A multi-microgrid system distributed real-time scheduling method and system

Abstract

The application belongs to the technical field of electrical engineering, and discloses a kind of multi-microgrid system distributed real-time scheduling method and system, method includes: according to the technical parameters of each element of multi-microgrid system and Bellman optimality principle, the centralized scheduling model of multi-microgrid system is established;With each upper grid node on main line as split node, the alternating current flow constraint in centralized scheduling model is split using ADMM algorithm, and the distributed scheduling model of multi-microgrid system is established;Combined with energy transaction mechanism, the optimization problem corresponding to the scheduling model of each microgrid is further split into upper grid interaction subproblem and microgrid internal optimization subproblem, and the distributed scheduling model with enhanced privacy protection is obtained;ADP random optimization algorithm is used to solve, and the optimal real-time scheduling decision of the whole scheduling domain is obtained.Through the above method, distributed scheduling solution can be realized, data privacy can be improved, and calculation cost can be reduced.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field related to electrical engineering, and more particularly, relates to a multi-micro-grid system decentralized real-time scheduling method and system. BACKGROUND

[0002] With the development of distributed generation technology and demand side response, micro-grid has become an important form of energy efficient utilization. In practical application, multiple micro-grids with similar geographical locations can form a multi-micro-grid system to further improve the safety and economy of system operation. However, the operation of multi-micro-grid is influenced by external physical information attacks and internal operation randomness (new energy generation, load, electricity price, etc.). Therefore, it is of great significance to study a multi-micro-grid system decentralized real-time scheduling method to improve the privacy of multi-micro-grid system operation and cope with randomness, and ensure the safe and economic operation of multi-micro-grid system.

[0003] The approximate dynamic programming (ADP) method is widely used in real-time scheduling of multi-micro-grid as a stochastic optimization method. However, the current ADP method is a centralized scheduling method, which faces the problems of high communication and calculation cost, poor system privacy and reliability when applied to multi-micro-grid decentralized scheduling. SUMMARY

[0004] In view of the above defects or improvement needs of the prior art, the present application provides a multi-micro-grid system decentralized real-time scheduling method and system, which aims to realize decentralized scheduling planning of multi-micro-grid to improve privacy protection and reduce calculation cost.

[0005] To achieve the above-mentioned purpose, according to one aspect of the present application, a multi-micro-grid system decentralized real-time scheduling method is provided, which comprises:

[0006] S1, establishing a centralized scheduling model of the multi-micro-grid system according to the technical parameters of each element of the multi-micro-grid system and Bellman optimality principle;

[0007] S2, taking each upper grid node on the main line as a splitting node, using ADMM algorithm to split the alternating current flow constraint in the centralized scheduling model, establishing a decentralized scheduling model of the multi-micro-grid system, different micro-grids corresponding to independent scheduling models, the upper grid node including external grid node, micro-grid node and intermediate connection node;

[0008] S3, combining the energy trading mechanism, further splitting the optimization problem corresponding to the scheduling model of each micro-grid into upper grid interaction sub-problem and micro-grid internal optimization sub-problem, obtaining a decentralized scheduling model with enhanced privacy protection;

[0009] S4, solving the enhanced privacy protection distributed scheduling model by using an ADP random optimization algorithm to obtain an optimal real-time scheduling decision of the whole scheduling domain.

[0010] Optionally, in S1, the objective function of the centralized scheduling model of the multi-microgrid system is:

[0011] wherein J represents the objective function, is the operation cost of the mth microgrid at time t, is the post-decision value function and post-decision time state variable of the microgrid m at time t, p t is the electricity price at time t, is the active power transaction power of the multi-microgrid and the external grid at time t, m is the serial number of the microgrid, and M is the set of microgrids, is the power generation power of the gas-electricity unit g at time t, a g , b g , c g is the power generation cost coefficient of the gas-electricity unit g, G m is the set of gas-electricity units of the mth microgrid, C w , C s is the penalty coefficient of abandoned wind and light, is the available power generation power of the wind turbine w and the photovoltaic unit s at time t, is the actual power generation power of the wind turbine w and the photovoltaic unit s at time t, W m , S m , B m is the set of wind turbines, photovoltaic units and energy storages of the mth microgrid, is the approximate value function related to the energy storage b and the post-decision state of charge at time t.

[0012] Optionally, in S1, the constraints of the centralized scheduling model of the multi-microgrid system include: AC power flow constraints, node voltage upper and lower limit constraints, gas-electricity unit output upper and lower limit constraints, gas-electricity unit ramping constraints, wind power output upper and lower limit constraints, photovoltaic output upper and lower limit constraints, multi-microgrid system and external grid transaction power upper and lower limit constraints, energy storage operation constraints, and gas-electricity, wind power, photovoltaic, external grid and energy storage reactive power output upper and lower limit constraints.

