K12604_COVER_PRINT_REV.pdf 11/23/11 9:26 AM El-Mansi BIOLOGICAL SCIENCES & LIFE SCIENCES Third Edition Fermentation Microbiology and Biotechnology, Third Edition explores and illustrates the diverse array of metabolic pathways employed for the production of primary and secondary metabolites as well as biopharmaceuticals This updated and expanded edition addresses the whole spectrum of fermentation biotechnology, from fermentation kinetics and dynamics to protein and co-factor engineering C M Y CM The third edition builds upon the fine pedigree of its earlier predecessors and extends the spectrum of the book to reflect the multidisciplinary and buoyant nature of this subject area To that end, the book contains four new chapters: MY CY CMY K • • • • Functional Genomics Solid-State Fermentations Applications of Metabolomics to Microbial Cell Factories Current Trends in Culturing Complex Plant Tissues for the Production of Metabolites and Elite Genotypes Organized and written in a concise manner, the book’s accessibility is enhanced by the inclusion of definition boxes in the margins explaining any new concept or specific term The text also contains a significant number of case studies that illustrate current trends and their applications in the field With contributions from a global group of eminent academics and industry experts, this book is certain to pave the way for new innovations in the exploitation of microorganisms for the benefit of mankind Third Edition K12604 ISBN: 978-1-4398-5579-9 90000 781439 855799 Fermentation Microbiology and Biotechnology Fermentation Microbiology and Biotechnology Fermentation Microbiology and Biotechnology Third Edition Edited by E.M.T El-Mansi • C.F.A Bryce • B Dahhou S Sanchez • A.L Demain • A.R Allman Fermentation Microbiology and Biotechnology Third Edition This page intentionally left blank Fermentation Microbiology and Biotechnology Third Edition Edited by E.M.T El-Mansi • C.F.A Bryce • B Dahhou S Sanchez • A.L Demain • A.R Allman Boca Raton London New York CRC Press is an imprint of the Taylor & Francis Group, an informa business MATLAB® is a trademark of The MathWorks, Inc and is used with permission The MathWorks does not warrant the accuracy of the text or exercises in this book This book’s use or discussion of MATLAB® software or related products does not constitute endorsement or sponsorship by The MathWorks of a particular pedagogical approach or particular use of the MATLAB® software CRC Press Taylor & Francis Group 6000 Broken Sound Parkway NW, Suite 300 Boca Raton, FL 33487-2742 © 2012 by Taylor & Francis Group, LLC CRC Press is an imprint of Taylor & Francis Group, an Informa business No claim to original U.S Government works Version Date: 20111007 International Standard Book Number-13: 978-1-4398-5581-2 (eBook - PDF) This book contains information obtained from authentic and highly regarded sources Reasonable efforts have been made to publish reliable data and information, but the author and publisher cannot assume responsibility for the validity of all materials or the consequences of their use The authors and publishers have attempted to trace the copyright holders of all material reproduced in this publication and apologize to copyright holders if permission to publish in this form has not been obtained If any copyright material has not been acknowledged please write and let us know so we may rectify in any future reprint Except as permitted under U.S Copyright Law, no part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying, microfilming, and recording, or in any information storage or retrieval system, without written permission from the publishers For permission to photocopy or use material electronically from this work, please access www.copyright.com (http:// www.copyright.com/) or contact the Copyright Clearance Center, Inc (CCC), 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400 CCC is a not-for-profit organization that provides licenses and registration for a variety of users For organizations that have been granted a photocopy license by the CCC, a separate system of payment has been arranged Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe Visit the Taylor & Francis Web site at http://www.taylorandfrancis.com and the CRC Press Web site at http://www.crcpress.com A man lives not only his personal life as an individual, but also, consciously or unconsciously, the life of his epoch and his contemporaries Thomas Mann Professor Dr Mahmoud Ismael Taha, A chemist of exactitude and graceful humility (1924–1981) This edition is dedicated with affection and gratitude to the memory of the late Professor Dr Mahmoud Ismael Taha, who ignited in me a lifelong passion for biochemistry; he often reminded me that Louis Pasteur was a chemist E.M.T El-Mansi (Editor-in-Chief) This page intentionally left blank Contents Preface .ix Acknowledgments .xi Editors xiii Contributors xvii Chapter Fermentation Microbiology and Biotechnology: An Historical Perspective E.M.T El-Mansi, Charlie F.A Bryce, Brian S. Hartley, and Arnold L Demain Chapter Microbiology of Industrial Fermentation: Central and Modern Concepts E.M.T El-Mansi, F Bruce Ward, and Arun P Chopra Chapter Fermentation