A Distributed Scheduling Method for Active Distribution Networks Considering the Participation of Multiple Stakeholders

Abstract

The present invention belongs to the field of distributed optimal scheduling of distribution networks in power systems, and provides an active distribution network distributed scheduling method considering the participation of multiple stakeholders. By analyzing the costs of providing spinning reserve by each stakeholder in the distribution network, the present invention proposes a two-stage distributed scheduling method for active distribution networks considering the sharing of spinning reserve capacity. In this strategy, the spinning reserve capacity required to maintain the power balance of the distribution network is preferentially provided by the controllable distributed power sources within the distribution network, and the distributed solution of this scheduling method is realized using the consensus theory. The present invention can effectively reduce the dependence of the distribution network on the superior power grid, realize the localization of the supply and demand of spinning reserve and the shared responsibility for power balance, and each stakeholder independently controls its own output during the scheduling process, fully reflecting the wishes of distributed stakeholders.

Description

Technical Field

[0001] The present invention belongs to the field of distributed optimal scheduling of distribution networks in power systems. It involves the calculation of the spinning reserve cost of distributed power sources and related theories of distributed scheduling based on the consensus theory, and particularly relates to a method for distributed scheduling of active distribution networks considering the participation of multiple stakeholders in the sharing of spinning reserve capacity. Background Art

[0002] With the large-scale access of various types of distributed power sources, the distribution network has changed from a traditional receiving-end network to an active distribution network with active regulation capabilities and multi-source collaboration. These distributed power sources belong to different stakeholders, and information such as their cost parameters and output limits is their own business secrets, and the dispatching center has no right to obtain them, which makes the traditional centralized scheduling mode no longer applicable, and the research on distributed scheduling of distribution networks has emerged as the times require.

[0003] Different from centralized scheduling, distributed scheduling does not require the dispatching center to collect global information for unified calculation, and only needs to achieve the coordinated optimization of the system through simple information communication and a small amount of calculation among distributed entities, getting rid of the dependence of the system on the central controller, improving reliability while protecting the privacy of production information of each distributed entity. Among them, the distributed optimal scheduling based on the consensus algorithm has the advantages of high calculation efficiency, high reliability, and protection of private information, and has become an effective method for the optimal scheduling of active distribution networks under multi-source structures.

[0004] A large number of renewable distributed power sources represented by distributed photovoltaics and distributed wind power are connected to the distribution network, and the uncertainty of their generated power brings great challenges to the power balance control of the distribution network. Completely consuming renewable energy greatly increases the uncertainty of the net load. When making a daily dispatch plan, it is necessary to reserve a sufficient amount of spinning reserve capacity to ensure that the system maintains power balance. Under the traditional mode, the spinning reserve demand of the distribution network is provided by the superior power grid. However, with the increase in demand, the superior power grid needs to reserve more spinning reserve capacity, and the uncertainty of the power transmitted to the distribution network also increases greatly. It is necessary to study a method for providing spinning reserve capacity for the active distribution network to realize the localization of the supply and demand of spinning reserve capacity and reduce the dependence of the distribution network on the superior power grid. Summary of the Invention

[0005] The present invention provides a method for distributed scheduling of active distribution networks considering the participation of multiple stakeholders in the sharing of spinning reserve capacity. In order to reduce the dependence of the distribution network on the superior power grid, the spinning reserve capacity required inside the distribution network is preferentially provided by controllable distributed power sources within the distribution network. The dispatch plan is divided into two stages: Stage 1 is to determine the daily active power planned output of each stakeholder, and Stage 2 is to determine the spinning reserve capacity shared by each stakeholder.

[0006] To achieve the above object, the technical solution of the present invention is as follows:

[0007] An active distribution network distributed scheduling method considering the participation of multiple stakeholders includes the following steps:

[0008] Step 1. Report the day-ahead load forecast value and reserve demand, predict the day-ahead active power output of renewable distributed power sources, and report their predicted output and spinning reserve demand.

[0009] Step 2. Establish an active distribution network power balance scheduling model

[0010] (1) Objective function

[0011] The active distribution network power balance scheduling model takes the minimum total operating cost C of the distribution network p as the objective function, as shown in Equation (1).

[0012]

[0013] In the formula: T is the number of scheduling periods; N DG is the number of controllable distributed power sources in the distribution network; is the generation cost of controllable distributed power source i in the t-th period, and the expression is as shown in Equation (2); is the power purchase cost from the superior power grid in the t-th period of the distribution network, and the expression is as shown in Equation (3).

