Artificial Neural Networks based Multi-Objective Design Approach for Single-phase Inverters

Power decoupling functions of single-phase inverters are designed to minimise the total capacitance. This however might result in suboptimal design of the inverters because of complex tradeoff between inverter power density and efficiency. In this paper, an artificial neural network (ANN) based multi-objective design approach is proposed to first analyse the tradeoff and then optimise the design of a single-phase inverter with power decoupling function. The main contributions of this paper include 1) developing a systematic power loss and volume models by considering the dominant design and performance space parameters; 2) implementing the mathematical model to generate the training data set and develop an ANN based design approach; 3) comparing the results with a numerical models-based design method. The proposed design approach is implemented in MATLAB/Simulink. The simulation results show that the proposed ANN based design approach reduces the repeated usage of mathematical models, achieves optimal design and minimises the computational burden.