| dc.description.abstract | This work presents the development of an advanced, application-level control system, aimed at improving energy consumption levels in the itt-Chip / Unisinos cleanroom facilities. Heating Ventilation and Air Conditioning (HVAC) systems account for a large portion of semiconductor production costs due to the need for temperature, humidity and internal pressure control of controlled environments. Considering that most of the world’s buildings require climate control systems, there is a need for simplified HVAC systems. To address the great potential for new contributions to the semiconductor market, an application-level, multivariable predictive control system was developed to meet the control specifications required by the clean rooms of the technological institute. To obtain the synthesis of the predictive controller, an evaluation of the original control system was necessary, including process characterization and parametric modeling from response test analysis. A Simulink® application was developed using Model Predictive Control (MPC) and Open Platform Communications (OPC®) toolboxes for processing control laws and communication with actuators through local programmable logic controllers. At the application level, servo and regulatory behavior tests of the predictive controller, and the evaluation of an energy consumption indicator (ECI) were developed to compare the itt-Chip PID legacy control and the new advanced control strategy was implemented. The results indicated a good servo and effective regulatory behavior of the MPC control, and a reduction in energy consumption where the indicator read values between 80 and 90%, depending on the tuning chosen for the predictive control. | pt_BR |