[0014]

[0015]

[0016] In the formula: is the cost coefficient of controllable distributed power source i; is the active power output of controllable distributed power source i in the t-th period; is the square of the active power output of controllable distributed power source i in the t-th period; ep t is the selling electricity price of the superior power grid in the t-th period; P t grid is the power purchase from the superior power grid in the t-th period of the distribution network.

[0017] (2) Constraint conditions

[0018] The distribution network power balance constraint is as shown in Equation (4).

[0019]

[0020] In the formula: P t L is the day-ahead load forecast demand in the t-th period; is the day-ahead forecast output of renewable distributed power source k in the t-th period; K is the number of renewable distributed power sources.

[0021] The output constraints and ramping constraints of controllable distributed power sources are shown in Equations (5) and (6) respectively.

[0022]

[0023]

[0024] In the formula: and are the upper and lower limits of the output of controllable distributed power source i respectively; r i,d and r i,u are the downward ramping rate and upward ramping rate of controllable distributed power source i respectively.

[0025] Step 3. Establish an active distribution network spinning reserve capacity sharing model

[0026] The spinning reserve cost consists of two parts: capacity cost and energy cost.

[0027] The described capacity cost: For controllable distributed power sources, the capacity cost is the cost consumed by the unit running unloaded, and the cost function is shown in Equation (7); for flexible loads, the capacity cost is the compensation cost for providing curtailable capacity, and the cost function is shown in Equation (8).

[0028]

[0029]

[0030] In the formula: represents the capacity cost coefficient of controllable distributed power source i; is the spinning reserve capacity provided by controllable distributed power source i in the t-th period; represents the capacity cost coefficient of flexible load j; is the spinning reserve capacity provided by flexible load j in the t-th period.

[0031] The described energy cost: For controllable distributed power sources, the energy cost is the difference between the cost generated by the actual regulated power consumption of the unit and the cost generated by the day-ahead output plan, and the spinning reserve energy cost is as shown in Equation (9); for flexible loads, the energy cost expression is as shown in Equation (10).

[0032]

[0033]

[0034] In the formula: is the cost coefficient of flexible load j.

[0035] Since the actual amount of spinning reserve called is uncertain in advance, the capacity cost and energy cost cannot be directly added as the basis for capacity sharing when making the day-ahead plan. It is necessary to consider the probability that the provided spinning reserve capacity will be called to determine the willingness to participate in capacity sharing. To reflect the willingness of each stakeholder to participate in capacity sharing, the present invention introduces a spinning reserve willingness factor β. For controllable distributed power sources, the equivalent cost function is as shown in Equation (11); for flexible loads, the equivalent cost function is as shown in Equation (12).

[0036]

[0037]

[0038] In the formula: and respectively represent the willingness of controllable distributed power source i and flexible load j to participate in the sharing of spinning reserve capacity, and the value ranges from 0 to 1, which is determined by the stakeholders participating in capacity sharing themselves. The larger the value, the greater the possibility that the spinning reserve is predicted to be called by the entity, and the higher the willingness to participate in capacity sharing.

[0039] (1) The objective function of the active distribution network spinning reserve capacity sharing model described above is as follows:

[0040] The principle of spinning reserve sharing is that the spinning reserve capacity is preferentially provided by the controllable distributed power sources and flexible loads within the distribution network. Under the condition of meeting the regulation capacity constraint, the capacity sharing is aimed at minimizing the total equivalent cost Cr of each controllable distributed power source and flexible load.

[0041]

[0042] In the formula: T is the number of scheduling periods; N DG is the number of controllable distributed power sources in the distribution network; N FL is the number of flexible loads in the distribution network; and respectively represent the equivalent cost functions of controllable distributed power source i and flexible load j in the t-th period. According to formulas (7)-(12), their expressions can be deduced as shown in Equations (14) and (15) respectively:

[0043]

[0044]

[0045] (2) The constraint conditions of the active distribution network spinning reserve capacity sharing model described above are as follows:

[0046] The balance constraint of the rotational reserve capacity of the distribution network is shown in Equation (16).

