Active power distribution network planning method based on light quantum acceleration

Abstract

The invention belongs to the technical field of power distribution network planning, and relates to a light quantum acceleration-based active power distribution network planning method, which comprises the following steps of: 1, constructing a light quantum acceleration-based active power distribution network planning framework according to a planning target and an actual condition; 2, constructing a distribution network planning model based on light quantum acceleration; 3, abstraction processing is carried out on the active power distribution network planning model to form a compact model, and a light quantum embedded ADMM algorithm PQA-ADMM is constructed for the compact model; according to the invention, through decoupling and modeling of a discrete main problem and a continuous sub-problem of active power distribution network planning, calculation tasks performed by a power distribution network operator in a classical environment and a quantum environment are determined; under the architecture of an ADMM algorithm, a solver such as GUROBI or CPLEX is used for solving linear sub-problems in a classical environment, light quantum calculation is used for solving a discrete main problem in a quantum environment, dimensionality reduction of a mixed integer programming model is achieved, the mixed integer programming model is decomposed into a plurality of sub-problems, and finally accelerated calculation is achieved.

Description

Technical Field

[0001] The invention belongs to the technical field of distribution network planning, and relates to an active distribution network planning method based on photon acceleration. Background Art

[0002] Active distribution network (ADN) planning is essential to achieve an economically efficient transition to a highly reliable and modern power system. The large-scale integration of renewable energy and distributed devices brings great uncertainty and operational risks to ADN planning. In active distribution network planning, the optimal installation location, capacity, and operation strategy of each device are mainly determined based on the economic feasibility of distribution network planning and year-round operation. The introduction of a large number of binary and integer variables, such as investment and state variables, significantly increases the combinatorial difficulty in distribution network planning, which often results in unacceptably long solver solution times. Therefore, it is urgent to effectively layout various types of distributed resources and integrate efficient solution algorithms to promote the high-speed solution of distribution network planning models.

[0003] ADN planning is regarded as an NP-hard problem, and its solution time grows exponentially with the growth of integer variables and binary variables. Existing technologies usually study accelerated planning algorithms, such as convex difference planning algorithms, successive convex relaxation planning algorithms, etc., or use intelligent optimization algorithms with local optimization characteristics, such as particle swarm optimization algorithms, ant colony optimization algorithms and other solution algorithms to meet the rapid solution of distribution network planning.

[0004] However, existing active distribution network planning algorithms have the disadvantages of local optimization or iteration failure.

[0005] Therefore, a distribution network planning method is needed to solve the above technical problems by avoiding the exponential increase of solution time and realizing accelerated calculation. Summary of the invention

[0006] The technical solution adopted by the present invention to solve the technical problem is: an active distribution network planning method based on photon acceleration, comprising the following steps:

[0007] Step 1: According to the planning objectives and actual conditions, an active distribution network planning framework based on photonic quantum acceleration is constructed to clarify the tasks in the classical computing environment and photonic quantum computing environment;

[0008] Step 2: Construct a distribution network planning model based on photon acceleration. The distribution network planning model includes objective function, distribution network installation constraints, distribution network operation constraints, and second-order cone relaxation constraints.

[0009] Step 3: Abstract the active distribution network planning model to form a compact model, and construct the photonic quantum embedded ADMM algorithm PQA-ADMM for the compact model.

[0010] Preferably, in step 1, constructing an active distribution network planning framework specifically includes the following sub-steps:

[0011] Step 1-1: Collect and convey the historical output of new energy and various loads to the distribution network operator through measuring devices and information feedback. After mastering the historical data and making predictions, the distribution network operator comprehensively considers the planning cost and operating cost and issues control instructions on whether to install distributed resources. Under the condition of satisfying the topological constraints and operating constraints, coordinate the various resources in the distribution network and formulate a specific optimal operation strategy for the distribution network to maximize the satisfaction of the planning objectives and realize the economical and efficient operation of the distribution network.

