SCI和EI收录∣中国化工学会会刊

›› 2008, Vol. 16 ›› Issue (1): 62-66.

• • 上一篇    下一篇

An Intelligent Neural Networks System for Adaptive Learning and Prediction of a Bioreactor Benchmark Process

邹志云1,2, 于德弘2, 冯文强2, 于鲁平2, 郭宁2   

  1. 1. School of Mechanical Engineering, Xi’an Jiaotong University, Xi’an 710049, China;
    2. Beijing Research Institute of Pharmaceutical Chemistry, Beijing 102205, China
  • 收稿日期:2007-01-15 修回日期:2007-09-10 出版日期:2008-02-28 发布日期:2008-02-28
  • 通讯作者: ZOU Zhiyun, E-mail: zouzhiyun@sohu.com
  • 基金资助:
    China Scholarship Council Grant (No.21302095).

An Intelligent Neural Networks System for Adaptive Learning and Prediction of a Bioreactor Benchmark Process

ZOU Zhiyun1,2, YU Dehong2, FENG Wenqiang2, YU Luping2, GUO Ning2   

  1. 1. School of Mechanical Engineering, Xi’an Jiaotong University, Xi’an 710049, China;
    2. Beijing Research Institute of Pharmaceutical Chemistry, Beijing 102205, China
  • Received:2007-01-15 Revised:2007-09-10 Online:2008-02-28 Published:2008-02-28
  • Supported by:
    China Scholarship Council Grant (No.21302095).

摘要: The adaptive learning and prediction of a highly nonlinear and time-varying bioreactor benchmark process is studied using Neur On-Line, a graphical tool kit for developing and deploying neural networks in the G2 real time intelligent environment, and a new modified Broyden, Fletcher, Goldfarb, and Shanno (BFGS) quasi-Newton algorithm. The modified BFGS algorithm for the adaptive learning of back propagation (BP) neural networks is developed and embedded into Neur On-Line by introducing a new search method of learning rate to the full memory BFGS algorithm. Simulation results show that the adaptive learning and prediction neural network system can quickly track the time-varying and nonlinear behavior of the bioreactor.

关键词: intelligent system, neural networks, adaptive learning, adaptive prediction, bioreactor process

Abstract: The adaptive learning and prediction of a highly nonlinear and time-varying bioreactor benchmark process is studied using Neur On-Line, a graphical tool kit for developing and deploying neural networks in the G2 real time intelligent environment, and a new modified Broyden, Fletcher, Goldfarb, and Shanno (BFGS) quasi-Newton algorithm. The modified BFGS algorithm for the adaptive learning of back propagation (BP) neural networks is developed and embedded into Neur On-Line by introducing a new search method of learning rate to the full memory BFGS algorithm. Simulation results show that the adaptive learning and prediction neural network system can quickly track the time-varying and nonlinear behavior of the bioreactor.

Key words: intelligent system, neural networks, adaptive learning, adaptive prediction, bioreactor process