[0047]

[0048] In the formula: ΔP t grid is the rotational reserve capacity provided by the superior power grid in the t-th time period. If the dispatchable entities within the active distribution network can meet the positive rotational reserve demand of the system, then ΔP t grid = 0. If it cannot be met, the shortage part will be supplemented by the superior power grid; is the reserve demand of the conventional load in the t-th time period; is the reserve demand of the renewable distributed power source k in the t-th time period.

[0049] The regulation ability constraint and the ramping constraint of the controllable distributed power source are shown in Equation (17) and Equation (18) respectively.

[0050]

[0051]

[0052] The flexible load curtailment constraint is shown in Equation (19).

[0053]

[0054] In the formula: and are the upper and lower limits of the curtailment of the flexible load j respectively.

[0055] Step 4. Use the consensus algorithm to perform distributed solution on the two-stage day-ahead dispatch model of the active distribution network described in Step 2 (the first stage) and Step 3 (the second stage), and obtain the day-ahead active power planned output provided by each stakeholder and the shared rotational reserve capacity when the system cost is minimized.

[0056] In a multi-agent system, the consensus algorithm (or protocol) is an interaction rule used to specify the information exchange between an agent and its neighbors. The advantage of this algorithm is that it does not require a central processor. Each agent only needs to interact with adjacent nodes through a small amount of local node information, and update the local state parameters through consensus iteration, so that the state parameters of each node in the topological network converge to a stable common value. For any adjacent nodes i and j in the system, as the number of iterations k increases, when |x i [k] - x j [k]| → 0, it is considered that the consensus variables of all nodes reach consensus. The iterative process of the consensus algorithm is as follows:

[0057]

[0058] where: n is the number of nodes; k = 0, 1, … is the iteration number; x i [k + 1] is the state variable of node i after (k + 1) iterations; x j [k] is the state variable of node j after k iterations; s[k] is the deviation correction amount; μ is the deviation correction coefficient. d ij is the state transition matrix coefficient, representing the communication weight between nodes, and its construction method is:

[0059]

[0060] where: n i 、n j represent the number of neighbors of nodes i and j; N i represents the set composed of the neighbor nodes of node i.

[0061] 4.1 Solution of the Active Distribution Network Power Balance Scheduling Model

[0062] The main idea of the distributed economic dispatch based on the consensus algorithm is to use the partial derivative λ of each cost function with respect to the active power output as the consensus variable, and use the distributed consensus algorithm to obtain the equal incremental cost rate under the optimal economic dispatch. Among them, a leader needs to be selected to perform pinning control on the entire system based on the system power deviation and the electricity price deviation, that is:

[0063]

[0064] where: k is the iteration number; λ i [k + 1] is the incremental cost of the interest subject i after (k + 1) iterations; λ j [k] is the incremental cost of the interest subject j in the kth iteration; μ1, μ2 are the deviation correction coefficients; ΔP[k] is the power deficit in each iteration; Δe[k] is the difference between the incremental cost of the interest subject in each iteration and the electricity price of the superior power grid.

[0065] Each interest subject updates its own consensus variable value through Equation (22) and calculates the output power through Equation (23):

[0066]

[0067] where: P i DG [k] is the active power output of the controllable distributed power source i after the kth iteration calculation; and are respectively the upper and lower limits of the output of the controllable distributed power source i in the tth period; and are respectively the upper and lower limits of the corresponding consensus variables. The calculation method is:

[0068]

[0069] For the upstream power grid, its consistency variable is always equal to the electricity price. Therefore, there is no need to iterate on the consistency variable. Only the output needs to be adjusted according to the system power difference, that is:

[0070] P t grid [k + 1] = P t grid [k] + μ1ΔP[k] (25)

[0071] Where: P t grid [k + 1] is the active power output of the upstream power grid after the (k + 1)-th iteration in the t-th time period; P t grid [k] is the active power output of the upstream power grid in the k-th iteration in the t-th time period.

[0072] Each stakeholder calculates its own output according to equations (22)-(25) until the system satisfies power balance, that is, ΔP[k] = 0, and the consistency variables of all agents converge to the consistent optimal value.

[0073] 4.2 Solution of the spinning reserve capacity sharing model for the active distribution network

[0074] For the spinning reserve sharing model, the consistency variable η is selected as the partial derivative of the objective function with respect to each variable, and the deviation correction amount is the reserve imbalance. The iterative processes of the leader and each follower's consistency variables are shown in equation (26):

[0075]

[0076] Where: η i [k + 1] is the consistency variable of stakeholder i after the (k + 1)-th iteration; η j [k] is the consistency variable of stakeholder j in the k-th iteration; ΔR[k] is the reserve deficit in each iteration.

