SDM-CADMM distributed method for unit commitment P2P energy transaction

Abstract

The invention provides an SDM-CADMM distributed method for unit commitment P2P energy transaction, and relates to the field of electrical engineering, and the method comprises the steps: building a P2P energy transaction model of producers and consumers based on social welfare maximization; remodeling the P2P energy transaction model in the matrix form; and aiming at the non-convexity problem caused by integer variables contained in the remodeled P2P energy transaction model, designing an SDM-CADMM algorithm, and solving the mixed integer nonlinear P2P energy transaction problem. According to the method, the social welfare maximization problem in mixed integer nonlinear programming (MINLP) can be effectively solved in a distributed mode, parallel computing is achieved, privacy of two transaction parties is better protected, meanwhile, the non-convexity problem caused by integer variables in unit commitment is effectively solved, and the social welfare maximization efficiency is improved. Therefore, the optimal solution of the mixed integer nonlinear programming problem with the cost minimization as the target is obtained, and the optimality of the transaction result is better ensured.

Description

Technical Field

[0001] The present invention relates to the field of electrical engineering, and in particular to a distributed algorithm for peer-to-peer (P2P) energy trading of unit combinations, and in particular to an SDM-CADMM distributed method for P2P energy trading of unit combinations. Background Art

[0002] The advancement of modern technology has promoted the rise of P2P networks. This completely decentralized network architecture enables market participants to easily share their resources. P2P networks allow users with distributed energy to trade directly. Each user is both a producer and a consumer, and the status of each user in the network is equal. As an emerging energy supply model, P2P energy trading can not only ensure the equal status of both parties to the transaction and user privacy, and enhance the enthusiasm of producers and consumers to participate in the energy market; it can also reduce electricity demand, minimize reserve requirements, and improve the resilience of the power system.

[0003] Traditional energy trading problems mostly use centralized optimization methods, which perform comprehensive calculations and scheduling through a central control system to directly obtain the optimal trading plan. However, with the rapid development of renewable energy, more and more distributed power generation equipment is connected to the power grid, which significantly increases the time and computing resource consumption of centralized optimization methods. These problems make it difficult for centralized methods to effectively deal with large-scale grid-connected distributed energy, forcing energy trading systems to develop in a more open and decentralized direction to improve the flexibility, reliability and real-time response capabilities of the system.

[0004] In order to solve the above problems, some studies have begun to use distributed methods to optimize trading schemes. Distributed Optimization (DO) methods can not only achieve parallel computing, but also maintain the optimality of trading results. They are an effective means to design new energy trading mechanisms. However, current research mainly focuses on solving linear or convex global optimization problems. When the unit commitment (UC) problem is considered in the system, the introduced binary variables will cause the problem to become non-convex, and the traditional DO algorithm usually cannot guarantee convergence to the optimal result. Summary of the invention

[0005] In order to solve the non-convexity problem caused by integer variables in unit combination, the present invention provides a SDM-CADMM distributed method for unit combination P2P energy trading, constructs a social welfare maximization problem involving the user side and the power generation side; designs an SDM-CADMM algorithm to solve the MINLP problem in a distributed manner; in the iterative process, the method converts the non-convex set into an approximate set in the convex hull, and obtains the optimal solution to the MINLP problem when the algorithm converges, thereby ensuring the optimality of the transaction result. The specific steps of the method are as follows:

[0006] S1: Establish a P2P energy trading model between producers and consumers based on maximizing social welfare;

[0007] S2: Reshape the P2P energy trading model in matrix form;

[0008] S3: Aiming at the non-convexity problem caused by the integer variables contained in the reshaped P2P energy trading model, the SDM-CADMM algorithm is designed to solve the mixed integer nonlinear P2P energy trading problem.

[0009] A storage device stores instructions and data for implementing an SDM-CADMM distributed method for unit combination P2P energy trading.

[0010] A SDM-CADMM distributed device for unit combination P2P energy trading includes: a processor and a storage device; the processor loads and executes instructions and data in the storage device to implement an SDM-CADMM distributed method for unit combination P2P energy trading.

