Tatiana Ilkova

AN APPROACH FOR OPTIMAL CONTROL OF FED- BATCH FERMENTATION PROCESS WITH MULTI CONTROL VARIABLES

ABSTRACT In this paper a fermentation process of E. Coli production is presented by using real laboratory experimental data. The adequate model of the process is proved. An effective algorithm for process optimization in the presence of three optimal control variables: rotation speed, gas flow rate and substratum floating rate is developed by using a modified approach of method of the dynamic programming. The analysis of results shows that the complex investigation with more control variables vastly raises efficiency of the process.

Optimization of a Whey Bioprocess using Neuro-Dynamic Programming Strategy

ABSTRACT A method for finding the optimal feeding profile for whey fermentation by strain Kluyveromyces lactic MC 5 in a fed-batch bioreactor was developed. The optimal profile maximizes the process effectiveness and minimizes the bioprocess duration. The method is based on Neuro-dynamic Programming (NDP), wherein the optimal control decision is parameterized in the form of a cost-to-go function. The suggested method employs simulations from a heuristic feeding strategy as an initial point to generate the cost-to-go to experimental data. A neural network is applied to obtain cost-to-go as a function of system state. Iterations of the Bellman equation are included to improve the cost function. Thus, the obtained approach guarantees optimal control of the bioreactor when disturbances are present. The developed approach was compared with other methods—the Pontryagins Maximum Principle and Fuzzy Sets Theory. The NDP method provided better results than the other methods.

Dynamic and Neuro-Dynamic Optimization of a Fed-Batch Fermentation Process

A fed-batch fermentation process is examined in this paper for experimental and further dynamic optimization. The optimization of the initial process conditions is developed for to be found out the optimal initial concentrations of the basic biochemical variables --- biomass, substrate and feed substrate concentration. For this aim, the method of dynamic programming is used. After that, these initial values are used for the dynamic optimization carried out by neuro-dynamic programming. The general advantage of this method is that the number of the iterations in the cost approximation part is decreased.

An Approach for Modeling of Aerobic Fed-Batch Fermentation Process

ABSTRACT In this paper is development an model for small scale E. Coli fermentation and identifying the model parameters for an effective and reliable of the reactor dynamic with laboratory data from DFG project (5). Model parameters were identified by non-linear regression technique assisted by computer program. Parametric sensitivity analysis indicated specific grown rate of biomass to the most sensitive model parameter. Statistical validity of the model indicated confidence on the prediction of the model.

Using Intercriteria Analysis for Assessment of the Pollution Indexes of the Struma River

In this paper we are presenting the recently proposed approach Intercriteria Analysis (ICrA) for assessment of the pollution index of the Struma River in Bulgaria. The approach is based on the apparatus of the index matrices and the intuitionistic fuzzy sets. At the first we have investigated all indexes at the all measurement point with ICrA and we have searched the dependences between points. Results show the measurement points are dependent criteria and we have ignored some over others. At the second we have applied the ICrA to establish the pollution relations and the model structure based on different criteria involved in the Struma River. The investigations show that there are three positive consonances and dissonances between criteria. Using of a Modification of the Time Series Analysis (MTSA) method we have developed an adequate mathematical model of the pollution dynamic as function of time.

Modelling and Fuzzy-Decision-Making of Batch Cultivation of Saccharomyces cerevisiae using Different Mixing Systems

531 This study is focussed on the modelling and fuzzy-decision-making of impulse mix- ing and vibromixing for a Saccharomyces cerevisiae batch cultivation. Different substrate inhibition models (Monod, Aiba, Andrews, Haldane, Luong, Edward, and Han-Leven- spiel) have been considered in order to explain the cell growth kinetics. The results ob- tained (correlation coefficient, Fisher coefficient, relative error and statistics l) show that all growth rate models are adequate and can be used for modelling. The investigations have shown that the most suitable for both mixing systems (according to the best statis- tical indicators) is the Luong growth rate model, which will be used for the process modelling. A fuzzy-decision-making procedure is developed with the initial conditions (maximal rotation speed for the impulse mixing and maximal amplitude for the vibromix- ing systems). The developed optimisation and results obtained have shown that the im- pulse mixing systems have better productiveness and better glucose assimilation. In addi- tion, it is easier to realize this system.

Neuro-fuzzy based model of batch fermentation of Kluyveromyces marxianus var. lactis MC5

In this work a neuro-fuzzy based model of a whey batch fermentation process by a strain Kluyveromyces marxianus var. lactis MC5 is presented. A three-layered neuro-fuzzy network is realized. The simulation results are compared with conventional models (based on mass balance and differential equations). The neuro-fuzzy model provides a better fitness and allows inclusion of linguistic variables (such as colour, smell, taste, morphophysiology, etc.). The accuracy is approximately equal to this achieved by a conventional neural network. The proposed approach is flexible (with regard to the process model) and quite robust (with regard to the possible uncertainties and to the optimization surface). Future work will focus on applying this approach for modelling of different biotechnological processes.

Multiple Objective Optimisation of Batch Cultivation of Saccharomyces Cerevisiae in Mixing Systems

ABSTRACT Multiple objective optimisation is a natural extension of the traditional optimisation of a single objective function. On the one hand, if the multiple objective functions are commensurate, minimizing a single objective function, it is possible to minimize all the criteria and the problem can be solved using traditional optimisation techniques. On the other hand, if the objective functions are incommensurate or competing, then the minimization of one objective function requires a compromise in another objective function. Here we discuss the problems of multiple objective optimisation of batch cultivation of Saccharomyces cerevisiae in different mixing systems (impulse and vibromixing). The multiple objective optimisation problems are transformed to a single objective function with weight coefficients. A combined algorithm is applied for solving the single optimisation. The applied multiple objective optimisation of the process showed a vast increase in the productivity and, respectively, decrease in the residual substrate concentration.