CN-122000863-A - Power demand prediction method, equipment and medium for power distribution network based on big data analysis

CN122000863ACN 122000863 ACN122000863 ACN 122000863ACN-122000863-A

Abstract

The invention discloses a power demand prediction method, equipment and medium for a power distribution network based on big data analysis, which belong to the technical field of power demand prediction and comprise the steps of dividing the power distribution network into a plurality of power nodes, collecting power data of the power nodes, and screening to obtain power abnormal nodes; the method comprises the steps of collecting demand information of power abnormal nodes, extracting to obtain power demand events according to the demand information, obtaining event information of the power demand events, evaluating to obtain propagation effect conditions of the power demand events according to the event information, collecting real-time prediction demand values of the power abnormal nodes, determining power change values of other power nodes in combination with the propagation effect conditions, collecting real-time environment data of the power nodes, predicting to obtain basic power demand values according to the real-time environment data, and superposing the power change values to obtain real-time power demand values. The method and the system improve the accuracy of power demand prediction of the power distribution network based on big data analysis.

Inventors

LI MIN
LAI MINGYUAN
ZHONG XIANG
ZHANG JUN
LIAO JUNHUA
TAN QIWEN
Luo Tianlu
WAN SONG
WEI LISHAN
TAN XIAOHONG
LI RONGYAO
Tao Yigang

Assignees

广西电网有限责任公司

Dates

Publication Date: 20260508
Application Date: 20251215

Claims (10)

