Neural Predictive Control of IUT based on Focused Time Delay Structure

Neural network Controller methodology is a nonlinear control fashion equipped with a novel method of Neural Predictive Controller (NPC) as an intelligent optimizer that in this cased based on the Focused Time Delay Neural Network (FTDNN) for modeling the nonlinear system and performing the optimization procedure. In case of prediction and control, two individual strategies are concerned for the current projects. The first is FTDNN procedure modeling the dynamics of system. The other is an optimization unit expected for minimization of optimization index. In this regards the Intelligent Universal Transformer (IUT) which will be raised in the Advanced Distribution of Automation (ADA) are discussed. IUT is an electrical key point introducing as a heart of ADA in case of Intelligent Electrical Devices (IED). ADA is the state of art, comprising the flexible electrical architecture and open communication construction empowered synergistically each other to contribute the tomorrow’s distribution automation. IUT construction relay on the bases of power electronic devices that employs the modern technologies of high voltage-low current solid-State equipment to cope with the deficiencies of current traditional transformers. So the IUT topology emphasizes on rectifiers, converters and PWM inverters in both input-output stages. These could be controlled via the intelligent control fashion for enhancing the robustness and stability of system in condition of variation take place. The proposed predictive control technique is a moderated control strategy using artificial neural networks to investigate the three phase power PWM converters of IUT for designing the current and voltages regulators in input-output stages. These result in the smooth regulation in IUT control and improve the system characteristics under load and source disturbances. At first the FTDNN model the power inverter dynamics of IUT. Then the optimization procedure takes place to minimize the optimization index for constructing the duty cycle of inverters as a control signals. Keywords— ADA, IUT, NPC, FTDNN