Double-layer optimization scheduling method considering electricity-storage-carbon cooperation and user requirements

Abstract

The invention discloses a double-layer optimization scheduling method considering electricity-storage-carbon cooperation and user demands, and the method comprises the steps: building a constraint model for the dynamic segmented adjustment of a carbon transaction price along with the carbon emission according to the carbon emission quotas of thermal power, wind power and photovoltaic power and based on a reference carbon transaction cost model; taking the lowest comprehensive power generation cost as a target, jointly optimizing output plans of thermal power, wind power, photovoltaic power and energy storage, coupling the constraint model to limit carbon emission of the thermal power generating unit, and constructing an electricity-storage-carbon collaborative optimization model; a demand response mechanism is constructed based on the user energy consumption mode satisfaction degree and the energy consumption cost satisfaction degree, a user load curve is adjusted and optimized through the electricity price, and a user demand optimization model is established; solving the double-layer optimization model hierarchically by adopting an improved algorithm; and compared with a reference carbon transaction model and a stepped carbon transaction model, the constraint force is stronger, and collaborative optimization of power generation economy, low-carbon property and user satisfaction is realized under the condition of meeting system power balance and power grid security constraint conditions.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and more specifically, to a two-layer optimal scheduling method considering the coordination of electricity, energy storage, and carbon and user needs. Background Art

[0002] With the vigorous development of new energy and the improvement of users' living standards, the randomness and volatility of new energy have affected the consumption of new energy and restricted its development. At the same time, the user-side problems in the transformation of the power grid also have an impact. Demand response is the behavior of users to actively change their consumption patterns according to electricity prices. Under the condition of meeting the low-carbon coordination of thermal power, wind power, photovoltaic power, and energy storage, improving the satisfaction of users' electricity demand is of great significance for the stability of the power system. Currently, in the research on the coordinated optimization of thermal power, wind power, and photovoltaic power, considering the carbon trading cost, an optimal scheduling model for the uncertainty of wind power and photovoltaic power is established. For the power generation side and the load side, an optimal strategy for the demand response of adjustable load resources considering user satisfaction is proposed.

[0003] Currently, for the coordinated optimization of the power supply side and the load side, the optimization method focuses on the load type and does not consider user satisfaction enough. For the power generation side of thermal power, wind power, and photovoltaic power, although the carbon trading cost model can reduce carbon emissions to a certain extent, the benchmark carbon trading cost model or the ladder carbon trading cost model still has insufficient binding force on high-carbon-emission thermal power units, and a dynamic constraint carbon trading model is rarely used to restrict the carbon emissions of thermal power units. Summary of the Invention

[0004] Aiming at the fact that the prior art does not jointly consider the coordinated operation of electricity, energy storage, and carbon and the maximization of user demand satisfaction, the present invention provides a two-layer optimal scheduling method considering the coordination of electricity, energy storage, and carbon and user needs.

[0005] The following are the technical solutions of the present invention:

[0006] A two-layer optimal scheduling method considering the coordination of electricity, energy storage, and carbon and user needs, comprising the following steps:

[0007] S1. According to the carbon emission quotas of thermal power, wind power, and photovoltaic power, based on the benchmark carbon trading cost model, construct a constraint model in which the carbon trading price is dynamically adjusted in segments according to the carbon emissions.

[0008] S2. With the goal of minimizing the comprehensive power generation cost, jointly optimize the output plans of thermal power, wind power, photovoltaic power, and energy storage, and couple the constraint model to limit the carbon emissions of thermal power units to construct an electricity-energy storage-carbon coordinated optimization model.

[0009] S3. Based on the satisfaction of users' energy usage patterns and the satisfaction of energy usage costs, construct a demand response mechanism, optimize the user load curve through electricity price adjustment, and establish a user demand optimization model.

[0010] S4. Use the improved algorithm to solve the bi-level optimization model layer by layer. Solve the electricity-storage-carbon collaborative optimization model through the improved genetic algorithm in the upper layer, and solve the user demand optimization model through the improved particle swarm algorithm in the lower layer. Iteratively update until the comprehensive optimal conditions of the system are met.

[0011] Preferably, in step S1, according to the carbon emission quotas of thermal power, wind power, and photovoltaic power, it is determined by the following formula:

[0012] The carbon emission quota EA1 of thermal power determined by the baseline method is:

[0013]

[0014] In the formula: δ EF is the baseline carbon emission factor, T is the scheduling period, N is the number of thermal power units, P G,i,t is the power of thermal power unit i at time t, ξ is the carbon quota correction coefficient of wind power and photovoltaic power, α is the reserve coefficient of wind power output, P w,t is the output of the wind turbine at time t, β is the reserve coefficient of photovoltaic power output, P p,t is the output of photovoltaic power at time t;

[0015] The carbon emission quotas EA2 and EA3 of wind power and photovoltaic power are:

[0016]

[0017]

[0018] Based on the baseline carbon trading cost model, construct a constraint model for the dynamic piecewise adjustment of carbon trading price with carbon emissions. The specific steps are as follows: The initial carbon emission quota adopts the form of mainly free allocation and supplemented by auction allocation. The proportion of auction quotas is ε, and the auction quota quantity is:

[0019] EA a = ε(EA1 + EA2 + EA3) (4)

[0020] The free quota quantity is:

[0021] EA f = (1 - ε)(EA1 + EA2 + EA3) (5)

[0022] The total carbon emissions of thermal power units within a certain period range are:

[0023]

[0024] T is the scheduling period, N is the number of thermal power units e i is the carbon emission coefficient of thermal power unit i, P G,i,t is the power of thermal power unit i at time t;

[0025] The carbon trading price is c1, the quota auction price is c2, and the carbon trading cost is F c1 is:

[0026] F c1 = c1(E p - EA f - EA a_ ) + c2EA a (7)

[0028] The carbon trading price in the constraint model is not fixed and increases with the increase of carbon emissions. When the carbon emissions are lower than the preset threshold, the carbon trading cost increases according to a linear function; when the carbon emissions exceed the threshold, the carbon trading cost increases according to a non-linear function; the carbon trading cost F under dynamic constraints c is:

[0029]

[0030] l is the interval length of the carbon trading cost under dynamic constraints, λ is the dynamic constraint price growth coefficient, E t is the carbon trading volume, which is the difference between the carbon emissions of the thermal power unit and the quota, E t = E p - EA f - EA a , when 5l < E t , a quadratic function is adopted to constrain the carbon emissions of the unit's power generation through high trading costs.