Kinetics: Central and Modern Concepts 37 Jens Nielsen Chapter Microbial Synthesis of Primary Metabolites: Current Trends and Future Prospects .77 Arnold L Demain and Sergio Sanchez Chapter Microbial and Plant Cell Synthesis of Secondary Metabolites and Strain Improvement 101 Wei Zhang, Iain S Hunter, and Raymond Tham Chapter Applications of Metabolomics to Microbial “Cell Factories” for Biomanufacturing: Current Trends and Future Prospects 137 David M Mousdale and Brian McNeil Chapter Flux Control Analysis and Stoichiometric Network Modeling: Basic Principles and Industrial Applications 165 E.M.T El-Mansi, Gregory Stephanopoulos, and Ross P Carlson Chapter Enzyme and Cofactor Engineering: Current Trends and Future Prospects in the Pharmaceutical and Fermentation Industries 201 George N Bennett and Ka-Yiu San vii viii Chapter Contents Conversion of Renewable Resources to Biofuels and Fine Chemicals: Current Trends and Future Prospects 225 Aristos A Aristidou, Namdar Baghaei-Yazdi, Muhammad Javed, and Brian S. Hartley Chapter 10 Functional Genomics: Current Trends, Tools, and Future Prospects in the Fermentation and Pharmaceutical Industries 263 Surendra K Chikara and Toral Joshi Chapter 11 Beyond Cells: Culturing Complex Plant Tissues for the Production of Metabolites and Elite Genotypes 295 Pamela J Weathers, Melissa J Towler, and Barbara E Wyslouzil Chapter 12 Cell Immobilization and Its Applications in Biotechnology: Current Trends and Future Prospects 313 Ronnie G Willaert Chapter 13 Biosensors in Bioprocess Monitoring and Control: Current Trends and Future Prospects 369 Chris E French and Chris Gwenin Chapter 14 Solid-State Fermentation: Current Trends and Future Prospects 403 Lalita Devi Gottumukkala, Kuniparambil Rajasree, Reeta Rani Singhania, Carlos Ricardo Soccol, and Ashok Pandey Chapter 15 Bioreactors: Design, Operation, and Applications 417 Anthony R Allman Chapter 16 Control of Industrial Fermentations: An Industrial Perspective 457 Craig J.L Gershater and César Arturo Aceves-Lara Chapter 17 Monitoring and Control Strategies for Ethanol Production in Saccharomyces Cerevisiae 489 Gilles Roux, Zetao Li, and Boutaib Dahhou Appendix: Suppliers List 519 Index 521 Preface I beseech you to take interest in these sacred domains, so expressively called laboratories Ask that, there be more and that they be adorned for these are the temples of the future, wealth and well being Louis Pasteur Microorganisms, free-living and immobilized, are widely used industrially as catalysts in the biotransformation of many chemical reactions, especially in the production of stereospecific isomers The high specificity, versatility, and the diverse array of microbial enzymes (proteomes) are currently being exploited for the production of important primary metabolites including amino acids, nucleotides, vitamins, solvents, and organic acids, as well as secondary metabolites such as antibiotics, hypercholesterolemia agents, enzyme inhibitors, immunosuppressants, and antitumor therapeutics Recent innovations in functional genomics, proteomics, metabolomics, bioinformatics, biosensor technology, nanobiotechnology, cell and enzyme immobilization, and synthetic biology and in silico research are currently being exploited in drug development programs to combat disease and hospital-acquired infections as well as in the formulation of a new generation of therapeutics The third edition builds upon the fine pedigree of its earlier predecessors and extends the spectrum of the book to reflect the multidisciplinary and buoyant nature of this subject area To that end, four new chapters have been commissioned: • • • • Functional Genomics Solid-State Fermentations Applications of Metabolomics to Microbial Cell Factories Current Trends in Culturing Complex Plant Tissues for the Production of Metabolites and Elite Genotypes More exciting advances and discoveries are yet to be unraveled, and the best is yet to come as we enter a new era in which the exploitations of microorganisms continue to astonish the world community, especially the use of renewable resources and the generation of new therapeutics to combat disease are recognized as an urgent need To that end, Professor Brian S Hartley predicts the emergence of a new era in which “biorefineries” play a central role in climate control and the balance of geochemical cycles in our ecosystem To aid learning and to make the text more lively and interactive, boxes highlighting the definitions of new and central concepts are shown in the margin, a feature that is now synonymous with our book We very much hope that the third edition will be assimilated and appreciated by those actively engaged in the pursuit of advancing our field, and to that end, the editor-in-chief wishes to stress his readiness to receive your feedback, including suggestions by authors who wish to add or extend the knowledge base of our book, which is becoming increasingly global with every edition ix 508 Fermentation Microbiology and Biotechnology, Third Edition These kinds of residuals are very significant and helpful in the classification procedure because they are robust to measurement noise The goal is to diagnose the kind of fault by evaluating the