[0012] Step 1-2: Use the standardized alternating direction multiplier method ADMM algorithm to decompose the distribution network planning into a main problem and sub-problems. The main problem with discrete variables is optimized and solved in a quantum computing environment; the convex optimization sub-problem with continuous variables is optimized and solved in a classical environment; the optimal solution that meets the convergence conditions is obtained through iterative calculation between sub-problems;

[0013] Steps 1-3: For optimization problems in classical environments, directly call the commercial solver Gurobi / Cplex to obtain the optimal solution; for optimization problems in quantum environments, reconstruct the main problem into a problem that can be deployed and embedded on a quantum computer; in a quantum computing environment, obtain the solution set of the final optimization problem.

[0014] Preferably, in step 2, the objective function includes:

[0015] min C Total =C IN +C OP (1)

[0016]

[0017] C OM =λC IN (4)

[0018]

[0019] In formula (1) to formula (9): C Tota l represents the annual comprehensive cost of ADN planning, C IN represents the construction cost, C OP represents the operating cost, C OM represents the operation and maintenance cost, C PU Indicates the cost of purchasing electricity, C CU It represents the penalty fee for abandoning wind and solar power. represents the carbon penalty fee, C DSM represents the demand-side management fee, CLOSS Indicates network loss cost, Indicates the present value annual value conversion coefficient of the planning object a, Ω U and Ω T Denote the typical scenario day and the set of operation time periods, respectively. u represents the probability of the u-th scene appearing, Ψ S , N and DSM are the substation node set, load node set and node set participating in demand side management, Ψ PV , WT , SV , MT and ESS They are the set of wind power nodes to be built, the set of photovoltaic nodes to be built, the set of reactive power compensation equipment nodes to be built, the set of gas turbine nodes to be built, and the set of energy storage nodes to be built, c PV 、c WT 、c SV and c MT Respectively represent the unit investment and construction costs of each type of equipment, c ESS,P and c ESS,E They represent the unit power and unit capacity investment and construction costs of energy storage respectively, λ represents the conversion coefficient between construction cost and operation and maintenance cost, and denote the 0-1 construction decision variables of the i-th wind power, i-th photovoltaic, i-th reactive power compensation and i-th gas turbine, respectively. and are continuous decision variables representing the installed power and capacity of the i-th energy storage, respectively. represents the active power transmitted by the substation to the upper grid in the tth period of the uth scenario, They represent the active power output of photovoltaic, wind power and gas turbine in the tth period of the uth scenario, respectively, DSM is the unit compensation cost of demand-side response, represents the interruption amount of the load of the ith demand-side management node, represents the unit price of electricity purchased by the i-th substation, and represents the penalty cost for wind and solar curtailment at node i at time t in scenario u, and They represent the abandoned wind power and abandoned solar power at node i at time t in scenario u, represents the network loss power in the uth scenario and the tth period, C carbon represents carbon emissions, γ PV , γ WT , γ MT , γ SThey respectively represent the carbon emissions per unit of electricity of photovoltaic power, wind power, gas turbines and thermal power of the upper main grid.

[0020] Preferably, in step 3, the compact model includes:

[0021]

[0022] subject to:Gm=b,g(n)≤0

[0023] h(m,n)≤0 (35)

[0024] In formula (34)-formula (35), p(m) represents discrete installation cost; q(n) represents continuous installation cost and operation cost, m∈M and n∈N represent discrete variable sets and continuous variable sets respectively, Gm=b represents a set of inequalities that only contain discrete variable equality constraints or are constrained to equality through slack variables in the model, g(n)≤0 represents a set of equality or inequality constraints that only contain continuous variables in the model, and h(m,n)≤0 represents a set of equality or inequality constraints that contain both continuous and discrete variables in the model.

