为预测高原环境下缸内直喷(Gasoline Direct Injection, GDI)汽油车CO和PN的瞬时排放量,开发并评估了一套基于深度学习的排放预测模型。利用便携式车载排放测试系统对一辆GDI汽油车进行实际道路排放测试;加入奇异谱分析对原始时间序列...为预测高原环境下缸内直喷(Gasoline Direct Injection, GDI)汽油车CO和PN的瞬时排放量,开发并评估了一套基于深度学习的排放预测模型。利用便携式车载排放测试系统对一辆GDI汽油车进行实际道路排放测试;加入奇异谱分析对原始时间序列进行处理,剔除时间序列中的异常值;利用XGBoost模型对GDI汽油车的CO和PN的瞬时排放进行初步预测,并利用SVR模型进行残差修正得到最终的预测结果。将预测结果与实际道路排放试验中使用PEMS设备测量的实际值进行比较,结果表明,XGBoost-SVR排放预测模型能较好地预测GDI汽油车瞬时CO和PN的排放,相比单一的XGBoost模型,RMSE分别提高了22.9%和39.7%,决定系数R2均大于0.9,支持预测结果的可靠性。该模型对监测高原环境下GDI汽油车实际道路排放具有一定的工程意义。展开更多
To thrive in a world that is undergoing changes not seen in a century,China and Arab states must enhance solidarity and collaboration.CHINA and Arab states are all developing countries.Combined,they account for one-si...To thrive in a world that is undergoing changes not seen in a century,China and Arab states must enhance solidarity and collaboration.CHINA and Arab states are all developing countries.Combined,they account for one-sixth of the world’s land mass,one-fourth of the world’s population,and one-eighth of the world economy.展开更多
文摘为预测高原环境下缸内直喷(Gasoline Direct Injection, GDI)汽油车CO和PN的瞬时排放量,开发并评估了一套基于深度学习的排放预测模型。利用便携式车载排放测试系统对一辆GDI汽油车进行实际道路排放测试;加入奇异谱分析对原始时间序列进行处理,剔除时间序列中的异常值;利用XGBoost模型对GDI汽油车的CO和PN的瞬时排放进行初步预测,并利用SVR模型进行残差修正得到最终的预测结果。将预测结果与实际道路排放试验中使用PEMS设备测量的实际值进行比较,结果表明,XGBoost-SVR排放预测模型能较好地预测GDI汽油车瞬时CO和PN的排放,相比单一的XGBoost模型,RMSE分别提高了22.9%和39.7%,决定系数R2均大于0.9,支持预测结果的可靠性。该模型对监测高原环境下GDI汽油车实际道路排放具有一定的工程意义。
文摘To thrive in a world that is undergoing changes not seen in a century,China and Arab states must enhance solidarity and collaboration.CHINA and Arab states are all developing countries.Combined,they account for one-sixth of the world’s land mass,one-fourth of the world’s population,and one-eighth of the world economy.