[0031] Preferably, the objective function of the power-storage-carbon collaborative optimization model includes: the lowest comprehensive power generation cost, the power generation cost of the thermal power unit, the penalty cost for wind and light abandonment, and the operation cost of the energy storage device;

[0032] The objective function of the lowest comprehensive power generation cost is:

[0033] minF t = F f + F w + F p + F b + F loss + F c (9)

[0034] F t is the total power generation cost under power-storage-carbon collaboration, F f is the power generation cost of the thermal power unit, F w is the penalty cost for wind abandonment, F p is the penalty cost for light abandonment, F b is the operation cost of the energy storage device, F loss is the network loss cost, Fc is the carbon trading cost under dynamic constraints;

[0035] The power generation cost of a thermal power unit includes the coal consumption cost F of power generation f1 , and the start-stop cost F of the unit f2 , and the specific formula is:

[0036]

[0037]

[0038] P g,i,t is the output power of thermal power unit i at time t, a i , b i and c i are the cost coefficients of the quadratic function of the coal consumption cost with respect to the output power, S on,i,t , S off,i,t is the action power of unit i to start and stop at time t, η on,i , η off,i are the cost coefficients of unit i to start and stop;

[0039] The penalty cost for curtailment of wind and solar power is:

[0040]

[0041]

[0042] P wq,t , P pq,t are the curtailment of wind power and curtailment of solar power at time t, μ w , μ p are the penalty cost coefficients for curtailment of wind and solar power.

[0043] The operating cost of the energy storage device is:

[0044] F b = ν d P d + ν s P s (14)

[0045] ν d , ν s are the cost coefficients of the energy storage device for discharging and charging, P d and P s are the discharging and charging powers of the energy storage device.

[0046] Preferably, the constraint conditions of the objective function include: the safety stability and reactive power constraint of the output adjustment of the thermal power unit, the start and stop constraint of the thermal power unit, the output constraint of wind power and photovoltaic power, the operation constraint of the energy storage device, the grid operation safety constraint, and the system power balance constraint;

[0047] The safety stability of the output adjustment and the reactive power constraint of the thermal power unit are as follows:

[0048]

[0049] P g,i,min and P g,i,max are the minimum and maximum output powers of the thermal power unit respectively, and R i,min and R i,max are the minimum and maximum values of the ramp of the thermal power unit respectively. Q i,t is the reactive power generated by the thermal power unit i at time t, and Q i,min and Q i,max are the minimum and maximum values of the reactive power of the unit i;

[0050] The start-up and shutdown constraints of the thermal power unit are as follows:

[0051]

[0052] τ i,t is the variable of the operation state of the thermal power unit from shutdown to startup in the t time period, and T on,i,t-1 is the continuous operation time of the thermal power unit i in the t-1 time period, and T off,i,t-1 is the shutdown time of the thermal power unit i in the t-1 time period;

[0053] The output constraints of wind power and photovoltaic are as follows:

[0054] 0 ≤ P w,t ≤ P w，max (17)

[0055] 0 ≤ P p,t ≤ P p,max (18)

[0056] P w，max and P p,max are the upper limits of the output of wind power and photovoltaic respectively;

[0057] The operation constraints of the energy storage device are as follows:

[0058]

[0059] P s,t is the charging power of the energy storage device at time t, and P s,min and P s,max are the minimum and maximum power values of the charging of the energy storage device respectively. Y s,t and Y d,t are the charging and discharging states of the energy storage device in the t time period respectively, between 0 and 1. S soc,t is the state of charge of the energy storage device at time t, and S soc,min and S soc,maxare the minimum and maximum values of the state of charge of the energy storage device, respectively, S soc,s,t , S soc,e,t are the initial and final states of charge of the energy storage device within a time period t, respectively;

[0060] The power grid operation safety constraint is:

[0061] (P G,zl + P w,zl + P p,zl + P s,zl ) 2 + Q G,zl ≤ (S l,max ) 2 (20)

[0062] P G,zl , P w,zl , P p,zl , P s,zl are the active power transmitted by thermal power, wind power, photovoltaic power, and energy storage on the branch, respectively, Q G,zl is the reactive power transmitted by thermal power on the branch, S l,max is the maximum transmission capacity on the branch;

[0063] The system power balance constraint is:

[0064]

[0065]

[0066] α d is the discharge efficiency of the energy storage device, J is the total number of loads, P load,j,t is the power value of load j in time period t, P loss,t is the active power loss of the branch in time period t, Q i,t is the reactive power output of unit i in time period t, Q load,j,t is the reactive power of load j in time period t, Q loss,t is the reactive power loss of the branch in time period t.

[0067] Preferably, in step S3, the user demand optimization model is constructed by the following indicators:

[0068] The satisfaction degree of the user's energy consumption mode is defined as the ratio of the actual power consumption after the user participates in demand response to the original load; the satisfaction degree of the energy consumption cost is defined as the ratio of the real-time electricity price to the initial electricity price; the user demand optimization model considers the highest level of user satisfaction, and the user satisfaction objective function is the highest satisfaction level value of the energy consumption mode and the energy consumption cost. The user satisfaction objective function is:

[0069] maxF s = R 1,c+R 2,c (23)

[0070] R 1,c is the satisfaction degree of the user's energy consumption mode, R 2,c is the satisfaction degree of the user's energy consumption cost.

[0071] During the actual electricity consumption process, the user participates in the load demand response of the price regulation mechanism, optimizes the user's energy consumption mode. The load participating in the demand response includes the transferred load and the curtailed load. The satisfaction degree of the user without time-of-use pricing during free electricity consumption is 1. The satisfaction degree of the user with time-of-use pricing will decrease with the adjustment of electricity consumption behavior after participating in the demand response. The satisfaction degree of the user's energy consumption mode is:

[0072]

[0073] P c,t is the original load of the user at time t, P′ c,t is the electricity consumption of the user after participating in the demand response,

[0074] Although the user's energy consumption cost changes after participating in the demand response, the real-time electricity price is not consistent with the initial electricity price, which affects the energy consumption cost. The initial energy consumption cost satisfaction degree of the user during free electricity consumption is 1. The value of the energy consumption cost satisfaction degree after participating in the demand response is:

[0075]

[0076] ρ e is the initial value of the electricity price, ρ e ' is the real-time electricity price after participating in the demand response;

[0077] The constraint of the demand response, i.e., the electricity price response, is:

[0078]

[0079] D j,t is the demand response volume of load j at time t, Δρ e,max is the maximum value of the electricity price adjustment.

