文章摘要
李庆伟,申志文.基于改进差分极端学习机的燃煤锅炉NOx预测[J].,2022,21(1):98-104
基于改进差分极端学习机的燃煤锅炉NOx预测
Prediction of NOx emission of coal-fired boilers based on improved differential evolution algorithm and extreme Llearning machine
投稿时间:2020-01-28  修订日期:2020-11-07
DOI:10.13738/j.issn.1671-8097.020016
中文关键词: 差分进化算法  极端学习机  NOx排放  预测模型
英文关键词: differential evolution  extreme learning machine  NOx Emission  prediction model
基金项目:国家重点基础研究发展计划(973计划);上海市青年科技英才扬帆计划资助
作者单位E-mail
李庆伟 上海电力大学 liqingweish@163.com 
申志文 上海电力大学  
摘要点击次数: 8
全文下载次数: 9
中文摘要:
      低NOx燃烧优化是一种简单、高效、廉价的燃煤电站NOx减排方法。建立NOx预测模型是该方法的重要组成部分。极端学习机(Extreme learning machine,ELM)是一种简单有效的建模方法,但随机生成的输入权值和隐层阈值会影响ELM的泛化性能和逼近能力。针对该问题,利用一种改进的差分进化算法(Improved differential evolution,IDE)优化输入权值和隐层阈值,得到了改进的NOx预测模型(Improved differential evolution- extreme learning machine, IDE- ELM),该模型可以有效克服参数寻优过程中的早熟问题。利用IDE-ELM模型预测了某330 MW的NOx排放量,并与标准ELM和DE-ELM预测结果进行对比。为消除启发式算法随机性的影响,每个实验独立重复101次。实验结果表明,IDE-ELM模型有效提升了ELM算法的泛化能力和稳定性。
英文摘要:
      Low NOx combustion optimization is a simple, efficient and inexpensive NOx emission reduction technology for coal-fired power plants. Establishing NOx Emission prediction model is an important part of low NOx combustion optimization. Extreme learning machine (ELM) is a simple and effective modeling method. However, randomly generated input weights and hidden layer thresholds would affect the performance of ELM. To solve this problem, this paper employed an improved differential evolution (IDE) to optimize the input weights and hidden layer thresholds. The proposed model was applied to some 330MW boiler and compared with the standard ELM and DE-ELM. Each model was repeated 101 runs independently. Results show that IDE-ELM improves the generalization ability and stability of ELM.
查看全文   查看/发表评论  下载PDF阅读器
关闭