CN-122026332-A - Scheduling method and system for active support type virtual power plant and electronic equipment

CN122026332ACN 122026332 ACN122026332 ACN 122026332ACN-122026332-A

Abstract

The invention relates to a scheduling method, a scheduling system and electronic equipment of an active support type virtual power plant, wherein the method comprises the steps of constructing an active support type virtual power plant system model, generating power prediction data of various resources in a subsequent period, carrying out coordinated optimization on various resources by adopting a double-layer distributed optimization algorithm of an alternating direction multiplier method based on the power prediction data, carrying out inner layer local optimization processing on consistency constraint in an area, outer layer global coordination processing on supply and demand balance constraint among areas, carrying out alternating iteration until convergence to generate an optimal scheduling plan, and issuing the optimal scheduling plan for scheduling. The method can ensure that the generated scheduling plan meets the operation boundary and economic requirements of each virtual power plant on a microscopic scale, meets the supply and demand balance conditions of the power system on a macroscopic scale, and remarkably improves the calculation efficiency and scheduling stability of the virtual power plant when dealing with large-scale renewable energy access.

Inventors

WANG HUIDONG
WEN FUSHUAN
LI LEI
YU JIALI
WANG JIAYING
LU PENGFEI
JIANG CHI
NI LINNA
HUANG RONGGUO
SUN GANG
CHEN YUHAO
WU YUEBO
WANG ZIRONG
ZHANG JINGCHEN
CHENG YING

Assignees

国网浙江省电力有限公司营销服务中心
国网浙江省电力有限公司

Dates

Publication Date: 20260512
Application Date: 20260409

Claims (10)

