摘要
有效的测量明胶浓度是研究明胶生产工艺中一项非常重要的课题,如何准确、快速地测量出明胶的浓度是一项重要的任务。目前明胶浓度的检测手段多为离线人工检测,不能实现胶液浓度的在线实时测量。在研究自适应小波神经网络(AWNN)的基础上,从提高明胶浓度软测量模型的实时性和鲁棒性着手,采用混合递阶遗传算法(HGA)对模型的参数进行优化。仿真结果得到的训练精度和预测精度分别为0.45和0.31,满足精度要求,算法在实现明胶浓度的在线测量上具有一定的实用性。
The gelatin concentration is an important area in gelatin production.How to accurately and quickly measure the gelatin concentration is an important task,but the equipment for monitoring the gelatin concentration on-line is so expensive and difficult to maintain that is not suitable for application,and the off-line sampling and monitoring method with low accuracy is applied in general.Based on the study of adaptive wavelet neural network(AWNN),a hierarchy genetic algorithm(HGA) is proposed to training network.The results of simulation indicates that this algorithm is realized in the on-line measurement on the gelatin concentration,and it has certain usability.
出处
《电子测量与仪器学报》
CSCD
2010年第8期775-779,共5页
Journal of Electronic Measurement and Instrumentation
关键词
明胶
自适应
混合递阶遗传算法
gelatin
adaptive wavelet neural network
hierarchy genetic algorithm