[0011] Compared with the prior art, the technical solution provided by the present invention has the following beneficial effects: the traditional P2P energy trading does not consider the unit combination problem, while the present invention takes the unit combination and the uncertainty of wind power generation into consideration, and establishes a day-ahead P2P energy trading model based on social welfare maximization. When solving the P2P energy trading problem, the traditional method usually assumes that the problem is linear or convex. However, the non-convexity problem has not been solved. Therefore, the present invention proposes an SDM-CADMM algorithm that does not rely on a central coordinator, which can effectively solve the social welfare maximization problem in mixed integer nonlinear programming (MINLP) in a distributed manner to achieve parallel computing, rather than centralized scheduling for unified management, so as to better protect the privacy of both parties to the transaction. The present invention effectively solves the non-convexity problem caused by integer variables in the unit combination, thereby obtaining the optimal solution of the mixed integer nonlinear programming (MINLP) problem with cost minimization as the goal, thereby better ensuring the optimality of the transaction results. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The present invention will be further described below with reference to the accompanying drawings and embodiments, in which:

[0013] Figure 1 It is a flow chart of an SDM-CADMM distributed method for unit combination P2P energy trading in an embodiment of the present invention.

[0014] Figure 2 It is a schematic diagram of the operation of the hardware device in the embodiment of the present invention. DETAILED DESCRIPTION

[0015] In order to have a clearer understanding of the technical features, purposes and effects of the present invention, specific embodiments of the present invention are now described in detail with reference to the accompanying drawings.

[0016] Example 1

[0017] Please refer to Figure 1 , Figure 1 : is a flow chart of an SDM-CADMM distributed method for unit combination P2P energy trading in an embodiment of the present invention, which specifically includes:

[0018] S1: Establish a P2P energy trading model between producers and consumers based on maximizing social welfare; the specific steps are as follows:

[0019] S1.1: Establish a bilateral transaction model between producers and consumers:

[0020]

[0021] Among them, formula (1) is the calculation of the net injected power of node n at time t, formula (2) is the active output range constraint of renewable energy under scenario s, formula (3) indicates that the net injected power of node n at time t is equal to the sum of its transaction volume with all adjacent individuals, formula (4) indicates that prosumer n sells electricity to adjacent prosumer m, in which case prosumer n is the producer and prosumer m is the consumer; formula (5) indicates that prosumer n purchases electricity from adjacent prosumer m, in which case prosumer n is the consumer and prosumer m is the producer; formula (6) indicates that in energy trading, the sum of the electricity purchased and sold between prosumer n and prosumer m is 0.

[0022] Among them, E n,t represents the net injected power of each prosumer n at time t, E n,t ={E n1,t ,...,E nm,t ,m∈ω n}, E n,t ≥0 means individual n is a producer, E n,t ≤0 means individual n is a consumer, represents the predicted output value of wind farm n at time t under scenario s, represents the amount of wind curtailment of wind farm n at time t under scenario s, represents the output of thermal power unit n at time t in scenario s, represents the electricity load of user n in scenario s, the set Ω represents all producers and consumers, and the set Ω p Represents all producers, set Ω c represents all consumers, set Λ is the transaction time set, set S is the scenario set, set ω n is the set of neighbors of prosumer n.

[0023] S1.2: Taking the generator set as the producer, establish its safe operation constraints in the energy trading process, that is, the unit combination problem; the safe operation constraints are:

[0024]

[0025] Among them, formula (7) is the upper and lower limit constraints of the output of thermal power unit n at time t, formula (8) represents the upper limit constraint of the climbing of thermal power unit n under scenario s, formula (9) is the lower limit constraint of the climbing of thermal power unit n under scenario s, formula (10) is the start-stop mutual exclusion constraint of thermal power unit n at time t, formula (11) associates the operating status and operation of thermal power unit n at time t and t-1, formula (12) is the upper limit constraint of the continuous start-stop duration of thermal power unit n at time t and t-1, and formula (13) is the lower limit constraint of the continuous start-stop duration of thermal power unit n at time t and t-1.