1. A method for scheduling an actively supported virtual power plant, the method comprising: Constructing a system model of an active support type virtual power plant, and respectively generating power prediction data of a subsequent period according to renewable resources, conventional unit resources, energy storage resources and loads in the system model; based on the power prediction data, performing coordinated optimization on the various resources by adopting a double-layer distributed optimization algorithm based on an alternate direction multiplier method, wherein the double-layer distributed optimization algorithm comprises an inner-layer local optimization process constrained in a processing region and an outer-layer global coordination process constrained between the processing regions; the inner layer local optimization process uses a first Lagrange multiplier for representing marginal coordination price in a region, calculates a local scheduling scheme meeting consistency constraint based on the power data, and updates the first Lagrange multiplier based on consistency deviation of the local scheduling scheme; the outer global coordination process uses a second Lagrange multiplier for representing inter-region marginal adjustment prices, and updates the second Lagrange multiplier based on inter-region supply-demand deviations generated by the local scheduling scheme; Alternately carrying out the inner layer local optimization process and the outer layer global coordination process until the local scheduling scheme and the first Lagrangian multiplier and the second Lagrangian multiplier are converged to generate an optimal scheduling plan of each virtual power plant; and generating a control instruction according to the optimal scheduling plan, and performing rolling scheduling on each virtual power plant.
2. The scheduling method of an active support type virtual power plant according to claim 1, wherein power prediction data of a subsequent period is generated for renewable resources, conventional unit resources, energy storage resources and loads in the system model respectively, specifically, the differential prediction model is configured according to respective physical characteristics and operation rules, so as to generate the power prediction data of the subsequent period.
3. The scheduling method of an actively supported virtual power plant of claim 1, wherein calculating a local scheduling scheme that satisfies a consistency constraint based on the power data, and updating the first lagrangian multiplier based on a consistency deviation of the local scheduling scheme, comprises: Constructing an inner Lagrange function, and calculating the sum of the operation cost and consistency deviation penalty cost of all the virtual power plants according to the first Lagrange multiplier, the operation cost function of each virtual power plant and the penalty term of the consistency constraint in the region by the inner Lagrange function; On the premise of meeting the internal operation constraint of each virtual power plant, aiming at minimizing the inner Lagrangian function, and solving the power variable of each virtual power plant; And updating the first Lagrangian multiplier according to the deviation of the power variable of each virtual power plant and the regional coordination target.
4. The method for scheduling an actively supported virtual power plant according to claim 1, wherein updating the second lagrangian multiplier based on the inter-region supply-to-demand bias generated by the local scheduling scheme comprises: constructing an outer Lagrangian function, and calculating the punishment cost of the system supply and demand deviation according to the punishment items of the second Lagrangian multiplier, the system supply and demand deviation and the supply and demand balance constraint among the areas by the outer Lagrangian function; acquiring the consistency variable of the unconstrained region based on the local scheduling scheme, and projecting the consistency variable of the unconstrained region through supply-demand balance constraint with the aim of minimizing the outer Lagrange function to obtain the consistency variable meeting the supply-demand balance; and updating the second Lagrangian multiplier according to the deviation of the sum of the consistency variables and the total system demand.
5. The method for dispatching an active support type virtual power plant according to claim 1, wherein the inner layer local optimization process and the outer layer global coordination process are alternately performed until the local dispatching scheme and the first lagrangian multiplier and the second lagrangian multiplier converge, specifically comprising: After each round of alternate updating, calculating a consistency residual error representing the violation degree of the consistency constraint in the area, a balance residual error representing the violation degree of the supply and demand balance constraint of the whole system and a dual residual error representing the change degree of the iteration variable; judging whether the consistency residual error, the balance residual error and the dual residual error are smaller than a preset threshold value, if yes, the judging algorithm converges, and if not, the next round of alternate updating is continued.
6. The method for scheduling an actively supported virtual power plant of claim 1, further comprising: In a rolling period, real-time monitoring actual operation data of each virtual power plant, and calculating power deviation between the actual operation data and the optimal scheduling plan; and compensating the power deviation in real time by utilizing internal resources of the virtual power plant.
7. The method for dispatching an actively supported virtual power plant of claim 6, wherein said power bias is compensated in real time using internal resources of the virtual power plant, comprising: Calculating comprehensive allocation weights of the virtual power plants based on the prediction precision and the adjustment capacity of the virtual power plants; distributing the total power deviation of the system to each virtual power plant according to the comprehensive distribution weight; and each virtual power plant performs power consumption or supplement according to the allocated deviation task.
8. The method for dispatching an active support type virtual power plant according to claim 7, wherein the comprehensive allocation weight is calculated by: Calculating a corresponding prediction precision weight according to the historical prediction accuracy of each virtual power plant, wherein the lower the historical prediction accuracy is, the larger the prediction precision weight is; Calculating corresponding adjusting capacity weights according to the remaining adjusting capacity of each virtual power plant, wherein the larger the remaining adjusting capacity is, the larger the adjusting capacity weights are; and carrying out weighted summation on the prediction precision weight and the adjustment capability weight to obtain the comprehensive allocation weight.
9. A dispatch system for an actively supported virtual power plant, comprising: The prediction module is used for constructing a system model of the active support type virtual power plant and respectively generating power prediction data of a subsequent period according to renewable resources, conventional unit resources, energy storage resources and loads in the system model; the optimization solving module is used for carrying out coordination optimization on the various resources by adopting a double-layer distributed optimization algorithm based on an alternate direction multiplier method based on the power prediction data, wherein the double-layer distributed optimization algorithm comprises an inner-layer local optimization process constrained in a processing area and an outer-layer global coordination process constrained between the processing areas; the inner layer local optimization process uses a first Lagrange multiplier for representing marginal coordination price in a region, calculates a local scheduling scheme meeting consistency constraint based on the power data, and updates the first Lagrange multiplier based on consistency deviation of the local scheduling scheme; the outer global coordination process uses a second Lagrange multiplier for representing inter-region marginal adjustment prices, and updates the second Lagrange multiplier based on inter-region supply-demand deviations generated by the local scheduling scheme; The optimization solving module is further used for alternately performing the inner layer local optimization process and the outer layer global coordination process until the local scheduling scheme and the first Lagrangian multiplier and the second Lagrangian multiplier converge to generate an optimal scheduling plan of each virtual power plant; and the scheduling execution module is used for generating a control instruction according to the optimal scheduling plan and performing rolling scheduling on each virtual power plant.
10. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the method of any one of claims 1 to 8 when the program is executed by the processor.

