CN-122000940-A - Power distribution network voltage control method based on running state estimation

CN122000940ACN 122000940 ACN122000940 ACN 122000940ACN-122000940-A

Abstract

The invention relates to a power distribution network voltage control method based on running state estimation, and belongs to the field of power distribution network voltage control. The method comprises the steps of firstly establishing an optimal control model aiming at minimizing voltage deviation and reactive power consumption, and defining a state space and an action space of a system. The method is characterized by introducing a graph attention network strategy and dynamically calculating the attention weight of the topological nodes of the power distribution network so as to capture the electric coupling relation among the nodes. And then, inputting the weight information and the current and predicted states of the power grid into a pipeline model prediction controller for rolling optimization, and outputting an optimal reactive power control instruction. According to the invention, by combining a graph attention mechanism and model predictive control, the controller can coordinate nodes with close electrical distances preferentially, explicit input of network topology priori knowledge is not needed, and voltage control efficiency and robustness in a photovoltaic power generation and load rapid change scene are improved remarkably.

Inventors

HUANG XINGHUA
DOU XIAOBO
DING JINGXUAN
LIN YURUN
CHEN JUNMING
WU PANPAN
LIAO YANBIN
ZHANG GONGLIN
YIN ZILI
WANG LIN
ZHUANG GUOXIAN
SU JIE
Han Youyang
FAN YUANLIANG
GAO LEYAN

Assignees

国网福建省电力有限公司电力科学研究院
国网福建省电力有限公司
国网福建省电力有限公司漳州供电公司
国网福建省电力有限公司泉州供电公司
东南大学

Dates

Publication Date: 20260508
Application Date: 20260205

Claims (10)

1. A power distribution network voltage control method based on operational state estimation, comprising: Establishing a power distribution network voltage optimization control model, wherein an objective function of the power distribution network voltage optimization control model is used for adjusting reactive power to maintain voltage within a safety range and minimize voltage deviation; analyzing and defining a state space and an action space of safe operation of the power distribution network; based on the graph attention network strategy, updating the dynamic attention weight of the power distribution network topological node in the current state in real time; And inputting the dynamic attention weight, the current running state of the power distribution network and the predicted next-moment state into a pipeline model prediction controller, and outputting the reactive power control quantity.
2. The power distribution network voltage control method based on operation state estimation according to claim 1, wherein an objective function of the power distribution network voltage optimization control model adopts an optimization method based on an obstacle function for punishing voltage deviation deviating from a safety range.
3. A method of controlling the voltage of a power distribution network based on an estimation of the operating state according to claim 1 or 2, characterized in that the objective function is expressed as follows: Wherein, the Representing nodes Is used for the voltage of the per unit value, For a differentiable barrier function, for penalizing voltage deviations from a range of 0.95, 1.05 per unit values, Is an electric generator Is provided for the reactive power output of the (c), The weight is penalized for reactive power, In order to be the number of nodes, The number of the generators is controllable; The barrier function adopts a form comprising an L1 norm, an L2 norm, a Courant-Beltrami function, a bowl function or a lamp function, and is selected according to the smoothness requirement and the punishment characteristic; The voltage control converges to a safe operation state by optimizing the objective function driving system as shown in the following formula: 。
4. The method for controlling voltage of a power distribution network based on operation state estimation according to claim 1, wherein the state s (t) of the state space at time t aggregates the measured values of the multi-dimensional power system of photovoltaic power generation power, load active and reactive demands, generator reactive output, voltage amplitude and phase angle, s (t) is expressed as follows: Wherein, the The active power generated by the photovoltaic system at the moment t, For the active power demand of the node at time t, For the reactive power demand of the node at time t, Reactive power generated by the generator at the moment t; The next state is indicated to be the next state, Representing the external power injection vector, Representing a nonlinear alternating current power flow equation; the control actions of the action space are normalized to reactive power set points under inverter capacity constraints, in particular by an action parameter in the range of [ -1, 1] The product of the remaining capacity of the inverter is defined as follows: Wherein, the For the inverter to be rated for apparent power capacity, Represents the active output of the jth photovoltaic power generation unit at the time t, And the maximum available active power output upper limit of the jth photovoltaic power generation unit under the current operation condition is represented, and the value of the maximum available active power output upper limit is determined by the photovoltaic installed capacity, the operation working condition and the rated capacity of the inverter.
5. The power distribution network voltage control method based on running state estimation according to claim 1, wherein updating node weights based on graph attention network strategies comprises the steps of carrying out linear transformation coding of belt layer normalization on local observation to obtain initial node characteristics, calculating attention coefficients between a node and a neighbor node and carrying out normalization, aggregating neighbor characteristics by using the normalized attention weights, and splicing outputs of all attention heads and generating actions through final transformation.
6. The power distribution network voltage control method based on operation state estimation according to claim 1, wherein the pipeline model prediction controller performs rolling optimization with minimum overall voltage deviation of the power distribution network as an optimization target.
7. The method according to claim 6, wherein the rolling optimization is required to meet power regulation constraints, response time constraints, climbing rate constraints, duration constraints, regulation cost constraints of the distributed power source, and voltage safety constraints and branch power constraints of the power distribution network nodes.
8. The method for controlling voltage of a power distribution network based on operation state estimation according to claim 1, wherein the graph attention network is a multi-head graph attention mechanism capable of aggregating information from adjacent distributed power sources based on a physical topology of the power network.
9. The method of claim 1, further comprising issuing a first instruction in the sequence and performing a smooth transition on a real-time scale after the pipeline model predictive controller obtains an optimal control sequence of the adjustable device.
10. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the steps of the method according to any one of claims 1 to 9 when the program is executed.