[0077] Each controllable distributed power source and flexible load update their own consistency variable values through equation (26), and calculate the spinning reserve capacity to be shared through equations (27) and (28) respectively:

[0078]

[0079]

[0080] Where: ΔP i DG [k] is the spinning reserve capacity to be shared by controllable distributed power source i after the k-th iteration calculation; and are the upper and lower limits of the output of the controllable distributed power source i in the t-th period; η i,t,max and η i,t,min are the upper and lower limits of the corresponding controllable distributed power source consistency variables; is the spinning reserve capacity that the flexible load j should share after the k-th iteration calculation; and are the upper and lower limits of the output of the flexible load j in the t-th period; η j,t,max and η j,t,min are the upper and lower limits of the corresponding flexible load consistency variables.

[0081] After a finite number of iterations through equations (26)-(28), the consistency variables can tend to be consistent, realizing the distributed solution of the model.

[0082] The present invention proposes an active distribution network distributed scheduling method considering the participation of multiple stakeholders, and its beneficial effects are as follows:

[0083] The present invention establishes a distributed scheduling model considering the willingness of each stakeholder to participate in the sharing of spinning reserve capacity, and uses the consensus algorithm to achieve distributed solution. By introducing the willingness factor, the sharing of reserve capacity considering the participation willingness of multiple stakeholders is realized, and the sharing result can not only reflect the true costs of each stakeholder, but also fully reflect the participation willingness. Description of the Drawings

[0084] Figure 1 is the communication topology structure diagram;

[0085] Figure 2 is the flow chart of the scheduling method;

[0086] Figure 3 is the day-ahead power prediction value of the load and renewable distributed power sources;

[0087] Figure 4 is the spinning reserve demand value of the load and renewable distributed power sources;

[0088] Figure 5 is the day-ahead active power output scheduling result;

[0089] Figure 6 is the iteration process of the consistency variable, the iteration process of the active power output, and the power balance iteration process in each period;

[0090] Figure 7 is the spinning reserve capacity sharing result under each scenario; Figure 7 (a) is the spinning reserve capacity sharing result of Scenario 1; Figure 7 (b) is the spinning reserve capacity sharing result of Scenario 2; Figure 7 (c) is the spinning reserve capacity sharing result of Scenario 3;

[0091] Figure 8 is the sharing ratio of the spinning reserve capacity of each entity under each scenario in the 7th period; Figure 8 (a) is the sharing ratio of the spinning reserve capacity of each entity in Scenario 1 of the 7th period; Figure 8 (b) is the sharing ratio of the spinning reserve capacity of each entity in Scenario 2 of the 7th period; Figure 8 (c) is the sharing ratio of the spinning reserve capacity of each entity in Scenario 3 of the 7th period. Detailed implementation manner

[0092] The present invention will be further described below in conjunction with specific embodiments.

[0093] Taking the distribution system in a certain area in Northeast China as an example, the dispatchable entities include 3 controllable distributed power sources and 2 flexible loads. The superior power grid participates in dispatching as an independent dispatchable entity. The communication topology is as Figure 1 shown, and DG2 is the leader node, responsible for calculating the deviation correction amount during the iterative calculation process. The basic parameters of the controllable distributed power sources and flexible loads are shown in Table 1, and the selling electricity price of the superior power grid is shown in Table 2.

[0094] Table 1 Parameters of controllable distributed power sources and flexible loads

[0095]

[0096] Table 2 Selling electricity price of the superior power grid

[0097]

[0098] Figure 2 is the flow chart of the dispatching method of the present invention, and the specific steps are as follows:

[0099] First step, sorting out the relevant data of the system, sorting out the day-ahead power prediction values and spinning reserve demand values of the load and renewable distributed power sources. There is 1 distributed photovoltaic power station and 1 distributed wind power station in the system. The system takes one hour as a dispatching period. The day-ahead 24-hour power prediction values of the load and each renewable distributed power source are as Figure 3 shown, and the spinning reserve demand value is as Figure 4 shown.