given residuals Therefore, two behavior models (M1, M2) are implemented Each of them is designed to be sensitive to one fault, as shown in Figure 17.19 These models are generated as explained in Figure 17.3 (Section 17.3.2.1) The models’ output (f1, f2) denotes the presence or not of fault For each fault the training data are collected in the presence of measurement Gaussian noise with zero mean and variance equal to 0.5 The effect of this noise on residuals is very important compared with faults Thus, residual evaluation is delicate The validation has been done on multiple data with different noise of variance 0.2, 0.3, 0.5, and The results obtained with variance equal to are presented As can be seen from Figure 17.20, residuals are generated as growth rate (μm) changes from 0.38 to 0.40 h–1 As dictated by the design, a fault is detected should any residuals (e.g., γ1) depart from a zero value Unfortunately, this behavior is not clearly observed in the residuals γ1 and γ2 plotted in Figure 17.20 because of a high level of noise in the measurements In this case the composite residuals can be used, but just for fault detection The residual evaluation using behavioral models M1 γ1 γ2 S1 S2 Decision making M2 Amplitude Amplitude Amplitude Amplitude FIGuRE 17.19 Fault decision f2 Decision making using behavioral models γ1 –5 50 100 150 200 250 300 50 100 150 200 250 300 50 100 150 200 250 300 50 100 150 200 250 300 γ2 –5 10 S1 10 S2 State f1 Fault in Ks No fault 50 100 150 Time (h) FIGuRE 17.20 Residuals valuation with fault in μm (with noise) 200 No fault 250 300 Monitoring and Control Strategies for Ethanol Production in Saccharomyces Cerevisiae 509 allows us to detect and isolate faults as shown in the bottom of Figure 17.20 by the transient from state (no fault) to state (fault in μm) The fault is now simulated by changing the parameter Ks from to 4.7 g/L at t = 130 h and from 4.7 to g/L at t = 250 h The corresponding residuals are illustrated in Figure 17.21 Similarly, the behavioral models based on residual evaluation show clearly the transient from state (no fault) to state (fault in Ks) at t = 130 h and from state to state at t = 250 h 17.5.2.3 Principle of Fault-Tolerant Control The control law is of the form u = K ( x , ϕ ), where the controller parameter vector is ϕ ∈R q Using this controller, the closed-loop system is x = f ( x , θ, ϕ ) and y = Cx The choice of the controller parameter vector ϕ is called controller configuration One assumes that this choice is related with a cost function J(θ, ϕ ) Let p1 (θ, ϕ ), p2 (θ, ϕ ) … pi (θ, ϕ ), and pm (θ, ϕ ) are m parameters depending of the closed-loop system with respective to the constraint condition These m parameters can take values of eigenvalues (these are a special set of scalars associated with a linear system of equations that are sometimes also known as characteristic roots, characteristic values or proper values; these values describe the dynamic of the process) of the closed-loop system or other values following the application context Then the objective of controller parameter configuration is that the closed-loop system satisfies J (θ, ϕ ) (17.8) pi (θ, ϕ ) ∈ Ω i , ∀i = 1, , m (17.9) Amplitude Amplitude Amplitude Equation 17.8 can be analytic or nonanalytic Equation 17.9 represents the constraint condition of the controller parameters choice Ωi represents a certain domain in the complex plane For example, if pi is an eigenvalue, then Ωi can be chosen as the left-half s plane The constraint condition (Equation 17.9) implies a set of crucial indexes that should be satisfied The closed-loop system is called a system with good stability if Equation 17.9 is satisfied γ1 –5 0 State Amplitude –5 10 0 10 γ2 100 150 200 250 300 50 100 150 200 250 300 50 100 150 200 250 300 50 100 150 200 250 300 S1 S2 0 50 Fault in Ks No fault 50 100 150 Time (h) FIGuRE 17.21 Residuals evaluation with fault in KS (with noise) 200 No fault 250 300 510 Fermentation Microbiology and Biotechnology, Third Edition After the occurrence of the fault, we accommodate the fault by controller reconfiguration We change the controller parameter vector according to the system information We use fault detection, isolation, and identification procedures to get this information Then, in different periods the controller parameter vector is chosen as follows: Before the fault occurrence: The system parameter vector is known and equal to θ0 Δ is only the point θ0 The controller parameter vector should be optimized in the domain ΨM(θ0) The optimum controller parameter vector in this case as φ0 ΨM(θ0) contains all possible maximum available domain ΨM(Δi), where Δi is any domain containing the point θ0 The optimum result in this case is the best one of the results corresponding to all possible ΨM(Δi ) After the fault occurrence and before the fault detection: The system parameter vector in this case is θf, and the corresponding maximum available domain is ΨM(θ f ) Because we not know the fault occurrence, the controller parameter vector φ0 is still used, which may be not in the domain ΨM(θ f ) So, Equation 17.9 may be not satisfied, therefore the system may be unstable After the fault is detected but it is not isolated: