CN-121984008-A - Virtual power plant distributed scheduling optimization method and system oriented to zero-carbon intelligent park availability driving

CN121984008ACN 121984008 ACN121984008 ACN 121984008ACN-121984008-A

Abstract

The invention belongs to the technical field of energy scheduling optimization, and particularly relates to a virtual power plant distributed scheduling optimization method and system for zero-carbon intelligent park availability driving. The invention describes node power and load randomness by using an opportunity constraint model, and adopts a semi-invariant and quantile approximation technology to convert probability constraint into a deterministic form capable of being rapidly solved. On the basis, a utility function integrating availability, energy supply quality and economy is constructed, a two-stage distributed strategy of 'neighborhood energy borrowing-collaborative refinement' is designed, and overload relief and load balancing are realized.

Inventors

HU JINPING
CHEN QINGSONG
LI RUNLIN
PENG SHIFEI
ZHANG BO
LI BAIPING
XU JINLIN

Assignees

云南华电金沙江中游水电开发有限公司
徐锦林

Dates

Publication Date: 20260505
Application Date: 20251222

Claims (8)

1. A virtual power plant distributed scheduling optimization method oriented to zero-carbon intelligent park availability driving is characterized by comprising the following steps: s1, establishing first-order linear mapping of node voltage and line power flow to node net injection power based on park power distribution network data; s2, uniformly representing the conventional distributed power supply output, energy storage charge and discharge power, renewable output, random load, demand response reduction power and unsatisfied power as net injection power of each node; s3, constructing comprehensive utility based on source side availability, function quality indexes and source side marginal energy supply cost, and generating the comprehensive utility as directional transfer cost of line flow; S4, taking voltage out-of-limit and line thermal stability out-of-limit as opportunity constraints, carrying out first-order linearization risk quantity at an uncertain vector mean value, and obtaining score sites by Cornish-Fisher quantile approximation based on semi-invariants, so that the opportunity constraints are converted into deterministic linear inequality; S5, in a safe and feasible region limited by opportunity constraint, combining minimized unsatisfied penalty, running economy and directional transfer cost, constructing a linear/second-order cone resolvable optimization model, and applying power balance and running boundary constraint; S6, solving an optimization model by adopting a two-stage distributed strategy of neighborhood energy borrowing-collaborative refinement, wherein the first stage rapidly distributes mutual power flow according to comprehensive utility in a preset neighborhood, and the second stage performs small-step replacement by combining energy storage and demand response in a local range; and S7, solving an optimization model on line with a rolling time sequence, outputting source load power, energy storage charge and discharge and interconnection power flow scheduling instructions in each period, and realizing availability-driven virtual power plant distributed scheduling optimization.
2. The method of claim 1, wherein the step of establishing a first order linear mapping of node voltage to line power flow versus node net injection power based on campus power distribution network data in S1 comprises: ; In the formula, Representing the voltage vector of the node at the time t; the line flow at the time t is represented; representing the net injection power of the node at the time t, xi, t representing the net injection power of the node i at the time t; 、 representing a relationship matrix between node voltage and power; Representing the reference voltage.
3. The method according to claim 2, wherein the step of representing the regular distributed power source output, the stored charge-discharge power, the renewable output, the random load, the demand response cut power and the unsatisfied power as the net injection power of each node in S2 comprises: ; In the formula, Representing the conventional distributed power output of the node i at the time t; 、 Respectively representing the charge/discharge power of the energy storage of the node i at the time t; the renewable energy random output of the node i at the time t is represented; representing the random load of the node i at the time t; the unsatisfied load power of the node i at the time t is represented; indicating that node i is responding to demand at time t by curtailing power.
4. The method of claim 3, wherein the method of constructing the integrated utility based on the source side availability, the function quality index, and the source side marginal energy supply cost and generating the integrated utility as the directional transfer cost of the line flow in S3 comprises: Comprehensive utility The expression is: ; In the formula, 、、 Respectively the source side availability, the function quality index and the source side marginal energy supply cost, 、、 Weight coefficients respectively representing availability, quality and cost source side availability, functional quality index and source side marginal energy supply cost; the comprehensive utility is generated into the directional transfer cost of the line tide, and the expression is as follows: ; In the formula, Representing corresponding lines The cost is transferred from node i to node j; Representing corresponding lines Is transferred to the node i; Representing utility-cost trade-off coefficients.
