Bioreactor profile control by a nonlinear auto regressive moving average neuro and two degree of freedom PID controllers |
| |
| Affiliation: | 1. Department of Electrical Engineering, University of Malaya, Faculty of Engineering, 50603 Kuala Lumpur, Malaysia;2. School of Microelectric Engineering, University Malaysia Perlis, Malaysia;3. Department of Petroleum and Chemical Engineering, Faculty of Engineering, Institut Teknologi Brunei, Brunei Darussalam;1. Department of Computer Science and Numerical Analysis, University of Cordoba, Campus de Rabanales, 14071 Cordoba, Spain;2. Department of Computer Science and Automatic Control, UNED, Juan del Rosal 16, 28040 Madrid, Spain;1. State Key Laboratory of Alternative Electrical Power System with Renewable Energy Sources, North China Electric Power University, Beijing 102206, PR China;2. School of Control and Computer, North China Electric Power University, Beijing 102206, PR China;3. Department of Chemical Engineering, Chung-Yuan Christian University, Chung-Li, Taiwan 320, ROC;1. Beijing Key Lab of Membrane Science and Technology, College of Chemical Engineering, Beijing University of Chemical Technology, 100029 Beijing, China;2. Department of Chemical Engineering, University of California, Davis, CA 95616, USA |
| |
| Abstract: | This paper presents the use of nonlinear auto regressive moving average (NARMA) neuro controller for temperature control and two degree of freedom PID (2DOF-PID) for pH and dissolved oxygen (DO) of a biochemical reactor in comparison with the industry standard anti-windup PID (AWU-PID) controllers. The process model of yeast fermentation described in terms of temperature, pH and dissolved oxygen has been used in this study. Nonlinear auto regressive moving average (NARMA) neuro controller used for temperature control has been trained by Levenberg–Marquardt training algorithm. The 2DOF-PID controllers used for pH and dissolved oxygen have been tuned by MATLAB's auto tune feature along with manual tuning. Random training data with input varying from 0 to 100 l/h have been obtained by using NARMA graphical interface. The data samples used for training, validation and testing are 20,000, 10,000 and 10,000 respectively. Random profiles have been used for simulation. The NARMA neuro controller and the 2DOF-PID controllers have shown improvement in rise time, residual error and overshoot. The proposed controllers have been implemented on TMS320 Digital Signal Processing board using code composure studio. Arduino Mega board has been used for input/output interface. |
| |
| Keywords: | Bioreactor profile Inverse neural network Process control NARMA neuro controller |
| 本文献已被 ScienceDirect 等数据库收录！ |
|