We have known the fault occurrence; we will change the controller parameter vector to accommodate the fault Because the fault has not been isolated, we not know the location of the system faulty parameter vector, so we considered that θ may be any point in its maximum possible domain ΔM The maximum available domain corresponds to ΔM is ΨM(ΔM) Because ΔM contains all possible Δ, ΨM(ΔM) will be subset of any ΨM(Δ); that is to say ΨM(ΔM) is the smallest of all possible maximum available domains ΨM(Δ) If the condition in Equation 17.9 is satisfied by the open-loop system, then ΨM(ΔM ) ≠ ϕ because it at least has an element φoff, where φoff is the controller parameter vector value, which places the system in a closed-loop state The controller parameter vector should be optimized in the domain ΨM(ΔM ), which is only a small subset of ΨM(θ f ), but not in the whole domain ΨM(θ f ) because we not know ΨM(θ f ) in this case We note that the optimum controller parameter vector in this case is φd The result in this case is worse than the one optimized using ΨM(θ f ), but it is the best choice in this case The condition in Equation 17.9 is satisfied because ΨM(ΔM ) is a subset of ΨM(θ f ) After the fault is isolated and is identified: We assume that the estimation of the system parameter vector is θ f Because we know that the best optimum result is based on ΨM(θf), we want to get the estimated value of θf as quickly as possible Usually a precise estimation cannot be obtained immediately; therefore, we assume that in the early period (when the fault is identified) the estimation is given with the error limits In other words, the estimation is not a point but a possible domain defined by a line segment Because the fault has been isolated, this domain will be in a parallel of the coordinate axes that passes the point θ0 We note this line segment as the domain Δii, then Δii will be a subset of ΔM and it contains the point θf Along with the operation of the fault identification, the domain Δii will become smaller and smaller, and at the end it converges to the point θf Accordingly, the correspondent maximum available domain ΨM(Δii) is a subset of ΨM(θ f ) and it contains the domain ΨM(ΔM ) The controller parameter vector should be optimized in ΨM(Δii) We note the optimum controller parameter vector in this case as φii The result in this case will be worse than the one optimized in ΨM(θ f ) but better than the ones optimized in ΨM(ΔM) Along with the operation of the fault identification, ΨM(Δii) will become bigger and bigger and converges to ΨM(θ f ) So, the optimum result will become increasingly better and better and converge to the one that best corresponds to ΨM(θ f ) The objective of the control is to make the specific growth rate μ(t) of the system in Equation 17.3 follow the rate μr(t) of a given reference model This is done by manipulating the dilution rate D(t) The reference model is chosen as (Li and Dahhou 2006) Monitoring and Control Strategies for Ethanol Production in Saccharomyces Cerevisiae dCr (t ) = µ r (t )Cr (t ) − r (t )Cr (t ) dt dCr (t ) µ r (t )Cr (t ) + r (t ) Sin (t ) − r (t ) Sr (t ) = Yc / s dt dPr (t ) Yp / s µ r (t )Cr (t ) − r (t ) Pr (t ) = Yc / s dt 511 (17.10) It has a same structure as the process system model In the reference model in Equation 17.10, r(t) is input of the reference model and μr(t) is the same structure of μ(t) The control variable is chosen as (Li and Dahhou 2006) D(t ) = T (t )r (t ) − L (t )µˆ (t ) (17.11) with T (t ) = ar (t ) aˆ (t ) L (t ) = ar (t ) −1 aˆ (t ) (17.12) and µˆ (t ) = µˆ m S (t ) ˆ K s + S (t ) (17.13) The terms aˆ (t ) and ar(t) are given by (Zeng et al 1993b) ˆ µˆ (t ) C (t ) µ (t ) − Yp / s µ m S (t ) (17.14) µ (t ) C (t ) µ r (t ) − r r Yp / s µ mr Sr (t ) (17.15) aˆ (t ) = ar (t ) = The constraint condition of Equation 17.9 of the controller parameter choice is chosen as • Ensuring that the eigenvalue of the closed-loop system is in the left-half s plane or, in other words, the closed-loop system is stable; • Ensuring the gain kc can be limited in the area between the two values kc0 and kc1: kc (t ) = T (t ) δ(t ) L (t ) + (17.16) When the fault is associated with a parameter μm, we know that K s = K s0 For the parameter μm, we know from its estimation mˆ m and the corresponding bounds µ bm ≤ µ m ≤ µ ma that we can get δ(t ) = µˆ m µm (17.17) 512 Fermentation Microbiology and Biotechnology, Third Edition According to the limits it will be δ (t ) = µˆ m µˆ and δ max (t ) = mb µ am µm (17.18) When the fault is associated with a parameter Ks, we know that µ m = µ 0m For the parameter K s, we know from its estimation Kˆ s, and the corresponding limits K sb ≤ K s ≤ K sa, we can get δ(t ) = K s + S (t ) Kˆ + S (t ) (17.19) s According to the limits it will be δ (t ) = K sb + S (t ) K a + S (t ) and δ max (t ) = s ˆ K s + S (t ) Kˆ s + S (t ) (17.20) For different faults, using Equation 17.14 or Equation 17.16, we can calculate the controller parameters L(t) and T(t) To show the validity of the method, simulation experiments have been done on the alcoholic fermentation process The nominal value of the system parameter vector is [μm, Ks] = [0.38, 5.0] The parameter vector value of the