Multi-area power distribution system new energy consumption scheduling method considering feedback delay

Abstract

The invention provides a multi-region power distribution system new energy consumption scheduling method considering feedback delay, and the method comprises the steps: 1), constructing a double-layer communication topological structure, and enabling a first layer to be a communication link between a conventional power generation unit and a new energy power generation unit in a region; the second layer is composed of nodes with the maximum out-degree in all the areas; 2) establishing an online operation cost model based on double-layer topology, considering the time-varying new energy power generation condition, minimizing the cost of the power distribution system, ensuring that the power generation amount does not exceed the upper limit, and maintaining power balance; 3) establishing a Lagrange dual optimization model in order to meet the power balance requirement; 4) for the second-layer topology, making a multi-region power distribution system new energy consumption scheduling strategy considering feedback delay, and realizing region-level optimization; and 5) aiming at the first layer of topology, constructing a collaborative online optimization strategy fusing a distributed online gradient optimization technology and a projection operator, and giving out local power generation optimal solutions of all power generation units in each region.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution system scheduling and energy consumption, and particularly to a new energy consumption scheduling method for multi-region distribution systems considering feedback delay. Background Art

[0002] In recent years, with the proposal of the goals of carbon peak and carbon neutrality, how to effectively improve the consumption ratio of new energy in the distribution system while ensuring the stability and economy of the system has become one of the important issues faced by the power industry. With the progress and wide application of smart grid technology, the efficient operation of the distribution system increasingly depends on the ability to collect and process real-time data. However, in this process, the problem of feedback delay has gradually emerged and become an important factor affecting the accuracy and timeliness of scheduling decisions. Traditional power system scheduling methods often fail to fully consider the time delay phenomenon in data transmission and processing, which may lead to deviations between the execution of scheduling instructions and the actual situation, thereby affecting the operation efficiency of the entire system.

[0003] In the distribution system, the access of new energy increases the complexity and uncertainty of the system, such as the intermittency of photovoltaic output and the randomness of wind power generation, etc., which puts higher requirements on the coordinated scheduling including new energy consumption. The existence of these uncertainties, combined with the influence of feedback time delay, makes it difficult for traditional optimization models based on static or quasi-static assumptions to meet the dynamic response requirements of modern power systems. Therefore, exploring a coordinated scheduling solution method that can effectively solve the feedback delay problem and adapt to the uncertainty changes of the system after the access of new energy is of great significance for promoting the efficient utilization of new energy and ensuring the stable operation of the power system. This research not only helps to improve the flexibility and reliability of the power system, but also provides technical support for realizing the green and low-carbon energy transformation.

[0004] Literature 1, "Delay Stability Analysis of Distributed Economic Dispatch System in Active Distribution Network" (Power System Protection and Control, 2022, Vol. 50, No. 21, pp. 75-87). For the active distribution network with high-penetration distributed power sources, to solve the problem that the time delay in communication between units in the distributed mode may damage the stability of the system, a fully distributed economic dispatch strategy for the active distribution network introducing virtual distributed power sources is proposed. By using the Lyapunov stability theorem and the free-weight matrix method, a system delay stability criterion with lower conservatism is obtained, realizing the privacy protection function of power system communication information. However, the method mentioned in this literature only focuses on the time delay at the data transmission level and does not consider the time-delay characteristics of obtaining feedback optimization information in the whole scheduling optimization process. In addition, Literature 2, "Distribution Network Protection and Control Microservice Scheduling Model and Optimization Considering Virtualization Delay" (Journal of Shanghai Jiao Tong University, 2024, First Online). Aiming at the virtualization delay of the distribution network terminal based on the microservice and container architecture, a microservice scheduling framework considering virtualization delay and a microservice scheduling strategy based on queue state transition are constructed. However, this distribution network microservice scheduling optimization strategy only partially eliminates the influence brought by feedback delay and cannot essentially eliminate the disadvantage of delay. In addition, its centralized scheduling framework has certain limitations in scalability.

[0005] Therefore, it is necessary to develop a method that can not only accommodate the inherent time-varying new energy input in the multi-region distribution system but also perform collaborative scheduling and solution under the premise of considering the time delay of feedback optimization information, which is of great significance for improving the efficiency and reliability of the distribution network system. Summary of the Invention

[0006] Aiming at the deficiencies of the existing technology, the purpose of the present invention is to provide a new energy accommodation scheduling method for multi-region distribution systems considering feedback delay. Under the premise that the acquisition of optimized feedback information has a time delay, by constructing a two-layer communication topology for multi-region distribution systems and integrating distributed online gradient optimization technology and projection operators, fully considering the time-varying new energy generation with uncertainty characteristics and the requirements of new energy generation - traditional energy generation - global electricity load balance, local generation optimization solutions for multiple regions of the distribution system are given, realizing the accommodation of uncertain new energy generation at two levels, namely between regions and within regions, and reducing the overall operating cost of the distribution network.

[0007] To achieve the above object, the present invention adopts the following technical solutions: A new energy accommodation scheduling method for multi-region distribution systems considering feedback delay, comprising the following steps:

[0008] S1: Construct a two - layer communication topology for a multi - regional distribution system. The first - layer topology consists of the inherent communication links between traditional energy generation units and new - energy generation units within each region. The second - layer topology consists of the nodes with the highest out - degree in each region, and the second - layer topology is fully connected;

[0009] S2: Establish an online operating cost model for a multi - regional distribution system considering time - varying new - energy generation based on the two - layer communication topology. It is necessary to minimize the cost of the entire distribution system within a specific time, and at the same time, it must be ensured that the power generation of traditional energy generation units does not exceed their capacity limits and meets the requirements of new - energy generation - traditional - energy generation - global electricity load balance;

[0010] S3: Establish a Lagrangian dual unconstrained optimization model corresponding to the online operation optimization problem of a multi - regional distribution system considering time - varying new - energy generation based on the two - layer communication topology to meet the requirements of new - energy generation - traditional - energy generation - global electricity load balance;

[0011] S4: Construct a new - energy consumption scheduling strategy for a multi - regional distribution system considering feedback delay for the second - layer topology composed of the nodes with the highest out - degree in each region, give the distribution optimization solutions for each region in the distribution system, and achieve new - energy consumption at the regional level;

[0012] S5: For the first - layer topology composed of the inherent communication links between traditional energy generation units and new - energy generation units within each region, construct an operation collaborative online optimization strategy for the generation units in each region of the distribution system that integrates distributed online gradient optimization technology and projection operators, and give the local generation optimization solutions for all generation units in each region of the distribution system.