[0026] in, They represent the upper limit of the output of thermal power unit n, P n Indicates the lower limit of the output of thermal power unit n, P n,t represents the lower limit of the output of thermal power unit n at time t, It represents the upper limit of the output of thermal power unit n at time t; represents the output of thermal power unit n at time t in scenario s, represents the output of thermal power unit n at time t-1 under scenario s; I n,t ∈{0,1} represents the start and stop status of thermal power unit n at time t, I n,t =0 means that the thermal power unit n is in shutdown state at time t, I n,t =1 means that the thermal power unit n is in the on state at time t; u n,t / v n,t ∈{0,1} respectively represent the start / shutdown operation of thermal power unit n at time t, u n,t =1 means that the thermal power unit n starts at time t, u n,t=0 means that the thermal power unit n has not started at time t, v n,t =1 means that the thermal power unit n is shut down at time t, v n,t =0 means that the thermal power unit n did not shut down at time t. They represent the upper limit of the ramp rate of thermal power unit n, Δ p n They represent the lower limit of the ramp rate of thermal power unit n, I n,t-1 represents the start / stop status of thermal power unit n at time t-1, I n,τ represents the start and stop status of thermal power unit n at time τ, T is the last time period, Indicates the minimum duration of continuous startup of thermal power units. They respectively represent the minimum duration of continuous shutdown of thermal power units.

[0027] S1.3: Taking the power users as consumers, establish the power load range constraints in the energy trading process; the power load range constraints are:

[0028]

[0029] Among them, formula (14) is the upper and lower limit constraints of the power load of user n at time t, and formula (15) represents the total energy demand of user n at all times.

[0030] Among them, D n,t represents the electricity load of user n at time t, They represent the upper limit of the power load of user n at time t, D n,t They represent the lower limit of the power load of user n at time t, It is the sum of the electricity load of user n at all times.

[0031] S1.4: Taking cost minimization as the objective function, a P2P energy trading model is constructed that includes the social welfare maximization problem on the user side and the power generation side. The P2P energy trading model is:

[0032]

[0033] st(1)-(15)(17)

[0034]

[0035] Among them, formula (16) is the objective function to achieve cost minimization, formula (17) includes various constraints on the user side and the power generation side, and formula (18) is the expected social welfare of all producers and consumers. represents the minimum value of social welfare expected by all producers and consumers. Formula (19) is the cost-utility function of producer n, that is, C n (Pn,t ) represents the cost-utility function of producer n, and formula (20) represents the utility function of consumer n, that is, U n (D n,t ) represents the utility function of consumer n; a n , b n d n is a normal number, is a positive constant of the cost-utility function, is a positive constant of the utility function, c n is the startup cost coefficient of thermal power unit n.

[0036] The P2P energy trading model for cost minimization problem can be written in the following matrix form:

[0037]

[0038]

[0039] z=(z n ) n∈Ω ,z∈Z (24)

[0040] Among them, formula (21) is the matrix form of the cost minimization objective function, and formula (16) is f n (x n ) in its specific form; X n is x n The feasible set of x n is a feasible solution for the nth prosumer, forming a non-convex set formed by the mixed integer nonlinear constraints (17), x n represents the set of all decision variables, I n represents the start and stop state vector of thermal power unit n, u n represents the startup operation vector of thermal power unit n, v n represents the shutdown operation vector of thermal power unit n, represents the output vector of thermal power unit n in scenario s, E n represents the net injected power vector of each prosumer n; Q n represents the constant matrix of the nth producer and consumer, Z represents the coupling relationship between producers and consumers, z represents the introduced auxiliary variable, Z={z=(z n ) n＝1,2,..Ω.}, z n Represents the auxiliary variable introduced by the nth prosumer.

[0041] S2: Reshape the P2P energy trading model in matrix form. The specific steps are as follows:

[0042] Step 2.1: Convert the non-convex set into a convex set. The specific model is as follows

[0043]

[0044] z=(z n ) n∈Ω ,z∈Z (28)

[0045] Among them, conv(X n ) represents a non-convex set X n The convex hull points of .

[0046] S2.2: Lagrangian dual function of the original problem

[0047]

[0048] λ∈z T :={v:v T z=0,z∈Z} (30)

[0049]

[0050] Wherein, formula (29) is the convex relaxed dual function of the original problem, that is, φ(λ) represents the convex relaxed dual function of the original problem, and λ is the set of dual multiplier vectors; Represents the equality constraint Q corresponding to the nth producer and consumer n x n =z n The transpose of the dual multiplier vector, Q n represents the constant matrix of prosumers n, x n represents the set of all decision variables, z n Auxiliary variable representing prosumer n, z T represents the transpose of the auxiliary variable set, n represents the nth producer and consumer, v represents the set of dual multiplier vectors, and v T represents the transpose of v.