reference model is [μmr, Ksr] = [0.30, 5.0] The fault takes place at t = 100 h The faulty parameter vector is [μm, Ks] = [0.38, 2.0] Non-FTC control: To make the comparison, in the first we give two results of non-FTC control One is fixed parameter control and another is adaptive control In the non-FTC control situation, the result of controller parameter calculation is not modified • Fixed parameter control: In the controller parameter calculation, the process parameter vector [μm, Ks] is substituted by its nominal value [m 0m , K s0 ] Figure 17.22 presents the control result It shows that after the fault occurrence at tf = 100h the control effect becomes very bad There is a large follow error between μ(t) and μr(t) Figure 17.23 shows the system gain kc(t) It shows that after the fault occurrence the gain kc(t) has great variation and its deviation from is large This is caused by the parameter vector difference between θf of the postfault system and θ0, which is used to calculate the controller parameters and the variable μ(t) • Adaptive control: In the controller parameter calculation, the process parameter vector [μm, Ks] is substituted by its estimation value [mˆ m , Kˆ s ] This estimation is provided by 0.16 µ(t) µr(t) 0.14 0.12 0.1 0.08 0.06 0.04 0.02 50 75 100 125 150 175 FIGuRE 17.22 The control result (fixed parameter control) 200 225 250 275 300 (h) Monitoring and Control Strategies for Ethanol Production in Saccharomyces Cerevisiae 513 1.2 0.8 0.6 0.4 0.2 50 kc(t) 75 100 125 150 175 200 225 250 275 300 (h) 200 225 250 275 300(h) FIGuRE 17.23 The system gain (fixed parameter control) 0.25 µ(t) µr(t) 0.2 0.15 0.1 0.05 50 75 100 125 150 175 FIGuRE 17.24 The control result (adaptive control) a parameter-state joint estimate procedure For details of the parameter-state joint estimate, the reader is refereed to Zeng et al (1993b) Figure 17.24 shows the control result It shows that after the fault occurrence the control effect is very bad Figure 17.25 shows that the system gain kc(t) shows great variation and the deviation from is large FTC control: In the FTC control manner, kc0 is chosen as 0.85 • FTC control with fault isolation and identification but without constraint to the controller parameter L(t): In the controller parameter calculation, the process parameter vector [μm, Ks] is substituted by its estimation value [mˆ m , Kˆ s ] This estimation is provided by the intervals that are based on the fault isolation and identification method Figure 17.26 shows the control result It shows that the control effect is much better than the ones of Figures 17.22 and 17.24 However, in the beginning period after the fault occurrence there is also an evident control error Figure 17.27 shows the gain kc(t) It shows that kc(t) has not been limited by the condition kc(t) ≥ 0.85 • FTC control with fault isolation, identification, and with constraint to the controller parameter L(t): In the first we calculate the controller parameters by the same way as in the previous case (without constraint), then the calculated result is modified according to the fault-tolerant control procedure (Li and Dahhou 2006) Figure 17.28 presents the control result It shows that the effect is much better than the ones of preceding examples In the begin period after the fault occurrence, the control error is much smaller than the one of Figure 17.26 Figure 17.29 presents the gain kc(t) It shows that, in the beginning period after the fault occurrence, the deviation of kc(t) from is much smaller than the case of Figure 17.27, and it accords with the condition kc(t) ≥ 0.85 And in the following time, kc(t) always equals 1, and the control arrives at its optimum 514 Fermentation Microbiology and Biotechnology, Third Edition kc(t) 2.5 1.5 0.5 50 75 100 125 150 175 200 225 250 275 300(h) 200 225 250 275 300(h) FIGuRE 17.25 The system gain (adaptive control) 0.16 0.14 0.12 0.10 0.08 µ(t) µr(t) 0.06 0.04 50 75 100 125 150 175 FIGuRE 17.26 The control result (without constraint of the controller parameter L(t)) 1.2 1.0 0.8 0.6 0.4 0.2 50 kc(t) 75 100 125 150 175 200 225 250 275 300(h) 275 300(h) FIGuRE 17.27 The system gain (without constraint of the controller parameter L(t)) 0.16 0.14 0.12 0.1 0.08 0.06 0.04 50 µ(t) µr(t) 75 100 125 150 175 200 225 250 FIGuRE 17.28 The control result (with constraint of the controller parameter L(t)) Monitoring and Control Strategies for Ethanol Production in Saccharomyces Cerevisiae 515 1.2 0.8 0.6 0.4 0.2 50 kc(t) 75 100 125 150 175 200 225 250 275 300(h) FIGuRE 17.29 The system gain (with constraint of the controller parameter L(t)) 17.6 CONCLuSIONS In this chapter, we tried to highlight the stages to be followed and the difficulties engineers can meet to ensure good control of the biotechnological processes and, in particular, the alcoholic fermentation processes We developed two approaches of modeling according to the objective considered: one is quantitative and the other is qualitative The models obtained from mass balance considerations are called nonstructured models and those obtained from the methods of classification are called behavioral models The nonstructured models were used for the development of the algorithms of the software