[0013] As a preferred technical solution of the present invention: The present invention constructs a two - layer communication topology for a multi - regional distribution system, and the specific steps are as follows:

[0014] S1.1: Obtain the communication link information of the entire distribution system;

[0015] S1.2: Perform regional division of the distribution system. The division basis is:

[0016] 1) When the manager has a clear division requirement, divide according to the requirement;

[0017] 2) When the manager has no clear division requirement, divide all the generation units in the distribution system into n m regions in an approximately average manner, and the generation units within each region need to be as similar as possible;

[0018] S1.3: Based on the divided regions, cut off the communication links between the generation units in any two regions;

[0019] S1.4: Construct the communication topology of each region based on the remaining communication links within each region, and merge the topologies within the obtained n m regions to obtain the first-layer topology composed of the inherent communication links between the traditional energy generation units and the new energy generation units within each region;

[0020] S1.5: For the topologies within the obtained n m regions, calculate the out-degree of each power generation unit within the region, that is, the number of links for sending information outward, and select the n m traditional energy generation units with the highest out-degree, and connect these power generation units pairwise to obtain the second-layer topology composed of the nodes with the highest out-degree in each region;

[0021] As a preferred technical solution of the present invention: The present invention establishes an online operating cost model for a multi-region power distribution system considering time-varying new energy generation based on a two-layer communication topology, as specifically shown in formula (1):

[0022]

[0023] In formula (1), t is the optimization period, and T is the specific total number of optimization periods; n m is the number of regions obtained by dividing the power distribution system through step S1.2; The i-th region in the power distribution system contains power generation units, including traditional energy generation units and new energy generation units; p ij (t) is the power generation of the j-th traditional energy generation unit in the i-th region of the power distribution system during the optimization period t; r ij (t) is the uncertain new energy input of the j-th new energy generation unit in the i-th region of the power distribution system during the optimization period t; is the power generation cost function of the j-th traditional energy generation unit in the i-th region of the power distribution system; is the uncertain new energy usage cost function of the j-th new energy generation unit in the i-th region of the power distribution system;

[0024] In formula (1), the power generation cost function of the j-th traditional energy generation unit in the i-th region of the power distribution system The specific expression is as shown in formula (2):

[0025]

[0026] In formula (2), a ij (t), b ij (t) and c ij(t) are the quadratic, linear, and constant term coefficients of the power generation cost function of the j-th conventional energy power generation unit in the i-th area of the power distribution system;

[0027] In formula (1), the uncertain new energy usage cost of the j-th new energy power generation unit in the i-th area of the power distribution system Specifically, it is shown in formula (3):

[0028]

[0029] In formula (3), k ij is the loss cost coefficient for using uncertain new energy;

[0030] To ensure that the power generation of the conventional energy power generation units in each area does not exceed their capacity limits, that is, for any conventional energy power generation unit, its p ij (t) needs to satisfy:

[0031]

[0032] In formula (4), is the power generation upper limit of the j-th conventional energy power generation unit in the i-th area;

[0033] In addition, the entire power distribution system also needs to meet the new energy power generation - conventional energy power generation - global electricity load balance requirement considering the integrated line loss, that is

[0034]

[0035] In formula (5), D(t) is the time-varying global electricity demand of the power distribution system, and are the power distribution transmission line losses of the j-th conventional energy power generation unit and the j-th new energy power generation unit in the i-th area, and are the corresponding power distribution transmission line loss coefficients;

[0036] As a preferred technical solution of the present invention: A Lagrangian dual unconstrained optimization model corresponding to the online operation optimization problem of a multi-area power distribution system considering time-varying new energy power generation based on a two-layer communication topology is established to meet the new energy power generation - conventional energy power generation - global electricity load balance requirement, specifically as follows:

[0037] At the optimization time period t, the following Lagrangian function is defined for all conventional energy power generation units:

[0038]

[0039] In formula (6), μ ij(t) is the Lagrangian dual variable related to new energy power generation - traditional energy power generation - global power load balance requirements; is defined as the local power demand of the j-th traditional energy power generation unit in the i-th region. At the optimization time period t, the following Lagrangian function is defined for all new energy power generation units:

[0040]

[0041] In formula (7), is defined as the local power demand of the j-th new energy power generation unit in the i-th region. Combining with formula (6), and needs to satisfy:

[0042]

[0043] It should be noted that for the power generation unit with the highest out-degree in each region of the second-layer topology, its local power demand or is not 0, and the local power demands of all other power generation units or are all 0.

[0044] Based on formulas (6) and (7), at the optimization time period t, the following Lagrangian dual unconstrained optimization model corresponding to the online operation optimization problem of the multi-region distribution system considering time-varying new energy power generation based on the double-layer communication topology can be constructed:

[0045]

[0046] As a preferred technical solution of the present invention: A new energy consumption scheduling strategy for a multi-region distribution system considering feedback delay is constructed for the second-layer topology composed of the traditional energy power generation units with the highest out-degree in each region, and the distribution optimization solutions of each region in the distribution system are given to achieve new energy consumption at the regional level, specifically as follows:

[0047] Assume that the first traditional energy power generation unit in each region of the distribution system is the node with the highest out-degree in that region. Taking the i-th region as an example, the implementation steps of the new energy consumption scheduling strategy for the multi-region distribution system considering feedback delay are specifically as follows:

[0048] S4.1: Set the optimization time index to 0, that is, t = 0; Initialize the local optimization variables and auxiliary variables:

[0049]

[0050] S4.2: Read the μ i1 (t) stored after the last update between regions and within regions.