[0051] S2.3: Augmented Lagrangian function of the original problem

[0052]

[0053] Wherein, formula (33) is the augmented Lagrangian function form of the original problem. L(x,z,λ) represents the augmented Lagrangian function of the original problem, ρ represents the penalty parameter, z n→m represents the transaction volume between prosumer n and its neighboring prosumer m, z m→n represents the transaction volume between prosumer m and its neighboring prosumer n, Represents the average transaction volume between prosumer n and prosumer m.

[0054] S2.4: Approximation of the augmented Lagrangian function of the original problem

[0055]

[0056] Among them, formula (35) is the mirror function of all augmented Lagrangian functions, that is represents the mirror function of all augmented Lagrangian functions, represents the mirror function of the augmented Lagrangian function of prosumer n at the kth iteration. Formula (36) is the objective function f n The feasible solution for the nth prosumer in the kth iteration is The first-order Taylor expansion at , formula (37) is the tangent plane function of all augmented Lagrangian functions, that is, represents the tangent plane function of all augmented Lagrangian functions, represents the tangent plane function of the augmented Lagrangian function of producer / consumer n at the kth iteration. Formula (38) is The tangent plane approximation function centered at is represents the tangent plane approximation function, and the approximation function satisfies Denotes the objective function f n At the point The first-order derivative at , represents the feasible solution in the kth iteration, represents the auxiliary variable introduced by prosumer n at the kth iteration.

[0057] S3: Aiming at the non-convexity problem caused by the integer variables in the reshaped P2P energy trading model, a SDM-CADMM algorithm that does not rely on a central coordinator is designed to solve the mixed integer nonlinear P2P energy trading problem. The specific SDM-CADMM algorithm is as follows:

[0058] S3.1: Update the feasible solution x of the nth producer and consumer n

[0059]

[0060] Among them, x n represents the set of all decision variables, represents the auxiliary variable introduced by prosumer n at the k-1th iteration, represents the transpose of the dual multiplier vector of the nth prosumer, is the convex hull feasible domain conv(X n ) is the inner approximation set of the nth producer and consumer, which is composed of a finite convex set composition, represents the number of data in the convex set, α i represents the weight of the convex set, is a convex set The i-th vertex in n represents the set of all decision variables.

[0061] S3.2: Update the auxiliary variable z of the nth prosumer n

[0062]

[0063] in, represents the auxiliary variable of the nth prosumer in the kth iteration.

[0064] S3.3: Update the convex hull feasible domain D of the nth producer and consumer n

[0065]

[0066] Wherein, formula (41) represents the minimization At the point The gradient at is conv(X n ) is a vertex of . represents the feasible domain of the convex hull conv(X n )’s inner approximate set of the nth producer and consumer, It represents the feasible domain of the convex hull conv(X n )’s inner approximate set of the nth producer and consumer, Representation function At point x n The gradient at is the convex hull conv(X n ) is a vertex of .

[0067] S3.4: Convergence judgment is performed through the operations of steps S3.1-S3.4

[0068]

[0069] Among them, ∈ is the threshold for judging whether the SDM-CADMM algorithm is terminated, ∈≥0, when When , the algorithm converges. It represents the difference between the tangent plane function value and the mirror function value of the nth augmented Lagrangian function at the kth iteration.

[0070] S3.5: Update the dual multiplier λ of the nth producer / consumer n and mirror function

[0071]

[0072] When ηk ≥γ, update λ n and

[0073]

[0074] Among them, γ∈(0,1) represents the given strict step size condition; η k represents the ratio of the difference between the tangent plane function value of the augmented Lagrangian function at the kth iteration and the k-1th iteration and the difference between the mirror function value of the augmented Lagrangian function at the kth iteration and the tangent plane function value of the augmented Lagrangian function at the k-1th iteration; Represents the updated dual multiplier vector.

[0075] Otherwise, λ n and remain unchanged,

[0076]

[0077] in, represents the transaction volume between prosumer n and its neighboring prosumer m in the kth iteration, represents the transaction volume between prosumer m and its neighboring prosumer n in the kth iteration; represents the mirror function of the augmented Lagrangian function at the k+1th iteration, Represents the dual multiplier vector of the nth producer and consumer at the k+1th iteration.