sensors, controls, and fault detection and isolation The behavioral models were used for the determination of physiological states and the fault detection and isolation problem These behavioral models use the environmental variables such as temperature, pH, and stirrer speed, as well as the expert system’s knowledge of the processes to determine the physiological states The validation of the nonstructured model obtained was carried out to determine the best minimization of a mathematical criterion by using experimental data available for fermentation in discontinuous mode We preferred this mode because it is richer in kinetic information than the continuous mode, which provides only stationary states The classification method was applied to the alcoholic fermentation process It was supposed that the behavior during batch processing influences the phenomena observed during the continuous phase Therefore, a good knowledge of the physiological states during the batch phase is of prime importance for the biologist In agreement with the expert system describing the process, a total of four signals were selected, which, according to its knowledge, contain the most relevant information to determine the physiological states Considering the lack of reliable and inexpensive sensors in the field of biotechnological processes, we were motivated by the development of software sensors We used the nonstructured model for the development of these types of algorithms The experimental results showed that from the measurement of the substrate concentration, one can estimate the biomass and product concentrations with the help of an adaptive observer that jointly estimates the state and the parameters This algorithm was validated based on three types of experiments (batch, fed-batch, and continuous modes) and gave satisfactory results Experimental results obtained from the control of alcoholic fermentation in continuous mode are presented The objective was the regulation and the tracking of the substrate concentration inside of the bioreactor while acting on dilution rate The results obtained are also satisfactory We obtained these results by the application of an indirect adaptive control scheme using an estimator with freezing parameter when the condition of persistent excitation is not satisfied This situation can be caused by a change of the system dynamics, and thus we concluded that it was necessary to develop 516 Fermentation Microbiology and Biotechnology, Third Edition a supervisory system to supervise the control algorithms and to detect and isolate the faults coming from process dynamics at the same time In the supervision system, the two approaches of modeling are used for the determination of physiological states and for faults detection and isolation For the determination of physiological states, the classification method based on fuzzy techniques was used This method enabled us to construct a behavioral model that was used in a supervision system as a reference model for online monitoring task The nonstructured model obtained was used for development of FDI algorithms The method used is based on adaptive observers This method gives satisfactory results in noisefree conditions, but it proves limited in the presence of noise To solve this problem, a combined analytical and knowledge-based method was proposed for FDI The FTC method proposed in this work has two specifics: It uses the idea of active FTC and of passive FTC Therefore, in the early period after the fault occurrence, when the information of the system fault is insufficient, the crucial performance of the system can also be maintained Along with increasing the amount of information, the control effect of the FTC system becomes progressively more perfect The realization of this objective benefits from a reconfigurable controller, the design of which is based on the available controller parameter space analysis The use of interval-based fault isolation and identification approach This interval-based isolation and identification approach is very quick This is the more important factor to make the FTC system produce a beneficial effect This interval-based isolation and identification approach provides superior/inferior limits of the faulty parameter, thus making the design of the above reconfigurable controller possible This FTC method has been used in the simulation of a fermentation microbial-specific growth rate process; the results show the good performance of the method SuMMARY The development of an integrated process control methodology requires the design of a supervisory block containing all available information and procedures This supervisory block recognizes specific indicators/parameters and acts, if necessary, on the process or by informing the operator Two approaches to modeling were implemented: the behavioral model and the nonstructured model Whereas the former is based on the use of a fuzzy classification technique, the latter is based on mass balance considerations The behavioral model is obtained by using the online measurement of environmental variables (temperature, pH, stirrer speed, etc.) and describes the physiological states of the process The nonstructural