- Email: [email protected]
On-line recognition and control of physiological state by computing intelligence in fermentation processes
VOL. 86, 1998
Abstracts of the Articles Printed in Seibutsu-kogaku Kaishi Vol. 76, No. 8 (1998)
On-Line Recognition and Control of Physiological State by Computing Intelligence in Fermentation Processes. -MonographHIROSHISHIMZU(Department of Biotechnology, Graduate School oj Engineering, Osaka University, 2-1 Yamadaoka, Suita, Osaka 5650871) Seibutsu-kogaku 76: 338-348. 1998. The application of computing intelligence, specifically fuzzy logic and a neural network, to feature capturing, state recognition, fault diagnosis, and control in bioprocesses is discussed through three examples of Saccharomyces cerevisiae fermentations. The ethanol concentration in a fed-batch culture was controlled by a newly developed fuzzy controller that is also to diagnose the status of the glucose concentration in the medium, i.e., the depletion or overfeeding of glucose, using not only the ethanol concentration data but also the carbon dioxide concentration in the exhaust gas. As a result, the control performance was much improved and overaction of the controller was avoided. A fuzzy physiological state recognition system was developed as a powerful tool to capture the physiological states in the fedbatch fermentation process. The error vector was newly defined in the macroscopic elemental balance. The physiological states were characterized by a database of error vectors, and membership functions for state recognition were constructed based on the error vectors. The physiological states could be recognized, including an abnormal case in which aerobic ethanol production occurred with low growth. A technique for the integration of fuzzy logic and mathematical and stoichiometric analysis is also discussed. A novel neural network-an autoassociative neural network (AANN)-was applied to fault diagnosis in an a-amylase production process using a temperature-sensitive mutant of Saccharomyces cerevisiae. Faults and uncertainties, such as faulty sensors and plasmid instability, significantly affected the performance of the optimized process. The autoassociative neural network was trained so that the network inputs were reproduced at the output layer. The features of the time courses of the state variables in “good” fermentations were captured by the AANN, and the data from “bad” fermentations could be discriminated successfully. By implementing corrective action after fault detection, performance was recovered and the final production amount was increased.
Enhancement of an Immobilized Enzyme Reaction Using Heat-Generating Carriers under an Alternating Magnetic Field. YAWZOSAKAI*and KEISIJKEWATANABE (Department of Applied Chemistry, Faculty of Engineering, Utsunomiya University, 2753 Ishiimachi Utsunomiya 321-8585) Seibutsu-kogaku 76: 325-330. 1998. A magnetic heating method was investigated with the aim of enhancing an immobilized enzyme reaction. A polyacrylamide gel carrier containing an appropriate amount of ferromagnetic powder (57% stainless steel or 45% y-Fe,O,) was found to generate heat in an alternating magnetic field (O-1000 Oe, O-l.5 kHz). In the case of gel containing stainless steel powder, the heat generated (Ws, in W or J/s) was given as a function of W,, ~fTp.~, where f and Hare the frequency and magnetic field intensity, respectively. Sucrose hydrolysis was investigated using gel containing ferromagnetic powder, in which invertase was entrapped. Enhancement of the enzyme reaction was proportional to jItP in the condition without thermal inactivation of the enzyme. This enhancement effect was due to an increase in the local temperature around the enzyme immobilized within the gel. * Corresponding author. Deciding the Temperature Course during Sake Mashing Using a GAFNN for Quality Control of Sake. TAIZO HANAI,’ NAOYASUUEDA,’ HIROYUICIHONDA,’ HI~AO TOHYAMA,~and TAKESHIKOBAYASHI’*(Department of Biotechnology, Graduate School of Engineering, Nagoya University, Chikusa-ku, Nagoya 464-8603’ and Sekiya Brewing Co. Ltd., Taguchi Shitaracho, Kitashitara-gun, Aichi 441-23OP) Seibutsu-kogaku 76: 331-337. 1998. Simulation models for Baumt and alcohol concentration from the 11th day to the end of the sake mashing were constructed using a fuzzy neural network (FNN). The models could simulate the time courses of Baumt and alcohol concentration in 17 actual sake mashings. Average errors at the ends of the mashings were 0.22 and 0.40% for Baumt and alcohol concentration, respectively. By applying a genetic algorithm (GA) with the simulation models, temperature time courses were calculated with good accuracy, and the target values for Baumt! and alcohol concentration on the final day could be achieved. To make a variety of sakes with different qualities, temperature courses were calculated against 3 target values: higher (+0.3), ordinary (O.O), and lower (-0.3) final day Baumes. The calculated temperature courses were found to be similar to a Toji’s (expert’s) strategy for making decisions on temperature. By applying this procedure, quality control of sake can be realized. * Corresponding author.
