Search

CN-122026621-A - Data-driven-based power grid medium-and-long-term power outage strategy decision method

CN122026621ACN 122026621 ACN122026621 ACN 122026621ACN-122026621-A

Abstract

The invention discloses a method for deciding a medium-long-term power outage strategy based on a data driven power grid, which comprises the following steps of S1, enabling a data acquisition module to access multi-source heterogeneous data in real time through a standardized interface, enabling the data acquisition module to comprise but not limited to power grid topological structure data, equipment running state data, meteorological environment data, user electricity consumption behavior data and new energy output prediction data, S2, enabling a data preprocessing module to preprocess the acquired data, enabling a prediction analysis module to output equipment fault probability distribution for 1-12 months in the future, S3, enabling a decision optimization module to take a prediction analysis result as input, constructing a power outage loss function, solving an optimal power outage strategy scheme, S4, enabling an execution feedback module to push the generated power outage strategy scheme to a power grid dispatching system, and monitoring the power grid running state in the strategy execution process in real time. The invention solves the problem of excessive dependence of the traditional decision mode on the physical model and the manual experience, and realizes the dynamic and accurate formulation of the power outage plan.

Inventors

  • LIU XUFEI
  • LIANG HONGBO
  • ZHANG ZHEMING
  • CHEN SHILONG
  • CHEN XI
  • CHENG QIAN
  • YANG JIJU
  • WANG XIYING
  • TAO YIRAN
  • WANG SHULING
  • CHEN RUCHANG
  • GAO DAOCHUN
  • SHAN WENLIN
  • Zeng Pijiang
  • QIN WENCHAO

Assignees

  • 云南电网有限责任公司

Dates

Publication Date
20260512
Application Date
20251220

Claims (10)

  1. 1. The method for deciding the medium-and-long-term power failure strategy of the power grid based on data driving is characterized by comprising the following steps of: The method comprises the following steps that S1, a data acquisition module (10) accesses multi-source heterogeneous data in real time through a standardized interface, wherein the data comprise, but are not limited to, power grid topological structure data, equipment running state data, meteorological environment data, user electricity behavior data and new energy output prediction data; S2, after the data preprocessing module (20) preprocesses the acquired data, the predictive analysis module (30) outputs the probability distribution of the equipment faults for 1-12 months in the future; S3, a decision optimization module (40) takes a prediction analysis result as input, constructs a power outage loss function, and solves an optimal power outage strategy scheme, wherein the scheme comprises a power outage period, a power outage range and a rush repair resource allocation plan; And S4, pushing the generated power failure strategy scheme to a power grid dispatching system by an execution feedback module (50), and monitoring the power grid running state in the strategy execution process in real time.
  2. 2. The method for deciding the medium-long term power outage strategy of the power grid based on data driving of claim 1, wherein the data source of the data acquisition module (10) comprises real-time power flow data provided by a main network EMS system, equipment topological relation data stored by a distribution network GIS system, temperatures issued by a meteorological department, precipitation, wind speed data and user power load data acquired by a smart electric meter.
  3. 3. The method for deciding a medium-long term outage strategy of a power grid based on data driving according to claim 1, wherein the outage loss function adopted by the S3 decision optimization module (40) is as follows: Wherein K is a user type or user number index, K is a total number of user types/total number of users, T is a time period index, T is a total time period number, c k is a unit electric energy value coefficient of a kth user, P k is average load power of the kth user, and T k is actual power failure duration of the kth user in a single time period.
  4. 4. The power grid medium-long-term power failure strategy decision-making system based on data driving is characterized by comprising a data acquisition module (10), a data preprocessing module (20), a prediction analysis module (30), a decision optimization module (40) and an execution feedback module (50) which are sequentially in communication connection.
  5. 5. The power grid medium-long-term power outage strategy decision system based on data driving is characterized in that a data acquisition module (10) is used for acquiring data, a data preprocessing module (20) is used for preprocessing the acquired data, a prediction analysis module (30) is used for outputting equipment fault probability distribution of 1-12 months in the future, a decision optimization module (40) is used for constructing a power outage loss function, and an execution feedback module (50) is used for pushing a generated power outage strategy scheme to a power grid dispatching system.
  6. 6. The medium-long term outage strategy decision system of a power grid based on data driving of claim 4, wherein said data acquisition module (10) performs data interaction with a data source through IEC 61850 standard protocol, and said execution feedback module (50) establishes two-way communication with a power grid dispatching system.
  7. 7. The medium-long term outage policy decision system of a data driven power grid according to claim 4, wherein said data preprocessing module (20) comprises an outlier rejection unit (201) and a feature extraction unit (202), said outlier rejection unit (201) is configured to identify and remove outliers in original data, and said feature extraction unit (202) extracts key features from preprocessed data.
  8. 8. The medium-long term outage strategy decision system of a data-driven power grid according to claim 4, wherein said decision optimization module (40) further comprises a power grid security constraint verification sub-module (401), said power grid security constraint verification sub-module (401) being configured to perform security verification on a generated outage strategy scheme.
  9. 9. The medium-long term outage policy decision system of a data-driven power grid of claim 7, wherein said outlier rejection unit (201) uses a 3 sigma criterion to statistically identify significant outliers, and directly marks samples exceeding a range of data mean + -3 standard deviations as outliers.
  10. 10. The medium-long term power outage strategy decision system of a power grid based on data driving of claim 9, wherein said outlier rejection unit (201) directly marks samples exceeding the range of standard deviation of +/-3 times of the mean value of the data as abnormal data, and the abnormal data comprises obvious error data of abrupt voltage change and overrun current.