[0078] Example 2

[0079] A SDM-CADMM distributed device 201 for unit combination P2P energy trading, such as Figure 2 As shown, it includes: a processor 202 and a storage device 203; the processor 202 loads and executes instructions and data in the storage device 203 to implement an SDM-CADMM distributed method for unit combination P2P energy trading.

[0080] Example 3

[0081] A storage device stores instructions and data for implementing an SDM-CADMM distributed method for unit combination P2P energy trading.

[0082] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A SDM-CADMM distributed method for unit combination P2P energy trading, characterized by: include: S1: Establish a P2P energy trading model between producers and consumers based on maximizing social welfare; S2: Reshape the P2P energy trading model in matrix form; S3: Aiming at the non-convexity problem caused by integer variables in the reshaped P2P energy trading model, the SDM-CADMM algorithm is designed to solve the mixed integer nonlinear P2P energy trading problem; The designed SDM-CADMM algorithm is: S3.1: Update the feasible solution x of the nth producer and consumer n : in, represents the augmented Lagrangian function, x n represents the set of all decision variables, represents the auxiliary variable introduced by prosumer n at the k-1th iteration, represents the dual multiplier vector of the nth producer and consumer at the kth iteration, is the convex hull feasible domain conv(X n ) is the inner approximation set of the nth producer and consumer, which is composed of a finite convex set composition, Represents a convex set The amount of data in α i represents the weight of the convex set, is the i-th vertex in the convex set; S3.2: Update the auxiliary variable z of the nth prosumer n : in, represents the auxiliary variable of the nth prosumer in the kth iteration, Q n represents the constant matrix of the nth prosumer, represents the decision variable of the nth prosumer in the kth iteration, z n represents the auxiliary variable introduced by the nth prosumer; S3.3: Update the convex hull feasible domain D of the nth producer and consumer n : in, represents the augmented Lagrangian function At point x n The gradient at represents the auxiliary variable introduced by producer-consumer n at the kth iteration, is the convex hull feasible domain conv(X n ), Denotes the objective function f n At the point The first-order derivative at , represents the set of dual multiplier vectors of producer-consumer n at the kth iteration, ρ represents the penalty parameter, represents the feasible solution in the kth iteration, represents the transaction volume between prosumer n and its neighboring prosumer m at the kth iteration, represents the transaction volume between prosumer m and its neighboring prosumer n at the kth iteration, represents the feasible domain of the convex hull conv(X n )’s inner approximate set of the nth producer and consumer, It represents the feasible domain of the convex hull conv(X n )’s inner approximation set of the nth prosumer; S3.4: Through the operations of steps S3.1-S3.4, make a convergence judgment: in, It represents the difference between the tangent plane function value and the mirror function value of the nth augmented Lagrangian function at the kth iteration. When , the algorithm converges, ∈ is the threshold for judging whether the SDM-CADMM algorithm is terminated, and ∈≥0; S3.5: Update the dual multiplier λ of the nth producer / consumer n and mirror function in, represents the updated dual multiplier vector; η k represents the ratio of the difference between the tangent plane function value of the augmented Lagrangian function at the kth iteration and the k-1th iteration and the difference between the mirror function value of the augmented Lagrangian function at the kth iteration and the tangent plane function value of the augmented Lagrangian function at the k-1th iteration; represents the tangent plane approximation function of the nth prosumer, represents the mirror function of the augmented Lagrangian function of the nth prosumer at the kth iteration, represents the mirror function of all augmented Lagrangian functions of the nth prosumer; When η k ≥γ, update λ n and Among them, γ∈(0,1) represents the given strict step size condition; Otherwise, λ n and Remain unchanged, that is: in, represents the mirror function of the augmented Lagrangian function at the k+1th iteration, Represents the dual multiplier vector of the nth producer and consumer at the k+1th iteration.

2. The SDM-CADMM distributed method for unit combination P2P energy trading according to claim 1, characterized in that: The specific steps of step S1 are as follows: S1.1: Establish a bilateral transaction model between producers and consumers; S1.2: Taking the generator set as the producer, establish its safe operation constraints in the energy trading process, that is, the unit combination problem; S1.3: Taking electricity users as consumers, establish constraints on their electricity load range in the energy trading process; S1.4: Construct a P2P energy trading model with cost minimization and social welfare maximization as the objective function.

