基于县域尺度神经网络模型的中国钢铁工业污染物排放预测研究  

Prediction of Pollutant Emissions from China's Steel Industry Based on County-scale Neural Network Model

作　　者：徐佳雨 XU Jiayu

机构地区：[1]南京大学环境学院,南京210000

出　　处：《科技创新与应用》2024年第33期9-12,共4页Technology Innovation and Application

摘　　要：实现精细尺度的钢铁工业排放预测对我国的区域污染控制和产业政策调整具有重要意义。该研究通过融合多源数据构建神经网络模型并实现2025—2060年的县域尺度活动水平预测,建立中国钢铁工业的大气污染物排放预测清单。结果表明,粗钢产量将在2025年达到最高值,且随后呈现缓慢下降趋势。同时,未来中西部地区将形成大型县域钢铁生产基地。2060年,我国粗钢产量预计为7.6亿t,较2015年下降4%,共排放SO_(2)、NO_(X)、PM_(2.5)及CO_(2)分别为10.6万t、8.6万t、25.5万t和5.9亿t,较2015年分别下降88%、89%、76%和52%。Realizing fine-scale emission prediction from the iron and steel industry is significant for regional pollution control and industrial policy adjustment in China.In this study,we calculate future air pollutant emission inventory from the iron and steel industry in China by using multi-source data to obtaining activity level predictions from 2025 to 2060 based on county-scale neural networks.The results show that crude steel production will peak in 2025 and then show a slow downward trend.Meanwhile,counties with large steel production will be formed in the Midwest in the future.The crude steel production in China is expected to be 760 million tons in 2060,with a decrease of 4%from 2015.The emission of SO_(2),NO_(X),PM_(2.5) and CO_(2) from the iron and steel industry will be 106 kilotons,86 kilotons,255 kilotons and 590 million tons,respectively,with a decrease of 88%,89%,76%and 52%from 2015.

关 键 词：县域尺度预测 神经网络模型 钢铁排放 排放因子 区域污染 

分 类 号：TP274.2[自动化与计算机技术—检测技术与自动化装置]