[0013] Optionally, in S2, when the AC power flow constraints in the centralized scheduling model are split by using the ADMM algorithm, the x update sub-problem, the y update sub-problem, the λ update problem and the convergence judgment condition related to each upper-level grid node in the distributed scheduling model of the multi-microgrid system are as follows:

[0014] x update sub-problem:

[0015]

[0016] x t the decision variable of the y-update subproblem at time x t the decision variable of the y-update subproblem at time t t the Lagrange multiplier, the decision variable of the λ-update subproblem at time t

[0017] the augmented Lagrange function in the ADMM algorithm at time t the feasible region of the x-update subproblem in the ADMM algorithm at time t; the superscript ns is the iteration number of the ADMM algorithm

[0018] the set of neighboring nodes of node i in the upper-level grid the total active and reactive power of node i in the upper-level grid, and ρ is a positive number approaching 0 the set of decision variables of node i in the upper-level grid the set of decision variables of node i in the upper-level grid the replicated variable of node j

[0019] the total generation cost of node i in the upper-level grid, and for the microgrid nodes for the external grid nodes for the intermediate connection nodes

[0020] the square of the voltage and current of node i at time t the active and reactive power flow of line i composed of node i and its neighboring nodes at time t the lower and upper limits of the square of the voltage of node i the set of nodes in the upper-level grid the internal constraint convex set of the subproblem of node i, and for the microgrid nodes contains: the internal AC power flow constraint of the microgrid, the upper and lower limits of the node voltage, the upper and lower limits of the gas turbine output, the gas turbine ramping constraint, the upper and lower limits of the wind power output, the upper and lower limits of the photovoltaic output, the energy storage operation constraint, and the upper and lower limits of the reactive power output of the gas turbine, wind power, photovoltaic, and energy storage, and for the external grid nodes contains: the upper and lower limits of the active and reactive power trading between the multi-microgrid system and the external grid, and for the intermediate connection nodes requires that the active and reactive power of the node be 0

[0021] the y-update subproblem

[0022]

[0023] where, is the feasible region of the y update subproblem in the ADMM algorithm at time t, r i , x i is the resistance and reactance of line i, is the set of upstream grid lines, superscript A i denotes the parent node of node i, C i is the set of child nodes of node i.

[0024] λ update problem:

[0025]

[0026] Convergence criterion:

[0027]

[0028] where, r t , s t are the initial residual and dual residual at time t.

[0029] Optionally, in S3, the solution objective of the microgrid internal optimization subproblem is to obtain the total active and reactive output power of the entire microgrid, and the objective function is a subset of the x update subproblem objective function, only contains and the corresponding constraint contains the microgrid internal constraint in the x update subproblem.

[0030] The solution objective of the microgrid upstream grid interaction subproblem is to optimize the communication of adjacent nodes to obtain power flow information, thereby meeting the alternating current power flow constraint in a decentralized manner, and the objective function is a subset of the x update subproblem objective function, only contains the corresponding constraint contains the alternating current power flow constraint and the upper and lower limits of the node voltage in the x update subproblem.

[0031] Optionally, in S4, the ADP stochastic optimization algorithm uses a piecewise linear function to fit the value function in the Bellman optimality principle, and the proposed piecewise linear function expression is as follows:

[0032]

[0033] where, PW is the total number of segments of the piecewise linear function, is the slope of the bth piecewise linear function of energy storage b at time t in the pwth segment, is the SOC value of energy storage b at time t in the pwth segment, is the maximum and minimum state of charge of energy storage b.

[0034] Optionally, before solving the scheduling model by using the ADP random optimization algorithm in S4, the slope of the piecewise linear function in the ADP random optimization algorithm is trained as follows:

[0035] S4-1, initialize the slope of the piecewise linear function and set the number of training times, let n = 1;

[0036] S4-2, based on the day-ahead forecast values of wind power, photovoltaic power, load and electricity price, a set of training scenarios is generated by using the Monte Carlo method;

[0037] S4-3, let t = Δt;

[0038] S4-4, determine the time state variable of the multi-microgrid system at time t, solve the decentralized scheduling model for strengthening privacy protection, and obtain the decision at time t;

[0039] S4-5, add a disturbance ΔSOC to the SOC of the energy storage, and re-solve the decentralized scheduling model for strengthening privacy protection;

[0040] S4-6, if the output of each element of the multi-microgrid system obtained in S4-5 is consistent with that obtained in S4-4, update the value function slope in the first mode; if the output of each element of the multi-microgrid system obtained in S4-5 is different from that obtained in S4-4, update the value function slope in the second mode; the first mode is to calculate the sampling estimation value of the slope through the marginal influence of the energy storage SOC on the value function of the energy storage, and then update the slope, and the second mode is to calculate the sampling estimation value of the slope through the decentralized calculation of the marginal electricity price of the negative node on the energy storage side in the microgrid, and then update the slope;

[0041] S4-7, calculate the state of the multi-microgrid system at the next time;

[0042] S4-8, let t = t + Δt, repeat steps S4-4 to S4-7 until t = T;

[0043] S4-9, let n = n + 1, repeat steps S4-2 to S4-8 until n = N;

[0044] S4-10, output the trained slope of the piecewise linear function.

[0045] Optionally, in S4, the step of solving the decentralized scheduling model for strengthening privacy protection comprises:

[0046] S2S3-1, initialize the state of the multi-microgrid system;

[0047] S2S3-2, for the x update sub-problem, sequentially solve the microgrid internal optimization problem and the upper-level grid interaction sub-problem, respectively;

[0048] S2S3-3, Solve the y-update subproblem;

[0049] S2S3-4, Solve the λ update problem;

[0050] S2S3-4. Perform convergence determination. If the initial residual and dual residual are not within the allowable error range, return to S2S3-2 to continue the distributed solution; if both the initial residual and dual residual are within the allowable error range, output the distributed scheduling decision for the multi-microgrid.

