Design of AI Controllers for Multivariable Bioprocesses

Microcontroller-based systems require the design of a hardware/software interface that enables software running on the microcontroller to control external devices. This paper developed such a automatical system containing a programmable central microprocessor, ATmega32L, for controlling temperature and pH values of fermentor. With a Recursive Learning Fuzzy Algorithm, a very optimistic temperature control is completed. Results shown us the situation that difference between desired value and room temperature was lower than 10℃, exists 0.15℃ overshoot at the inception of reaching set point, but after 4 minutes it maintained around the set point at 99.90 percent accuracy. And when the desired value was more higher than room temperature, overshoot would reduce in 0.1℃. because air may seize more heat from liquid in fermentor. In another side, an on-off control model experimented on pH control is easier and faster then temperature control. Only about 1 minute will be consumed on changing 1 unit pH value. But less stability lead to a 0.2 overshoot and ±0.1 range wave around the set point.