Smartenance DC-DC On-Board Power Converters

Publikation: Bidrag til tidsskrift/Konferencebidrag i tidsskrift /Bidrag til avisTidsskriftartikelForskningpeer review


—Traditional marine power grids, using diesel motors/generators, pollute the atmosphere with their CO2 emissions and adverse particles. Nowadays, green energy supplies like solar systems and fuel cells play significant roles in the electrification of marine propulsion plants. In general, marine power systems with green energy units are often considered as a kind of mobile DC microgrids under constant power loads (CPLs). However, since the CPLs impose nonlinearity and negative impedance instability on marine power systems, traditional linear controllers no longer stabilize the system under various disturbances. This motivated this study to develop a non-integer model predictive control for the accurate control of DC/DC converters with CPLs to stabilize the main bus voltage and current in the full-electric ferry ship. In practice, the stability of ship systems is highly threatened due to the negative-independence effect of CPLs. To address the adverse challenges of CPLs, the deep deterministic policy gradient (DDPG) with actor and critic neural networks (NNs) is applied to effectively design the coefficients embedded in the non-integer MPC. In this approach, the actor NN provides the regulatory signals to adjust the noninteger MPC controller while the critic NN is trained to evaluate the quality of actions of actor NN. The experimental results with comparative analysis are carried out to validate the accuracy of the non-integer MPC under the change in the CPL’s power. The results demonstrated that the proposed controller offers a higher level of robustness than the state-of-art-schemes.
TidsskriftIEEE Transactions on Circuits and Systems II: Express Briefs
Sider (fra-til)191-195
Antal sider5
StatusUdgivet - jan. 2023


Dyk ned i forskningsemnerne om 'Smartenance DC-DC On-Board Power Converters'. Sammen danner de et unikt fingeraftryk.