; ω is the carbon trading price in yuan / t; d is the length of the carbon emission range in t; and τ is the increase in carbon trading price for each step up in carbon emissions in yuan. With the greatest renewable energy absorption capacity and the lowest cost, The wind and solar power absorption capacity is represented by the amount of wind and solar power curtailment during the dispatch cycle. The higher the amount of curtailment, the weaker the wind and solar power absorption capacity. Since wind power and photovoltaic power generation do not consume fuel, only operation and maintenance costs are considered here. The specific calculation formula is as follows: ， Among them, C w,t C cw,t - Wind farm operation and maintenance costs, wind curtailment penalty costs, in yuan; C v,t C cv,t - Photovoltaic power plant operation and maintenance costs, curtailment penalty costs, in yuan / MW; k w k cw -Wind farm operating cost coefficient, wind curtailment penalty cost coefficient, yuan / MW; k v k cv - Photovoltaic power plant operating cost coefficient and curtailment penalty cost coefficient, in yuan; , -Predicted wind power and photovoltaic power generation at time t, in MW; The system aims to maximize its benefits and minimize its costs. （5） In the formula, p price -Grid electricity price, yuan; C sy C cb - The discharge revenue and charging cost of an energy storage power station, in yuan; ν d ν c - The discharge and charging efficiency of the energy storage power station, %; P sd,t ,P sc,t - The discharge and charge power of the electrochemical energy storage power station at time t, in MW; c sc - Charging cost coefficient for energy storage power stations, in yuan / MW; c sd - Discharge cost coefficient, RMB / MW; The sum of the four objective functions C1 to C4 is taken as the overall objective function to be scheduled.

7. The multi-objective cooperative optimal scheduling method as described in claim 6, wherein, The following constraints were added to the constraint conditions module: Establish system power balance constraints: （6） In the formula, -Active power loss of the system at time t, in MW; Operating constraints for thermal power units include: Unit output constraints （7） In the formula, , -Minimum and maximum output of the j-th thermal power unit, in MW; Unit ramp-up constraints, （8） In the formula, - The gradeability of thermal power unit j, in MW / h; Unit start-up and shutdown status constraints. （9） In the formula, , - The start-up and shutdown actions of the jth thermal power unit at time t; Wind power output constraints （10） Energy storage constraints include: Energy storage constraints (11) Energy storage charging and discharging constraints, (12) In the formula, S t - Energy storage capacity, MW; θ i - Self-loss rate; φ sc,t - Charging efficiency, % φ sd,t -Discharge efficiency, % u sc ,u sd,t - Charging state, discharging state; S t,min S t,max - Capacity upper limit, capacity lower limit, MW; P sc,max - Maximum charging power, MW; P sd,max - Maximum discharge power, MW.

