摘要
在实际化工生产过程中 ,存在许多参数难以或无法用仪表直接检测 ,同时它们又需要加以严格控制 ,因此实际需要促进了软测量技术的产生和发展 .而用神经网络来建立软测量的模型 ,比其他方法有着不可比拟的优点 ,近年来 ,随着科技的进步 ,又出现了神经网络结合其他方法的软测量方法 .如神经网络与最小二乘法相结合 ,神经网络与模糊数学相结合 ,在实际应用中都得到了很好的效果 .本文采用PCA改进的BP算法建立了盐酸浓度的软测量模型 .仿真结果表明 。
In practical chemical industry producing process,there are many variables that are uneasy to or can't directly be measured,but they must be under control strictly,so pratical need made soft sensing technique come about and progress.Soft sensing model based on artificial neural networks has many advantages that others haven't.Recent years,with the progress of technology,soft sensing based on neural networks with other methods comes about.eg,soft-sensing based on neural networks and PLS,based on neural networks and fuzzy-mathematics.they all gain better results.Based on the improved BP algorithm with PCA,a soft sensing model of concentration of hydroehloric acid was established.Simulation showed that this model raises the measure precision and increases the training speed.
出处
《沈阳化工学院学报》
2002年第1期39-43,共5页
Journal of Shenyang Institute of Chemical Technolgy