[0025] Preferably, in step 3, the process of constructing a photon-embedded ADMM algorithm for a compact model includes the following sub-steps:

[0026] Step 3-1: Initialize continuous and discrete variables and the maximum number of iterations k max , initial value of penalty function and dual variable, set the number of iterations k = 1, and start the PQA-ADMM algorithm;

[0027] Step 3-2: In the kth loop, solve the continuous subproblem and update the continuous variable solution set.

[0028]

[0029] Step 3-3: In the kth loop, solve the discrete QUBO subproblem as follows:

[0030]

[0031] Converting the constrained distribution network planning model into a QUBO model requires embedding the constraints into the objective function in the form of multiple penalty terms, thereby achieving the construction of the optimization problem;

[0032] Step 3-4: Calculate the degree of convergence of the problem; calculate the original residual and dual residual at the kth cycle:

[0033]

[0034] The original residual reflects the convergence consistency of discrete and continuous problems, and the dual residual reflects the degree of convergence after two consecutive iterations; the convergence criteria are as follows:

[0035]

[0036] If the convergence condition is met, the iteration ends and the data is output. If not, continue to execute steps 3-5;

[0037] Step 3-5: Update the dual variables as follows:

[0038]

[0039] Step 3-6: According to the following formula:

[0040]

[0041] Update the penalty function to ensure that the original residual and the dual residual converge synchronously, and then perform the next round of iteration.

[0042] Preferably, in the sub-step 3-3, the process of converting each sub-discrete problem into a QUBO model includes:

[0043]

[0044]

[0045] H M =p(m) (56)

[0046] The optical quantum computing system is used to find the minimum Hamiltonian. After obtaining the optimal solution to the problem, the optimization result of the discrete variables is obtained from the mapping relationship.

[0047] H obj =H PV +H WT +H SV +H MT +H PER +H C +H ρ +H M (57).

[0048] The beneficial effects of the present invention are:

[0049] 1. This invention clarifies the computational tasks performed by distribution network operators in classical and quantum environments by decoupling and modeling the discrete main problem and continuous subproblems of active distribution network planning. Under the framework of the ADMM algorithm, linear subproblems are solved using solvers such as GUROBI or CPLEX in a classical environment, and discrete main problems are solved using optical quantum computing in a quantum environment, thereby achieving dimensionality reduction of the mixed integer programming model, decomposing it into multiple subproblems, and ultimately achieving accelerated computing.

[0050] 2. The present invention uses a decomposition algorithm to decompose and calculate the main problem into sub-problems. In the future when discrete variables increase sharply, the exponential increase in solution time can be avoided. If quantum computing resources are sufficient, the acceleration effect will be more obvious. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 This is a photon acceleration framework diagram of an active distribution network planning method based on photon acceleration of the present invention;

[0052] Figure 2 It is a schematic diagram of the steps of the present invention. DETAILED DESCRIPTION

[0053] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the relevant technologies in the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0054] refer to Figure 1-2 As shown, in this embodiment, the active distribution network planning method based on photon acceleration includes the following steps:

[0055] Step 1: According to the planning objectives and actual conditions, construct an active distribution network planning framework based on photon acceleration, and clarify the tasks in the classical computing environment and photon computing environment. (This framework includes the decomposition calculation principle of the hybrid photon-classical algorithm)

[0056] The framework diagram is as follows Figure 1 As shown in the figure, the historical output of new energy and various loads is collected and communicated to the distribution network operator through measuring devices, information feedback, etc. After mastering the historical data and making predictions, the distribution network operator comprehensively considers the planning cost and operating cost factors, and issues control instructions on whether to install distributed resources such as distributed photovoltaic (PV), wind power (WG), energy storage (ESS) and gas turbine (MT). Under the condition of meeting the topological constraints and operating constraints, various resources in the distribution network are coordinated to formulate specific optimal operation strategies for the distribution network, so as to meet the planning goals to the greatest extent and realize the economic and efficient operation of the distribution network.