[0080] Preferably, in step S4: The improved genetic algorithm adopts a decision space dimension reduction strategy, compresses the high-dimensional scenario into a list of candidate optimizations, and generates the initial population through multi-scenario collaborative coding; during the decision space dimension reduction process, each electricity-storage-carbon collaborative scenario S is calculated as 4 candidate electricity-storage-carbon collaborative optimization lists with a length of L e , which are X i1 , X i2 , X i3 and X i4 , and the candidate electricity-storage-carbon collaborative optimization lists X i1 , X i2 , Xi3 and X i4 is the current optimal solution. Combine the candidate lists to obtain the final decision space O i The formula for

[0081] O i = X i1 ∪ X i2 ∪ X i3 ∪ X i4 (27)

[0082] Repeat the above process. The total decision space O after dimensionality reduction is described as:

[0083] O = {O1, O2, …, O n} (28)

[0084] n is the number of electricity-storage-carbon collaborative scenarios. The length of the finally recommended electricity-storage-carbon collaborative optimization combination is L. To maintain a similar selection ratio in the search space before and after dimensionality reduction, the formula is:

[0085] S / L e = L e / L (29)

[0086]

[0087] During the multi-scenario collaborative optimization scheduling process, the genetic algorithm encodes the initial values to form an initial population. The initial population consists of a group of chromosomes, and each chromosome represents a solution to the problem, which is represented by a k-source vector. During the population generation process, duplicate genes are avoided in the same chromosome. After the population is formed, it is converted from the objective function to the fitness function. The formula for the fitness function is:

[0088]

[0089] The gene operation process includes selection, crossover, and mutation operations. In the selection operation, the selection of an individual is associated with the fitness. The higher the fitness, the higher the selection probability. The probability of an individual being selected is:

[0090]

[0091] where v i is the fitness value of the population individual, and n is the number of population individuals;

[0092] During the crossover operation, the k-th chromosome r k and the l-th chromosome r l simultaneously exchange at the j position. The formula is: r

[0093] The formula is: r k,j = r k,j (1 - b)+ r l,jb (33)

[0094] r l,j = r l,j (1 - b)+r k,j b (34)

[0095] where b is a random number in [0, 1];

[0096] During the mutation operation, the formula for the mutation of the j-th gene of the i-th individual is:

[0097]

[0098] where r min and r max are the upper and lower limits of gene r i,j , ξ is a random number between 0 and 1, g is the number of iterations, and G max is the maximum number of evolutions.

[0099] Preferably, in step S4, the improved particle swarm algorithm dynamically adjusts the inertia weight using the beta distribution, the learning factor decreases linearly with the number of iterations, and a weighted combination of the historical best position and the global best position is introduced into the particle position update formula; when using the improved particle swarm algorithm to solve the user demand optimization model, the update formulas for speed and position are:

[0100] v j (d + 1)= ωv j (d)+c1r1(p j (d)-x j (d))+c2r2(g j (d)-x j (d)) (36)

[0101] x j (d + 1)= x j (d)+v j (d + 1) (37)

[0102] where: w is the dynamically adjusted inertia weight; d represents the number of iterations; p j (d) is the best position found so far by the j-th particle, g j (d) is the best position found by all particles in the entire population so far, j represents the corresponding particle number; x j (d) and v j (d) are the position and velocity of the particle respectively; c1 and c2 are linearly decreasing learning factors; r1 and r2 are random numbers between (0, 1);

[0103] The dynamic adjustment formula for the inertia weight is:

[0104]

[0105] Among them, ω max is the maximum value of the inertia weight, ω min is the minimum value of the inertia weight, d max is the maximum number of iterations, d is the current number of iterations, and σ is the inertia adjustment factor; betarnd(p, q) is a function for generating random numbers that follow a beta distribution, and p and q are the shape parameters of the beta distribution;

[0106] The linearly decreasing learning factor is:

[0107]

[0108] Among them, c max is the maximum value of the learning factor, c min is the minimum value of the learning factor, d max is the maximum number of iterations, and d is the current number of iterations;

[0109] Combined with the coordinated optimization model of thermal power, wind power, photovoltaic and energy storage and the user demand optimization model for hierarchical solution, in the double-layer optimal scheduling process, first, the upper-layer model is solved to obtain the minimum operating cost of the power generation side and the unit output plan, including the active power, reactive power and reactive power loss parameters of thermal power, wind power, photovoltaic and energy storage; secondly, the lower-layer model adjusts the load demand based on the feedback of the user satisfaction level, generates real-time electricity price signals, and outputs the load after demand response; finally, the upper-layer model is iteratively solved again according to the real-time electricity price signals and the adjusted load to calculate the minimum operating cost of the system, the optimal power flow distribution and the maximum user satisfaction until the system comprehensive cost and user satisfaction reach equilibrium.

[0110] An electronic device includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the steps of the above-mentioned double-layer optimal scheduling method considering the coordination of electricity, energy storage and carbon and user demands are implemented.

[0111] A computer-readable storage medium stores computer-executable instructions, and when the instructions are executed by a processor, the steps of the above-mentioned double-layer optimal scheduling method considering the coordination of electricity, energy storage and carbon and user demands are implemented.

[0112] Compared with the prior art, the beneficial effects of the technical solution of the present invention are:

[0113] A two - layer optimal scheduling method considering the coordination of electricity - storage - carbon and user needs has clear logic. By using a constraint model in which the carbon trading price is dynamically adjusted in segments according to carbon emissions, it restricts the high carbon emissions of thermal power units. Compared with the benchmark carbon trading model and the stepped carbon trading model, it has stronger binding force. It conducts two - layer collaborative optimization on the electricity - storage - carbon collaborative optimization model and the user - demand satisfaction model. On the basis of the coordinated development of wind - solar - storage - carbon, it greatly improves the user - demand satisfaction. The electricity - storage - carbon collaborative optimization model is solved by an improved genetic algorithm, and the user - demand satisfaction model is solved by an improved particle swarm algorithm, which improves the speed and efficiency of iterative operations.