[0051] Optionally, in S4, obtaining the optimal real-time scheduling decision for the entire scheduling domain includes solving the optimal real-time scheduling decision for each time step in sequence to form the optimal real-time scheduling decision for the entire scheduling domain.

[0052] The decision-making process at time t is as follows:

[0053] Based on the real information of wind power, photovoltaic power, load, and electricity price at time t, determine the real time state S of the multi-microgrid system at time t. t tem ;

[0054] The piecewise linear function is determined based on the slope of the trained piecewise linear function;

[0055] Solve the distributed scheduling model with enhanced privacy protection to obtain the optimal scheduling decision at time t.

[0056] According to another aspect of the present invention, a distributed real-time scheduling system for multiple microgrid systems is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the method described in any of the preceding claims.

[0057] In summary, compared with the prior art, the technical solutions conceived in this invention have the following main advantages:

[0058] This invention provides a distributed real-time scheduling method for multi-microgrid systems. Based on the technical parameters of each component in the multi-microgrid system and the Bellman optimality principle, a scheduling model for the multi-microgrid system is established. Then, based on the ADMM algorithm, AC power flow constraints are decomposed into multiple microgrids, establishing a distributed scheduling model for the multi-microgrid system. Furthermore, by incorporating an energy trading mechanism, the optimization problem for each microgrid is further decomposed, enhancing privacy protection and improving computational efficiency. Finally, the ADP algorithm is used to solve the distributed scheduling model, obtaining an approximately globally optimal real-time scheduling scheme for the multi-microgrid system. Compared to centralized scheduling models, this invention, by constructing and solving a distributed scheduling model, offers higher data privacy protection and lower computational cost.

[0059] Further, the application also proposes a decentralized value function slope training method, which can embed the experience information of the randomness of the entire system only according to the transaction energy information and the internal information of the micro-grid, and strengthen the privacy protection in the ADP offline training, and the trained ADP value function can cope with randomness in real-time scheduling. BRIEF DESCRIPTION OF DRAWINGS

[0060] Figure 1 A flowchart of the decentralized real-time scheduling method of the multi-micro-grid system provided by the embodiment of the application is shown.

[0061] Figure 2 A multi-micro-grid system structure diagram applied to the embodiment of the application is shown.

[0062] Figure 3 The output of wind power, photovoltaic power, load and electricity price in 24 hours in the multi-micro-grid system applied to the embodiment of the application, wherein (a), (b), (c) and (d) correspond to wind power, photovoltaic power, load and electricity price respectively.

[0063] Figure 4 The convergence of the multi-micro-grid decentralized scheduling model based on ADMM provided by the embodiment of the application is shown.

[0064] Figure 5 The difference between the slope of the piecewise linear function obtained by decentralized calculation of the ADP algorithm and the slope obtained by centralized calculation is shown.

[0065] Figure 6 The effect comparison diagram of the fully decentralized ADP algorithm (FD-ADP) provided by the embodiment of the application and the decentralized short-sighted (FD-Myopic) and decentralized model predictive control (FD-MPC) is shown. DETAILED DESCRIPTION

[0066] In order to make the purpose, technical scheme and advantages of the application clearer, the application is further described in detail below in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the application and do not limit the application. In addition, the technical features involved in each embodiment of the application described below can be combined with each other as long as they do not conflict with each other.

[0067] As Figure 1 A flowchart of the decentralized real-time scheduling method of the multi-micro-grid system provided by the embodiment of the application is shown, and the steps thereof are described in detail below.

[0068] S1, according to the technical parameters of each element of the multi-micro-grid system and the Bellman optimality principle, a centralized scheduling model of the multi-micro-grid system is established.

[0069] Specifically, the technical parameters of each element, specifically include:

[0070] 1) The number of micro-grid nodes N bus , the number of lines of the multi-micro-grid N branch , the line first and last node number, the node range contained by each micro-grid;

[0071] 2) The node number where the gas-electricity station is located, the upper and lower limits of the output of the gas-electricity unit and The upper and lower ramp rates of the gas-electricity unit and the cost coefficients a g , b g and c g of the gas-electricity unit;

[0072] 3) The node number where the external power grid is located, the upper limit of the active and reactive power exchange power between the multi-micro-grid system and the external power grid

[0073] 4) The node where the wind turbine, load and photovoltaic unit are located, and the available output of the wind turbine and photovoltaic unit at the current moment and The load P and Q at the current moment;

[0074] 5) The node where the energy storage unit is located, the theoretical maximum charge and discharge power of the energy storage The upper and lower limits of the capacity of the energy storage unit and and the initial electric quantity of the energy storage unit