[0057] Considering the application of quantum advantage to distribution network planning, the standardized alternating direction method of multipliers (ADMM) algorithm is used to decompose the distribution network planning into a main problem and sub-problems. The main problem with discrete variables is optimized and solved in a quantum computing environment, involving distribution network installation constraints; the convex optimization sub-problem with continuous variables is directly optimized and solved in a classical environment, involving distribution network operation constraints and second-order cone relaxation constraints. The optimal solution that meets the convergence conditions is obtained through iterative calculation between sub-problems.

[0058] For optimization problems in classical environments, mature commercial solvers Gurobi / Cplex can be directly called to obtain the optimal solution; for optimization problems in quantum environments, the main problem needs to be reconstructed into a problem that can be deployed and embedded in a quantum computer, such as the quadratic unconstrained optimization problem (QUBO). In the quantum computing environment, users interact with the coherent optical quantum computing machine through a Web user interface (UI). In this process, the optical quantum computer completes measurement, modulation, and interference operations through optical and electrical modules, guiding the Hamiltonian of the photon to evolve to the minimum. Its final phase state, namely the phase 0 state and the π state, corresponds to the {-1, +1} spin state of the photon quantum bit, respectively, and then establishes a connection with the {0, 1} decision variable. This mechanism can be used to construct a mapping relationship between binary decision variables and photon quantum bits, establish a close connection between the Ising model and the QUBO model, and obtain the solution set of the final optimization problem.

[0059] Step 2: Construct a distribution network planning model based on photon acceleration, which includes objective function, distribution network installation constraints, distribution network operation constraints, and second-order cone relaxation constraints. Among them, the annual comprehensive cost is considered as the objective function, including installation costs, operation and maintenance costs, wind and solar abandonment penalty costs, carbon penalty costs, demand response costs, and electricity purchase costs from the superior power grid. The installation constraints of each component include the installation of PV, WG, MT, and ESS. The distribution network operation constraints include PV, WG, ESS, DSM, MT, voltage and current, and Distflow operation constraints (the installation part fits the main problem, and the operation part fits the sub-problem, which is convenient for subsequent solution)

[0060] (1) Objective function

[0061] In order to characterize the economic feasibility of distribution network planning, the planning goal should be determined as the optimal economic cost, which mainly includes construction cost and operation cost. The operation cost should include operation and maintenance cost, wind and solar power abandonment penalty cost, demand response cost, power purchase cost from the upper grid and carbon penalty cost, as shown in formula (1-9):

[0062] min C Total =C IN +C OP (1)

[0063]

[0064] C OM =λC IN (4)

[0065]

[0066] Where: C Total The annual comprehensive cost of ADN planning, which specifically includes the construction cost C IN and operating costs C OP , operating costs include operation and maintenance costs C OM , Electricity purchase cost C PU 、Construction cost C IN , Penalty fees for wind and solar power abandonment C CU Carbon penalty fees Demand side management costs C DSM and network loss cost C LOSS , represents the present value to annual value conversion coefficient of the object a to be planned; r represents the depreciation rate; Represents the economic service life cycle of the object a to be planned; Ω U and Ω T Represent the typical scenario day and operation time set respectively; D u represents the probability of the u-th scene appearing; Ψ S ,Ψ N and DSM They are the substation node set, the load node set and the node set participating in demand side management; PV ,Ψ WT ,Ψ SV ,Ψ MT and ESS They are wind power node set to be built, photovoltaic node set to be built, reactive power compensation equipment node set to be built, gas turbine node set to be built and energy storage node set to be built; c PV 、c WT 、c SV and c MT and are the unit investment and construction costs of each type of equipment respectively; c ESS,P and c ESS,E are the unit power and unit capacity investment and construction costs of energy storage, respectively; λ is the conversion coefficient between construction cost and operation and maintenance cost. and are the 0-1 construction decision variables for the i-th wind power, i-th photovoltaic, i-th reactive power compensation and i-th gas turbine respectively; and They represent the continuous decision variables of the installed power and capacity of the i-th energy storage respectively; is the active power delivered by the substation to the upper grid in the tth period of the uth scenario; are the active power output values ​​of photovoltaic, wind power and gas turbine in the tth period of the uth scenario respectively; c DSM Unit compensation costs for demand-side response; is the interruption amount of the load of the ith demand side management node; is the unit price of electricity purchased by the ith substation, and the actual time-of-use electricity price is taken; and is the penalty fee for wind and solar power abandonment at node i at time t in scenario u; and are the abandoned wind power and abandoned solar power at node i at time t in scenario u, respectively. loss is the unit grid loss electricity price, is the network loss power in the uth scenario and the tth period, C carbon Carbon emissions: γ PV , γ WT , γ MT , γ S They are the carbon emissions per unit of electricity from photovoltaic power, wind power, gas turbines and thermal power from the upper-level main grid.

