EVALUATION OF A MODEL BASED CONTROL STRATEGY FOR A NON- ISOTHERMAL SYSTEM WITH MULTIPLE INLETS AND OUTLETS

Resumo

With the rising development of embedded technologies, the use of a controller with a
traditional control strategy, such as a PID, can be replaced by advanced control methods. Advanced
control methods guarantee greater efficiency in complex situations which enhances process security
and final product quality. The present work investigated how model-based control methods performed
in a laboratorial tank system scenario. For this purpose, a theoretical and stochastic model was
developed which was based on mass, energy, momentum, and analysis of the data collected. After

validation, a model-based control strategy with a generic and then optimal approach was applied off-
line, with ISE tuning. To solve the model, an explicit Runge-Kutta’s method and parameter

optimization with PSO method was used. Finally, the influence of data reconciliation processing on
control development was investigated. It was concluded that both model-base strategies were more
efficient regarding the stability and proximity of the controlled variable to the setpoint, especially
when the data reconciliation was present.