摘要
提出一种基于工业色谱仪的软测量建模方法,并针对碳五馏分分离过程中的精馏脱炔烃塔塔底成分估计问题,建立了合适的工业软测量模型。介绍了工业色谱仪在线质量检测原理和LM-BP神经网络模型的建立,并利用工业色谱仪在线检测的质量数据进行系统的在线和周期性模型更新,提高了软测量模型的在线估计精度。研究结果表明,基于工业色谱仪的LM-BP神经网络模型是一种有效的软测量建模方法。
The soft sensor modeling method based on gas chromatograph was proposed.In accordance with the task of composition estimation for alkynes removal distillation tower in the process of C5 separation,appropriate soft sensor model was established.The on-line quality detection principle of gas chromatograph and LM-BP neural network modeling was introduced.By adopting on-line quality data from gas chromatograph,on-line optimization of the system and the update of the model were carried out.Thus on-line estimation accuracy of soft sensor mode was enhanced.The result shows LM-BP neural network based on gas chromatograph is an efficient modeling method for soft sensor.
出处
《化工自动化及仪表》
CAS
北大核心
2010年第10期37-40,46,共5页
Control and Instruments in Chemical Industry
关键词
工业色谱仪
软测量
精馏脱炔烃塔
LM-BP神经网络
建模
gas chromatograph
soft sensor
alkynes removal distillation tower
LM-BP neural network
modeling
