Performance Improvement of Renewable Energy Sources Inverter for Interface with Smart Grid

Abstract

Now a days, Renewable Energy Sources (RES) are the most important resources for generating pollution-free electricity. This work completely concentrates on the integration of solar system with smart grid. Different types of inverters are there, such as capacitor-clamped type, diode clamped model and cascaded H-bridge multilevel inverter. However, cascaded H-bridge multilevel inverter is more appropriate for photovoltaic applications because it uses separate dc sources for each H-bridge module. Solar source acts as a dc source for modules. The proposed work introduces a three phase cascaded H-bridge five level inverter for the integration of RES with smart grid. The presence of computational components within the multilevel inverter can produce the measurement afflictions. In order to reduce this problem, an artificial intelligence based total harmonic distortion (THD) of multi-level inverters (MLI) is proposed. Two types of controllers are used in this work for tuning MLI. First, PV integration with smart gird in which Neural Network (NN) is optimized with Ant Lion Optimizer (ALO) as the control strategy. The integration of Adaptive Neuro Fuzzy Inference System (ANFIS) and Particle Swarm Optimization (PSO) is the second controller. The first controller reduces the THD of 1.8% which is further reduced by next controller design. The grid voltage and difference voltage are provided as input for ANFIS to deliver harmonic-free control output. The ANFIS is optimized by PSO in the training stage to handle the switching angle of MLI for generating the control voltage with reduction in harmonics. To assess THD level, the execution of proposed work is implemented in MATLAB/Simulink. To validate the effectiveness of the proposed PSO-ANFIS strategy, the rate of error convergence is compared with GA-ANFIS, ANFIS and Artificial Neural Network (ANN). On comparison with the various control strategies like PI, FLC and ALO-NN, the proposed work yields the THD of 37.29% (without controller) and 1.25% (with PSO-ANFIS). Hence, the pr