E. Eyng and A.M.F. Fileti (Brazil)
Measurement uncertainty, Artificial neural network control, Bioethanol.
Bioethanol lost by evaporation from the sugar cane fermentation tanks may be recovered from the gas mixture using a staged absorption column. This equipment also can be used to take off the humidity of CO2, a by-product of this process. In the present work, the development of a nonlinear feedforward-feedback controller (ANN controller), based on a neural network inverse model, was proposed and tested to manipulate the absorbent flow rate in order to control the residual ethanol concentration in the effluent gas phase at the first absorption column, and the residual water at the second one. Simulation studies were carried out, where sensor uncertainties were introduced to system control input variables. Three uncertainty levels were tested: 5%, 10% and 15%. The ANN controller outperformed the Proportional Integral Derivative (PID) controller, nevertheless, under the influence of measurement uncertainties, the ANN and PID controller performance deteriorated. For the CO2 treatment absorption column, the PID controller did not work properly under uncertainty levels of 10% or 15%, while the ANN controller just deteriorated under uncertainty level of 15%. The results demonstrate that the developed neural control system is a robust and reliable tool to control the absorption columns studied.
Important Links:
Go Back