1. A power demand prediction method for a power distribution network based on big data analysis is characterized by comprising the following steps of, Dividing the power distribution network into a plurality of power nodes, collecting power data of the power nodes, and screening according to the power data to obtain power abnormal nodes; Acquiring demand information of the power abnormal node, and extracting to obtain a power consumption demand event according to the demand information; acquiring event information of an electricity demand event, and evaluating and acquiring a propagation effect condition of the electricity demand event according to the event information; Acquiring a real-time prediction demand value of the power abnormal node, and determining a power variation value of the residual power node by combining the propagation effect condition; and acquiring real-time environment data of the power node, predicting to obtain a basic power demand value according to the real-time environment data, and superposing the power change value to obtain the real-time power demand value.
2. The method for predicting power demand of a power distribution network based on big data analysis of claim 1, wherein said screening for power anomaly nodes comprises, Collecting power data of the power nodes, judging whether the power grid nodes have the phenomenon of insufficient power supply according to the power data, and judging the power nodes as abnormal power nodes if the power grid nodes have the phenomenon of insufficient power supply; if the phenomenon of insufficient power supply does not exist, extracting real-time load of the power node according to the power data; and acquiring the historical load of the power node, comparing the real-time load with the historical load, judging whether the real-time load is abnormal, and if the real-time load is abnormal, judging that the power node is an abnormal power node.
3. The method for predicting power demand of a power distribution network based on big data analysis of claim 2, wherein the obtaining the power demand event comprises collecting demand information of power abnormal nodes, collecting a corresponding use time point of electric equipment, and judging whether the use time point accords with the use time point; if the historical electricity utilization time point is met, counting the use quantity of the electric equipment corresponding to the historical electricity utilization time point, recording the use quantity as the historical use quantity, collecting the real-time use quantity of the electric equipment, and calculating the difference value of the historical use quantity and the real-time use quantity to obtain an equipment use difference value; If the use time point is not met, the real-time quantity of the electric equipment is used as the equipment use difference value, the equipment with the equipment use difference value reaching the preset use difference value threshold is recorded as abnormal equipment, and the power consumption requirement event is obtained by combining the abnormal equipment and the node position.
4. The method for predicting power demand of a power distribution network based on big data analysis of claim 3, wherein said combining the abnormal equipment and the node location to obtain the power demand event comprises collecting equipment efficacy of the abnormal equipment, and determining whether the abnormal equipment usage is related to the geographic location according to the equipment efficacy; Screening the events related to the geographic position from the basic events and recording the events as geographic events, extracting the geographic features of the geographic events, acquiring the position features of the node positions, and taking the geographic events with the geographic features consistent with the position features as electricity consumption requirement events; If the equipment usage is irrelevant to the geographic position, screening an event related to the abnormal equipment from the basic event as an abnormal event, collecting equipment usage characteristics corresponding to the abnormal event, and confirming and obtaining a power consumption demand event according to the equipment usage characteristics.
5. The method for predicting power demand of a power distribution network based on big data analysis of claim 4, wherein said obtaining power demand event according to equipment usage characteristics comprises obtaining feature similarities of equipment usage characteristics corresponding to all abnormal events, and judging whether the feature similarities are in a preset similarity range; If the power influence value is in the preset similarity range, acquiring a power influence value corresponding to the abnormal event when the history occurs; Acquiring a real-time change value of the power abnormal node, calculating to obtain a difference value between the power influence value and the real-time change value, acquiring occurrence probability of an abnormal event, and comprehensively determining to obtain a power demand event; If the similarity is not in the preset similarity range, extracting the equipment operation characteristics of the abnormal equipment, comparing the similarity of the equipment operation characteristics and the equipment use characteristics and marking the similarity as the operation similarity; and selecting the abnormal event with the maximum operation similarity as the electricity consumption requirement event.
6. The method for predicting power demand of a power distribution network based on big data analysis of claim 5, wherein the obtaining the propagation effect of the power demand event comprises determining whether the power demand event is a geographic event; If the event is a geographic event, judging whether the electricity demand event is diffused or not; if the electricity demand event is diffused, judging whether the electricity demand event depends on the geographic characteristics; if the geographic features are relied on, the position features of the rest power nodes are obtained, the similarity of the geographic features and the position features is obtained through comparison and is recorded as geographic similarity, and the range of the power nodes with the geographic similarity reaching a preset geographic similarity threshold is used as a node propagation range; If the geographic characteristics are not relied on, historical diffusion data of electricity demand events are collected, and a propagation effect condition is obtained according to the historical diffusion data; And if the event is not a geographic event, estimating and obtaining the propagation effect condition of the electricity demand event according to the event information.
7. The method for predicting power demand of a power distribution network based on big data analysis of claim 6, wherein the obtaining propagation effect according to the historical diffusion data comprises collecting historical diffusion data of power demand events, wherein the historical diffusion data comprises a historical diffusion range, a historical diffusion speed and a historical diffusion force; Judging whether message diffusion of an electricity demand event affects electricity demand, if so, collecting message diffusion data of the electricity demand event, wherein the message diffusion data comprises a message diffusion range, a message diffusion speed and a message diffusion strength; combining the historical diffusion data and the message diffusion data to comprehensively obtain a propagation effect condition; the evaluation of the propagation effect of the electricity demand event according to the event information comprises, If the event is not a geographic event, collecting event information of a power consumption requirement event, wherein the event information comprises event occurrence time and event attention crowd; Judging whether the electricity demand event is an emergency event, if not, collecting the historical propagation time length of the electricity demand event, and judging whether the event occurrence time length reaches the historical propagation time length; If the historical propagation time is not reached, calculating a propagation time difference value between the historical propagation time and the event occurrence time; establishing a correlation table of the propagation time difference value and the propagation effect condition, and searching according to the correlation table to obtain the propagation effect condition; if the electricity demand event is an emergency event, collecting an information diffusion channel of the electricity demand event, and combining event attention crowd to obtain a propagation effect condition; the situation that the spreading effect is obtained by combining the event concern crowd comprises that the average speed of the event concern crowd receiving information in the statistical information spreading channel is used as the node spreading speed; extracting crowd features according to the event-concerned crowd, and confirming the distribution range of the event-concerned crowd according to the crowd features to serve as a node transmission range; And counting the number proportion of event concern groups of different power nodes, and taking the number proportion as the node propagation strength.
8. The method for predicting power demand of a power distribution network based on big data analysis of claim 7, wherein determining the power variation value of the remaining power nodes comprises marking the power nodes within the node propagation range as power prediction nodes, and calculating the distances between the power anomaly nodes and the power prediction nodes; acquiring an electric power abnormal value of the electric power abnormal node according to the electric power change time point obtained by calculation of the node propagation speed and the distance; Acquiring a real-time point, calculating a time difference between the power change time point and the real-time point, and obtaining a real-time prediction demand value according to the time difference; And when the real-time prediction demand value reaches a preset demand value threshold value, calculating according to the power abnormal value and the node propagation strength to obtain a power change value of the power prediction node at a power change time point.
9. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor, when executing the computer program, implements the steps of a method for predicting power demand of a power distribution network based on big data analysis according to any of claims 1 to 8.
10. A computer readable storage medium having stored thereon a computer program, which when executed by a processor implements the steps of a method for predicting power demand of a power distribution network based on big data analysis according to any one of claims 1 to 8.

