一种基于时间序列的配电网物资预测方法

在国家电网当下背景与发展要求下,针对电网物资库存的管理现状,迫切需要对物资需求进行精准预测。以目前的电网数据条件、数据挖掘技术条件为基础进行可行性研究验证;最后通过时间序列、 LSTM神经网络等算法对物资领用历史数据进行建模分析,得到一种有效的物资预测研究方法。

A Distribution Network Material Forecasting Method Based on Time Series

Under the current context and development requirements of the SGCC, it is urgent to accurately forecast the material demand given the existing management of grid material inventory. Based on the present situation of grid data and data mining technology, the feasibility study is verified;finally, the time series forecasting and LSTM are adopted to build a model and analyze historical reception data to conclude an effective material forecasting study method.