Neural Network and Fuzzy Control Based 11-Level Cascaded Inverter Operation

This paper presents a combined control and modulation technique to enhance the power quality (PQ) and power reliability (PR) of a hybrid energy system (HES) through a single-phase 11-level cascaded H-bridge inverter (11CHBI). The controller and inverter specifically regulate the HES and meet the load demand. To track optimum power, a Modified Perturb and Observe (MP&O) technique is used for HES. Ultra-capacitor (UCAP) based energy storage device and a novel current control strategy are proposed to provide additional active power support during both voltage sag and swell conditions. For an improved PQ and PR, a two-way current control strategy such as the main controller (MC) and auxiliary controller (AC) is suggested for the 11CHBI operation. MC is used to regulate the active current component through the fuzzy controller (FC), and AC is used to regulate the dc-link voltage of CHBI through a neural network-based PI controller (ANN-PI). By tracking the reference signals from MC and AC, a novel hybrid pulse width modulation (HPWM) technique is proposed for the 11-CHBI operation. To justify and analyze the MATLAB/Simulink software-based designed model, the robust controller performance is tested through numerous steady-state and dynamic state case studies.

Automated summary

This paper presents a combined control and modulation technique using a single-phase 11-level cascaded H-bridge inverter to enhance the power quality and power reliability of a hybrid energy system. The technique employs a Modified Perturb and Observe technique to track optimum power and proposes an ultra-capacitor-based energy storage device and a novel current control strategy to provide additional active power support during voltage sag and swell conditions. A two-way current control strategy and a hybrid pulse width modulation technique are suggested for improved power quality and reliability. The robust controller performance is tested using MATLAB/Simulink software-based designed model.