model is obtained from mass balance considerations and is used for the development of software sensors, control scheme, and FDI algorithms The whole of the results obtained and the algorithms developed using these modeling approaches are used in a supervisory block to ensure an effective monitoring of the process The application of this methodology to fermentation processes gave satisfactory results REFERENCES Aguilar-Martin, J 1996 Knowledge-based real time supervision Tempus-Modify Workshop Budapest, Hungary Antsaklis, P.J., and K.M Passino, eds 1993 An Introduction to Intelligent and Autonomous Control Norwell, MA: Kluwer Academic Publishers Bâati, L., G Roux, B Dahhou, and J.L Uribelarrea 2004 Unstructured modeling growth of Lactobacilius acidophilus as a function of the temperature Math Comput Simul 65:137–45 Monitoring and Control Strategies for Ethanol Production in Saccharomyces Cerevisiae 517 Ben-Youssef, C 1996 Filtrage, estimation et commande adaptative d’un procédé de traitement des eaux usées Ph.D thesis Institut National Polytechnique de Toulouse Ben-Youssef, C., B Dahhou, F.Y Zeng, and J.L Rols 1996 Estimation and filtering of nonlinear systems: Application to a wastewater treatment process Int J Syst Sci 27:497–505 Dahhou B., G Chamilothoris, and G Roux 1991a Adaptive predictive control of a continuous fermentation process Int J Adapt Cont Signal Process 5:351–62 Dahhou, B., G Roux, and A Cheruy 1993 Linear and nonlinear adaptive control of alcoholic fermentation process Int J Adapt Cont Signal Process 7:213–33 Dahhou, B., G Roux, and I Queinnec 1991b Adaptive control of a continuous fermentation process Presented at the Symposium on Modeling and Control of Technological Systems, Lille, France, pp 738–43 Dahhou, B., G Roux, I Queinnec, and J.B Pourciel 1991c Adaptive pole placement control of a continuous fermentation process Int J Syst Sci 22:2625–38 Dojat, M., N Ramaux, amd D Fontaine 1998 Scenario recognition for temporal reasoning in medical domains Artif Intell Med 14:139–55 Kabbaj, N 2004 Développement d’algorithmes de détection et d’isolation de défauts pour la supervision des bioprocédés Ph.D thesis Université de Perpignan Kabbaj, N., A Doncescu, B Dahhou, and G Roux 2002 Wavelet based residual evaluation for fault detection and isolation Presented at the 17th IEEE International Symposium on Intelligent Control, Vancouver, British Columbia, Canada Kabbaj, N., M Polit, B Dahhou, and G Roux 2001 Adaptive observers based fault detection and isolation for an alcoholic fermentation process Presented the 8th IEEE International Conference on Emerging Technologies and Factory Automation Antibes-Juan les Pins, France Kotch, G.-G 1993 Modular reasoning A new approach towards intelligent control Ph.D thesis Swiss Federal Institute of Technology Zurich, Suisse Li, Z., and B Dahhou 2006 An observers based fault isolation approach for nonlinear dynamic systems Presented at the Second IEEE-EURASIP International Symposium on Control, Communications and Signal Processing 2006 (ISCCSP’06), Marrakech, Morocco Monod, J 1942 Recherche sur la Croissance des Cultures Bactériennes Edition Hermes: Paris M’Saad, M., I.-D Landau, and M Duque 1989 Example applications of the partial state reference model adaptive control design technique Int J Adapt Cont Signal Process 3:155–65 M’Saad, M., I.-D Landau, and M Samaan 1990 Further evaluation of the partial state reference model adaptive control design Int J Adapt Cont Signal Process 4:133–46 Nakkabi, Y., N Kabbaj, B Dahhou, G Roux, and J Aguilar 2003 A combined analytical and knowledge based method for fault detection and isolation Presented at the 9th IEEE International Conference on Emerging Technologies and Factory Automation Vol 2, pp 161–6, Lisbon, Portugal Nakkabi,Y., A Doncescu, G Roux, M Polit, and V Guillou 2002 Application of data mining in biotechnological process Presented at the Second IEEE International Conference on Systems, Man, and Cybernetics, Hammamet, Tunisia Nejjari, F., B Dahhou, A Benhammou, and G Roux 1999a Nonlinear multivariable adaptive control of an activated sludge wastewater treatment process Int J Adapt Cont Signal Process 13:347–65 Nejjari, F., G Roux, B Dahhou, and A Benhammou 1999b Estimation and optimal control design of a biological and wastewater treatment process Int J Math Comp Simul 48:269–80 Piera, N., P Desroches, and J Aguilar-Martin 1989 Lamda: An incremental conceptual clustering system Report 89420 LAAS/CNRS Queinnec, I., C Destruhaut, J.