Distribution transformer power failure electric quantity loss prediction method based on deep learning

Abstract

The invention discloses a distribution transformer power failure electric quantity loss prediction method based on deep learning. The method includes: classifying the distribution transformer load curves based on a fuzzy C-means clustering algorithm to obtain various clustering center load curves, calculating the amplification of various distribution transformer loads and corresponding various center loads in each time period, and selecting the time period of which the amplification rate is smaller than a threshold value as a time period to be predicted; predicting a load value of a to-be-predicted time period by utilizing the trained gating cycle unit neural network to serve as power failure loss electric quantity; and solving the target function, and obtaining the electric duration whenthe target function is optimized. According to the invention, the calculation of power failure loss electric quantity is realized, data support is provided for improving the reliability of power supply, and the optimal management of planned power failure is realized based on user load curve clustering.

Description

technical field [0001] The invention relates to a method for predicting power loss of distribution transformer power failure based on deep learning, and belongs to the technical field of power system load forecasting. Background technique [0002] The losses caused by power outages to the national economy far exceed the losses of the power system itself. World-class urban power distribution networks require users in urban core areas to have an average annual power outage of no more than 5 minutes, and a reliability rate of power supply of about 99%. In 2017, the average power outage time of users in my country was 16.27 hours per household. Many urban power grids in China still need to accelerate the optimization and upgrading of distribution grids, apply high-end equipment, improve lean operation and maintenance and intelligent interaction levels, and reduce fault handling time from hours to minutes. level, reduce power outage time, and improve average power supply reliabil...

AI Technical Summary

Problems solved by technology

[0008] Purpose: In order to overcome the existing problems in the prior art, such as slow processing speed, large memory usage, and difficulty in processing high-dimensional features when processing massive data, the present invention provides a method for predicting the power loss of distribution transformer power outages based on deep learning. First, Cluster analysis of distribution transformer load characteristics, compare the load increase in each time period, and then use the Pearson correlation coefficient algorithm to extract the historical load with high correlation with the predicted time period load as the training volume of the input feature, and the actual value of the predicted time period as the output The training amount of the feature is to train the neural network of the gated cyclic unit to obtain relevant parameters, and then predict the power loss of the power outage in the last time period, and then obtain the duration of the power outage through the objective function, and complete the arrangement plan for future power outages

Images