Description

Power demand prediction method, equipment and medium for power distribution network based on big data analysis Technical Field The invention relates to the technical field of power demand prediction, in particular to a power demand prediction method, equipment and medium for a power distribution network based on big data analysis. Background In the field of power systems, stable operation of a power distribution network is important for guaranteeing social production and people's life, and accurate power demand prediction is a basis for realizing optimal scheduling and reasonable resource allocation of the power distribution network. Traditional power demand prediction methods generally rely on static data such as historical load data, weather information, economic indicators and the like for model construction. The method can reflect the basic requirement rule of electric power to a certain extent, but because the model is solidified and lacks a real-time response mechanism to dynamic emergency, the method is poor in performance when coping with scenes such as extreme weather, equipment faults, emergency events or abrupt changes of energy supply and demand, and the like, and a prediction strategy cannot be dynamically adjusted to adapt to the emergency. This results in inaccurate predictions of the power demands of the distribution network, further affecting the stable operation of the distribution network. Disclosure of Invention The present invention has been made in view of the above-described problems. Therefore, the method aims to solve the problems that the power demand of the power grid is not accurately predicted, so that the stable operation of the power distribution network is further affected, and the like. In order to solve the technical problems, the invention provides a power demand prediction method of a power distribution network based on big data analysis, which comprises the following steps of, The power distribution network is divided into multiple power nodes, power data of the power nodes are collected, power abnormal nodes are obtained through screening according to the power data, demand information of the power abnormal nodes is collected, a power demand event is obtained through extraction according to the demand information, event information of the power demand event is obtained, propagation effect conditions of the power demand event are obtained through evaluation according to the event information, real-time prediction demand values of the power abnormal nodes are collected, power change values of the remaining power nodes are determined according to the propagation effect conditions, real-time environment data of the power nodes are collected, basic power demand values are obtained through prediction according to the real-time environment data, and real-time power demand values are obtained through superposition of the power change values. As an optimal scheme of the power demand prediction method of the power distribution network based on big data analysis, the power demand prediction method comprises the steps of screening to obtain power abnormal nodes, Collecting power data of the power nodes, judging whether the power grid nodes have the phenomenon of insufficient power supply according to the power data, and judging the power nodes as abnormal power nodes if the power grid nodes have the phenomenon of insufficient power supply; if the phenomenon of insufficient power supply does not exist, extracting real-time load of the power node according to the power data; and acquiring the historical load of the power node, comparing the real-time load with the historical load, judging whether the real-time load is abnormal, and if the real-time load is abnormal, judging that the power node is an abnormal power node. The power demand event acquisition method based on big data analysis comprises the steps of collecting demand information of power abnormal nodes, collecting a corresponding use time point of electric equipment, and judging whether the use time point accords with the use time point or not; if the historical electricity utilization time point is met, counting the use quantity of the electric equipment corresponding to the historical electricity utilization time point, recording the use quantity as the historical use quantity, collecting the real-time use quantity of the electric equipment, and calculating the difference value of the historical use quantity and the real-time use quantity to obtain an equipment use difference value; If the use time point is not met, the real-time quantity of the electric equipment is used as the equipment use difference value, the equipment with the equipment use difference value reaching the preset use difference value threshold is recorded as abnormal equipment, and the power consumption requirement event is obtained by combining the abnormal equipment and the node position. The power demand event obtained by combining the abnormal equ