[0051] S4.3: Send the locally optimized information μ i1 (t) read in step S4.2 to all the other power generation units in the second-layer topology.

[0052] S4.4: Receive the locally optimized information μ k1 (t) sent by all the other power generation units in the second-layer topology.

[0053] S4.5: Obtain the optimized feedback information with delay obtained by the calculation unit τ i1 Calculate the time delay for the feedback information of the first power generation unit in the i-th area of the distribution system.

[0054] S4.6: Generate the updated optimization step size α i1 (t) = 1 / (t); α i1 (t) can also be generated according to other rules, but it must meet the decreasing requirement.

[0055] S4.7: Based on μ k1 (t) received in step S4.4 and the optimized feedback information with delay obtained in step S4.5 Perform an update operation on the locally optimized dual auxiliary variable x i1 (t + 1), and the update basis is as follows:

[0056]

[0057] S4.8: Perform a projection operation on x i1 (t + 1) calculated in step S4.7 to solve the locally optimized variable μ i1 (t + 1), and the update basis is as follows:

[0058]

[0059] where is the projection operation on the constraint set, and the location of the constraint set is:

[0060]

[0061] S4.9: Based on the locally optimized variable μ i1 (t + 1) obtained in step S4.8, the locally optimized power generation p i1 (t + 1) of the first conventional energy power generation unit in the i-th area of the distribution system can be calculated, and the calculation basis is as follows:

[0062]

[0063] The min-max operation in formula (13) can ensure that the power generation of traditional energy power generation units does not exceed their capacity limits.

[0064] S4.10: Obtain the local power demand of this area at the current moment according to the measurement information

[0065] S4.11: According to the locally optimized power generation p i1 (t + 1) obtained in step S4.9 and the local power demand of this area obtained in S4.10 Calculate and solve the optimization gradient as follows:

[0066]

[0067] S4.12: Store the updated μ i1 (t + 1), and determine the condition t < T. If it is satisfied, repeat step S4.2. If it is not satisfied, end all steps.

[0068] As a preferred technical solution of the present invention: For the first-layer topology composed of the inherent communication links between the traditional energy power generation units and the new energy power generation units in each area, an operation collaborative online optimization strategy for the power generation units in each area of the distribution system that integrates the distributed online gradient optimization technology and the projection operator is constructed, and the local power generation optimization solutions for all power generation units in each area of the distribution system are given, as follows:

[0069] For the first-layer local topology composed of the inherent communication links between the traditional energy power generation units and the new energy power generation units in the i-th area, define the corresponding communication weight matrix A i (t) = [a ijk (t)]:

[0070]

[0071] In formula (8), N ij (t) represents the set of all power generation units within this area that have communication links with the j-th power generation unit in the i-th area of the distribution system;

[0072] Taking the j-th traditional energy power generation unit and the k-th new energy power generation unit in the i-th area of the distribution system as an example, the solution steps of the operation collaborative online optimization strategy for the power generation units in each area of the distribution system that integrates the distributed online gradient optimization technology and the projection operator are as follows:

[0073] S5.1: Set the optimization time index to 0, that is, t = 0; Initialize the local optimization variables and auxiliary variables:

[0074]

[0075] μ ik z(0)=1 ik z(0)=1

[0076] where z ij (t) is an auxiliary vector variable used to estimate and compensate for the imbalance error caused by the directed communication topology.

[0077] S5.2: Obtain the out-degree neighbor set N ij (t), N ik (t) and determine the communication weight coefficients {a ijk (t)}.

[0078] S5.3: Read the locally optimized variables and auxiliary variable information μ ij (t), μ ik (t) and z ij (t), z ik (t) stored after the previous update.

[0079] S5.4: Send the μ ij (t), μ ik (t) and z ij (t), z ik (t) read in step S5.3 to the power generation units in the out-degree neighbor set.

[0080] S5.5: Receive the locally optimized information μ iy (t), μ io (t) and z iy (t), z io (t) sent by the other yth traditional energy power generation units and oth new energy power generation units within this area.

[0081] S5.6: Obtain the delayed optimization feedback information obtained by the calculation unit

[0082] S5.7: Generate the updated optimization step size α ij α(t + 1) = 1 / (t + 1), α ik α(t + 1) = 1 / (t + 1); α ij α(t + 1), α ik α(t + 1) can also be generated according to other rules, but it must satisfy the decreasing constraint.

[0083] S5.8: Based on the μ iy (t), μ io (t) and z iy (t), z io (t) received in step S5.5 and the Optimize the local dual auxiliary variable x ij (t + 1), x ik (t + 1) for update operation, and the update basis is as follows:

[0084]

[0085] S5.9: Based on x ij (t + 1), x ik (t + 1), perform a projection operation to solve the local optimization variable μ ij (t + 1), μ ik (t + 1), and the update basis is as follows:

[0086]

[0087] S5.10: Based on z ij (t), z ik (t), update the auxiliary variable z ij (t + 1), z ik (t + 1) for update operation, and the update basis is as follows:

[0088]

[0089] S5.11: Based on the local optimization variable μ ij (t + 1), the power generation unit p of the jth traditional energy source in the ith region can be calculated ij (t + 1), and the calculation basis is as follows:

[0090]

[0091] S5.12: Combine the power generation unit p ij (t + 1) of the jth traditional energy source in the ith region obtained in step S3.11 with the measured uncertain new energy input r ik (t + 1) of the kth new energy power generation unit in the ith region to calculate the optimization gradient for solution, and the calculation basis is as follows:

[0092]

[0093] S5.13: Store μ ij (t), μ ik (t) and z ij (t), z ik (t), and determine the condition t < T. If it is satisfied, repeat step S3.2; if not, end all steps.

