Abstract
A crude-oil desalting unit was modeled and automatic control implemented. This control system involved three variables, namely, pressure, pH, and interfacial level. Three-term proportional, integral, and derivative (PID) controllers were used to control these operating variables. Crude oil flow (throughput), brine discharge rate, and caustic injection rate were manipulated for control. Controllers were tuned and performances recorded for set-point perturbation events. Delays, overshoots, and oscillations in PID control performance were observed. In order to improve the performance, an artificial neural network (ANN) control logic has been implemented for each of the variables. In this network, two input nodes, six hidden nodes, and one output node were used for control purpose. Normalized values of the process measurement and the set point were entered as the inputs that activated the hidden nodes and fired output signals for the manipulated variable. Three such ANN controllers were used in the system. Performance of the ANN controllers was found to be excellent as compared to PID controllers.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call