-B Pourciel, and G Goma 1992 An effective automated glucose sensor for fermentation monitoring and control World J Microbiol Biotechnol 8:7–13 Waissman-Vilanova, J 2000 Construction d’un modèle compartemental pour la supervision de procédés: Application une station de traitement des eaux Ph.D thesis Institut National Polytechnique de Toulouse Waissman-Vilanova, J., R Sarrate-Estruch, B Dahhou, and J Aguilar-Martin 1999 Building an automaton for condition monitoring in a biotechnological process Presented at the 5th European Control Conference, Karlsruhe, Alemagne Zeng, F.Y., B Dahhou, and M.T Nihtila 1993a Adaptive control of nonlinear fermentation process via MRAC technique J Appl Math Model 17:58–69 Zeng, F.Y., B Dahhou, M.T Nihtila, and G Goma 1993b Microbial specific growth rate control via MRAC method Int J Syst Sci 24:1973–85 Zhang, Q 2000 A new residual generation and evaluation method for detection and isolation of faults in nonlinear systems Int J Adapt Control Signal Process 14:759–73 This page intentionally left blank Appendix: Suppliers List This suppliers list is not exhaustive and no recommendation is implied by inclusion Instrumentation, sensors, and software, etc http://www.aber-instruments.co.uk/ http://www.adc-service.co.uk/page5.html http://www.broadleyjames.com/ http://www.buglab.com/ http://www.foxylogic.com/ http://www.cerexinc.com/ http://www.wireworkswest.com/fermworks/ http://www.finesse.com/ http://www.fluorometrix.com/products/cellphase/ http://www.flownamics.com/ http://www.gelifesciences.com http://www.hamiltoncomp.com http://invitrogen.com http://www.ksr-kuebler.com/website/index.php http://www.red-y.com/en/ http://uk.mt.com/home http://www.magellaninstruments.com/ http://www.millipore.com http://www.novabiomedical.com http://www.processmeasurement.uk.com/ http://www.ptiinstruments.com/products http://www.oceanoptics.com/homepage.asp http://www.pall.com/ http://www.hitec-zang.com/html/biotech.htm http://ravenbiotech.com/ http://www.russellph.com/foam2.htm http://www.schleicher-schuell.com http://www.watson-marlow.com/ http://www.ysi.com/lifesciences.htm biomass monitor gas analyzers pH and pO2 electrodes, instrumentation external optical density sensor control software optical density monitor with sample chamber fermentation software for process plants sensors, controllers, software external pO2 analyzer sample probes and analyzers single-use systems pO2 sensors media level sensors for larger fermentors mass flow measurement and control Mettler Toledo—sensors, balances, etc perfusion/filtration system filters, tangential flow filtration, single use analytical devices e.g., ammonia, lactate, etc Monitek optical density systems methanol sensor optical sensors for pH, pO2, etc filters, microfiltration, etc RQ monitoring methanol sensing and control antifoam sensor tangential flow filtration peristaltic pumps glucose, carbon dioxide sensors, etc Fermentation equipment http://www.abec.com/design.htm http://www.applikon.com/ http://www.bioxplore.net/home-en.html http://www.sartorius.com http://www.sartorius-stedim.com (single-use systems) http://www.belach.se/kent_series.html http://www.biolafitte.com/Fermentors.htm (Pierre Guerin) http://www.broadleyjames.com/ http://www.bioengineering.ch/ http://www.biotroninc.com/eng (BioG range) http://www.dasgip.de/ 519 520 http://www.electrolab.co.uk/ http://www.infors-ht.com/ http://www.bemarubishi.co.jp/ http://www.newmbr.ch/ http://www.nbsc.com/Main.asp (New Brunswick Scientific) http://www.novaferm.se/ http://www.wavebiotech.com/products/ Appendix: Suppliers List This page intentionally left blank K12604_COVER_PRINT_REV.pdf 11/23/11 9:26 AM El-Mansi BIOLOGICAL SCIENCES & LIFE SCIENCES Third Edition Fermentation Microbiology and Biotechnology, Third Edition explores and illustrates the diverse array of metabolic pathways employed for the production of primary and secondary metabolites as well as biopharmaceuticals This updated and expanded edition addresses the whole spectrum of fermentation biotechnology, from fermentation kinetics and dynamics to protein and co-factor engineering C M Y CM The third edition builds upon the fine pedigree of its earlier predecessors and extends the spectrum of the book to reflect the multidisciplinary and buoyant nature of this subject area To that end, the book contains four new chapters: MY CY CMY K • • • • Functional Genomics Solid-State Fermentations Applications of Metabolomics to Microbial Cell Factories Current Trends in Culturing Complex Plant Tissues for the Production of Metabolites and Elite Genotypes Organized and written in a concise manner, the book’s accessibility is enhanced by the inclusion of definition boxes in the margins explaining any new concept or specific term The text also contains a significant number of case studies that illustrate current trends and their applications in the field With contributions from a global group of eminent academics and industry experts, this book is certain to pave the way for new innovations in the exploitation of microorganisms for the benefit of mankind Third Edition K12604 ISBN: 978-1-4398-5579-9 90000 781439 855799 Fermentation Microbiology and Biotechnology Fermentation Microbiology and Biotechnology Fermentation Microbiology and Biotechnology Third Edition Edited by E.M.T El-Mansi • C.F.A Bryce • B Dahhou S Sanchez • A.L Demain • A.R Allman ... in the quality and yields of traditional crops decreased demand for SCP production 6 Fermentation Microbiology and Biotechnology, Third Edition 1.6 FERMENTATION BIOTECHNOLOGY AND THE PRODuCTION... glycerol Fermentation Microbiology and Biotechnology: An Historical Perspective Box 1.1 “FILL AND SPILL” This pattern of fermentation is essentially a “batch fermentation or “fed-batch fermentation .. .Fermentation Microbiology and Biotechnology Third Edition This page intentionally left blank Fermentation Microbiology and Biotechnology Third Edition Edited