摘要
发酵过程中的一些关键生物参数难以实时在线测量,融合人工智能技术和面向对象的编程技术,设计并实现了一种基于Web的建模系统,集成各种神经网络及其优化算法对生物参数进行建模预估。利用Eclipse的开发平台,使用轻量级框架,分离了表示层、业务逻辑层和数据持久层,在业务逻辑层对神经网络算法进行了实现,并完成了Java与Matlab的交互。该系统可用于实际的软测量建模,还可用于对不同建模方法的分析、比较。对赖氨酸发酵过程的实际应用效果验证了软件的有效性和可行性。
Aiming at the problem that some key biological parameters are difficult to be measured online in fermentation process, combined artificial intelligence technology with object-oriented programming technology, a web-based modeling system was designed and realized, which supply all kinds of neural network algorithms and optimization design. Based on lightweight framework and making use of standard of Eclipse, the web layer, business logic layer and data persistence layer were separated. On business logic layer, neural network algorithms were implemented, and the interactive exchange between java and matlab was acbieved. The system can be applied to practical soft-sensing modeling and the analysis and compare among different modeling methods. The validity and feasibility of the software were verified by practical effect application in lysine fermentation process.
出处
《传感器与微系统》
CSCD
北大核心
2010年第8期135-137,140,共4页
Transducer and Microsystem Technologies
基金
国家"863"计划资助项目(2007AA04Z179)
教育部高等学校博士学科专项科研基金资助项目(20070299010)
关键词
SSH框架
MVC
神经网络
建模
发酵过程
struts spring hibernate（SSH） framework
neural network
modeling
fermentation process