[0114] The present invention uses a dynamic carbon trading mechanism to restrict the high carbon emissions of thermal power through quadratic penalty, combines the new - energy carbon quota correction coefficient and the reserve coefficient, effectively reduces carbon emissions and improves the new - energy consumption rate; the double - satisfaction index drives the electricity - price response mechanism, smooths the load curve and reduces the peak - valley difference, realizing the dual optimization of the user's electricity - using experience and the economic operation of the system. The decision - space dimension - reduction strategy of the improved genetic algorithm shortens the calculation time, and the beta - distribution particle swarm algorithm enhances the global optimization ability, significantly improving the solution efficiency of large - scale complex scenarios. Through the coupled optimization of multiple elements such as power - balance constraints, energy - storage charge - discharge limits, and power - grid transmission - capacity constraints, it collaboratively reduces network losses, extends the energy - storage life cycle, and ensures the safe and stable operation of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0115] Figure 1 is the flowchart of the method provided by the present invention;

[0116] Figure 2 is the flowchart of the hierarchical solution of the improved algorithm for solving the two - layer optimization model of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0117] The present invention will be further described below in conjunction with specific embodiments. The following embodiments are only used to more clearly illustrate the technical solutions of the present application and should not be used to limit the protection scope of the present application.

[0118] It should be understood that in various embodiments of the present invention, the magnitude of the sequence numbers in each process does not mean the order of execution. The execution order of each process should be determined by its function and internal logic and should not constitute any limitation to the implementation process of the embodiments of the present invention.

[0119] Embodiment: A two - layer optimal scheduling method considering the coordination of electricity - storage - carbon and user needs, as Figure 1 shown, includes the following steps:

[0120] Step 1: Considering the carbon - emission quotas of power generation such as thermal power, wind power, and photovoltaic power, based on the benchmark carbon - trading cost model, construct a constraint model in which the carbon - trading price is dynamically adjusted in segments according to carbon emissions;

[0121] In this embodiment, taking the improved IEEE 33-node system as an example, there is 1 thermal power unit, 1 wind power unit, 1 photovoltaic power station, 1 energy storage power station, 5 transmission lines and 30 load nodes in the system.

[0122] The carbon emission quota EA1 of the thermal power unit determined by the baseline method is:

[0123]

[0124] In the formula: δ EF is the baseline carbon emission factor, T is the scheduling period, N is the number of thermal power units, P G,i,t is the power of thermal power unit i at time t, ξ is the carbon quota correction coefficient of wind power and photovoltaic power, α is the reserve coefficient of wind power output, P w,t is the output of the wind power unit at time t, β is the reserve coefficient of photovoltaic power output, P p,t is the output of the photovoltaic power at time t.

[0125] The free carbon emission quotas EA2 and EA3 of wind power and photovoltaic power are

[0126]

[0127]

[0128] Step 2: Constrain the carbon emissions of thermal power units with a constraint model, and establish a collaborative optimization model for thermal power, wind power, photovoltaic power and energy storage;

[0129] The initial carbon emission quota adopts the form of mainly free allocation and supplemented by auction allocation. The proportion of the auction quota is ε, and the auction quota quantity is:

[0130] EA a = ε(EA1 + EA2 + EA3) (4)

[0131] The free quota quantity is:

[0132] EA f = (1 - ε)(EA1 + EA2 + EA3) (5)

[0133] The total carbon emissions of thermal power units within a certain cycle range are:

[0134]

[0135] T is the scheduling period, N is the number of thermal power units e i is the carbon emission coefficient of thermal power unit i, P G,i,t is the power of thermal power unit i at time t

[0136] The carbon trading price is c1, the quota auction price is c2, and the carbon trading cost is F c1 is:

[0137] F c1 = c1(E p - EA f - EA a_ ) + c2EA a (7)

[0138] Preferably, the establishment of the constraint model is specifically as follows:

[0139] The carbon trading price in the constraint model is not fixed and increases with the increase in carbon emissions. It is a linear growth model when the increase in carbon emissions is low, and a quadratic function is used for constraint after reaching a certain value. The carbon trading cost under dynamic constraint is:

[0140]

[0141] l is the interval length of the carbon trading cost under dynamic constraint, λ is the dynamic constraint price growth coefficient, E t is the carbon trading volume, which is the difference between the carbon emissions of the thermal power unit and the quota, E t = E p - EA f - EA a , when 5l < E t , a quadratic function is adopted to constrain the carbon emissions of the unit by high trading costs.

[0142] In this embodiment, the carbon trading benchmark price c1 is 100 yuan / t, the auction price c2 is 120 yuan / t, the auction ratio is 5%, the carbon trading interval l is 0.25t, the dynamic price constraint price growth coefficient λ is 0.15, the benchmark carbon emission factor δ EF is 0.85, the carbon quota correction coefficient ξ for wind power and photovoltaic is 0.2, and the carbon emission coefficient of the thermal power unit is 0.581t / MW·h.

[0143] Taking the lowest comprehensive power generation cost as the objective function, the objective function is:

[0144] minF t = F f + F w + F p + F b + F loss + F c (9)

[0145] F t is the total power generation cost under the electricity - storage - carbon coordination, F f is the power generation cost of the thermal power unit, F w is the penalty cost for wind curtailment, Fp is the curtailment penalty cost, F b is the operation cost of the energy storage device, F loss is the network loss cost, F c is the carbon trading cost under dynamic constraints.

[0146] The power generation cost of thermal power units includes the coal consumption cost F f1 , the start-up and shut-down cost F f2 , and the specific formula is:

[0147]

[0148]

[0149] P g,i,t is the output power of thermal power unit i at time t, a i , b i and c i are the cost coefficients of the quadratic function of coal consumption cost with respect to output power, S on,i,t , S off,i,t is the action power of unit i for start-up and shut-down at time t, η on,i , η off,i are the cost coefficients of unit i for start-up and shut-down.

[0150] The curtailment penalty cost for wind and light curtailment is:

[0151]

[0152]

[0153] P wq,t , P pq,t are the wind curtailment power and light curtailment power at time t, μ w , μ p are the curtailment penalty cost coefficients for wind and light curtailment.

[0154] The operation cost of the energy storage device is:

[0155] F b =ν d P d +ν s P s (14)

[0156] ν d , ν s are the cost coefficients for discharging and charging of the energy storage device, P d and P s are the discharging and charging powers of the energy storage device.

[0157] The constraint conditions of the model objective function include the operation, start-up and shutdown constraints of thermal power units, the output constraints of wind power and photovoltaic power, the operation constraints of energy storage devices, the operation safety constraints of the power grid, and the system power balance constraints.

[0158] The safety stability and reactive power constraints for the output adjustment of thermal power units are as follows:

[0159]

[0160] P g,i,min 、P g,i,max are the minimum and maximum output powers of the thermal power unit respectively, and R i,min 、R i,max are the minimum and maximum values of the ramp rate of the thermal power unit respectively. Q i,t is the reactive power generated by thermal power unit i at time t, and Q i,min 、Q i,max are the minimum and maximum values of the reactive power of unit i.

