A Neural Network Energy Management Controller Applied to a Hybrid Energy Storage System using Multi-Source Inverter

In this paper, a Neural Network Energy Management Controller (NN-EMC) is designed and applied to a Hybrid Energy Storage System (HESS) using the Multi-Source Inverter (MSI). Its aim is to manage the current sharing between a Li-ion battery and an Ultracapacitor by actively controlling the operating modes of the MSI. A discharge duty cycle that biases the use of one source over another is used as the control variable. To limit the battery wear and the input source power loss, an optimized solution is obtained with Dynamic Programming (DP). The NN-EMC is designed with an artificial neural network and trained with the optimized duty cycle obtained by DP. The DP/NN-EMC solution was compared to the battery-only Energy Storage System (ESS) and the HESS-MSI with 50% discharge duty cycle. Both the battery RMS current and peak battery current have been found to be reduced by 50% using the NN-EMC compared to the battery-only ESS for the New York City drive cycle.