3. The SDM-CADMM distributed method for unit combination P2P energy trading according to claim 2, characterized in that: In step S1.1, the bilateral transaction model is: Among them, formula (1) is the calculation of the net injected power of node n at time t; formula (2) is the active output range constraint of renewable energy under scenario s; formula (3) indicates that the net injected power of node n at time t is equal to the sum of its transaction volume with all adjacent individuals; formula (4) indicates that prosumer n sells electricity to adjacent prosumer m, in which case prosumer n is the producer and prosumer m is the consumer; formula (5) indicates that prosumer n buys electricity from adjacent prosumer m, in which case prosumer n is the consumer and prosumer m is the producer; formula (6) indicates that in energy trading, the sum of the electricity purchased and sold between prosumer n and prosumer m is 0; Among them, E n,t represents the net injected power of each prosumer n at time t, E n,t ={E n1 ,t,...,E nm,t ,m∈ω n }, E n,t ≥0 means individual n is a producer, E n,t ≤0 means individual n is a consumer, represents the predicted output value of wind farm n at time t under scenario s, represents the amount of wind curtailment of wind farm n at time t under scenario s, represents the output of thermal power unit n at time t in scenario s, represents the electricity load of user n in scenario s, the set Ω represents all producers and consumers, and the set Ω p Represents all producers, set Ω c represents all consumers, set Λ is the transaction time set, set S is the scenario set, set ω n is the set of neighbors of prosumer n.

4. The SDM-CADMM distributed method for unit combination P2P energy trading according to claim 3, characterized in that: In step S1.2, the safe operation constraints are: Among them, formula (7) is the upper and lower limit constraints of the output of thermal power unit n at time t, formula (8) represents the upper limit constraint of the climbing of thermal power unit n under scenario s, formula (9) is the lower limit constraint of the climbing of thermal power unit n under scenario s, formula (10) is the start-stop mutual exclusion constraint of thermal power unit n at time t, formula (11) associates the operating status and operation of thermal power unit n at time t and t-1, formula (12) is the upper limit constraint of the continuous start-stop duration of thermal power unit n at time t and t-1, and formula (13) is the lower limit constraint of the continuous start-stop duration of thermal power unit n at time t and t-1; in, They represent the upper limit of the output of thermal power unit n, P n Indicates the lower limit of the output of thermal power unit n, P n,t represents the lower limit of the output of thermal power unit n at time t, It represents the upper limit of the output of thermal power unit n at time t; represents the output of thermal power unit n at time t in scenario s, represents the output of thermal power unit n at time t-1 under scenario s; I n,t ∈{0,1} represents the start and stop status of thermal power unit n at time t, I n,t =0 means that the thermal power unit n is in shutdown state at time t, I n,t =1 means that the thermal power unit n is in the on state at time t; u n,t / v n,t ∈{0,1} respectively represent the start / shutdown operation of thermal power unit n at time t, u n,t =1 means that the thermal power unit n starts at time t, u n,t =0 means that the thermal power unit n has not started at time t, v n,t =1 means that the thermal power unit n is shut down at time t, v n,t =0 means that the thermal power unit n did not shut down at time t. Indicates the upper limit of the ramp rate of thermal power unit n, Δ p n Indicates the lower limit of the ramp rate of thermal power unit n, I n,t-1 represents the start / stop status of thermal power unit n at time t-1, I n,τ represents the start and stop status of thermal power unit n at time τ, T is the last time period, Indicates the minimum duration of continuous startup of thermal power units. They respectively represent the minimum duration of continuous shutdown of thermal power units.

5. The SDM-CADMM distributed method for unit combination P2P energy trading according to claim 4, characterized in that: In step S1.3, the power load range is constrained as follows: Among them, formula (14) is the upper and lower limit constraints of user n's power load at time t, and formula (15) represents the total energy demand of user n at all times; Among them, D n,t represents the electricity load of user n at time t, They represent the upper limit of the power load of user n at time t, D n,t They represent the lower limit of the power load of user n at time t, It is the sum of the electricity load of user n at all times.

