Dynamic Neural-based Model Predictive Voltage Controller for an Interleaved Boost Converter with Adaptive Constraint Tuning

in recent years, interleaved current-fed boost dc converters consisting of a voltage multiplier and an active clamp circuit have been interested because of their good features like low input currents, and output voltages ripple high voltage-gains. Then developing more effective modeling and control techniques is important to increase their performance. In modeling section, it has been tried to estimate system model using the real data and dynamic neural networks with reducing number of variables. In control section there are constraints on control signal and its changing rate and in the proposed dynamic neural-based model predictive control (NMPC) controller, control signal changes constraint has been calculated adaptively. The proposed NMPC has been applied on system in online form and experimental results have been compared to the basic NMPC and a PI controllers. The output voltage have been specified for two loads with different input voltage. The experimental results show that their current transient response has smaller current peak and output voltage has smaller overshoot when load and voltage change. Output voltage steady-state response also has smaller oscillations about 2 volts for using proposed NMPC.