[0067] (2) DG installation constraints

[0068]

[0069] Where: and are the rated capacity of each wind power and photovoltaic unit respectively; is the maximum active output of the load at node i; η is the maximum allowable penetration rate of distributed generation in the distribution network; and are the upper limits of wind power and photovoltaic installation quantities at node i respectively.

[0070] (3) ESS installation constraints

[0071]

[0072] Where: and is the upper limit of energy storage power and capacity installed at node i; ESS is the energy rate coefficient of the energy storage system; and Install caps on energy storage power and capacity in distribution networks.

[0073] (4) SV and MT installation constraints

[0074]

[0075] Where: and are the upper limits of the number of reactive compensation equipment and gas turbines installed at node i respectively.

[0076] (5) Substation power operation constraints

[0077]

[0078] Where: is the rated capacity of the transformer in the substation at node i.

[0079] (6) Voltage and current constraints

[0080]

[0081] Where: v i,u,t is the square value of the voltage amplitude at node i at time t in scenario u; U min,i and U max,i are the upper and lower limits of the voltage amplitude at node i; l ij,u,t is the square of the current amplitude at node i at time t in scenario u; I max,ij is the safety current corresponding to branch ij.

[0082] (7) Distflow Constraints

[0083]

[0084] Where: P ij,u,t and Q ij,u,t are the active and reactive powers flowing through branch ij at time t in scenario u; R ij and X ij are the resistance and reactance values ​​corresponding to branch ij; and are the active and reactive powers flowing into node i at time t in scenario u, respectively; are the active power injected by the energy storage at node i at time t in scenario u; They are the reactive power injected by substation, photovoltaic, wind power, gas turbine, energy storage, reactive compensation, and demand-side management at node i at time t in scenario u, and For active load and reactive load.

[0085] (8) DG operation constraints

[0086]

[0087] Where: DG includes PV and WT, and are the upper and lower limits of the active output of distributed generation at node i respectively; and They respectively represent the upper and lower limits of the power factor angle of the distributed power source; is the maximum power removal ratio of distributed generation at node i; It is the maximum available power of distributed power source.

[0088] (9) ESS operation constraints

[0089]

[0090]

[0091] Where: is the initial charge factor of the energy storage installed at node i; and are the lower and upper limit coefficients of the charge state of the energy storage at node i respectively.

[0092] (10)SV operation constraints

[0093]

[0094] Where: Installation capacity for a single SV.

[0095] (11)MT operation constraints

[0096]

[0097] Where: and They are the upper and lower limits of active output of MT at node i respectively; and They represent the upper and lower limits of the power factor angle of MT respectively; μ is the maximum climbing rate of MT.

[0098] (12) Demand response constraints

[0099]

[0100] Where: The maximum call power of demand-side management at node i at time t in scenario u; is the load fixed power factor, λ DSM is the load reduction factor.

