Experimental evaluation of model predictive control using data driven models

The Single Board Heater System (SBHS) can be chosen as the system which may find many analogous industrial applications where acute temperature control is demanded. The basic requirement in process control industry is the development of good model for the betterment of controller behavior, hence the fundamental task for any control problem is the development of model for the process to be controlled and design of controller using the model developed. The system identification of the process is done by using data driven or black box modeling. Model Predictive Control (MPC) is a control strategy that was developed in the process control industries to deal with constraints during operation of the plant and multivariable interactions. MPC can be formulated as : A nearly accurate model for given system, possible effects of the current and future manipulated input on the future plant dynamics can be predicted on-line and used to choose the input moves optimally [1]. The on-line predictions can be done over a window moving with time. The efficacy of these algorithms is tested through a Single Board Heater System which is temperature control process with heater and fan. The analysis of the control algorithms is done using LabVIEW and Matlab.