[0094] The new - energy consumption scheduling method for a multi - area distribution system considering feedback delay described in the present invention, when adopting the above - mentioned technical solution, has the following technical effects compared with the prior art:

[0095] 1. By constructing a two - layer communication topology for the multi - area distribution system, the scalability of the new - energy consumption scheduling method for the multi - area distribution system can be effectively enhanced, and at the same time, the ability to absorb uncertain new energy is enhanced from both the inter - area and intra - area perspectives;

[0096] 2. By designing an operation coordination online optimization strategy for the regional - level and intra - regional power generation units facing the two - layer communication topology, which integrates distributed online gradient optimization technology and projection operators, the local power generation optimization solutions of all power generation units in each region can be calculated based on the distributed time - varying uncertain new - energy power generation, and the operation cost of the distribution network can be reduced as a whole. Brief Description of the Drawings

[0097] Figure 1 is a schematic diagram of the method flow proposed by the present invention;

[0098] Figure 2 is an iterative flow chart of the collaborative scheduling solution method between regions and within regions of the distribution system considering feedback delay;

[0099] Figure 3 is a physical connection structure and regional division result diagram of the constructed simulation system;

[0100] Figure 4 is a two - layer communication network topology diagram in the constructed simulation system;

[0101] Figure 5 is a consistency convergence result diagram of the local optimization dual variables of 5 traditional - energy power generation units in the constructed simulation system;

[0102] Figure 6 is an evolutionary calculation result diagram of the local power generation online optimization solutions of 5 traditional - energy power generation units in the constructed simulation system;

[0103] Figure 7 is a satisfaction situation of the new - energy power generation - traditional - energy power generation - global power load balance requirement and a consumption result diagram of the time - varying new - energy power generation;

[0104] Figure 8 is a performance result diagram of the proposed new - energy consumption scheduling method for a multi - area distribution system considering feedback delay in terms of the Regret index. Detailed Embodiments

[0105] The present invention will be further described below with reference to the drawings.

[0106] The present invention proposes a new - energy consumption scheduling method for a multi - area distribution system considering feedback delay, which is divided into 5 steps as shown below, and is specifically described as follows: Figure 1 as follows:

[0107] S1: Construct a two - layer communication topology for the multi - area distribution system. The first - layer topology consists of the inherent communication links between the traditional energy generation units and new - energy generation units within each area. The second - layer topology consists of the nodes with the highest out - degree in each area, and the second - layer topology is fully connected;

[0108] S2: Establish an online operating cost model for the multi - area distribution system considering time - varying new - energy generation based on the two - layer communication topology. It is necessary to minimize the cost of the entire distribution system within a specific time. At the same time, it must be ensured that the power generation of the traditional energy generation units does not exceed their capacity limits and meets the requirements of new - energy generation - traditional energy generation - global power load balance;

[0109] S3: Establish a Lagrangian dual unconstrained optimization model corresponding to the online operation optimization problem of the multi - area distribution system considering time - varying new - energy generation based on the two - layer communication topology to meet the requirements of new - energy generation - traditional energy generation - global power load balance;

[0110] S4: Construct a new - energy consumption scheduling strategy for the multi - area distribution system considering feedback delay for the second - layer topology composed of the nodes with the highest out - degree in each area, give the distribution optimization solutions for each area in the distribution system, and achieve new - energy consumption at the regional level;

[0111] S5: For the first - layer topology composed of the inherent communication links between the traditional energy generation units and new - energy generation units within each area, construct an online operation collaborative optimization strategy for the generation units in each area of the distribution system that integrates distributed online gradient optimization technology and projection operators, and give the local generation optimization solutions for all generation units in each area of the distribution system.

[0112] The present invention first constructs a two - layer communication topology for the multi - area distribution system. The specific construction steps are as follows:

[0113] S1.1: Obtain the communication link information of the entire distribution system;

[0114] S1.2: Perform regional division of the distribution system. The division basis is:

[0115] 1) When the manager has a clear division requirement, divide according to the requirement;

[0116] 2) When the manager has no clear division requirement, divide all the generation units in the distribution system into n m regions in an approximately average manner, and the generation units within each region need to be as similar as possible;

[0117] S1.3: Based on the divided regions, cut off the communication links between any two power generation units within the regions;

[0118] S1.4: According to the remaining communication links within each region, construct the communication topology of each region, and merge the topologies of the obtained n m regions to obtain the first-layer topology composed of the inherent communication links between the traditional energy power generation units and new energy power generation units within each region;

[0119] S1.5: For the topologies of the obtained n m regions, calculate the out-degree of each power generation unit within the region, that is, the number of links for sending information outward, and select the n m traditional energy power generation units with the highest out-degree, and connect these power generation units pairwise to obtain the second-layer topology composed of the nodes with the highest out-degree in each region;

[0120] The present invention constructs an online operation cost model of a multi-region distribution system considering time-varying new energy power generation based on a two-layer communication topology, as specifically shown in formula (1):

[0121]

[0122] In formula (1), t is the optimization period, and T is the number of specific total optimization periods; n m is the number of regions obtained by dividing the distribution system through step S1.2; the i-th region in the distribution system contains power generation units, including traditional energy power generation units and new energy power generation units; p ij (t) is the power generation amount of the j-th traditional energy power generation unit in the i-th region of the distribution system during the optimization period t; r ij (t) is the uncertain new energy input of the j-th new energy power generation unit in the i-th region of the distribution system during the optimization period t; is the power generation cost function of the j-th traditional energy power generation unit in the i-th region of the distribution system; is the uncertain new energy usage cost function of the j-th new energy power generation unit in the i-th region of the distribution system;

[0123] In formula (1), the power generation cost function of the j-th traditional energy power generation unit in the i-th region of the distribution system The specific expression is as shown in formula (2):

[0124]

