Search

The modeling of biotechnological processes is a complex task, because of the high interaction degree of the phenomena, non-linearities, and uncertainties of reaction kinetics. A novel approach for bioprocess modeling, based on pseudo bond graphs, is proposed in this paper. Bond graph method is an appropriate tool for modeling non-linear processes, eventually along with their control systems. Firstly, the pseudo bond graph method is applied in order to model some prototype bioprocesses starting from the reactions schemes and taking into account the biochemical phenomena. Then the rules used for the design of these models are applied on two widespread wastewater treatment processes. The first is an activated sludge bioprocess taking place in a sequential batch reactor, and the second is a wastewater biodegradation with the production of methane gas, which is carried out inside a continuous stirred tank bioreactor. The obtained bond graph models and several simulations are conducted using the 20sim modeling and simulation environment. This modeling procedure represents a valuable illustration of the power of the bond graph technique, and can be used as a base for the development of models of bioprocesses with a high level of complexity. [ABSTRACT FROM PUBLISHER]

The paper addresses the modeling and control of the orthogonalization plants for textile industry. An innovative mathematical theory of the textile yarns correction phenomenon, particularly of the orthogonalization phenomenon is developed, as a ground for obtaining the mathematical model of the controlled plant. It is proven that the mathematical model corresponds to a nonlinear time-varying system with several time-delay variables with 2 inputs and 2ns outputs, where 2ns is the number of the slope sensors. The process analysis shows that we have a non-dynamical nonlinearity, which can be separated from the linear dynamical part. The multivariable closed-loop control system and the equivalent closed-loop control structure are developed. Also, a control algorithm which can be implemented on a numerical process computer is obtained. [ABSTRACT FROM AUTHOR]