[0075] As Figure 2 shown is a multi-micro-grid system structure diagram to which the embodiment of the present application is applied, in the embodiment, the upper and lower limits of the output of the gas-electricity unit are 60 kW and 10 kW (micro-grid 1-micro-grid 3) and 80 kW and 20 kW (micro-grid 4), the ramp rates are 40 kW / h (micro-grid 1-micro-grid 3) and 50 kW / h (micro-grid 4), and the cost coefficients are 0.00012 $ / kWh^2, 0.034 $ / kWh, 0.04$ (micro-grid 1-micro-grid 3) and 0.00008 $ / kWh^2, 0.038 $ / kWh, 0.08$ (micro-grid 4); the wind turbine capacities of the micro-grids 1, 3 and 4 are 30 kW, 40 kW and 40 kW; the photovoltaic unit capacities of the micro-grids 2, 3 and 4 are 30 kW, 30 kW and 40 kW; the maximum charge and discharge power of the energy storage unit of the micro-grid 4 is 45 kW, the efficiency is 0.95, and the initial, maximum and minimum electric quantities are 90 kWh, 180 kWh and 0 kWh.

[0076] Specifically, the objective function of the centralized scheduling model of the multi-micro-grid system based on the Bellman equation is:

[0077]

[0078] wherein, J represents the objective function, T is the dispatch domain of the multi-microgrid system, C t is the operation cost of the multi-microgrid system at time t, is the state variable and decision variable of the system at time t, is the operation cost of the mth microgrid at time t, p t is the electricity price at time t, is the active power trading power between the multi-microgrid and the external grid at time t, is the power generation of the gas-electricity unit g at time t, a g , b g , c g is the power generation cost coefficient of the gas-electricity unit g, G m is the set of gas-electricity units of the mth microgrid, C w , C s is the penalty coefficient of wind and light abandonment, is the available power generation of the wind turbine w and the photovoltaic unit s at time t, is the actual power generation of the wind turbine w and the photovoltaic unit s at time t, W m , S m , B m is the set of wind turbines, photovoltaic units and energy storages of the mth microgrid, is the post-decision value function and post-decision time state variable of the microgrid m at time t, is the approximate value function and post-decision state of charge related to the energy storage b at time t.

[0079] Specifically, the constraints of the centralized dispatch model of the multi-microgrid system include AC power flow constraints, node voltage upper and lower limit constraints, gas-electricity unit output upper and lower limit constraints, gas-electricity unit ramping constraints, wind power output upper and lower limit constraints, photovoltaic output upper and lower limit constraints, multi-microgrid system and external grid trading power upper and lower limit constraints, energy storage operation constraints, and gas-electricity, wind, photovoltaic, external grid and energy storage reactive power output upper and lower limit constraints.

[0080] The AC power flow constraints include upper-level grid layer AC power flow constraints and microgrid layer AC power flow constraints:

[0081]

[0082] Node voltage upper and lower limit constraints (upper-level grid layer and microgrid layer):

[0083]

[0084] Microgrid internal gas-electricity unit output upper and lower limit and ramping constraints:

[0085]

[0086] in, Let g be the downhill and uphill gradient rates of the thermal power unit.

[0087] Upper and lower limits of wind and solar power output within the microgrid:

[0088]

[0089] Upper and lower limits of reactive power output of generator units within the microgrid:

[0090]

[0091] in, Let t be the reactive power of generator unit o. Generator unit o includes thermal power unit g, wind power unit w, photovoltaic unit s, and energy storage unit b. These represent the lower and upper limits of the output of generator set o.

[0092] Constraints on energy storage operation within microgrids:

[0093]

[0094] in, Let be the charging and discharging power of energy storage b at time t. The charging and discharging efficiency of energy storage b. This represents the upper limit of the charging and discharging power of energy storage b. This represents the initial state of charge of the stored energy.

[0095] Upper and lower limits of active and reactive power trading between multi-microgrid systems and the external power grid:

[0096]

[0097] S2. Using each upstream grid node on the main line as a splitting node, the ADMM algorithm is used to split the AC power flow constraints in the centralized scheduling model to establish a distributed scheduling model for a multi-microgrid system. Different microgrids correspond to independent scheduling models. The upstream grid nodes include external grid nodes, microgrid nodes, and intermediate connection nodes.

[0098] Specifically, with Figure 2 For example, nodes 0, 1, 2, 3, 4, 5, and 6 are all upstream power grid nodes.

[0099] The centralized scheduling model constructed in S1 is split into an external power grid model, microgrid models, and intermediate connection node models, thereby splitting the centralized scheduling model into a distributed scheduling model.

[0100] Since the alternating current flow constraint is a continuous variable, mainly the splitting of the alternating current flow constraint is realized, and the ADMM algorithm is adopted for splitting, the core of which involves x updating sub-problems, y updating sub-problems, λ updating problems and the construction of convergence judgment conditions.