8. The multi-objective cooperative optimal scheduling method as described in claim 7, wherein, The module for generating the optimal scheduling scheme generates the optimal scheduling scheme according to the following steps: The first step is to initialize the population, determining the population size as N and the dimension m of each individual. Each individual is a vector composed of a set of parameters to be optimized, representing a candidate solution to an optimization problem, i.e., a single objective function value; the initial value of each individual in the search space is a vector x. i,j , that is, x i,j =Lb i,j +rand(x i,j (Ub) i,j -Lb i,j ), that is, the decision variables within a single objective function, and the total objective function value matrix X obtained from all candidate solutions: (13) Among them, Ub i,j Lb i,j The upper and lower boundary values ​​of the candidate solution represent the constraints of each decision variable; rand(x) i,j ) is a random uniform function, and rand(x) i,j X ∈ [0,1]; X is the matrix of all individual candidate positions, i.e., the total objective function value; X i Xi,j is the vector of the i-th individual, i.e., the objective function value of a single individual; Xi,j is the decision variable within a single objective function; N is the value of the overall objective function; i∈{1,2,⋯,N}, j is the j-th dimension of the i-th individual, and m is the individual dimension, i.e., the number of single objective functions; Where Xi,j are decision variables within a single objective function, including: T - time of day, T=24h; a j b j c j - Coal consumption coefficient for thermal power units, t / MW; P G,j,t - Output value of thermal power unit j at time t, in MW; S jt - Start-up and shutdown cost of thermal power unit j, in ten thousand yuan; u j,t - The start-up and shutdown status of thermal power unit j at time t; M L - Total carbon emission allowance for the system, t; M P — Total carbon emissions of the system within one cycle, t; ϖ j - Carbon emission allocation coefficient per unit electricity of thermal power unit, t / MW; , - Carbon emission allocation coefficients for photovoltaic and wind turbine units, t / MW; P G,j,t ,P v,t ,P w,t - Output of thermal power units, photovoltaic power units, and wind power units at time t, in MW; λ j - Carbon emission intensity, t / MW; ω - Carbon trading price, yuan / t; d - Length of carbon emission range, t; τ - Increase in carbon trading price for each carbon emission increment, yuan; k w k cw -Wind farm operating cost coefficient, wind curtailment penalty cost coefficient, yuan / MW; k v k cv - Photovoltaic power plant operating cost coefficient and curtailment penalty cost coefficient, in yuan; , -Predicted wind power and photovoltaic power generation at time t, in MW; p price -Grid electricity price, yuan; ν d ν c - The discharge and charging efficiency of the energy storage power station, %; P sd,t ,P sc,t - The discharge and charge power of the electrochemical energy storage power station at time t, in MW; c sc —Charging cost coefficient for energy storage power stations, yuan / MW; c sd - Discharge cost coefficient, RMB / MW; X i Let be the vector of the i-th individual, i.e., the objective function value of a single individual, namely: C1 - operating cost of thermal power unit, yuan; C2 - total carbon trading cost, yuan; C3 - new energy absorption capacity and cost, yuan; C4 - operating revenue and cost of energy storage system, yuan; X is the matrix of all individual candidate positions, which is the total objective function value, namely minC1+minC2+minC3+maxC4. This means that the coal consumption cost and start-up and shutdown cost of thermal power units are minimized, carbon emission cost is minimized, new energy absorption capacity is maximized and cost is minimized, and the operating benefits of energy storage system are maximized and cost is minimized. Achieving these four aspects simultaneously is considered as the realization of multi-objective coordinated low-carbon optimal scheduling in the desert area. The second step is fitness calculation. Each individual is a candidate solution to the problem. Based on its fitness value (in this invention, the smaller value of individual objective function values ​​C1~C3 and the larger value of C4 indicate better fitness), the individual objective function can be evaluated. The fitness value is calculated for each candidate position, resulting in: （14） Where F is the fitness vector, and F i It is the fitness vector of the i-th candidate position; Here, we explore and identify the leader individual; the leader individual is the one with the best fitness in the population, with either the minimum or maximum fitness value, to guide other individuals in optimization. We then update the individual positions to generate a new leader individual. （15） （16） （17） （18） Among them, X i It is the new position of the i-th generated individual; It is the j-th dimension of the new position; It is the objective function value at the new position; rand i,j It is a random number in the interval [0,1]; SW j It is the best candidate solution for the strongest individual; , Indicates the safety and danger signals at the new location; T is the total number of iterations, and t is the current iteration number; The third step involves using different update strategies to adjust the positions of individuals in the population based on the different value ranges of the safety and danger signals. Migration behavior is judged when the risk factor is too high (Dange_signal≧1), the walrus population will migrate to an area more suitable for the population's survival; assuming that each individual migrates from one region of the search space to another random location, the new location is generated by equation (19), that is: （19） （20） According to Equation 19, if