247
Abstracts of the Articles Printed in Seibutsu-kogaku Kaishi Vol. 76, No. 8 (1998)
On-Line Recognition and Control of Physiological State by Computing Intelligence in Fermentation Processes. -MonographHIROSHISHIMZU(Department of Biotechnology, Graduate School oj Engineering, Osaka University, 2-1 Yamadaoka, Suita, Osaka 5650871) Seibutsu-kogaku 76: 338-348. 1998. The application of computing intelligence, specifically fuzzy logic and a neural network, to feature capturing, state recognition, fault diagnosis, and control in bioprocesses is discussed through three examples of Saccharomyces cerevisiae fermentations. The ethanol concentration in a fed-batch culture was controlled by a newly developed fuzzy controller that is also to diagnose the status of the glucose concentration in the medium, i.e., the depletion or overfeeding of glucose, using not only the ethanol concentration data but also the carbon dioxide concentration in the exhaust gas. As a result, the control performance was much improved and overaction of the controller was avoided. A fuzzy physiological state recognition system was developed as a powerful tool to capture the physiological states in the fedbatch fermentation process. The error vector was newly defined in the macroscopic elemental balance. The physiological states were characterized by a database of error vectors, and membership functions for state recognition were constructed based on the error vectors. The physiological states could be recognized, including an abnormal case in which aerobic ethanol production occurred with low growth. A technique for the integration of fuzzy logic and mathematical and stoichiometric analysis is also discussed. A novel neural network-an autoassociative neural network (AANN)-was applied to fault diagnosis in an a-amylase production process using a temperature-sensitive mutant of Saccharomyces cerevisiae. Faults and uncertainties, such as faulty sensors and plasmid instability, significantly affected the performance of the optimized process. The autoassociative neural network was trained so that the network inputs were reproduced at the output layer. The features of the time courses of the state variables in “good” fermentations were captured by the AANN, and the data from “bad” fermentations could be discriminated successfully. By implementing corrective action after fault detection, performance was recovered and the final production amount was increased.
Enhancement of an Immobilized Enzyme Reaction Using Heat-Generating Carriers under an Alternating Magnetic Field. YAWZOSAKAI*and KEISIJKEWATANABE (Department of Applied Chemistry, Faculty of Engineering, Utsunomiya University, 2753 Ishiimachi Utsunomiya 321-8585) Seibutsu-kogaku 76: 325-330. 1998. A magnetic heating method was investigated with the aim of enhancing an immobilized enzyme reaction. A polyacrylamide gel carrier containing an appropriate amount of ferromagnetic powder (57% stainless steel or 45% y-Fe,O,) was found to generate heat in an alternating magnetic field (O-1000 Oe, O-l.5 kHz). In the case of gel containing stainless steel powder, the heat generated (Ws, in W or J/s) was given as a function of W,, ~fTp.~, where f and Hare the frequency and magnetic field intensity, respectively. Sucrose hydrolysis was investigated using gel containing ferromagnetic powder, in which invertase was entrapped. Enhancement of the enzyme reaction was proportional to jItP in the condition without thermal inactivation of the enzyme. This enhancement effect was due to an increase in the local temperature around the enzyme immobilized within the gel. * Corresponding author. Deciding the Temperature Course during Sake Mashing Using a GAFNN for Quality Control of Sake. TAIZO HANAI,’ NAOYASUUEDA,’ HIROYUICIHONDA,’ HI~AO TOHYAMA,~and TAKESHIKOBAYASHI’*(Department of Biotechnology, Graduate School of Engineering, Nagoya University, Chikusa-ku, Nagoya 464-8603’ and Sekiya Brewing Co. Ltd., Taguchi Shitaracho, Kitashitara-gun, Aichi 441-23OP) Seibutsu-kogaku 76: 331-337. 1998. Simulation models for Baumt and alcohol concentration from the 11th day to the end of the sake mashing were constructed using a fuzzy neural network (FNN). The models could simulate the time courses of Baumt and alcohol concentration in 17 actual sake mashings. Average errors at the ends of the mashings were 0.22 and 0.40% for Baumt and alcohol concentration, respectively. By applying a genetic algorithm (GA) with the simulation models, temperature time courses were calculated with good accuracy, and the target values for Baumt! and alcohol concentration on the final day could be achieved. To make a variety of sakes with different qualities, temperature courses were calculated against 3 target values: higher (+0.3), ordinary (O.O), and lower (-0.3) final day Baumes. The calculated temperature courses were found to be similar to a Toji’s (expert’s) strategy for making decisions on temperature. By applying this procedure, quality control of sake can be realized. * Corresponding author.
247