[0101] x updating sub-problems:

[0102]

[0103]

[0104] x t is the decision variable of the x updating sub-problems at time t, y t is the decision variable of the y updating sub-problems at time t, λ t is the Lagrange multiplier, and is the decision variable of the λ updating sub-problems at time t;

[0105] is the augmented Lagrange function in the ADMM algorithm at time t, is the feasible region of the x updating sub-problems in the ADMM algorithm at time t; the superscript ns is the iteration number of the ADMM algorithm;

[0106] is the neighbor node set of the node i in the upper grid node, is the total active and reactive power of the upper grid node i, and ρ is a positive number approaching 0, is the set of decision variables of the upper grid node i, is is the copy variable of the node j;

[0107] is the total generation cost of the upper grid node i, and for the micro-grid node, for the external grid node, for the intermediate connection node,

[0108] is the square of the voltage and current of the node i at time t, is the active and reactive power flow of the line i composed of the node i and its neighbor nodes at time t, is the lower limit and upper limit of the voltage square of the node i, is the set of nodes in the upper grid; is the internal constraint convex set of the node i sub-problems, and for the micro-grid node, The micro-grid internal alternating current flow constraint, the node voltage upper and lower limit constraint, the gas-electric unit output upper and lower limit constraint, the gas-electric unit climbing constraint, the wind power output upper and lower limit constraint, the photovoltaic output upper and lower limit constraint, the energy storage operation constraint, and the gas-electric, wind power, photovoltaic, and energy storage reactive power output upper and lower limit constraint are included, for the external power grid node, The multi-micro-grid system and external power grid transaction active and reactive power upper and lower limit constraint is included, for the intermediate connection node: The node active and reactive power is required to be 0;

[0109] The y update sub-problem is:

[0110]

[0111] wherein, is the feasible region of the y update sub-problem in the ADMM algorithm at time t, r i , x i is the resistance and reactance of the line i, is the set of upper-level power grid lines, and the superscript A i represents the parent node of node i, C i is the set of child nodes of node i;

[0112] The lambda update problem is:

[0113]

[0114] The convergence determination condition is:

[0115]

[0116] wherein, r t , s t are the initial residual and dual residual at time t.

[0117] Through the ADMM algorithm, the preliminary splitting of the centralized scheduling model can be realized to form a decentralized scheduling model.

[0118] S3, in combination with the energy transaction mechanism, the optimization problem corresponding to the scheduling model of each micro-grid is further split into an upper-level power grid interaction sub-problem and a micro-grid internal optimization sub-problem, to obtain a decentralized scheduling model with enhanced privacy protection.

[0119] Specifically, the solution target of the micro-grid internal optimization sub-problem is to obtain the total active and reactive output power of the entire micro-grid, and the objective function is a subset of the x update sub-problem objective function, only contains and The corresponding constraint condition includes the micro-grid internal constraint in the x update sub-problem.

[0120] The micro-grid internal optimization sub-problem can be specifically expressed as:

[0121]

[0122] The solution target of the micro-grid super-grid interaction sub-problem is to optimize the communication of adjacent nodes to obtain power flow information, so as to satisfy the alternating current power flow constraint in a decentralized manner, and the objective function is a subset of the x update sub-problem objective function, Only The corresponding constraint condition includes the alternating current power flow constraint in the x update sub-problem.

[0123] The micro-grid super-grid interaction sub-problem can be specifically expressed as:

[0124]

[0125] Through this step, each micro-grid model can be further split to further strengthen data privacy protection.

[0126] S4, using ADP random optimization algorithm to solve the decentralized scheduling model with strengthened privacy protection, and obtaining the optimal real-time scheduling decision of the whole scheduling domain.

[0127] Specifically, after the splitting steps of S2 and S3, a decentralized scheduling model with very strong data privacy can be obtained, and the ADP random optimization algorithm is used for solving, so that the optimal real-time scheduling decision of the whole scheduling domain can be obtained.

[0128] In a specific embodiment, the ADP random optimization algorithm uses a piecewise linear function to fit the value function in the Bellman optimality principle, and the expression of the proposed piecewise linear function is as follows:

[0129]

[0130]

[0131] Wherein, PW is the total number of segments of the piecewise linear function, is the slope of the bth energy storage piecewise linear function at the pwth segment at time t, is the SOC value of the bth energy storage at the pwth segment at time t, is the maximum and minimum state of charge of the energy storage b.

[0132] It can be understood that the slope in the piecewise linear function needs to be trained in advance, such as Figure 3 The output of wind power, photovoltaic, load and electricity price in the multi-micro-grid system used for training the slope of the embodiment of the application is shown in FIG. 1, and the slope of the piecewise linear function in the ADP random optimization algorithm is trained as follows in this embodiment:

[0133] S4-1, initialize the slope of piecewise linear function and set the training times, let n = 1;

[0134] S4-2, based on the day-ahead forecast values of wind power, photovoltaic, load and electricity price, a set of training scenarios are generated by using Monte Carlo method;

[0135] S4-3, let t = Δt;

[0136] S4-4, determine the time state variable of the multi-microgrid system at time t, solve the decentralized scheduling model for strengthening privacy protection, and obtain the decision of the multi-microgrid system at time t

[0137] S4-5, add a disturbance ΔSOC to the SOC of the energy storage, and re-solve the decentralized scheduling model for strengthening privacy protection;

[0138] S4-6, if the output of each element of the multi-microgrid system obtained in S4-5 is consistent with that obtained in S4-4, the slope of the value function is updated in the first mode; if the output of each element of the multi-microgrid system obtained in S4-5 is different from that obtained in S4-4, the slope of the value function is updated in the second mode; the first mode is to calculate the sampling estimation value of the slope through the marginal influence of the energy storage SOC on the value function of the energy storage itself, and then update the slope, and the second mode is to calculate the sampling estimation value of the slope through the decentralized calculation of the marginal electricity price of the negative node on the energy storage side in the microgrid, and then update the slope;

[0139] S4-7, calculate the state of the multi-microgrid system at the next time;

[0140] S4-8, let t = t + Δt, repeat steps S4-4 to S4-7 until t = T;

[0141] S4-9, let n = n + 1, repeat steps S4-2 to S4-8 until n = N;

[0142] S4-10, output the trained slope of piecewise linear function.