[0161] The constraint conditions for the start-up and shutdown of thermal power units are as follows:

[0162]

[0163] τ i,t is the variable of the operation state of the thermal power unit from shutdown to startup in the t time period, T on,i,t-1 is the continuous operation time of thermal power unit i in the t-1 time period, and T off,i,t-1 is the shutdown time of thermal power unit i in the t-1 time period.

[0164] The output constraints of wind power and photovoltaic power are as follows:

[0165] 0 ≤ P w,t ≤ P w，max (17)

[0166] 0 ≤ P p,t ≤ P p,max (18)

[0167] P w，max 、P p,max are the upper limits of the output of wind power and photovoltaic power respectively;

[0168] The operation constraints of the energy storage device are as follows:

[0169]

[0170] P s,t is the charging power of the energy storage device at time t, and P s,min 、P s,max are the minimum and maximum power values of the energy storage device charging respectively, and Y s,t 、Y d,tThey are the charging and discharging states of the energy storage device in period t, ranging from 0 to 1, S soc,t is the state of charge of the energy storage device at time t, S soc,min 、S soc,max are the minimum and maximum values of the state of charge of the energy storage device, S soc,s,t 、S soc,e,t are the initial and final states of charge of the energy storage device within a period t respectively.

[0171] The power grid operation safety constraint is:

[0172] (P G,zl +P w,zl +P p,zl +P s,zl ) 2 +Q G,zl ≤(S l,max ) 2 (20)

[0173] P G,zl 、P w,zl 、P p,zl 、P s,zl are the active power transmitted by thermal power, wind power, photovoltaic power and energy storage on the branch respectively, Q G,zl is the reactive power transmitted by thermal power on the branch, S l,max is the maximum transmission capacity on the branch

[0174] The system power balance constraint is:

[0175]

[0176]

[0177] α d is the discharge efficiency of the energy storage device, J is the total number of loads, P load,j,t is the power value of load j in period t, P loss,t is the active power loss of the branch in period t, Q i,t is the reactive power output of unit i in period t, Q load,j,t is the reactive power of load j in period t, Q loss,t is the reactive power loss of the branch in period t.

[0178] Step 3: Considering the user's energy consumption pattern and energy consumption cost requirements, establish a user demand optimization model;

[0179] The user demand optimization model considers the highest user satisfaction level, and its objective function is the highest satisfaction level value of the energy consumption pattern and energy consumption cost.

[0180] maxF s =R1,c +R 2,c (23)

[0181] R 1,c is the satisfaction degree of the user's energy consumption pattern, R 2,c is the satisfaction degree of the user's energy consumption cost.

[0182] Considering that in the actual electricity consumption process of users, the load demand response participating in the price adjustment mechanism is considered to optimize the user's energy consumption pattern. The load participating in the demand response includes transferred load and curtailed load. The satisfaction degree of users without time-of-use pricing during free electricity consumption is 1, and the satisfaction degree of time-of-use pricing users after participating in the demand response will decrease with the adjustment of electricity consumption behavior. The satisfaction degree of the user's energy consumption pattern is:

[0183]

[0184] P c,t is the original load of the user at time t, P′ c,t is the electricity consumption of the user after participating in the demand response;

[0185] Although the user's energy consumption cost changes after participating in the demand response, the real-time electricity price is not consistent with the initial electricity price, which will affect the energy consumption cost. The initial energy consumption cost satisfaction degree of users with free electricity consumption is 1, and the energy consumption cost satisfaction degree value after participating in the demand response is:

[0186]

[0187] ρ e is the initial value of the electricity price, ρ′ e is the real-time electricity price after participating in the demand response.

[0188] The constraint of demand response, that is, electricity price response, is:

[0189]

[0190] D j,t is the demand response volume of load j at time t, Δρ e,max is the maximum value of electricity price adjustment.

[0191] Step 4: Use the improved algorithm to hierarchically solve the bi-level optimization model. Solve the electricity-storage-carbon collaborative optimization model through the improved genetic algorithm in the upper layer, and solve the user demand optimization model through the improved particle swarm algorithm in the lower layer. Iteratively update until the system comprehensive optimal condition is met.

[0192] The improved genetic algorithm adopts the decision space dimension reduction strategy, compresses the high-dimensional scenario into a candidate optimization list, and generates the initial population through multi-scenario collaborative coding; during the decision space dimension reduction process, each electricity-storage-carbon collaborative scenario S is calculated as 4 with a length of L eThe list of candidate electricity-storage-carbon collaborative optimizations is X i1 , X i2 , X i3 and X i4 . The candidate electricity-storage-carbon collaborative optimization list X i1 , X i2 , X i3 and X i4 are the current optimal solutions. Combining the candidate lists gives the final decision space O i . The formula for

[0193] O i is: i1 O i2 = X i3 ∪ X i4 (27)

[0194] Repeating the above process, the total decision space O after dimensionality reduction is described as:

[0195]

[0196] n is the number of electricity-storage-carbon collaborative scenarios. The length of the finally recommended electricity-storage-carbon collaborative optimization combination is L. To maintain a similar selection ratio in the search space before and after dimensionality reduction, the formula is:

[0197] S / L e = L e / L (29)

[0198]

[0199] During the multi-scenario collaborative optimization scheduling process, the genetic algorithm encodes the initial values to form an initial population. The initial population consists of a group of chromosomes, and each chromosome represents a solution to the problem, represented by a k-source vector. During the population generation process, duplicate genes are avoided in the same chromosome. After the population is formed, the conversion from the objective function to the fitness function is carried out. The formula for the fitness function is:

[0200]

[0201] The gene operation process includes selection, crossover, and mutation operations. In the selection operation, the selection of an individual is associated with the fitness. The higher the fitness, the higher the selection probability. The selection probability of an individual is:

[0202]

[0203] where v i is the fitness value of the population individual, and n is the number of population individuals.

[0204] During the crossover operation, the k-th chromosome r k and the l-th chromosome rl The formula for swapping at position j simultaneously is:

[0205] r k,j = r k,j (1 - b)+r l,j b (33)

[0206] r l,j = r l,j (1 - b)+r k,j b (34)

[0207] where b is a random number in [0, 1].

