A neuro-fuzzy compensator based Smith predictive control for FOPLDT process

This paper discusses a new control approach for a first-order-plus-long-dead-time (FOPLDT) process. A standard Smith predictor has a poor performance in the presence of modelling errors in process gain, time constant or time delay. To improve the system performance a neuro-fuzzy-compensator-based Smith predictor is used to form an advanced controller for the FOPLDT process. This proposed neuro-fuzzy-compensator-based Smith predictor provides good robustness with respect to simultaneous variations in all three process parameters. Furthermore, in order to reduce the steady-state error, a PID controller is integrated into the proposed control system. Simulation results and analysis revealed that this new control scheme for FOPLDT process control presents robustness to modelling errors and superior performance to the standard Smith predictor.