the new position changes the value of the objective function, then the new position replaces the previous position. Wherein, It is the position generated by the migration of the i-th individual; It is the objective function value at the new position; X k， k∈{1,2,⋯,N}, and k≠i is the location chosen for the migration of the i-th individual, F k It is the objective function value of the migration location; During migration, the walrus herd expands its search area, engaging in a global search. However, with increasing iterations, the ordinary walrus algorithm tends towards local search development, especially after iteration T / 2, when the herd fully enters the development phase. In the early stages of the algorithm, the herd not only performs global search but also engages in significant development. Initially, the distribution of global and local searches is not well balanced, potentially leading to premature convergence and hindering the herd's exploration of the solution space, thus impacting the final optimization. To enhance the global search capability of the walrus algorithm, a spiral search strategy inspired by natural spiral motion is adopted. This strategy strengthens the algorithm's global optimization ability, ensuring convergence speed and increasing individual diversity. The spiral search formula is as follows: （21） Where X(t+1) represents the position of the i-th walrus in the (k+1)-th iteration; X*(t) represents the position of the i-th walrus in the k-th iteration; X(t) represents the position of a randomly selected walrus individual; D' represents the position difference between the randomly selected walrus individual X(t) and the i-th walrus individual X*(t) in the k-th iteration; here, c is a constant with a value of 1; l is a random number in the range [-1, 1]. In the spiral search formula, e cl A random nonlinear expansion of the search radius was achieved, and cos(2πl) can generate a circular search trajectory at any angle, ensuring the diversity of search directions and effectively covering the corner regions of the solution space. In the early stage, there are significant differences between walrus individuals. By randomly selecting individuals, population difference information is introduced to perturb the search direction and maintain population diversity. In the later stage, the differences between individuals decrease, and walruses begin to explore locally. Through the synergistic effect of dynamic radius, omnidirectional search, and group perturbation, the exploration of the search space by walruses is expanded, achieving a balance between exploration and development. Compared to migration, walrus groups tend to breed in the present when the risk factor is low (Dange_signal<1). Breeding behavior mainly consists of two types of behaviors: resting and foraging. When the safety factor is high (Safety_signal>0.5), they engage in resting behavior, and vice versa. In habitat behavior, walrus population members are divided into three categories: males, females, and juveniles, each defined as X. male X female X Juvenile They update their positions in different ways; X male Position updates are performed using Halton sequences, X female and X Juvenile The update method is shown in the following formula: （22） （23） （24） Where α is the iterative convergence factor, P represents the distress coefficient of juvenile walruses, which is a random number between 0 and 1; O represents the reference safe position, LF is a random number vector based on the Levy distribution, representing Levy movement; and foraging behavior includes two behaviors: escape and gathering. Escape behavior: Walruses, even when foraging underwater, are vulnerable to predators. They will flee their current area based on danger signals from their companions (Dange_signal > 0.5). The update method is shown in the following formula: （25） （26） Where R is as shown in Formula 26, and r1 and r4 are random numbers ranging from 0 to 1; In terms of gregarious behavior, walruses can cooperate in foraging and move based on the location of other walruses in the population, as shown in the following formula: （27） （28） （29） （30） Among them, X1 and X z These are two weighted factors influencing walrus aggregation behavior, X best (t) is the optimal solution, X second (t) is the suboptimal solution, a and b are the aggregation coefficients, r5 is a random number between 0 and 1, and θ is a random number between 0 and π. The fourth step is to recalculate the fitness values ​​and update the walrus positions. The number of iterations in the walrus optimization algorithm is determined based on the actual situation. In each iteration, the fitness of each individual is calculated, and the positions of the leader and migrating individuals are determined. When the maximum number of iterations is reached, the algorithm terminates and outputs the final optimal solution. （31） （32） （33） Among them, lls j and uls j These are the upper and lower bound values ​​of the j-th variable in the new generation during iterative updates; and These are the local upper and lower boundary values ​​of the j-th variable in the new generation during iterative updates.

9. An electronic device, comprising: The system includes a memory, a processor, and a multi-objective cooperative optimal scheduler stored in the memory and executable on the processor, wherein the multi-objective cooperative optimal scheduler, when executed by the processor, implements the steps of the multi-objective cooperative optimal scheduling method as described in any one of claims 1 to 4.

10. A computer storage medium, wherein, The computer storage medium stores a multi-objective cooperative optimal scheduler, which, when executed by the processor, implements the steps of the multi-objective cooperative optimal scheduling method as described in any one of claims 1 to 4.

Images