| 杨蓉,杨林,谭盛兰,等.基于遗传算法—优化长短期记忆神经网络的柴油机瞬态NOx排放预测模型研究[J].内燃机工程,2022,43(1):10-17. |
| 基于遗传算法—优化长短期记忆神经网络的柴油机瞬态NOx排放预测模型研究 |
| Prediction Model for Transient NOx Emission of Diesel Engine Based on GA-Long Short Term Memory（LSTM） Neural Network |
| DOI：10.13949/j.cnki.nrjgc.2022.01.002 |
| 关键词:柴油机 瞬态 NOx预测 长短期记忆神经网络 遗传算法 |
| Key Words:diesel engine transient NOx prediction long short term memory（LSTM） neural network genetic algorithm（GA） |
| 基金项目:国家重点研发计划项目（2017YFE0102800）；广西科技基地和人才专项项目（AD19110019）；广西创新驱动发展专项项目（AA18242045-3）；广西制造系统与先进制造技术重点实验室项目（19-050-44-S004） |
|
| 摘要点击次数: 1839 |
| 全文下载次数: 454 |
| 摘要:为实现选择性催化还原系统（selective catalytic reduction， SCR）尿素喷射量的精确控制，构建并评估了一套利用遗传算法（genetic algorithm， GA）优化长短期记忆（long short term memory， LSTM）神经网络的柴油机瞬态NOx排放预测模型。针对柴油机瞬态运行特点，选择柴油机NOx排放的主要影响因素进行相关性分析，确定模型的输入变量；为避免人为选取参数对神经网络预测性能的不良影响，采用遗传算法优化LSTM神经网络的参数，建立预测柴油机瞬态NOx排放的GA-LSTM模型；最后对预测模型进行了性能测试。结果表明，该模型具有较好的预测能力。 |
| Abstract:In order to realize the accurate control of the urea injection volume by selective catalytic reduction（SCR）， a set of NOx emission prediction model for diesel engine in transient mode was constructed and evaluated. The model mainly adopted the long short term memory（LSTM） neural network algorithm， and was optimized by genetic algorithm（GA）. According to the transient operation characteristics of diesel engine， the correlation analysis was carried out by selecting the main influencing factors of NOx emission， and the input variables of the model were determined. Then， in order to avoid the adverse effect of artificial selection parameters on the prediction performance of the neural network， the structural parameters of the LSTM neural network were optimized by genetic algorithm， and the GA-LSTM model for forecasting transient NOx emission of diesel engine was established. Finally， the performance of the model was tested. The results show that the model has good predictive ability. |
|
查看全文 HTML
查看/发表评论 下载PDF阅读器 |