[0100] Second step, establishing an active distribution network power balance dispatching model, and performing distributed solution using the consensus algorithm. The dispatching results of the day-ahead active power planned output of each controllable distributed power source and the superior power grid are as Figure 5 shown. The iterative process of the consensus variable, the iterative process of the active power output, and the power balance iterative process in each period are shown in Figure 6, it can be seen that the consistency variables in each period tend to be consistent after a finite number of iterations and finally converge to the selling price of electricity of the superior power grid. The power imbalance also tends to zero, and the actual output of each controllable distributed power source meets its own constraints. The total power generation costs of each stakeholder are shown in Table 3, and the scheduling results verify the effectiveness of the distributed consistency algorithm proposed in this paper.

[0101] Table 3 Active power generation costs of each stakeholder

[0102]

[0103] In the third step, the spinning reserve capacity sharing model of the active distribution network is solved distributively using the consistency algorithm, and the sharing results are as Figure 7 shown.

[0104] Three scenarios are set to analyze the influence of the reserve willingness factor β on the sharing results. In Scenario 1, the participation willingness of each stakeholder is the same; in Scenario 2, the participation willingness of controllable distributed power sources is increased; in Scenario 3, the participation willingness values of each stakeholder are randomly generated. The data of β for each scenario are shown in Table 4.

[0105] Table 4 Reserve willingness factors of each stakeholder

[0106]

[0107] The spinning reserve capacity sharing results under each scenario are as Figure 7 shown. According to the results, it can be seen that different willingness factors lead to significant differences in the reserve sharing results.

[0108] Taking the 7th period as an example, the spinning reserve demand is 185.4 kW, and all three scenarios meet the demand. The sharing ratios are as Figure 8 shown. In Scenario 1, the participation willingness of each subject is the same, so the true cost becomes the main factor affecting the sharing results of each stakeholder. From Figure 8It can be seen that the sharing ratios of FL1 and FL2 are relatively high, indicating that their standby equivalent costs are relatively low. Compared with Scenario 1, the willingness factors of the three DGs in Scenario 2 increase, and their sharing ratios also increase accordingly. In Scenario 3, the participation willingness of DG1, FL1, and FL2 increases, while the willingness of DG2 and DG3 decreases. At this time, the sharing amounts of DG2 and DG3 are 0, and the standby capacity is shared by the three stakeholders with higher participation willingness. This shows that when the dispatchable capacity is sufficient, the lower the self-cost and the greater the participation willingness, the larger the sharing ratio of the positive spinning reserve capacity. During 12:00 - 17:00, since the remaining dispatchable capacity of DG3 is insufficient after participating in power balance, even though its participation willingness increases in Scenario 2 compared with Scenario 1, the sharing ratio remains unchanged. During 08:00 - 12:00 and 17:00 - 21:00, due to the high selling price of electricity from the superior power grid, the proportion of the gas turbine in the distribution network participating in power balance increases, and the day-ahead planned output of DG3 has reached its output limit, leaving no dispatchable capacity to participate in the sharing of positive spinning reserve capacity. Therefore, regardless of the change in its participation willingness, its sharing amount is 0. In summary, the self-cost, participation willingness, and dispatchable capacity of each stakeholder will all affect the sharing result of the spinning reserve capacity.

[0109] The above-described embodiments only express the implementation manners of the present invention, but should not be construed as limiting the scope of the patent of the present invention. It should be noted that for those skilled in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all belong to the protection scope of the present invention.