Description

Scheduling method and system for active support type virtual power plant and electronic equipment Technical Field The invention belongs to the technical field of power dispatching, and particularly relates to a dispatching method, a dispatching system and electronic equipment of an active support type virtual power plant. Background The virtual power plant is used as an emerging distributed energy management technology, can uniformly aggregate and coordinate and control multiple types of energy resources with dispersed geographic positions, and plays an important role in wide area resource network coordination. The virtual power plant technology can realize the aggregation and unified scheduling of distributed resources, improve the resource utilization efficiency, enhance the system regulation capability and provide an effective way for solving the uncertainty problem caused by the large-scale grid connection of renewable energy sources. However, the intermittent and random nature of renewable energy sources presents unprecedented challenges to the safe and stable operation of power systems. Under the wide scale, the superposition of wind power and photovoltaic output uncertainty, load prediction error, equipment faults and other factors makes the traditional deterministic scheduling method difficult to effectively cope with complex and changeable operating environments. In recent years, scholars at home and abroad have made a great deal of research in the field of virtual power plant optimization scheduling. The existing method is mainly divided into three types of centralized optimization, layered optimization and distributed optimization. The centralized optimization method can obtain the global optimal solution, but faces the problems of high computational complexity, heavy communication load and the like, the distributed optimization method has good expandability and privacy protection characteristics, but has difficulty in processing global coupling constraint, and the hierarchical optimization method reduces the complexity through upper and lower layer decoupling, but can reduce convergence performance due to interlayer information transfer. Therefore, in the prior art, a virtual power plant scheduling scheme capable of effectively decoupling intra-area constraint and inter-area constraint and realizing rapid, stable coordination and optimization of wide area resources is needed. Disclosure of Invention It is therefore an object of the present invention to at least solve one or more of the above-mentioned problems occurring in the prior art, in other words, to provide a scheduling method, system and electronic device for an actively supported virtual power plant that meets one or more of the above-mentioned needs. In order to achieve the aim of the invention, the invention adopts the following technical scheme: in a first aspect, the present invention provides a method for scheduling an actively supported virtual power plant, the method comprising: constructing a system model of the active support type virtual power plant, and respectively generating power prediction data of a subsequent period according to renewable resources, conventional unit resources, energy storage resources and loads in the system model; Based on power prediction data, a double-layer distributed optimization algorithm based on an alternate direction multiplier method is adopted to coordinate and optimize various resources, wherein the double-layer distributed optimization algorithm comprises an inner-layer local optimization process constrained in a processing area and an outer-layer global coordination process constrained between the processing areas; The inner layer local optimization process uses a first Lagrange multiplier for representing marginal coordination price in the region, calculates a local scheduling scheme meeting consistency constraint based on the power data, and updates the first Lagrange multiplier based on consistency deviation of the local scheduling scheme; the outer global coordination process uses a second Lagrange multiplier for representing the inter-region marginal adjustment price, and updates the second Lagrange multiplier based on the inter-region supply and demand deviation generated by the local scheduling scheme; Alternately performing an inner layer local optimization process and an outer layer global coordination process until the local scheduling scheme and the first Lagrangian multiplier and the second Lagrangian multiplier are converged to generate an optimal scheduling plan of each virtual power plant; and generating a control instruction according to the optimal scheduling plan, and performing rolling scheduling on each virtual power plant. As a preferred embodiment, the power prediction data of the subsequent period is generated for the renewable resources, the conventional unit resources, the energy storage resources and the load in the system model respectively, specifically, the di