[0101] (13) Second-order cone relaxation constraint

[0102] ||[2P ij，u，t 2Q ij，u,t l ij，u，t -v i，u，t ]T ||2≤l ij，u，t +v i，u，t (33)

[0103] In fact, this model also supports the consideration of more factors, such as electric vehicles, mobile energy storage or intelligent soft switches, and even transmission network planning, just by replacing the corresponding constraints of this model.

[0104] Step 3, abstract the active distribution network planning model to form a compact model, and construct a photon quantum embedded ADMM algorithm (PQA-ADMM) for this compact model. The compact model can be divided into two parts, discrete and continuous problems, according to the discrete variables and continuous variables or constraints of the active distribution network planning model, and written in a format that can be processed by the standardized alternating direction multiplier method; the PQA-ADM algorithm construction process is improved based on the basic steps of the ADMM method and embedded with a quantum algorithm.

[0105] (1) Compact model for distribution network planning

[0106] The active distribution network planning model is abstracted to obtain its compact form:

[0107]

[0108] Where: p(m) corresponds to discrete installation cost; q(n) corresponds to continuous installation cost and operation cost; m∈M and n∈N represent the discrete variable set and continuous variable set respectively; Gm=b is the set of inequalities in the model that only contains discrete variable equality constraints or is constrained to equality through slack variables; g(n)≤0 is the set of equality or inequality constraints in the model that only contains continuous variables; h(m,n)≤0 is the set of equality or inequality constraints in the model that contains both continuous and discrete variables.

[0109] In order to process the above model into the standard form embedded in the optical quantum computer, namely the QUBO form, the set of equality constraints is added as an augmented term to the total cost function, and a new variable u∈R is introduced x , the entire active distribution network planning problem is relaxed into a soft constraint form:

[0110]

[0111] The entire planning model is divided into two parts: continuous and discrete, to match different solution formats:

[0112]

[0113] w=x+y (41)

[0114]

[0115] Therefore, the distribution network planning model can be transformed into a standard compact form that can be processed by ADMM as follows:

[0116]

[0117] (2) Constructing the augmented Lagrangian function

[0118] The ADMM algorithm is an operator partitioning algorithm with a long history. Under the premise of the convexity assumption, ADMM has the properties of residual, objective function and dual variable convergence. It is suitable for the distribution network planning model proposed in this paper, and the augmented Lagrangian function is formed as follows:

[0119]

[0120] (3) Photon quantum embedded ADMM algorithm construction process

[0121] Step 1: Initialize continuous and discrete variables and the maximum number of iterations k max , the initial value of the penalty function and the dual variable, set the number of iterations k = 1, and start the PQA-ADMM algorithm.

[0122] Step 2: In the kth loop, solve the continuous subproblem and update the continuous variable solution set.

[0123]

[0124] Step 3: In the kth loop, solve the discrete QUBO subproblem as follows:

[0125]

[0126] Converting the constrained distribution network planning model into a QUBO model requires embedding the constraints into the objective function in the form of multiple penalty terms to achieve the construction of the optimization problem. The following describes the process of converting each sub-discrete problem into a QUBO model:

[0127]

[0128]

[0129] H M =p(m) (56)

[0130] Construct the following formula and use the optical quantum computing system to find its minimum Hamiltonian. After obtaining the optimal solution to the problem, the optimization result of the discrete variables can be obtained from the mapping relationship.

[0131] H obj =H PV +H WT +H SV +HMT +H PER +H C +H ρ +H M (57)

[0132] Step 4: Calculate the degree of convergence of the problem. Calculate the original residual and dual residual at the kth cycle according to the following formula:

[0133]

[0134] The original residual reflects the convergence consistency of discrete and continuous problems, and the dual residual reflects the degree of convergence after two consecutive iterations. The convergence criteria are as follows:

[0135]

[0136] If the convergence condition of formula (59) is met, the iteration ends and the data is output. If not, continue to step 5.