Description

Power distribution network voltage control method based on running state estimation Technical Field The invention belongs to the field of power distribution network voltage control, and particularly relates to a power distribution network voltage control method based on running state estimation. Background With the large-scale access of Photovoltaic (PV) power generation, the characteristics of the instantaneous, spatial intermittent and non-uniformity of the photovoltaic power generation bring unprecedented challenges to voltage regulation, so that voltage fluctuation is severe and rapid, and the traditional control scheme is difficult to effectively cope with. Although the centralized optimization method can realize global optimization theoretically, the method faces the problems of topological structure change of the power distribution network, insufficient response to dynamic load and power generation change and the like in practical application. Disclosure of Invention The invention aims to overcome the problems in the prior art and provide a power distribution network voltage control method based on running state estimation, which enables a controller to coordinate nodes with close electrical distances preferentially by combining a graph attention mechanism and model predictive control without explicitly inputting network topology priori knowledge, thereby remarkably improving the voltage control efficiency and robustness in the situations of photovoltaic power generation and rapid load change. In order to achieve the above purpose, the technical scheme of the invention is that a power distribution network voltage control method based on running state estimation comprises the following steps: Establishing a power distribution network voltage optimization control model, wherein an objective function of the power distribution network voltage optimization control model is used for adjusting reactive power to maintain voltage within a safety range and minimize voltage deviation; analyzing and defining a state space and an action space of safe operation of the power distribution network; based on the graph attention network strategy, updating the dynamic attention weight of the power distribution network topological node in the current state in real time; And inputting the dynamic attention weight, the current running state of the power distribution network and the predicted next-moment state into a pipeline model prediction controller, and outputting the reactive power control quantity. Furthermore, an objective function of the power distribution network voltage optimization control model adopts an optimization method based on an obstacle function, and is used for punishing voltage deviation deviating from a safety range. Further, the objective function is expressed as follows: Wherein, the Representing nodesIs used for the voltage of the per unit value,For a differentiable barrier function, for penalizing voltage deviations from a range of 0.95, 1.05 per unit values,Is an electric generatorIs provided for the reactive power output of the (c),The weight is penalized for reactive power,In order to be the number of nodes,The number of the generators is controllable; The barrier function adopts a form comprising an L1 norm, an L2 norm, a Courant-Beltrami function, a bowl function or a lamp function, and is selected according to the smoothness requirement and the punishment characteristic; The voltage control converges to a safe operation state by optimizing the objective function driving system as shown in the following formula: 。 further, the state s (t) of the state space at the time t aggregates the measured values of the multidimensional power system of photovoltaic power generation power, load active and reactive demands, generator reactive output, voltage amplitude and phase angle, and s (t) is expressed as follows: Wherein, the The active power generated by the photovoltaic system at the moment t,For the active power demand of the node at time t,For the reactive power demand of the node at time t,Reactive power generated by the generator at the moment t; The next state is indicated to be the next state, Representing the external power injection vector,Representing a nonlinear alternating current power flow equation; the control actions of the action space are normalized to reactive power set points under inverter capacity constraints, in particular by an action parameter in the range of [ -1, 1] The product of the remaining capacity of the inverter is defined as follows: Wherein, the For the inverter to be rated for apparent power capacity,Represents the active output of the jth photovoltaic power generation unit at the time t,And the maximum available active power output upper limit of the jth photovoltaic power generation unit under the current operation condition is represented, and the value of the maximum available active power output upper limit is determined by the photovoltaic installed cap