Claims

1. An active distribution network distributed scheduling method considering the participation of multiple stakeholders, characterized in that, It includes the following steps: Step 1. Report the day-ahead load forecast value and reserve demand. Forecast the day-ahead active power output of renewable distributed power sources, and report their forecast output and spinning reserve demand; Step 2. Establish an active distribution network power balance scheduling model Step 3. Establish an active distribution network spinning reserve capacity sharing model The spinning reserve cost consists of two parts: capacity cost and energy cost; When making the daily plan, it is necessary to consider the probability that the provided spinning reserve capacity is called to determine the willingness to participate in capacity sharing; to reflect the willingness of each stakeholder to participate in capacity sharing, a spinning reserve willingness factor β is introduced; for controllable distributed power sources, the equivalent cost function is shown in Equation (11); for flexible loads, the equivalent cost function is shown in Equation (12). Wherein: and respectively represent the willingness of the controllable distributed power source i and the flexible load j to participate in the sharing of the spinning reserve capacity. The values range from 0 to 1 and are determined by the stakeholders participating in the capacity sharing themselves. The larger the value, the greater the possibility that the spinning reserve of the subject is predicted to be called, and the higher the willingness to participate in the capacity sharing; (1) The objective function of the active distribution network spinning reserve capacity sharing model is as follows: The spinning reserve sharing principle is that the controllable distributed power sources and flexible loads inside the distribution network provide spinning reserve capacity preferentially. Under the condition of meeting the regulation capacity constraint, the capacity sharing is targeted at minimizing the total equivalent cost Cr of each controllable distributed power source and flexible load; Where: T is the number of scheduling periods; N DG is the number of controllable distributed power sources in the distribution network; N FL is the number of flexible loads in the distribution network; and respectively represent the equivalent cost functions of the controllable distributed power source i and the flexible load j in the t-th period. According to formulas (7)-(12), their expressions can be deduced as shown in formulas (14) and (15): (2) The constraint conditions of the active distribution network spinning reserve capacity sharing model are as follows: The distribution network spinning reserve capacity balance constraint is shown in Equation (16); Where: ΔP t grid is the spinning reserve capacity provided by the upstream power grid in the t-th period. If the dispatchable entities within the active distribution network can meet the positive spinning reserve demand of the system, then ΔP t grid = 0. If it cannot be met, the shortfall will be supplemented by the upstream power grid; is the reserve demand of the conventional load in the t-th period; is the reserve demand of the renewable distributed power source k in the t-th period; The regulation capacity constraint and ramp constraint of the controllable distributed power sources are shown in Equation (17) and Equation (18) respectively; The flexible load curtailment amount constraint is shown in Equation (19); Where: and are the upper and lower limits of the reduction amount of flexible load j, respectively; Step 4. Use the consensus algorithm to perform distributed solution on the two-stage day-ahead scheduling model of the active distribution network described in Step 2 (the first stage) and Step 3 (the second stage), and obtain the day-ahead active power planned output and the shared spinning reserve capacity provided by each stakeholder when the system cost is minimized; In a multi-agent system, the consensus algorithm does not require a central processor. Each agent only needs to interact with a small amount of information from its local nodes and adjacent nodes, and update the local state parameters through consensus iteration, so that the state parameters of each node in the topological network can converge to a stable common value. For any two adjacent nodes i and j in the system, as the number of iterations k increases, when |x i [k] - x j [k]| → 0, it is considered that the consensus variables of all nodes reach consensus. The iterative process of the consensus algorithm is as follows: Where: n is the number of nodes; k = 0, 1, … is the number of iterations; x i [k + 1] is the state variable of node i after (k + 1) iterations; x j [k] is the state variable of node j after k iterations; s[k] is the deviation correction amount; μ is the deviation correction coefficient; d ij is the state transition matrix coefficient, representing the communication weight between nodes, and the construction method is as follows: where: n i , n j represent the number of neighbors of nodes i and j; N i represents the set composed of the neighbor nodes of node i; 4.1 Solution of the active distribution network power balance scheduling model The main idea of the distributed economic scheduling based on the consensus algorithm is to use the partial derivative λ of each cost function with respect to the active power output as the consensus variable, and use the distributed consensus algorithm to obtain the equal-cost incremental rate under the optimal economic scheduling; among them, a leader needs to be selected to perform pinning control on the entire system with the system power deviation and electricity price deviation, that is: where: k is the number of iterations; λ i [k + 1] is the incremental cost of the stakeholder i after k + 1 iterations; λ j [k] is the incremental cost of the stakeholder j in the k-th iteration; μ1 and μ2 are deviation correction coefficients; ΔP[k] is the power deficit in each iteration; Δe[k] is the difference between the incremental cost of the stakeholder in each iteration and the electricity price of the superior power grid; Each stakeholder updates its own consensus variable value through Equation (22) and calculates the output power through Equation (23): Where: P i DG [k] is the active power output of the controllable distributed power source i after the k-th iterative calculation; and are respectively the upper and lower limits of the output of the controllable distributed power source i in the t-th time period; and are respectively the upper and lower limits of the corresponding consistency variables; the calculation method is: For the superior power grid, its consensus variable is always equal to the electricity price. Therefore, there is no need to iterate the consensus variable, and only the output needs to be adjusted according to the system power difference, that is: P t grid [k + 1]=P t grid [k]+μ1ΔP[k] (25) Where: P t grid [k + 1] is the active power output of the superior power grid after the (k + 1)-th iteration in the t-th time period; P t grid [k] is the active power output of the superior power grid in the k-th iteration in the t-th time period; Each stakeholder calculates its own output according to Equations (22)-(25) until the system meets the power balance, that is, ΔP[k] = 0, and the consensus variables of all agents converge to the consistent optimal value; 4.2 Solution of the active distribution network spinning reserve capacity sharing model For the spinning reserve sharing model, the consensus variable η is selected as the partial derivative of the objective function with respect to each variable, and the deviation correction amount is the reserve imbalance amount. The iterative process of the leader and each follower's consensus variable is shown in Equation (26): where: η i [k + 1] is the consistency variable of the stakeholder i after k + 1 iterations; η j [k] is the consistency variable of the stakeholder j in the k-th iteration; ΔR[k] is the spare deficit for each iteration; Each controllable distributed power source and flexible load update their own consensus variable values through Equation (26), and calculate the shared reserve capacity through Equation (27) and Equation (28) respectively: where: ΔP i DG [k] is the spinning reserve capacity that the controllable distributed power source i should share after the k - th iteration calculation; and are the upper and lower limits of the output of the controllable distributed power source i in the t - th period respectively; η i,t,max and η i,t,min are the upper and lower limits of the corresponding controllable distributed power source consistency variables respectively; is the spinning reserve capacity that the flexible load j should share after the k - th iteration calculation; and are the upper and lower limits of the output of the flexible load j in the t - th period respectively; η j,t,max and η j,t,min are the upper and lower limits of the corresponding flexible load consistency variables respectively; After a finite number of iterations through Equations (26)-(28), the consensus variables can tend to be consistent, realizing the distributed solution of the model.