[0208] During the mutation operation, the formula for mutating the j-th gene of the i-th individual is:

[0209]

[0210] where r min and r max are the upper and lower limits of gene r i,j , ξ is a random number between 0 and 1, g is the iteration number, and G max is the maximum number of evolutions.

[0211] Using the improved particle swarm optimization algorithm to solve the user demand optimization model, the update formulas for velocity and position are:

[0212] v j (d + 1)= ωv j (d)+c1r1(p j (d)-x j (d))+c2r2(g j (d)-x j (d)) (36)

[0213] x j (d + 1)= x j (d)+v j (d + 1) (37)

[0214] where: ω is the dynamically adjusted inertia weight; d represents the iteration number; p j (d) is the optimal position found by the j-th particle so far, g j (d) is the optimal position found by all particles in the entire population so far, j represents the corresponding particle number; x j (d) and v j (d) are the position and velocity of the particle respectively; c1 and c2 are linearly decreasing learning factors; r1 and r2 are random numbers between (0, 1).

[0215] The dynamic adjustment formula for the inertia weight is:

[0216]

[0217] where ω max is the maximum value of the inertia weight, ω min is the minimum value of the inertia weight, d max is the maximum number of iterations, d is the current number of iterations, and σ is the inertia adjustment factor; betarnd(p, q) is a function for generating random numbers obeying the beta distribution, and p, q are the shape parameters of the beta distribution.

[0218] The linearly decreasing learning factor is:

[0219]

[0220] where c max is the maximum value of the learning factor, c min is the minimum value of the learning factor, d max is the maximum number of iterations, and d is the current number of iterations.

[0221] As Figure 2 shown, the process of the improved algorithm for hierarchical solution of the bi-level optimization model in the present invention includes: parameter initialization and data generation, dynamic constraint processing and hierarchical model solution, user satisfaction evaluation and load adjustment feedback, and termination condition determination and result output. The specific steps are as follows:

[0222] Initial parameter generation:

[0223] The input is to determine the core parameters of the improved genetic algorithm and particle swarm algorithm, including population size, crossover and mutation rate, inertia weight, and learning factor range;

[0224] Data generation is to initialize the predicted output data of thermal power, wind power, and photovoltaic power, the initial state of charge of energy storage devices, and the original load curve of users.

[0225] Dynamic constraint processing:

[0226] Carbon trading cost calculation is to dynamically adjust the carbon trading price according to the difference between the actual carbon emissions and quotas of thermal power units, and use a piecewise function mechanism to impose cost constraints on the over-emission part;

[0227] Safety constraint loading is to integrate operating conditions such as the output limit and ramp rate constraint of thermal power units, the safe range of energy storage charge and discharge power, and the grid power flow transmission capacity limit.

[0228] Hierarchical model solution:

[0229] Upper layer (electricity-storage-carbon collaborative optimization): The solution objective is to optimize the collaborative output plan of thermal power, wind power, photovoltaic power, and energy storage through an improved genetic algorithm to minimize the comprehensive power generation cost; a decision space dimensionality reduction strategy is adopted to screen high-dimensional scenarios, and the optimal solution is evaluated based on the fitness function.

[0230] Lower layer (user demand optimization): The solution objective is to optimize the user load curve through an improved particle swarm algorithm, generate real-time electricity price signals, and adjust electricity consumption demand; the algorithm parameters are dynamically adjusted to balance user satisfaction and system economy.

[0231] User satisfaction evaluation and feedback:

[0232] Satisfaction calculation includes satisfaction with the energy consumption method: the matching degree of energy consumption habits is measured by the ratio of the adjusted electricity consumption of users to the original load; satisfaction with costs: the acceptance of cost changes is reflected by the ratio of the real-time electricity price to the initial benchmark electricity price.

[0233] Load adjustment judgment: According to the comprehensive level of satisfaction and the load peak-valley difference, decide whether to trigger electricity price adjustment and load re-optimization.

[0234] Result output:

[0235] On the power generation side, the optimized output plans of thermal power, new energy, and energy storage, as well as the dynamic carbon trading cost, are output;

[0236] On the user side, the adjusted load curve, real-time electricity price signal, and user two-dimensional satisfaction index are output.

[0237] In this embodiment, after the double-layer optimization model is scheduled, the maximum load before considering user satisfaction is 2647 kW, and the maximum peak-valley difference is 927 kW. After considering user satisfaction, the maximum load is 2389 kW, and the maximum peak-valley difference is 684 kW. The satisfaction of users' energy consumption methods after hierarchical optimization is 0.8746, the satisfaction with energy consumption costs is 0.8921, and the overall satisfaction is 1.7667. The double-layer optimization scheduling method not only achieves a high level of user demand satisfaction but also, under the premise of meeting the stability requirements of the power system, realizes the reasonable utilization of wind power and photovoltaic units. The output of thermal power units changes due to the change in load, reducing the overall carbon emissions and meeting the low-carbon requirements of the system.

[0238] In this embodiment, a constraint model for dynamically and segmentally adjusting the carbon trading price according to carbon emissions is constructed, a mechanism combining free quotas and auction quotas is introduced, and a quadratic function is used to impose non-linear penalties on carbon emissions exceeding the threshold. Compared with the traditional stepped carbon price model, it significantly enhances the binding force on high carbon emissions of thermal power units.

[0239] Jointly optimize thermal power / new energy output, energy storage charging / discharging, and carbon trading costs through the upper-layer model; the lower-layer model quantifies user needs using two indicators of energy usage satisfaction and cost satisfaction, and optimizes the load curve by combining with electricity price adjustment. Break through the limitation that energy storage operation and user needs are not coupled in the existing technology. The upper layer adopts an improved genetic algorithm with decision space dimensionality reduction to compress the high-dimensional scenario search space; the lower layer introduces a particle swarm optimization algorithm with beta distribution to dynamically adjust the inertia weight to improve the global search efficiency. Compared with the traditional genetic algorithm and ADMM algorithm, the solution speed and convergence accuracy are significantly improved.

[0240] Through four major innovations, namely, a dynamic carbon constraint mechanism, a two-layer collaborative optimization framework, intelligent algorithm improvement, and a user demand quantification model, the present invention solves the contradiction between the rigid carbon emission reduction constraint and the flexible demand for thermal power peak shaving in the existing technology. Through a dynamic segmented pricing mechanism, carbon constraints are strengthened while ensuring the consumption of new energy; at the same time, the contradiction between improving user satisfaction and optimizing system economy is solved, and the balance between user experience and operating cost is achieved using a dual-objective demand response model. The hierarchical improvement algorithm is used to significantly reduce the computational complexity and support real-time scheduling requirements. It is applicable to scenarios such as high-proportion new energy power grids, carbon-electricity collaborative markets, and park-level integrated energy systems, providing a "low-carbon - economic - user-friendly" trinity optimal scheduling solution for the new power system.

