Design of profile controller for biochemical reactor

The biotechnology industry is growing sharply due to progress in understanding of complex biological systems and high demand for chemical and biologically manufactured products. Bioreactors are used for the production of materials like ethanol, or extraction of enzymes from microorganisms, animal or plant cells. In order to maximize the productivity, the bioreactor needs optimal conditions for process parameters such as pH, temperature, and dissolved oxygen (DO). The goal of achieving high performance controller is the answer to these demands. There are many types of process controllers like the ON/OFF controller, PID controller, and controller based on an artificial intelligence. Neural network and PID controllers have been used together to learning system features, and to reduce the residual error as well as to replace currently PID controllers in the proposed design. This research focuses on designing a profile controller for pH, DO, and temperature that affects production of ethanol. The temperature is controlled by the following a cooling agent, and the pH is controlled by adding the appropriate amount of base or acid, while the dissolved oxygen is controlled by changing the speed of the stirrer. The inputs to the bioreactor are cooling agent (Fag), flow of base (Fb), and stirring speed of the liquid (Nstir). The parameters that need to follow a given profile are temperature, pH, and dissolved oxygen. This thesis presents the use of Inverse Neural Networks (INN) for temperature control of a biochemical reactor and its effect on ethanol production. The process model is derived indicating the relationship between temperature, pH and dissolve oxygen. Using the fundamental model obtained data sets; an inverse neural network has been trained by using the back-propagation learning algorithm. Two types of temperature profile are used to compare the performance of the controllers. The controllers have been simulated to have a quantitative comparison with two types of the controllers and show the effectiveness of the INN controller versus the conventional PID controller. The results obtained by the neural network based INN controller and by PID controller are presented and compared. There is an improvement in performance of INN controller in ISE over PID controller.

Preview PDF FK 2010 44R.pdf Download (1MB) | Preview

Actions (login required)

View Item