[0143] The first mode is to calculate the sampling estimation value of the slope through the marginal influence of the energy storage SOC on the value function of the energy storage itself, which can be specifically represented as:

[0144]

[0145] wherein the superscript n is the training times of the ADP algorithm, is the sampling estimation value of the slope of the energy storage b at the pw' segment at time t, the segment pw' represents the segment in which the piecewise linear function is located, and α is the iteration step length.

[0146] ​The second mode is to calculate the sampling estimation value of the slope of the marginal price of the negative node of the energy storage side in the micro-grid by distributed calculation, and then update the slope, which can be specifically represented as:

[0147]

[0148] Wherein, is the node marginal price of the energy storage b node of the micro-grid at time t, is the node set on the path from the source node i to the energy storage b node in the micro-grid m.

[0149] In an embodiment, in steps S4-4 and S4-5, the privacy-protected distributed scheduling model is re-solved, specifically:

[0150] S2S3-1, initialize the state of the multi-micro-grid system;

[0151] S2S3-2, for x update sub-problems, respectively solve the micro-grid internal optimization problem and the upper grid interaction sub-problem in sequence;

[0152] S2S3-3, solve the y update sub-problem;

[0153] S2S3-4, solve the lambda update problem;

[0154] S2S3-4, perform convergence judgment, if the initial residual error and the dual residual error are not within the allowable error range, return to S2S3-2 to continue distributed solving; if the initial residual error and the dual residual error are within the allowable error range, output the multi-micro-grid distributed scheduling decision.

[0155] When the slope is trained, the solution of the scheduling decision at each time can be obtained, specifically:

[0156] Based on the real information of wind power, photovoltaic, load and price at time t, the real time state of the multi-micro-grid system at time t is determined According to the trained piecewise linear function slope, the multi-micro-grid distributed scheduling optimization framework based on ADMM is used to obtain the optimal scheduling decision at time t According to the real information of wind power, photovoltaic, load and price at the next time, the real state of the multi-micro-grid system at the next time is obtained, and then the approximate optimal scheduling decision of the multi-micro-grid system at the next time is obtained; in sequence, the approximate optimal scheduling strategy of the multi-micro-grid system in the whole scheduling domain is obtained.

[0157] Correspondingly, the present application also protects a multi-micro-grid system distributed real-time scheduling system, comprising a memory and a processor, the memory stores a computer program, wherein the processor implements the steps of the above method when executing the computer program.

[0158] Specifically, this system can be installed on computing devices such as desktop computers, laptops, handheld computers, and cloud servers. The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The memory can be used to store computer programs and / or modules. The processor performs various functions of the electronic device by running or executing the computer programs and / or modules stored in the memory, and by accessing data stored in the memory.

[0159] To demonstrate the effectiveness of this scheme, the following verification is also performed, in which the method of this scheme is defined as a fully decentralized ADP algorithm.

[0160] like Figure 4 The figure shows the convergence of the distributed scheduling model for multi-microgrids based on ADMM provided in this embodiment of the invention. As can be seen from the figure, the ADMM algorithm converges after very few iterations in 24 hours of scheduling. The objective function obtained by distributed calculation is on the order of 1e-5 compared with the objective function of centralized scheduling, indicating that the distributed scheduling model for multi-microgrids based on ADMM can obtain centralized global optimal scheduling by relying only on local information.

[0161] Figure 5 The figure shows the difference between the slope of the piecewise linear function calculated by the fully distributed ADP algorithm provided in this embodiment of the invention and the slope calculated by the centralized algorithm. As can be seen from the figure, the error between the slope calculated by the distributed algorithm and the slope calculated by the centralized algorithm is on the order of 1e-4, indicating that the distributed slope update method in the fully distributed ADP algorithm can achieve the same effect as the centralized update.

[0162] Figure 6 The figure shows a comparison of the performance of the fully distributed ADP algorithm (FD-ADP), distributed myopic (FD-Myopic), and distributed model predictive control (FD-MPC) provided in the real-time example of this invention. It can be seen from the figure that the FD-ADP algorithm has the smallest optimization error, which is on the order of 1e-3, indicating that the fully distributed ADP algorithm can obtain the near-optimal scheduling decision in the entire scheduling domain.

[0163] In general, the present application provides a multi-micro-grid system distributed real-time scheduling method, according to the technical parameters of each element of the multi-micro-grid system and the Bellman optimality principle, a scheduling model of the multi-micro-grid system is established; and based on the ADMM algorithm, the alternating current power flow constraint is split into multiple micro-grids, and a distributed scheduling model of the multi-micro-grid system is established; further combining the energy transaction mechanism, the optimization problem of each micro-grid is split, the privacy protection is strengthened and the calculation efficiency is improved; finally, a completely distributed ADP algorithm is proposed to solve the scheduling model, a piecewise linear function is used to approximate the value function, an ADP distributed value function slope updating method is designed, and the trained distributed value function is applied to the real-time scheduling to assist in obtaining the approximate global optimal real-time scheduling scheme of the multi-micro-grid system.

