%0 Journal Article
%T 基于时间序列的全国新型能源发电量模型建立与预测分析
Time Series-Based Modeling and Forecast Analysis of National New Energy Power Generation
%A 宋婉宁
%A 覃丽娟
%A 吕萌
%A 周影慧
%A 高婉琪
%A 陈佳男
%A 白晓东
%J Hans Journal of Data Mining
%P 40-54
%@ 2163-1468
%D 2025
%I Hans Publishing
%R 10.12677/hjdm.2025.151004
%X 研究新型能源的发电能力可以为我国发电量的可持续发展提供有效策略。本文选取我国总发电量和新型能源发电量作为研究对象，采用ARIMA模型、二次移动平均预测模型、Holt线性指数平滑预测模型以及灰色预测模型，对我国发电量的时间序列进行模型拟合、预测及精度分析。结果显示，对于总发电量序列，二次移动平均、Holt线性指数平滑和灰色预测模型的组合模型具有较高的拟合精度；对于水电、核电发电量序列，二次移动平均预测模型表现出较好的拟合效果；对于风电、太阳能发电量序列，灰色预测模型的拟合精度最佳，大多数数据的预测值与拟合值的相对误差不超过10%。此外，研究预测2024至2025年间风电和太阳能发电量将快速增长，推动我国总发电量持续上升。
The study of power generation capacity from new energy sources offers effective strategies for the sustainable development of electricity generation in China. This paper analyzes China’s total electri- city generation and new energy generation using the ARIMA model, the quadratic moving average forecasting model, Holt’s linear exponential smoothing model, and the grey forecasting model. These models are applied to fit, forecast, and evaluate the accuracy of the time series data on electricity generation. Results indicate that, for the total electricity generation series, a combination model based on the quadratic moving average, Holt’s linear exponential smoothing, and grey forecasting models achieve high fitting accuracy. For hydropower and nuclear power generation, the quadratic moving average model provides the best fit, while the grey forecasting model is most accurate for wind and solar power generation, with most forecasted values exhibiting a relative error below 10%. Furthermore, the study predicts a rapid increase in wind and solar power generation between 2024 and 2025, contributing to a sustained annual rise in China’s total electricity output.
%K 新型能源发电，
%K ARIMA模型，
%K 二次移动平均，
%K Holt线性指数平滑，
%K 灰色预测
New Energy Power Generation
%K ARIMA Model
%K Quadratic Moving Average
%K Holt Linear Exponential Smoothing
%K Grey Forecasting
%U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=105199