[0125] In formula (2), a ij (t), b ij (t) and cij The quadratic term, linear term, and constant term coefficients of the power generation cost function of the j-th traditional energy power generation unit in the i-th area of the power distribution system are (t);

[0126] In formula (1), the cost of using uncertain new energy of the j-th new energy power generation unit in the i-th area of the power distribution system Specifically, as shown in formula (3):

[0127]

[0128] In formula (3), k ij is the loss cost coefficient of using uncertain new energy;

[0129] To ensure that the power generation of traditional energy power generation units in each area does not exceed their capacity limits, that is, for any traditional energy power generation unit, its p ij (t) needs to satisfy:

[0130]

[0131] In formula (4), is the upper limit of the power generation of the j-th traditional energy power generation unit in the i-th area;

[0132] In addition, the entire power distribution system also needs to meet the new energy power generation - traditional energy power generation - global electricity load balance requirements considering integrated line losses, that is

[0133]

[0134] In formula (5), D(t) is the time-varying global electricity demand of the power distribution system, and are the power distribution transmission line losses of the j-th traditional energy power generation unit and the j-th new energy power generation unit in the i-th area, and are the corresponding power distribution transmission line loss coefficients;

[0135] The present invention establishes a Lagrangian dual unconstrained optimization model corresponding to the online operation optimization problem of a multi-area power distribution system considering time-varying new energy power generation based on a two-layer communication topology to meet the new energy power generation - traditional energy power generation - global electricity load balance requirements, specifically as follows:

[0136] At the optimization time period t, the following Lagrangian function is defined for all traditional energy power generation units:

[0137]

[0138] In formula (6), μ ij(t) is the Lagrangian dual variable related to new energy power generation - traditional energy power generation - global power load balance requirements; It is defined as the local power demand of the j-th traditional energy power generation unit in the i-th region. At the optimization time period t, the following Lagrangian function is defined for all new energy power generation units:

[0139]

[0140] In formula (7), It is defined as the local power demand of the j-th new energy power generation unit in the i-th region. Combining with formula (6), and needs to satisfy:

[0141]

[0142] It should be noted that for the power generation units with the highest out-degree in each region of the second-layer topology, their local power demands or are not zero, and the local power demands of all other power generation units or are all zero.

[0143] Based on formulas (6) and (7), at the optimization time period t, the following Lagrangian dual unconstrained optimization model corresponding to the online operation optimization problem of the multi-region distribution system considering time-varying new energy power generation based on the double-layer communication topology can be constructed:

[0144]

[0145] The present invention constructs a new energy consumption scheduling strategy for a multi-region distribution system considering feedback delay for the second-layer topology composed of the traditional energy power generation units with the highest out-degree in each region, gives the distribution optimization solutions for each region in the distribution system, and realizes the new energy consumption at the regional level, specifically as follows:

[0146] Assume that the first traditional energy power generation unit in each region of the distribution system is the node with the highest out-degree in that region. Taking the i-th region as an example, the implementation steps of the new energy consumption scheduling strategy for the multi-region distribution system considering feedback delay are specifically as follows:

[0147] S4.1: Set the optimization time index to 0, that is, t = 0; initialize the local optimization variables and auxiliary variables:

[0148]

[0149] S4.2: Read μ i1 (t) stored after the last update between regions and within regions.

[0150] S4.3: Send the locally optimized information μ i1 (t) read in step S4.2 to all the remaining generating units in the second-layer topology.

[0151] S4.4: Receive the locally optimized information μ k1 (t) sent by all the remaining generating units in the second-layer topology.

[0152] S4.5: Obtain the optimized feedback information with delay obtained by the computing unit τ i1 Calculate the time delay for the feedback information of the first generating unit in the i-th area of the distribution system.

[0153] S4.6: Generate an updated optimization step size α i1 (t) = 1 / (t); α i1 (t) can also be generated according to other rules, but it must meet the decreasing requirement.

[0154] S4.7: Based on μ k1 (t) received in step S4.4 and the optimized feedback information with delay obtained in step S4.5 perform an update operation on the locally optimized dual auxiliary variable x i1 (t + 1), and the update basis is as follows:

[0155]

[0156] S4.8: Perform a projection operation on x i1 (t + 1) calculated in step S4.7 to solve the locally optimized variable μ i1 (t + 1), and the update basis is as follows:

[0157]

[0158] where is the projection operation on the constraint set, and the location of the constraint set is:

[0159]

[0160] S4.9: Based on the locally optimized variable μ i1 (t + 1) obtained in step S4.8, the locally optimized generated power p i1 (t + 1) of the first conventional energy generating unit in the i-th area of the distribution system can be calculated, and the calculation basis is as follows:

[0161]

[0162] The min-max operation in formula (13) can ensure that the power generation of traditional energy power generation units does not exceed their capacity limits.

[0163] S4.10: Obtain the local power demand of this area at the current moment according to the measurement information

[0164] S4.11: According to the locally optimized power generation p i1 (t + 1) obtained in step S4.9 and the local power demand of this area obtained in S4.10 Calculate and solve the optimization gradient :

[0165]

[0166] S4.12: Store the updated μ i1 (t + 1), and determine the condition t < T. If it is satisfied, repeat step S4.2; if not, end all steps.

[0167] The present invention constructs an operation collaborative online optimization strategy for power generation units in each area of the distribution system by integrating distributed online gradient optimization technology and projection operators for the first-layer topology composed of the inherent communication links between traditional energy power generation units and new energy power generation units in each area, and gives the local power generation optimization solutions for all power generation units in each area of the distribution system, as follows:

[0168] For the first-layer local topology composed of the inherent communication links between traditional energy power generation units and new energy power generation units in the i-th area, define the corresponding communication weight matrix A i (t) = [a ijk (t)]:

[0169]

[0170] In formula (8), N ij (t) represents the set of all power generation units within this area that have communication links with the j-th power generation unit in the i-th area of the distribution system;

[0171] Taking the j-th traditional energy power generation unit and the k-th new energy power generation unit in the i-th area of the distribution system as an example, the solution steps of the operation collaborative online optimization strategy for power generation units in each area of the distribution system by integrating distributed online gradient optimization technology and projection operators are as follows:

[0172] S5.1: Set the optimization time index to 0, i.e., t = 0; initialize the local optimization variables and auxiliary variables:

[0173]

[0174] μ ik z(0)=1 ik μ(0)=1

[0175] where z ij (t) is an auxiliary vector variable used to estimate and compensate for the imbalance error caused by the directed communication topology.