Description

Data-driven-based power grid medium-and-long-term power outage strategy decision method Technical Field The invention belongs to the technical field of power system management, and particularly relates to a data-driven power grid medium-and-long-term power outage strategy decision method. Background The current power grid outage strategy decision mainly depends on three types of technical paths, namely a digital simulation method based on a physical model, wherein the digital simulation method is used for carrying out stability evaluation by constructing a refined mathematical model of a power grid topological structure, for example, a traditional online stability evaluation system adopted in early stage by Chinese electric department is characterized in that the power grid running state is restored through digital simulation, the accuracy of the method is high, a complex physical process can be reproduced, the time consumption is long, the calculation force requirement is high, and the method is difficult to adapt to the rapidly-changed power grid form. The method has the advantages that the flexibility is high, emergency situations can be treated, but the subjectivity is high, the decision quality is limited by experience level, and the multi-objective optimization effect is difficult to quantitatively evaluate. And thirdly, a single prediction model assists decision, such as a load prediction model based on a gray system theory or a fault early warning algorithm of a basic neural network. The method has the advantages of simple algorithm and low deployment cost, but has weak capability of adapting to complex power grids, and lacks multi-factor coupling analysis and prediction error accumulation along with time. These three types of techniques are therefore each advantageous in certain scenarios, but all suffer from significant limitations. The prior art system has three core defects in the strategy decision of power failure in the middle and long time of a power grid, namely, the data utilization rate is insufficient, the traditional method depends on a single data source (such as SCADA system real-time data), and multi-dimensional information such as weather, geography, user behavior and the like cannot be effectively integrated, so that the decision basis is one-sided. Secondly, the multi-factor coupling analysis capability is lost, the traditional model often processes links such as load prediction, fault evaluation, risk prevention and control and the like in an isolated manner, and a dynamic association mechanism among all elements cannot be established. Finally, the dynamic adjustment mechanism is stiff, and the existing system mostly adopts a fixed period (such as 15 minutes) scanning mode, so that decisions cannot be updated in real time according to the running state of the power grid. Therefore, the invention provides a data-driven power grid medium-and-long-term power outage strategy decision method for solving the problems set forth in the background art. Disclosure of Invention Aiming at the problems of the background technology, the invention aims to provide a data-driven power grid medium-and-long-term power outage strategy decision method, solves the problem of excessive dependence of a traditional decision mode on a physical model and artificial experience, realizes dynamic and accurate formulation of a power outage plan, can be widely applied to annual power outage planning of a provincial power grid dispatching center, quarter overhaul arrangement of an urban power distribution network and differential power outage strategy formulation of a new energy enrichment region, and is particularly suitable for complex power grid operation and maintenance requirements under a high-proportion new energy grid-connected scene. In order to achieve the technical purpose, the invention adopts the following technical scheme: A method for deciding a medium-and-long-term power failure strategy of a power grid based on data driving comprises the following steps: The method comprises the following steps that S1, a data acquisition module accesses multi-source heterogeneous data in real time through a standardized interface, wherein the data comprise, but are not limited to, power grid topological structure data, equipment running state data, meteorological environment data, user electricity behavior data and new energy output prediction data; S2, after the data preprocessing module preprocesses the acquired data, the predictive analysis module outputs the probability distribution of the equipment faults for 1-12 months in the future; S3, the decision optimization module takes a prediction analysis result as input, constructs a power outage loss function, and solves an optimal power outage strategy scheme, wherein the scheme comprises a power outage period, a power outage range and a rush repair resource allocation plan; and S4, pushing the generated power failure strategy scheme to a power grid dispatchin