The research results of teachers of Shanghai Electric Power University were hired by the top journals of smart grids

近日，上海电力大学自动化工程学院崔承刚副教授关于“直流微电网智能控制”的论文“Implementation of Transferring Reinforcement Learning for DC-DC Buck Converter Control via Duty Ratio Mapping”被国际顶级智能电网期刊IEEE Transactions on Industrial Electronics（ IEEE TIE, SCI area, IF = 8.162).

In recent years, Reinforcement Learning (RL) control methods have become research hotspots in many areas. In the field of electrical electronics, compared with traditional control methods, strengthening learning to have adaptive self-learning ability can more effectively control the voltage control of the DC-DC converter with a constant power load. However, there is no deviation between the simulation environment and the actual system during the training process, which makes the performance decline when the intelligent body training shows good control performance while migrating to the actual environment. To this end, this article proposes a migration mapping relationship (DRM), which first migrates the trained intelligence to the actual system for the first time, which not only ensures the control of the transient performance of the control, but also ensures its steady -state performance. At the same time, the intelligent body has a certain adaptive ability for changes in circuit parameters, that is, the same intelligence can effectively control the parameters of different circuits (mainly inductors and capacitors).

This article first uses power electronics and strengthening the learning toolbox pre -trained intelligent body in Simulink, and then reconstruction of the smart body. Then collect the voltage current relationship of the actual system in the actual system and build the duty cycle mapping table. Finally, the table will be It is applied to the control of the actual system and realizes the significant improvement of the DC-DC converter's transient and steady-state control performance.

Associate Professor Cui Chenggang's work "VOLTAGE Regulation of DC-DC Buck Converters Feeding CPLS VIA Deep Reinforcement" (https://ieeexplueeeeeeeeeeeeeeeeeeeeeeeeeeeon )5959595959595959595959595959595959595959595959595959595959595959595 was Published in August 2021, IEEE TRANSACUITS and Systems II: Express Briefs (IEEE TCI, IF = 3.691) was published in another electrical electronic journal.

The above work has been funded by the National Natural Fund project. (Correspondent: Fan Lida, Shanghai Electric Power University)

- END -