[0176] S5.2: Obtain the out-degree neighbor set N ij (t), N ik (t) and determine the communication weight coefficient {a ijk (t)}.

[0177] S5.3: Read the locally optimized variables and auxiliary variable information μ ij (t), μ ik (t) and z ij (t), z ik (t) stored after the previous update.

[0178] S5.4: Send the μ ij (t), μ ik (t) and z ij (t), z ik (t) read in step S5.3 to the power generation units in the out-degree neighbor set.

[0179] S5.5: Receive the locally optimized information μ iy (t), μ io (t) and z iy (t), z io (t) sent by the other yth traditional energy power generation unit and the oth new energy power generation unit within this area.

[0180] S5.6: Obtain the delayed optimization feedback information obtained by the computing unit

[0181] S5.7: Generate the updated optimization step size α ij α(t + 1) = 1 / (t + 1), α ik α(t + 1) = 1 / (t + 1); α ij (t + 1), α ik (t + 1) can also be generated according to other rules, but it must satisfy the decreasing constraint.

[0182] S5.8: Based on the μ iy (t), μ io (t) and z iy (t), z io (t) received in step S5.5 and the Optimize the local dual auxiliary variable x ij (t + 1), x ik Perform an update operation on (t + 1), and the update basis is as follows:

[0183]

[0184] S5.9: Based on x obtained in step S5.8 ij (t + 1), x ik Perform a projection operation on (t + 1) to solve the local optimization variable μ ij (t + 1), μ ik (t + 1), and the update basis is as follows:

[0185]

[0186] S5.10: Based on z received in step S5.5 ij (t), z ik Perform an update operation on the auxiliary variable z ij (t + 1), z ik (t + 1), and the update basis is as follows:

[0187]

[0188] S5.11: Based on the local optimization variable μ obtained in step S5.9 ij (t + 1), the power generation unit p of the jth traditional energy source in the ith region can be calculated ij (t + 1), and the calculation basis is as follows:

[0189]

[0190] S5.12: The power generation unit p of the jth traditional energy source in the ith region obtained in step S3.11 ij (t + 1) and the measured uncertain new energy input r of the kth new energy power generation unit in the ith region ik (t + 1) are used to calculate the optimization gradient The calculation is as follows:

[0191]

[0192] S5.13: Store μ ij (t), μ ik (t) and z ij (t), z ik (t), and determine the condition t < T. If it is satisfied, repeat step S3.2; if not, end all steps.

[0193] The following gives a simulation example of a simulation system based on the IEEE 14-bus system:

[0194] The simulation system based on the IEEE 14-bus system includes 5 traditional energy units (nodes 1, 2, 3, 6, 8) and 9 new energy generation units (nodes 4, 5, 7, 9, 10, 11, 12, 13, 14). The physical connection structure of these units is as Figure 3 shown. As Figure 3 shown, the simulation system is divided into 5 regions in total. The first region includes nodes 1, 11, 12; the second region includes nodes 2, 5; the third region includes nodes 3, 4, 9; the fourth region includes nodes 6, 11, 12; the fifth region includes nodes 7, 8, 10. The second-layer topology is composed of nodes 1, 2, 3, 6, 8. The topologies of the first and second layers are as Figure 4 shown. The usage cost coefficient of all new energy generation units is set to k ij = 5. The parameters a ij (t), b ij (t), c ij (t) and the upper limit of power generation capacity of each traditional energy generation unit are selected as shown in Table 1.

[0195] Table 1

[0196]

[0197] For all traditional energy generation units in the second-layer topology and the power loss of the distribution transmission line For the uncertain power input r ij (t) of all new energy generation units and the power loss of the distribution transmission line are selected as shown in Table 2.

[0198] Table 2

[0199]

[0200] Figure 5 Shows the consistency convergence result diagram of the local optimization dual variables of 5 traditional energy generation units. Figure 6 Shows the evolution calculation result diagram of the online optimization solution of the local power generation of 5 traditional energy generation units. Figure 7 Shows the satisfaction of the new energy generation - traditional energy generation - global power load balance requirement and the consumption result of the time-varying new energy power generation. Figure 8 Shows the performance of the proposed new energy consumption scheduling method for multi-region distribution systems considering feedback delay in terms of the Regret metric.

[0201] The present invention takes into account the feedback delay, time-varying new energy consumption demands, and time-varying electricity consumption demands in a multi-region power distribution system. By constructing a two-layer communication topology for the multi-region power distribution system, the scalability of the new energy consumption scheduling method for the multi-region power distribution system is effectively enhanced, and at the same time, the ability to absorb uncertain new energy is enhanced from both inter-region and intra-region perspectives. By designing an operation collaborative online optimization strategy for the regional-level and intra-regional power generation units facing the two-layer communication topology, which integrates the distributed online gradient optimization technology and the projection operator, the local power generation optimization solutions of all power generation units in each region can be calculated based on the distributed time-varying uncertain new energy power generation, and the operating cost of the distribution network is overall reduced.