[0164] Moreover, the multi-micro-grid system distributed real-time scheduling method provided by the present application can take into account the influence of alternating current power flow constraints, and through the interaction and transaction of energy information between each micro-grid, the optimal operation of the multi-micro-grid system is achieved in a distributed manner, with the advantages of privacy protection and convenient calculation.

[0165] Further, the multi-micro-grid system distributed real-time scheduling method provided by the present application derives the physical meaning of the ADP piecewise linear value function slope under the alternating current power flow constraint condition, and designs a distributed value function slope updating method, which can only rely on the transaction energy information and the internal information of the micro-grid to embed the experience information of the randomness of the whole system, and strengthens the privacy protection in the ADP offline training, and the trained ADP value function can cope with randomness in real-time scheduling.

[0166] The technical features of the above embodiments can be combined in any manner, and to make the description concise, not all possible combinations of the technical features in the above embodiments are described, but as long as the combinations of the technical features do not exist contradictory, they should be considered as the scope of the present application. It should be noted that the "in an embodiment of the present application", "for example", "such as" and the like are intended to illustrate the present application, and are not used to limit the present application.

[0167] The above embodiments only express several embodiments of the present application, and the description is more specific and detailed, but it should not be understood as limiting the scope of the application. It should be noted that for those skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are within the scope of the present application.

Claims

1. A distributed real-time scheduling method for a multi-microgrid system, characterized in that, include: S1. Based on the technical parameters of each component in the multi-microgrid system and the Bellman optimality principle, establish a centralized dispatch model for the multi-microgrid system. S2. Using each upstream power grid node on the main line as a splitting node, the AC power flow constraints in the centralized scheduling model are split using the ADMM algorithm to establish a distributed scheduling model for a multi-microgrid system. Different microgrids correspond to independent scheduling models. The upstream power grid nodes include external power grid nodes, microgrid nodes, and intermediate connection nodes. S3. Combining the energy trading mechanism, the optimization problem corresponding to the scheduling model of each microgrid is decomposed into an interaction sub-problem with the upper-level power grid and an internal optimization sub-problem within the microgrid, resulting in a distributed scheduling model with enhanced privacy protection. S4. The ADP stochastic optimization algorithm is used to solve the privacy-enhanced distributed scheduling model to obtain the optimal real-time scheduling decision across the entire scheduling domain. In S1, the objective function of the centralized scheduling model for the multi-microgrid system is: in, Describe the objective function. for Time of the first The operating costs of a microgrid for Moment Microgrid The post-decision value function and the post-decision time state variable. for Electricity price at any time for Active power exchange between microgrids and the external power grid at any given time. This is the serial number of the microgrid. A collection of microgrids, for Time-based pneumatic generator set Power generation capacity, , , For gas turbine generator sets The power generation cost coefficient, For the first A collection of microgrid gas turbine units This refers to the penalty coefficient for wind and solar power curtailment. for Wind turbine Photovoltaic units Available power generation capacity for Wind turbine Photovoltaic units Actual power generation For the first A collection of microgrid wind turbines, photovoltaic units, and energy storage. for Time and Energy Storage The relevant approximate function and the state of charge after the decision; In S1, the constraints of the centralized scheduling model of the multi-microgrid system include: AC power flow constraints, node voltage upper and lower limits constraints, gas turbine output upper and lower limits constraints, gas turbine ramping constraints, wind power output upper and lower limits constraints, photovoltaic output upper and lower limits constraints, power trading upper and lower limits constraints between the multi-microgrid system and the external power grid, energy storage operation constraints, and reactive power output upper and lower limits constraints of gas power, wind power, photovoltaic, external power grid, and energy storage. In S4, the ADP stochastic optimization algorithm uses a piecewise linear function to fit the value function in the Bellman optimality principle. The proposed piecewise linear function expression is as follows: in, The total number of segments in a piecewise linear function. for Energy storage at all times The piecewise linear function in the th... The slope of the segment, for Energy storage at all times In the Segmented SOC values, For energy storage The maximum and minimum states of charge; In S4, obtaining the optimal real-time scheduling decision for the entire scheduling domain includes solving the optimal real-time scheduling decision for each moment in sequence to form the optimal real-time scheduling decision for the entire scheduling domain. Among them, the The decision-making process at any given moment is as follows: based on Real-time information on wind power, solar power, load, and electricity prices is needed to determine the status of multi-microgrid systems. The real time state at any given moment ; The piecewise linear function is determined based on the slope of the trained piecewise linear function; Solving the distributed scheduling model with enhanced privacy protection yields the following results. Optimal scheduling decision at any given time.