[0137] Step 5: Update the dual variables as follows:

[0138]

[0139] Step 6: According to the following formula:

[0140]

[0141] Update the penalty function to ensure that the original residual and the dual residual converge synchronously, and then perform the next round of iteration.

[0142] In summary, the present invention clarifies the computing tasks performed by distribution network operators in classical and quantum environments by decoupling and modeling the discrete main problem and continuous subproblems of active distribution network planning. Under the framework of the ADMM algorithm, linear subproblems are solved using solvers such as GUROBI or CPLEX in a classical environment, and discrete main problems are solved using optical quantum computing in a quantum environment, thereby achieving dimensionality reduction of the mixed integer programming model, decomposing it into multiple subproblems, and ultimately achieving accelerated computing.

[0143] It should be emphasized that the above are only preferred embodiments of the present invention and do not impose any form of limitation on the present invention. Any simple modification of the above embodiments based on the technical essence of the present invention also falls within the protection scope of the present invention. Other equivalent changes and modifications still fall within the scope of the technical solution of the present invention.

Claims

1. An active distribution network planning method based on photon acceleration, characterized in that: The following steps are involved: Step 1: According to the planning objectives and actual conditions, an active distribution network planning framework based on photonic quantum acceleration is constructed to clarify the tasks in the classical computing environment and photonic quantum computing environment; Step 2: Construct a distribution network planning model based on photon acceleration, wherein the distribution network planning model includes an objective function, distribution network installation constraints, distribution network operation constraints, and second-order cone relaxation constraints; Step 3: Abstract the active distribution network planning model to form a compact model, and construct the photonic quantum embedded ADMM algorithm PQA-ADMM for the compact model.

2. According to claim 1, the active distribution network planning method based on photon acceleration is characterized in that: In step 1, constructing an active distribution network planning framework specifically includes the following sub-steps: Step 1-1: Collect and convey the historical output of new energy and various loads to the distribution network operator through measuring devices and information feedback. After mastering the historical data and making predictions, the distribution network operator comprehensively considers the planning cost and operating cost and issues control instructions on whether to install distributed resources. Under the condition of satisfying the topological constraints and operating constraints, coordinate the various resources in the distribution network and formulate a specific optimal operation strategy for the distribution network to maximize the satisfaction of the planning objectives and realize the economical and efficient operation of the distribution network. Step 1-2: Use the standardized alternating direction multiplier method ADMM algorithm to decompose the distribution network planning into a main problem and sub-problems. The main problem with discrete variables is optimized and solved in a quantum computing environment; the convex optimization sub-problem with continuous variables is optimized and solved in a classical environment; the optimal solution that meets the convergence conditions is obtained through iterative calculation between sub-problems; Steps 1-3: For optimization problems in classical environments, directly call the commercial solver Gurobi / Cplex to obtain the optimal solution; for optimization problems in quantum environments, reconstruct the main problem into a problem that can be deployed and embedded on a quantum computer; in a quantum computing environment, obtain the solution set of the final optimization problem.