[0202] The above are only specific embodiments of the present invention and are not used to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for accommodating new energy in a multi-regional power distribution system considering feedback delay, characterized in that: The following steps are involved: S1: Construct a two-layer communication topology for a multi-regional power distribution system. The first layer of topology consists of inherent communication links between traditional energy generation units and new energy generation units in each region, and the second layer of topology consists of nodes with the highest out-degree in each region. The second layer of topology is fully connected. S2: Establish an online operation cost model for a multi-regional distribution system based on a two-layer communication topology and taking into account time-varying renewable energy generation. It is necessary to minimize the cost of the entire distribution system within a specific time, while ensuring that the power generation of traditional energy generation units does not exceed their capacity limit and meets the requirements of renewable energy generation-traditional energy generation-global power load balance; S3: Establish a Lagrangian dual unconstrained optimization model corresponding to the online operation optimization problem of a multi-regional distribution system considering time-varying renewable energy generation based on a two-layer communication topology to meet the requirements of renewable energy generation-traditional energy generation-global power load balance; S4: Construct a new energy consumption dispatching strategy for a multi-regional distribution system considering feedback delay for the second-layer topology composed of the nodes with the highest out-and-in degrees in each region, and provide the distribution optimization solution for each region in the distribution system to realize new energy consumption at the regional level; S5: For the first-layer topology composed of inherent communication links between traditional energy power generation units and new energy power generation units in each region, a coordinated online optimization strategy for the operation of power generation units in each region of the distribution system is constructed by integrating distributed online gradient optimization technology and projection operator, and the local power generation optimization solution of all power generation units in each region of the distribution system is given.

2. According to claim 1, a method for accommodating new energy in a multi-regional power distribution system considering feedback delay is characterized in that: The two-layer communication topology for the multi-area power distribution system described in step S1 has the following specific steps: S1.1: Obtain the communication link information of the entire power distribution system; S1.2: Perform distribution system zoning based on: 1) When managers have clear division requirements, division should be carried out according to the requirements; 2) When the manager does not clearly divide the demand, all the generating units in the distribution system are divided into n m The power generation units in each area need to be as close as possible; S1.3: Based on the divided areas, the communication link between the power generation units in any two areas is cut off; S1.4: Construct the communication topology of each area based on the communication links retained in each area, and convert the obtained n m The topology in each region is merged to obtain the first layer topology consisting of inherent communication links between traditional energy generation units and new energy generation units in each region; S1.5: For the obtained n m The topology in the region is calculated, and the out-degree of each power generation unit in the region is calculated, that is, the number of links that send information outward, and the n with the highest out-degree are selected. m Traditional energy generation units are connected in pairs to obtain the second-layer topology consisting of the nodes with the highest out-degree in each area.

3. According to a method for accommodating new energy in a multi-regional power distribution system considering feedback delay according to claim 1, it is characterized in that: The online operation cost model of the multi-regional distribution system considering time-varying renewable energy generation based on the double-layer communication topology described in step S2 is specifically shown in formula (1): In formula (1), i is the optimization period, T is the total number of specific optimization periods; n m is the number of areas divided by the power distribution system in step S1.2; the i-th area in the power distribution system contains Power generation units, including Traditional energy generation units and New energy power generation units; ij (t) is the power generation of the jth traditional energy generation unit in the i-th area of ​​the distribution system in the optimization period t; r ij (t) is the uncertain new energy input of the jth new energy generation unit in the i-th area of ​​the distribution system in the optimization period t; is the power generation cost function of the jth traditional energy generation unit in the i-th area of ​​the distribution system; is the uncertainty renewable energy use cost function of the j-th renewable energy generation unit in the i-th region of the distribution system; In formula (1), the power generation cost function of the jth traditional energy generation unit in the i-th area of ​​the distribution system is The specific expression is shown in formula (2): In formula (2), a ij (t),b ij (t) and c ij (t) is the coefficient of the quadratic term, linear term and constant term of the power generation cost function of the jth traditional energy power generation unit in the i-th area of ​​the distribution system; In formula (1), the uncertainty cost of renewable energy use for the jth renewable energy generation unit in the i-th region of the distribution system is The specific formula is as shown in formula (3): In formula (3), k ij is the loss cost coefficient of using uncertain new energy; To ensure that the power generation of traditional energy generation units in each area does not exceed its capacity limit, that is, for any traditional energy generation unit, its p ij (t) Must meet the following requirements: In formula (4), is the upper limit of power generation of the jth traditional energy power generation unit in the i-th region; In addition, the entire distribution system needs to meet the requirements of new energy generation-traditional energy generation-global power load balance with integrated line loss, that is, In formula (5), D(t) is the time-varying global power demand of the distribution system. and is the distribution transmission line loss of the j-th traditional energy generation unit and the j-th new energy generation unit in the i-th region, and is the corresponding distribution transmission line loss coefficient.

4. The method for accommodating new energy in a multi-regional power distribution system considering feedback delay according to claim 1, characterized in that: The establishment of the Lagrangian dual unconstrained optimization model corresponding to the online operation optimization problem of the multi-regional distribution system considering time-varying renewable energy generation based on the double-layer communication topology described in step S3 is used to meet the requirements of renewable energy generation-traditional energy generation-global power load balance, which is as follows: In the optimization period t, the following Lagrangian function is defined for all traditional energy generation units: In formula (6), μ ij (t) is the Lagrangian dual variable related to the balance requirement between new energy generation, traditional energy generation and global electricity load; It is defined as the local electricity demand of the jth traditional energy generation unit in the i-th region; in the optimization period t, the following Lagrangian function is defined for all new energy generation units: In formula (7), Defined as the local electricity demand of the jth renewable energy generation unit in the i-th region, combined with formula (6), and Need to meet: It is worth noting that for the power generation unit with the highest outgoing degree in each region in the second-layer topology, its local electricity demand or Not 0, local power demand of all other power generation units or All are 0; Based on formulas (6) and (7), in the optimization period t, the following Lagrangian dual unconstrained optimization model corresponding to the online operation optimization problem of a multi-regional distribution system considering time-varying renewable energy generation based on a two-layer communication topology can be constructed:

5. The method for accommodating new energy in a multi-regional power distribution system considering feedback delay according to claim 1, characterized in that: Step S4 constructs a multi-regional distribution system new energy consumption scheduling strategy considering feedback delay for the second-layer topology composed of traditional energy generation units with the highest out-and-in degrees in each region, provides a distribution optimization solution for each region in the distribution system, and realizes regional-level new energy consumption, as follows: Assuming that the first traditional energy generation unit in each area of ​​the distribution system is the node with the highest out-degree in the area, taking the i-th area as an example, the execution steps of the new energy consumption scheduling strategy of the multi-area distribution system considering feedback delay are as follows: S4.1: Set the optimization time index to 0, i.e. t = 0; initialize local optimization variables and auxiliary variables: S4.2: Read the μ stored after the last inter-region and intra-region update i1 (t); S4.3: Send the local optimization information μ read in step S4.2 to all other power generation units in the second layer topology. i1 (t); S4.4: Receive local optimization information μ sent from all other power generation units in the second layer topology k1 (t); S4.5: Obtaining delayed optimization feedback information obtained by the computing unit τ i1 Calculate the time delay for the feedback information of the first generating unit in the ith area of ​​the distribution system; S4.6: Generate updated optimization step size α i1 (t) = 1 / (t); α i1 (t) It may also be generated according to other rules, but must meet the decreasing requirements; S4.7: Based on μ received in step S4.4 k1 (t) and the delayed optimization feedback information obtained in step S4.5 For local optimization of the auxiliary variable x i1 (t+1) Perform update operation, the update basis is as follows: S4.8: For x calculated in step S4.7 i1 (t+1) Perform projection operation to solve the local optimization variable μ i1 (t+1), the update is based on the following: in It is a projection operation on the constraint set. The constraint set is positioned as follows: S4.9: Based on the local optimization variable μ obtained in step S4.8 i1 (t+1), the local optimal power generation p of the first traditional energy generation unit in the i-th area of ​​the distribution system can be calculated i1 (t+1), calculated as follows: The operation of taking the smaller and the larger in formula (13) can ensure that the power generation of the traditional energy generation unit does not exceed its capacity limit; S4.10: Obtain the local electricity demand in the area at the current moment based on the measurement information S4.11: According to the local optimized power generation p obtained in step S4.9 i1 (t+1) and the local electricity demand in this area obtained in S4.10 To optimize the gradient Perform calculations to solve: S4.12: Store the updated μ i1 (t+1), and determine the condition t<T. If it is satisfied, repeat step S4.2; if it is not satisfied, end all steps.

6. A method for accommodating new energy in a multi-regional power distribution system considering feedback delay according to claim 1, characterized in that: According to step S5, for the first layer topology composed of inherent communication links between traditional energy generation units and new energy generation units in each region, an online optimization strategy for the operation coordination of generation units in each region of the distribution system integrating distributed online gradient optimization technology and projection operator is constructed, and the local power generation optimization solution of all generation units in each region of the distribution system is given, which is as follows: For the first layer of local topology consisting of inherent communication links between traditional energy generation units and new energy generation units in the i-th region, the corresponding communication weight matrix A is defined as i (t) = [a ijk (t)]: In formula (8), N ij (r) represents the set of all power generation units in the region that have communication links with the jth power generation unit in the i-th region of the power distribution system; Taking the jth traditional energy generation unit and the kth new energy generation unit in the i-th area of ​​the distribution system as an example, the solution steps of the operation collaborative online optimization strategy of the generation units in each area of ​​the distribution system integrating the distributed online gradient optimization technology and the projection operator are as follows: S5.1: Set the optimization time index to 0, i.e. t = 0; initialize local optimization variables and auxiliary variables: μ ik (0)=1,of ik (0)=1, where z ij (t) is an auxiliary vector variable used to estimate and compensate for the imbalance error caused by the directed communication topology; S5.2: Get the out-degree neighbor set N ij (t), N ik (t) and determine the communication weight coefficient {a ijk (t)}; S5.3: Read the local optimization variables and auxiliary variable information stored after the last update μ ij (t), μ ik (t) and z ij (t),z ik (t) S5.4: Send the μ read in step S5.3 to the power generation unit in the out-degree neighbor set. ij (t), μ ik (t) and z ij (t),z ik ((t); S5.5: Receive local optimization information μ sent from the yth traditional energy generation unit and the oth new energy generation unit in the region iy (t), μ io (t) and z iy (t), z io (t); S5.6: Obtaining delayed optimization feedback information obtained by the computing unit S5.7: Generate updated optimization step size α ij (t+1)=1 / (t+1),α ik (t+1)=1 / (t+1); α ij (t+1),α ik (t+1) can also be generated according to other rules, but it must satisfy the decreasing constraint; S5.8: Based on μ received in step S5.5 iy (t), μ io (t) and z iy (t),z io (t) and obtained in step S5.5 For local optimization of the auxiliary variable x ij (t+1), x ik (t+1) Perform update operation, the update basis is as follows: S5.9: Based on x obtained in step S5.8 ij (t+1), x ik (t+1) Perform projection operation to solve the local optimization variable μ ij (t+1), μ ik (t+1), the update is based on the following: S5.10: Based on z received in step S5.5 ij (t), z ik (t), for the auxiliary variable z ij (t+1), z ik (t+1) Perform update operation, the update basis is as follows: S5.11: Based on the local optimization variable μ obtained in step S5.9 ij (t+1), the jth traditional energy generation unit p in the i-th region can be calculated ij (t+1), calculated as follows: S5.12: The jth traditional energy generation unit p of the i-th region obtained in step S3.11 is ij (t+1) and the uncertainty new energy input r measured by the k-th new energy generation unit in the i-th region ik (t+1) for the optimization gradient The calculation is based on the following: S5.13: Store the updated μ in this round ij (t), μ ik (t) and z ij (t), z ik (t), and determine the condition t<T. If it is satisfied, repeat step S3.

2. If not, end all steps.

Images