Claims

1. A two - layer optimal scheduling method considering the coordination of electricity - storage - carbon and user demands, characterized in that: It includes the following steps: S1. According to the carbon emission quotas of thermal power, wind power and photovoltaic power, based on the benchmark carbon trading cost model, construct a constraint model in which the carbon trading price is dynamically segmented and adjusted with the carbon emission volume; S2. With the goal of minimizing the comprehensive power generation cost, jointly optimize the output plans of thermal power, wind power, photovoltaic power and energy storage, and couple the constraint model to limit the carbon emissions of thermal power units, and construct an electricity - storage - carbon collaborative optimization model; S3. Based on the satisfaction of users' energy - using patterns and the satisfaction of energy - using costs, construct a demand response mechanism, optimize the user load curve through electricity price adjustment, and establish a user demand optimization model; S4. Use an improved algorithm to hierarchically solve the two - layer optimization model. Solve the electricity - storage - carbon collaborative optimization model through the upper - layer improved genetic algorithm, and solve the user demand optimization model through the lower - layer improved particle swarm algorithm, and iterate and update until the system comprehensive optimal conditions are met.

2. The two-layer optimal scheduling method considering the coordination of electricity-storage-carbon and user needs as described in claim 1, characterized in that In the step S1, according to the carbon emission quotas of thermal power, wind power and photovoltaic power, it is determined by the following formula: The carbon emission quota EA1 of thermal power determined by the baseline method is: where: δ EF is the reference carbon emission factor, T is the scheduling period, N is the number of thermal power units, P G,i,t is the power of thermal power unit i at time t, ξ is the carbon quota correction coefficient for wind power and photovoltaic power, α is the reserve coefficient of wind power output, P w,t is the output of the wind turbine at time t, β is the reserve coefficient of photovoltaic power output, P p,t is the output of the photovoltaic power at time t; The carbon emission quotas EA2 and EA3 of wind power and photovoltaic power are: Based on the benchmark carbon trading cost model, construct a constraint model in which the carbon trading price is dynamically segmented and adjusted with the carbon emission volume. The specific steps are as follows: The initial carbon emission quota adopts the form of mainly free allocation and supplemented by auction allocation. The proportion of auction quotas is ε, and the auction quota volume is: EA a = ε(EA1 + EA2 + EA3) (4) The free quota volume is: EA f = (1 - ε)(EA1 + EA2 + EA3) (5) The total carbon emission volume of thermal power units within a certain cycle range is: Let \(T\) be the scheduling period and \(N\) be the number of thermal power units. i \(\xi_i\) is the carbon emission coefficient of thermal power unit \(i\), and \(P\) G,i,t is the power of thermal power unit \(i\) at time \(t\). The carbon trading price is c1, the quota auction price is c2, and the carbon trading cost F c1 is as follows: F c1 = c1(E p - EA f - EA a_ ) + c2EA a (7) Carbon trading cost F under dynamic constraints c is as follows: Let \(l\) be the interval length of the carbon trading cost under dynamic constraints, \(\lambda\) be the dynamic constraint price growth coefficient, and \(E\) t be the carbon trading volume, which is the difference between the carbon emissions of thermal power units and the quotas, \(E\) t = \(E\) p - \(E_A\) f - \(E_A\) a , 5l < E t When this is the case, a quadratic function is adopted to restrict the carbon emissions of unit power generation through high transaction costs.

3. A two-layer optimal scheduling method considering the coordination of electricity-storage-carbon and user needs as described in claim 1, characterized in that In the step S2, the objective function of the electricity - storage - carbon collaborative optimization model includes: the lowest comprehensive power generation cost, the power generation cost of thermal power units, the penalty cost for wind and photovoltaic curtailment, and the operation cost of energy storage devices; The lowest comprehensive power generation cost is: minF t = F f + F w + F p + F b + F loss + F c (9) F t is the total power generation cost under the coordination of electricity - energy storage - carbon, F f is the power generation cost of thermal power units, F w is the curtailment of wind penalty cost, F p is the curtailment of solar penalty cost, F b is the operating cost of energy storage devices, F loss is the network loss cost, F c is the carbon trading cost under dynamic constraints; The power generation cost of a thermal power unit includes the coal consumption cost F for power generation f1 , the start-up and shutdown cost F of the unit f2 , and the specific formula is as follows: P g,i,t is the output power of thermal power unit i at time t, a i , b i and c i are the cost coefficients of the quadratic function of coal cost with respect to output power, S on,i,t , S off,i,t is the on / off action power of unit i at time t, η on,i , η off,i are the cost coefficients for the start-up and shut-down of unit i; The penalty cost for wind and photovoltaic curtailment is: P wq,t and P pq,t are the wind curtailment power and solar curtailment power during the t period, and μ w and μ p are the penalty cost coefficients for wind curtailment and solar curtailment; The operation cost of energy storage devices is: F b = ν d P d + ν s P s (14) ν d 、ν s are the cost coefficients for the discharge and charge of the energy storage device, P d and P s are the discharge and charge powers of the energy storage device.