2. The distributed real-time scheduling method for multi-microgrid systems according to claim 1, characterized in that, In S2, when the ADMM algorithm is used to decompose the AC power flow constraints in the centralized scheduling model, the distributed scheduling model of the multi-microgrid system related to each upstream grid node... Update subproblems Update subproblems The update problem and convergence criteria are as follows: Update subproblem: For a moment Update the decision variables of the subproblem. for time Update the decision variables of the subproblem. As a Lagrange multiplier, as time Update the decision variables of the subproblem; for The augmented Lagrangian function in the time-major ADMM algorithm, for In the time-based ADMM algorithm Update the feasible region of the subproblem; superscript This represents the number of iterations in the ADMM algorithm. Nodes in the upper-level power grid The set of neighboring nodes, For the upper-level power grid node Total active and reactive power generation For a positive number that approaches 0, For the upper-level power grid node The set of decision variables, for At the node The copied variable; For the upper-level power grid node The total cost of electricity generation for microgrid nodes, For external power grid nodes, For intermediate connection nodes, ; for Time Node The square of voltage and current for Time Node The line formed by it and its neighboring nodes The trend of meritorious and ineffective actions For nodes Lower and upper limits of voltage squared It is the set of nodes in the upper-level power grid; For nodes For microgrid nodes, the internal constraint convex set of the subproblem is... Includes: internal AC power flow constraints of the microgrid, node voltage upper and lower limits constraints, gas turbine output upper and lower limits constraints, gas turbine ramping constraints, wind power output upper and lower limits constraints, photovoltaic output upper and lower limits constraints, energy storage operation constraints, and upper and lower limits constraints for reactive power output of gas power, wind power, photovoltaic, and energy storage. For external grid nodes... Includes: upper and lower limits of active and reactive power transactions between multi-microgrid systems and the external power grid, for intermediate connection nodes: The node's active and reactive power must be zero. Update subproblem: in, for In the time-based ADMM algorithm Update the feasible region of the subproblem. For the line resistance, reactance This refers to a collection of upstream power grid lines, indicated by the superscript. Represents a node The parent node, For nodes The set of child nodes; Updated issue: Convergence criteria: in, for The initial residual and dual residual at time step.

3. The distributed real-time scheduling method for multi-microgrid systems according to claim 2, characterized in that, In S3, the objective of solving the microgrid internal optimization subproblem is to obtain the total active and reactive power output of the entire microgrid, and its objective function is: Update a subset of the objective function of the subproblem. Only contains and The corresponding constraints include Update the microgrid internal constraints in the subproblem; The objective of solving the interaction subproblem of the microgrid's supergrid is to optimize communication between adjacent nodes to obtain power flow information, thereby satisfying AC power flow constraints in a decentralized manner. Its objective function is: Update a subset of the objective function of the subproblem. Only contains , , , The corresponding constraints include Update the AC power flow constraints and node voltage upper and lower limits constraints in the subproblem.

4. The distributed real-time scheduling method for multi-microgrid systems according to claim 1, characterized in that, In S4, before using the ADP stochastic optimization algorithm to solve the scheduling model, the slope of the piecewise linear function in the ADP stochastic optimization algorithm is first trained as follows: S4-1. Initialize the slope of the piecewise linear function and set the number of training iterations, let n=1; S4-2. Based on the day-ahead forecasts of wind power, photovoltaic power, load, and electricity price, a set of training scenarios is generated using the Monte Carlo method. S4-3, Order ; S4-4. Identifying Multi-Microgrid Systems The time-state variables at each moment are used to solve a distributed scheduling model with enhanced privacy protection, yielding the following results. Decision-making at any moment; S4-5. Add a perturbation ΔSOC to the SOC of energy storage and re-solve the distributed scheduling model with enhanced privacy protection. S4-6. If the output of each component in the multi-microgrid system obtained in S4-5 is consistent with that obtained in S4-4, the slope of the value function is updated using the first method; if the output of each component in the multi-microgrid system obtained in S4-5 is different from that obtained in S4-4, the slope of the value function is updated using the second method; the first method is to calculate the sampled estimate of the slope by using the marginal impact of the energy storage SOC on the value function of the energy storage itself, and then update the slope; the second method is to calculate the sampled estimate of the slope by using the distributed calculation of the marginal electricity price of the negative node on the energy storage side of the microgrid, and then update the slope. S4-7 Calculate the state of the multi-microgrid system at the next moment; S4-8, Order Repeat steps S4-4 to S4-7 until... ; S4-9, Order Repeat steps S4-2 to S4-8 until... ; S4-10, Output the slope of the trained piecewise linear function.

5. The distributed real-time scheduling method for multi-microgrid systems according to claim 4, characterized in that, In S4, the steps for solving the privacy-enhancing distributed scheduling model include: S2S3-1. Initialize the state of the multi-microgrid system; S2S3-2, For Update the subproblems by solving the microgrid internal optimization subproblem and the upper-level power grid interaction subproblem in sequence; S2S3-3, Solve Update subproblems; S2S3-4, Solve Update issue; S2S3-5. Perform convergence determination. If the initial residual and dual residual are not within the allowable error range, return to S2S3-2 to continue the distributed solution; if both the initial residual and dual residual are within the allowable error range, output the distributed scheduling decision for the multi-microgrid.

6. A distributed real-time dispatching system for multiple microgrid systems, comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1 to 5.

Images