5. The method according to claim 4, wherein the step S4 is characterized in that the step S4 uses voltage out-of-limit and line thermal stability out-of-limit as opportunity constraints, and the step S4 is characterized in that the step S4 is performed by obtaining score points by using Cornish-Fisher quantile approximation based on semi-invariants, so that the opportunity constraints are converted into deterministic linear inequality, and the method comprises the steps of: The deterministic equivalence of the line tidal current opportunity constraint is: ; in the formula, sl represents a line Is limited by the thermal stability limit of (2); Line indicating time t An expected value of the power flow; Representation line Standard deviation of tide distribution; Representing the confidence level obtained based on the Cornish-Fisher series approximation as The lower part of the dividing point is provided with a plurality of dividing points, Is the upper limit of allowable out-of-limit probability.
6. The method of claim 5, wherein the step of constructing a linear/second order cone resolvable optimization model and applying power balance and operation boundary constraints in the safe feasible region defined by the opportunity constraints in S5 by combining minimizing unsatisfied penalties, operation economics and directional transfer costs comprises: the objective function expression of the optimization model is: ; In the formula, 、、 Respectively representing the outsourcing electricity price, the selling electricity price and the marginal cost of the conventional unit in the park; 、 respectively representing forward and backward tide components, J represents an objective function of an optimization model, The penalty coefficient indicating the unsatisfied power is bt, the electric power purchased to the external network at the time T, st, the electric power sold to the external network at the time T, T, the time, N, the total number of nodes and L the total number of routes.
7. The method of claim 6, wherein the step S5 of constructing a linear/second order cone resolvable optimization model in a safe feasible region defined by opportunity constraints and combining minimizing unsatisfied penalties, operational economics and directional transfer costs, and applying power balance and operational boundary constraints further comprises: The constraint conditions are as follows: ; In the formula, Representation line The actual power flow at the moment t, Representation line Is the thermal stability limit or upper transmission capacity limit of (c), And Representing the state of charge or the energy storage capacity of the energy storage device at node i at time t and time t +1 respectively, Representing a time step or time interval of the schedule, Indicating the charging efficiency of the energy storage device, Indicating the discharge efficiency of the energy storage device, Representing a lower limit of the state of charge of the energy storage device at node i, Representing an upper limit on the state of charge of the energy storage device at node i.
8. A virtual power plant distributed scheduling optimization system oriented to zero-carbon intelligent park availability driving, which is used for realizing the method of any one of claims 1-7, and is characterized by comprising a mapping module, a net injection power acquisition module, a directional transfer cost generation module, an opportunity constraint conversion module, an optimization model construction module, a model solving module and an instruction generation module; the mapping module is used for establishing first-order linear mapping of node voltage and line power flow to node net injection power based on park power distribution network data; The net injection power acquisition module is used for uniformly representing the output power, the energy storage charging and discharging power, the renewable output power, the random load, the demand response reduction power and the unsatisfied power of the conventional distributed power supply as the net injection power of each node; The directional transfer cost generation module is used for constructing comprehensive utility based on source side availability, function quality indexes and source side marginal energy supply cost and generating the comprehensive utility into directional transfer cost of line tide; The opportunity constraint conversion module is used for taking voltage out-of-limit and line thermal stability out-of-limit as opportunity constraints, carrying out first-order linearization risk quantity at an uncertain vector mean value, and obtaining a score point by adopting Cornish-Fisher quantile approximation based on a semi-invariant so as to convert the opportunity constraints into deterministic linear inequality; The optimization model construction module is used for constructing a linear/second-order cone resolvable optimization model in a safe and feasible domain defined by opportunity constraint, combining minimized unsatisfied penalty, running economy and directional transfer cost, and applying power balance and running boundary constraint; The model solving module is used for solving the optimization model by adopting a two-stage distributed strategy of neighborhood energy borrowing-collaborative refinement, wherein the first stage rapidly distributes mutual power flow according to comprehensive utility in a preset neighborhood, and the second stage performs small-step replacement by combining energy storage and demand response in a local range; the instruction generation module is used for solving the optimization model on line with a rolling time sequence, outputting source load power, energy storage charge and discharge and interconnection tide scheduling instructions of each period, and realizing the distributed scheduling optimization of the virtual power plant driven by the availability.