6. The SDM-CADMM distributed method for unit combination P2P energy trading according to claim 5, characterized in that: In step S1.4, the P2P energy trading model is: st(1)-(15)(17) Among them, formula (16) is the objective function to achieve cost minimization, formula (17) includes various constraints on the user side and the power generation side, and formula (18) is the expected social welfare of all producers and consumers. represents the minimum value of social welfare expected by all producers and consumers. Formula (19) is the cost-utility function of producer n, that is, C n (P n,t ) represents the cost-utility function of producer n, and formula (20) is the utility function of consumer n, that is, U n (D n,t ) represents the utility function of consumer n; is a positive constant of the cost-utility function, is a positive constant of the utility function, c n is the startup cost coefficient of thermal power unit n; Convert the P2P energy trading model into a matrix form and we get: from=(from n ) n∈Ω ,z∈Z (24) Among them, formula (21) is the matrix form of the cost minimization objective function, and formula (16) is f n (x n ) in its specific form; X n is x n The feasible set of x n is a feasible solution for the nth prosumer, forming a non-convex set formed by the mixed integer nonlinear constraints (17), x n represents the set of all decision variables, I n represents the start and stop state vector of thermal power unit n, u n represents the startup operation vector of thermal power unit n, v n represents the shutdown operation vector of thermal power unit n, represents the output vector of thermal power unit n in scenario s, E n represents the net injected power vector of each prosumer n; Q n represents the constant matrix of the nth producer and consumer, Z represents the coupling relationship between producers and consumers, z represents the introduced auxiliary variable, Z={z=(z n ) n＝1,2,..Ω. }, z n Represents the auxiliary variable introduced by the nth prosumer.

7. The SDM-CADMM distributed method for unit combination P2P energy trading according to claim 1, characterized in that: In step S2, the specific implementation process of reshaping is as follows: S2.1.: Convert the non-convex set in the P2P energy trading model into a convex set and obtain: from=(from n ) n∈Ω ,z∈Z (28) Among them, conv(X n ) represents a non-convex set X n The vertices of the convex hull of ; S2.2: The Lagrangian dual function φ(λ) of the original problem is: λ∈z T :={v:v T z=0,z∈Z} (30) Wherein, formula (29) is the convex relaxed dual function of the original problem, that is, φ(λ) represents the convex relaxed dual function of the original problem, and λ is the set of dual multiplier vectors; Represents the equality constraint Q corresponding to the nth producer and consumer n x n =z n The transpose of the dual multiplier vector, Q n The constant matrix representing the number of prosumers n, z n Auxiliary variable representing prosumer n, z T represents the transpose of the auxiliary variable set, n represents the nth producer and consumer, v represents the set of dual multiplier vectors, and v T represents the transpose of v; S2.3: The augmented Lagrangian function of the original problem is: Wherein, formula (33) is the augmented Lagrangian function form of the original problem, L(x,z,λ) represents the augmented Lagrangian function of the original problem, ρ represents the penalty parameter, z n→m represents the transaction volume between prosumer n and its neighboring prosumer m, z m→n represents the transaction volume between prosumer m and its neighboring prosumer n, represents the average transaction volume between prosumer n and prosumer m; S2.4: The augmented Lagrangian function of the original problem is approximated as: Among them, formula (35) is the mirror function of all augmented Lagrangian functions, that is represents the mirror function of all augmented Lagrangian functions, represents the mirror function of the augmented Lagrangian function of prosumer n at the kth iteration. Formula (36) is the objective function f n The feasible solution for the nth prosumer in the kth iteration is The first-order Taylor expansion at , formula (37) is the tangent plane function of all augmented Lagrangian functions, that is, represents the tangent plane function of all augmented Lagrangian functions, represents the tangent plane function of the augmented Lagrangian function of producer / consumer n at the kth iteration. Formula (38) is The tangent plane function at the center is represents the tangent plane function, and the approximate function satisfies Denotes the objective function f n At the point The first-order derivative at , represents the feasible solution in the kth iteration, represents the auxiliary variable introduced by prosumer n at the kth iteration.

8. A storage device, characterized in that: The storage device stores instructions and data for implementing the SDM-CADMM distributed method for unit combination P2P energy trading as described in any one of claims 1 to 7.

9. An SDM-CADMM distributed device for unit combination P2P energy trading, characterized by: include: Processor and storage device; the processor loads and executes instructions and data in the storage device to implement the SDM-CADMM distributed method for unit combination P2P energy trading as described in any one of claims 1 to 7.

Images