4. A two-layer optimal scheduling method considering the coordination of electricity-storage-carbon and user demands as described in claim 3, characterized in that The constraint conditions of the objective function include: the safety stability and reactive power constraints of the output adjustment of thermal power units, the start - up and shutdown constraints of thermal power units, the output constraints of wind power and photovoltaic power, the operation constraints of energy storage devices, the grid operation safety constraints and the system power balance constraints; The safety stability and reactive power constraints of the output adjustment of thermal power units are: P g,i,min and P g,i,max are the minimum and maximum output powers of the thermal power unit respectively, and R i,min and R i,max are the minimum and maximum values of the ramp rate of the thermal power unit respectively. Q i,t is the reactive power generated by the thermal power unit i at time t, and Q i,min and Q i,max are the minimum and maximum values of the reactive power of the unit i; The start - up and shutdown constraints of thermal power units are: τ i,t is the variable of the operating state of the thermal power unit changing from shutdown to startup in the t period, T on,i,t-1 is the continuous operation time of thermal power unit i in the t-1 period, T off,i,t-1 is the outage time of thermal power unit i in the t-1 period; The output constraints of wind power and photovoltaic power are: 0 ≤ P w,t ≤ P w，max (17) 0 ≤ P p,t ≤ P p,max (18) P w，max and P p,max are the upper limits of the wind power and photovoltaic power outputs, respectively; The operation constraints of energy storage devices are: P s,t is the charging power of the energy storage device at time t, P s,min , P s,max are the minimum and maximum power values for charging the energy storage device, respectively. Y s,t , Y d,t are the charging and discharging states of the energy storage device during the t period, respectively, and are between 0 and 1. S soc,t is the state of charge of the energy storage device at time t, S soc,min , S soc,max are the minimum and maximum values of the state of charge of the energy storage device, respectively. S soc,s,t , S soc,e,t are the initial and final states of charge of the energy storage device within a t period, respectively; The grid operation safety constraints are: (P G,zl + P w,zl + P p,zl + P s,zl ) 2 + Q G,zl ≤ (S l,max ) 2 (20) P G,zl 、P w,zl 、P p,zl 、P s,zl are the active power transmitted by thermal power, wind power, photovoltaic power and energy storage on the branch, respectively. Q G,zl is the reactive power transmitted by thermal power on the branch, and S l,max is the maximum transmission capacity on the branch; The system power balance constraints are: α d is the discharge efficiency of the energy storage device, J is the total number of loads, P load,j,t is the power value of load j at time t, P loss,t is the active power loss of the branch at time t, Q i,t is the reactive power output of unit i at time t, Q load,j,t is the reactive power of load j at time t, Q loss,t is the reactive power loss of the branch at time t.

5. A two - layer optimal scheduling method considering the coordination of electricity - storage - carbon and user demands as described in claim 1, characterized in that In the step S3, the user demand optimization model is constructed through the following indicators: The user satisfaction objective function is: max F s = R 1,c + R 2,c (23) R 1,c is the satisfaction degree of the user's energy consumption mode, R 2,c is the satisfaction degree of the user's energy consumption cost; The satisfaction of users' energy - using patterns is: P c,t is the original load of the user during the t period, P’ c,t is the electricity consumption of the user after participating in demand response. The satisfaction level when users without time-of-use pricing use electricity freely is 1; The satisfaction value of energy - using costs after participating in demand response is: ρ e is the initial electricity price, ρ' e is the real-time electricity price after participating in demand response, and the satisfaction of the initial energy consumption cost for users' free electricity consumption is 1; The constraint of demand response, that is, electricity price response, is: D j,t is the demand response volume of load j in period t, and Δρ e,max is the maximum value of electricity price adjustment.

6. A two-layer optimal scheduling method considering the coordination of electricity-storage-carbon and user demands as described in claim 1, characterized in that In step S4, the improved genetic algorithm adopts a decision space dimensionality reduction strategy to compress a high-dimensional scenario into a list of candidate optimizations, and generates an initial population through multi-scenario collaborative coding. During the decision space dimensionality reduction process, each electricity-storage-carbon collaborative scenario S is calculated as four candidate electricity-storage-carbon collaborative optimization lists of length L e respectively as X i1 , X i2 , X i3 and X i4 . The candidate electricity-storage-carbon collaborative optimization lists X i1 , X i2 , X i3 and X i4 are the current optimal solutions. The formula for combining the candidate lists to obtain the final decision space O i is: O i = X i1 ∪ X i2 ∪ X i3 ∪ X i4 (27) Repeat the above process, and the reduced - dimension total decision - making space O is described as: O = {O1, O2, …, O n} (28) n is the number of electricity - storage - carbon collaborative scenarios, the length of the finally recommended electricity - storage - carbon collaborative optimization combination is L, and the formula for maintaining a similar selection ratio in the search space before and after dimensionality reduction is: S / L e = L e / L (29) In the process of multi-scenario collaborative optimization scheduling, the genetic algorithm encodes the initial values to form an initial population. The initial population consists of a group of chromosomes, and each chromosome represents a solution to the problem, which is represented by a k-source vector. During the population generation process, duplicate genes are avoided in the same chromosome. After the population is formed, it is converted from the objective function to the fitness function. The formula for the fitness function is: The gene operation process includes selection, crossover, and mutation operations. The selection of an individual is associated with the fitness. The higher the fitness, the higher the selection probability. The probability of an individual being selected is: where v i is the fitness value of the population individuals, and n is the number of population individuals; During the crossover operation, the k-th chromosome r k and the l-th chromosome r l are simultaneously exchanged at the j position, and the formula is: r k,j = r k,j (1 - b)+ r l,j b (33) r l,j = r l,j (1 - b)+ r k,j b (34) where b is a random number in [0, 1]; During the mutation operation, the j-th gene of the i-th individual mutates, and the formula is: where r min and r max are the upper and lower limits of gene r i,j , ξ is a random number between 0 and 1, g is the number of iterations, and G max is the maximum number of generations.

7. A two-layer optimal scheduling method considering the coordination of electricity-storage-carbon and user demands as described in claim 6, characterized in that In step S4, the improved particle swarm algorithm dynamically adjusts the inertia weight using the beta distribution, and the learning factor decreases linearly with the number of iterations. The update formula is: v j (d + 1) = ωv j (d) + c1r1(p j (d) - x j (d)) + c2r2(g j (d) - x j (d)) (36) x j (d + 1) = x j (d) + v j (d + 1)(37) where: w is the dynamically adjusted inertia weight; d represents the number of iterations; p j (d) is the optimal position found so far by the j-th particle, g j (d) is the optimal position found so far by all particles in the entire population, j represents the corresponding particle number; x j (d) and v j (d) are the position and velocity of the particle, respectively; c1 and c2 are linearly decreasing learning factors; r1 and r2 are random numbers between (0, 1); The dynamic adjustment formula for the inertia weight is: Among them, ω max is the maximum value of the inertia weight, ω min is the minimum value of the inertia weight, d max is the maximum number of iterations, d is the current number of iterations, and σ is the inertia adjustment factor; betarnd(p, q) is a function for generating random numbers obeying the beta distribution, and p, q are the shape parameters of the beta distribution; The linearly decreasing learning factor is: Among them, c max is the maximum value of the learning factor, and c min is the minimum value of the learning factor. d max is the maximum number of iterations, and d is the current number of iterations.

8. An electronic device, characterized in that, It includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, it implements the steps of the method according to any one of claims 1-7.

9. A computer-readable storage medium, characterized in that, The storage medium stores computer-executable instructions, and when the instructions are executed by the processor, it implements the steps of the method according to any one of claims 1-7.

Images