2. A distributed scheduling method for an active distribution network considering the participation of multiple interest subjects according to claim 1, characterized in that, The active distribution network power balance scheduling model established in Step 2 is specifically as follows: (1) Objective function The active distribution network power balance scheduling model aims at minimizing the total operation cost C of the distribution network, as shown in Equation (1); p The objective function is to minimize it, as shown in Equation (1). Where: T is the number of scheduling periods; N DG is the number of controllable distributed power sources in the distribution network; is the power generation cost of the controllable distributed power source i in the t-th period, and the expression is shown in Equation (2); is the power purchase cost of the distribution network from the superior power grid in the t-th period, and the expression is shown in Equation (3); In the formula: is the cost coefficient of the controllable distributed power source i; is the active power output of the controllable distributed power source i in the t-th period; is the square of the active power output of the controllable distributed power source i in the t-th period; ep t is the selling electricity price of the superior power grid in the t-th period; P t grid is the power purchased by the distribution network from the superior power grid in the t-th period; (2) Constraint conditions The distribution network power balance constraint is shown in Equation (4); Where: P t L is the day-ahead predicted demand of the load in period t; is the day-ahead predicted output of renewable distributed power source k in the t-th period; K is the number of renewable distributed power sources; The constraints on the output and ramping rate of the controllable distributed power sources are shown in Equations (5) and (6), respectively; Where: and are the upper and lower limits of the output of the controllable distributed power source i, respectively; r i,d and r i,u are the downward and upward ramping rates of the controllable distributed power source i, respectively.

3. A distributed scheduling method for an active distribution network considering the participation of multiple interest subjects according to claim 1, characterized in that, In the said Step 3: The capacity cost mentioned above: for controllable distributed power sources, the capacity cost is the cost consumed by the unit running idly, and the cost function is shown in Equation (7); for flexible loads, the capacity cost is the compensation cost for providing the curtailed capacity, and the cost function is shown in Equation (8). In the formula: represents the capacity cost coefficient of the controllable distributed power source i; is the spinning reserve capacity provided by the controllable distributed power source i in the t-th period; represents the capacity cost coefficient of the flexible load j; is the spinning reserve capacity provided by the flexible load j in the t-th period; The electricity cost mentioned above: For controllable distributed power sources, the electricity cost is the difference between the cost generated by the actual regulated electricity consumption of the unit and the cost generated by the day-ahead output plan, and the cost of spinning reserve electricity is as shown in Equation (9); for flexible loads, the electricity cost expression is as shown in Equation (10); Wherein: is the cost coefficient of flexible load j.

Images