3. According to claim 1, the active distribution network planning method based on photon acceleration is characterized in that: In step 2, the objective function includes: my C Total =C IN +C OP (1) C OM =λC IN (4) In formula (1) to formula (9): C Total represents the annual comprehensive cost of ADN planning, C IN represents the construction cost, C OP represents the operating cost, C OM represents the operation and maintenance cost, C PU Indicates the cost of purchasing electricity, C CU It represents the penalty fee for abandoning wind and solar power. represents the carbon penalty fee, C DSM represents the demand-side management fee, C LOSS Indicates network loss cost, Indicates the present value annual value conversion coefficient of the planned object a, Ω U and Ω T Denote the typical scenario day and the set of operation time periods, respectively. u represents the probability of the u-th scene appearing, Ψ S ,Ψ N and DSM are the substation node set, load node set and node set participating in demand side management, Ψ PV ,Ψ WT ,Ψ SV ,Ψ MT and ESS They are the set of wind power nodes to be built, the set of photovoltaic nodes to be built, the set of reactive power compensation equipment nodes to be built, the set of gas turbine nodes to be built, and the set of energy storage nodes to be built, c PV 、c WT 、c SV and c MT Respectively represent the unit investment and construction costs of each type of equipment, c ESS,P and c ESS,E They represent the unit power and unit capacity investment and construction costs of energy storage respectively, λ represents the conversion coefficient between construction cost and operation and maintenance cost, and denote the 0-1 construction decision variables of the i-th wind power, i-th photovoltaic, i-th reactive power compensation and i-th gas turbine, respectively. and They represent the continuous decision variables of the installed power and capacity of the i-th energy storage, represents the active power delivered by the substation to the upper grid in the tth period of the uth scenario, They represent the active power output of photovoltaic, wind power and gas turbine in the tth period of the uth scenario, respectively, DSM is the unit compensation cost of demand-side response, represents the interruption amount of the load of the ith demand-side management node, represents the unit price of electricity purchased by the i-th substation, and represents the penalty cost for wind and solar curtailment at node i at time t in scenario u, and They represent the abandoned wind power and abandoned solar power at node i at time t in scenario u, represents the network loss power in the uth scenario and the tth period, C carbon represents carbon emissions, γ PV , γ WT , γ MT , γ S They respectively represent the carbon emissions per unit of electricity of photovoltaic power, wind power, gas turbines and thermal power of the upper main grid.

4. The active distribution network planning method based on photon acceleration according to claim 1 is characterized in that: In step 3, the compact model includes: In formula (34)-formula (35), p(m) represents discrete installation cost; q(n) represents continuous installation cost and operation cost, m∈M and n∈N represent discrete variable sets and continuous variable sets respectively, Gm=b represents a set of inequalities that only contain discrete variable equality constraints or are constrained to equality through slack variables in the model, g(n)≤0 represents a set of equality or inequality constraints that only contain continuous variables in the model, and h(m,n)≤0 represents a set of equality or inequality constraints that contains both continuous and discrete variables in the model.

5. The active distribution network planning method based on photon acceleration according to claim 4 is characterized in that: In step 3, the process of constructing a photon-embedded ADMM algorithm for a compact model includes the following sub-steps: Step 3-1: Initialize continuous and discrete variables and the maximum number of iterations k max , initial value of penalty function and dual variable, set the number of iterations k = 1, and start the PQA-ADMM algorithm; Step 3-2: In the kth loop, solve the continuous subproblem and update the continuous variable solution set; Step 3-3: In the kth loop, solve the discrete QUBO subproblem as follows: Converting the constrained distribution network planning model into a QUBO model requires embedding the constraints into the objective function in the form of multiple penalty terms, thereby achieving the construction of the optimization problem; Step 3-4: Calculate the degree of convergence of the problem; calculate the original residual and dual residual at the kth cycle: The original residual reflects the convergence consistency of discrete and continuous problems, and the dual residual reflects the degree of convergence after two consecutive iterations; the convergence criteria are as follows: If the convergence condition is met, the iteration ends and the data is output. If not, continue to execute steps 3-5; Step 3-5: Update the dual variables as follows: Step 3-6: According to the following formula: Update the penalty function to ensure that the original residual and the dual residual converge synchronously, and then perform the next round of iteration.

6. The active distribution network planning method based on photon acceleration according to claim 5 is characterized in that: In the sub-step 3-3, the process of converting each sub-discrete problem into a QUBO model includes: H M =p(m) (56) The optical quantum computing system is used to find the minimum Hamiltonian. After obtaining the optimal solution to the problem, the optimization result of the discrete variables is obtained from the mapping relationship. H obj =H PV +H WT +H SV +H MT +H PER +H C +H ρ +H M (57)。

Images