Description

Virtual power plant distributed scheduling optimization method and system oriented to zero-carbon intelligent park availability driving Technical Field The invention belongs to the technical field of energy scheduling optimization, and particularly relates to a virtual power plant distributed scheduling optimization method and system for zero-carbon intelligent park availability driving. Background With the centralized access of a large-scale distributed power supply and flexible load in an intelligent park, a power distribution system is evolved from a traditional radiation type passive form into an active network structure with multiple ends and multiple directions of energy flows, and the operation risks such as power flow reversal, node voltage out-of-range and output fluctuation are accompanied, and meanwhile, the space-time mismatch and local constraint triggering frequency in the park are obviously improved. Under the situation, how to realize the rapid collaborative scheduling and uncertainty management of park Virtual Power Plants (VPPs) under the condition of high-proportion new energy sources and balance between service reliability, economy and fairness among nodes has become a key issue in the field of intelligent power distribution operation. Aiming at the operation uncertainty caused by high-permeability renewable output, the existing research widely adopts frames such as Monte Carlo sample simulation, random planning based on scenes, opportunistic constraint optimization and the like. The first two types of methods depend on a large number of samples or scene sets, are heavy in calculation burden and difficult to meet the aging requirement of online rolling scheduling, and the probability constraint is transferred into a deterministic form by the aid of series/moment expansion in the problems of optimal power flow and the like by means of opportunity constraint, so that solving cost is remarkably reduced. However, such equivalences often introduce high-order nonlinear inequalities involving decision variables, leading to increased model complexity and increased difficulty in distributed implementation. In addition, although the decomposition-coordination type distributed scheduling oriented to the VPP can relieve the computational bottleneck of centralized solution, there is still room for improvement on the robustness of high-frequency exchange and convergence of the full-quantity state. Therefore, there is a need for an availability-driven virtual power plant distributed scheduling optimization method that solves the problems of unbalanced supply and demand and operation uncertainty caused by distributed energy access in an intelligent park. Disclosure of Invention The invention provides a virtual power plant distributed scheduling optimization method and system for zero-carbon intelligent park availability driving, aiming at solving the problems of supply and demand unbalance and operation uncertainty caused by distributed energy access in an intelligent park. In order to achieve the above object, the present invention provides the following solutions: a virtual power plant distributed scheduling optimization method oriented to zero-carbon intelligent park availability driving comprises the following steps: s1, establishing first-order linear mapping of node voltage and line power flow to node net injection power based on park power distribution network data; s2, uniformly representing the conventional distributed power supply output, energy storage charge and discharge power, renewable output, random load, demand response reduction power and unsatisfied power as net injection power of each node; s3, constructing comprehensive utility based on source side availability, function quality indexes and source side marginal energy supply cost, and generating the comprehensive utility as directional transfer cost of line flow; S4, taking voltage out-of-limit and line thermal stability out-of-limit as opportunity constraints, carrying out first-order linearization risk quantity at an uncertain vector mean value, and obtaining score sites by Cornish-Fisher quantile approximation based on semi-invariants, so that the opportunity constraints are converted into deterministic linear inequality; S5, in a safe and feasible region limited by opportunity constraint, combining minimized unsatisfied penalty, running economy and directional transfer cost, constructing a linear/second-order cone resolvable optimization model, and applying power balance and running boundary constraint; S6, solving an optimization model by adopting a two-stage distributed strategy of neighborhood energy borrowing-collaborative refinement, wherein the first stage rapidly distributes mutual power flow according to comprehensive utility in a preset neighborhood, and the second stage performs small-step replacement by combining energy storage and demand response in a local range; and S7, solving an optimization model on line with a