摘要
随着社会经济的飞速发展,电网企业对于电力物资的需求日益增多,物资需求精益化管理面临着物资种类繁多、周期多变、数据失真、预测困难等诸多问题。为此,从物资需求特性出发,通过分析各类物资的需求特征,挖掘需求规律,并建立时间序列、LSTM(长短期记忆网络)神经网络模型、灰色预测等算法模型对电力物资开展精准预测,以保证电力企业正常生产,节约采购成本和库存成本,提高企业竞争力。
With the rapid development of social economy,power gird enterprises demand more and more electric power materials,and multiple material types,changeable cycle,data distortion and prediction difficulties are faced with lean management of material demand. Therefore,the paper analyzes demand characteristics of all materials in view of the demand characteristics,probes into the demand rules and establishes time series,LSTM neural network model and grey prediction model for accurate prediction of electric power materials,ensuring normal operation of electric power enterprises,saving purchase cost inventory cost,and improving competitiveness of enterprises.
作者
黄宏和
吴臻
琚军
章斌
王雪峰
潘永贺
郑辉
赵仲夏
HUANG Honghe;WU Zhen;JU Jun;ZHANG Bin;WANG Xuefeng;PAN Yonghe;ZHENG Hui;ZHAO Zhongxia(State Grid Zhejiang Quzhou Power Supply Company,Quzhou Zhejiang 324000,China)
出处
《浙江电力》
2019年第7期75-80,共6页
Zhejiang Electric Power
基金
国网浙江省电力有限公司衢州供电公司科技项目(5211QZ170003)
关键词
需求特性
时间序列
灰色预测
量化预测
demand characteristics
time series
grey prediction
quantitative prediction
