基于时序数据的尾气监测分析平台研究

Research on tail gas monitoring and analysis platform based on time-series data

为实现对非道路移动机械高排放污染物的有效监管,提高环境监测效率,研究基于时序数据的尾气监测分析平台从硬件到软件系统等的开发,针对大量时序数据,提出利用改进并进行参数优化后的长短期记忆(Long Short-Term Memory,LSTM)网络模型,对几类尾气排放数据进行预测分析,从而预测未来一段时间内高排放污染物的排放趋势。通过对公开数据集的多维时序数据的实际预测分析,对比模型优选前后的排放预测结果,分析该尾气监测分析平台的实用性以及该平台在非道路移动机械尾气监测、预警、预测与分析过程中存在的问题,提出合理化建议与应用展望。

For the road to achieve mobile machinery high discharge pollutants effective oversight,improve the efficiency of environmental monitoring,a monitoring and analysis platform based on time-series data,from hardware to software system development,was researched.To a large number of time-series data,using the improved and parameter optimized Long Short-Term Memory(LSTM)network model was proposed to forecast a few class exhaust data analysis and predict the future high emission pollutant emissions trend over a period of time.Based on public data sets of multi-dimensional time-series data of actual forecast analysis,the comparison of emission forecast results before and after the model optimization,analysis of the tail gas monitoring and analysis platform practicality and the problems existing in the process of the platform moving in the road mechanical tail gas monitoring,early warning and forecasting and analyzing were carried out,suggestions and application prospects were put forward reasonable.