multiloop control strategies of aerobic fermentation in a fed-batch bioreactor

• Main Authors: Yu-Cheng Chung, 鍾佑政
• Other Authors: I-Lung Chien
• Format: Others
• Language: zh-TW
• Published: 2004
• Online Access: http://ndltd.ncl.edu.tw/handle/78926371298496807659

碩士 === 國立臺灣科技大學 === 化學工程系 === 92 === For the process of aerobic fermentation, oxygen is also an important factor or essential substrate besides the carbon source substrate. Both carbon source substrate and dissolved oxygen are essential substrates to a fedbatch process, therefore, it is necessary to maintain the concentration of essential substrates at optimized operation to obtain maximum productivity. However, it is difficult to obtain most state variables using the method of online measurements. It’s less robust to utilize open-loop optimized operation. Run away due to model uncertainty will reduce the productivity significantly. Multiloop control strategy that combines a feedforward controller with a feedback controller is proposed in this study. The feedforward controller is an open-loop optimal controller, and the adequate pairs of the feedback controller are determined via the dynamic relative gain array (DRGA). For the nonlinearity of the bioprocess, a gain scheduling feedback controller that is employed by a multimodel through the process of fuzzy modeling, and is tuned using Internal Model Control （IMC） principle to reject loads which is unpredictable in the feedforward controller. The structure of the multimodel utilizes interpolation techniques using only two local models at initial and terminative operating times. This presented strategy will not only perform well for typical inhibited fermentation, but for other types of fermentation as well. It is more robust than another nonlinear controller that has the same closed-loop characteristic function. In addition, the override control strategy of the gas-loop is able to overcome the constraint of the volumetric mass transfer coefficient. The uncertainty of saturated and inhibition constants will give wrong setpoint value of the carbon source substrate concentration control loop. The relationship between the specific growth rate and the carbon source substrate concentration is employed to design a dynamic setpoint searching method, and thus can be used to solve the problem stated above.