基于组合模型的生物质燃料价格预测研究

Research on Biomass Fuel Price Prediction Based on Combined Model

我国生物质能源储量丰富,生物质燃料发电前景广阔,但国内生物质发电普遍存在亏损现象,燃料采购成本居高不下,严重阻碍了生物质发电的推广。对生物质燃料进行价格预测分析,对保障生物质燃料发电厂的利益、促进生物质燃料发电产业健康发展具有重要意义。文章利用江苏某生物质发电厂2018年5月至2020年4月共3年的生物质燃料采购数据,运用滑动平均、趋势法、ARIMA模型等多种技术手段构建生物质燃料价格预测模型。运用2020年5月至2021年4月数据对模型进行检验,预测值相对误差均在5%以下,预测误差较小,较为接近真实值。文章采用组合预测模型的方法,能更好地发挥各单一模型的优势,使误差最小化,提高预测正确率以及稳定性。

China is rich in biomass energy reserves and has broad prospects for biomass fuel power generation.However,there are widespread losses in domestic biomass power generation,and the fuel procurement cost remains high,which seriously hinders the promotion of biomass power generation.The price prediction and analysis of biomass fuel is of great significance to protect the interests of biomass fuel power plants and promote the healthy development of biomass fuel power generation industry.Based on the biomass fuel purchase data of a biomass power plant in Jiangsu from May 2018 to April 2020,this paper constructs a biomass fuel price prediction model by using a variety of technical means such as moving average,trend method and ARIMA model.The data from May 2020 to April 2021 are used to test the model.The relative errors of the predicted values are less than 5%,and the prediction error is small,which is close to the real value.In this paper,the combination forecasting model is the best way to make the first mock exam more effective,minimize the error and improve the accuracy and stability of prediction.