Multiobjective Vector Modulation for Improved Control of NPC-Based Multi-Source Inverters in Hybrid Traction Systems

Abstract

The concept of a multi-source inverter (MSI) based on NPC topology was recently introduced for hybrid-electric powertrains to connect several dc sources to the traction load in a single conversion stage without magnetics. This work proposes a new modulation algorithm that integrates the tasks of dc currents and ac voltages control at the level of pulsewidth modulation by means of the space-vector formalism. The proposed algorithm allows the MSI to have full control of the dc input currents and fundamental ac output voltages, and enables a higher number of feasible operating modes, including load power sharing between sources and controlled recharging of one source from the other, both in motion and at standstill. Hence, improved control capability of the load and sources is achieved with respect to the state-of-literature modulation approach. To validate the proposed modulation under different operating conditions, steady-state and dynamic tests are performed on a small-scale traction system fed by a primary dc supply and a battery pack. The results are compared with those of the baseline control approach and show that the proposed technique improves the smoothness and flexibility of the control action and reduces the distortion of currents and voltages while ensuring the same converter efficiency.