An FPGA implementation of the fast gradient method for solving the Model Predictive Pulse Pattern Control problem

Author

Peyrl, Helfried ; Junyi Liu ; Geyer, Tobias

Author_Institution

ABB Corp. Res., Baden-Dättwil, Switzerland

fYear

2013

fDate

17-19 Oct. 2013

Firstpage

1

Lastpage

6

Abstract

The computational complexity of Model Predictive Pulse Pattern Control (MP³C) and its real-time requirements have so far limited the deployment of this optimisation-based control method for power electronic systems. In this paper, we present an FPGA implementation, which uses the Fast Gradient Method for solving the quadratic program underlying MP³C. The proposed method can solve the control problem in less than the required 25 μs with sufficient accuracy.

Keywords

computational complexity; field programmable gate arrays; gradient methods; power electronics; predictive control; quadratic programming; sampled data systems; FPGA implementation; MP³C; computational complexity; fast gradient method; model predictive pulse pattern control problem; optimisation-based control method; power electronic systems; quadratic program; Field programmable gate arrays; Gradient methods; Stator windings; Switches; Trajectory; Vectors;

fLanguage

English

Publisher

ieee

Conference_Titel

Sensorless Control for Electrical Drives and Predictive Control of Electrical Drives and Power Electronics (SLED/PRECEDE), 2013 IEEE International Symposium on

Conference_Location

Munich

Print_ISBN

978-1-4799-0680-2

Type

conf

DOI

10.1109/SLED-PRECEDE.2013.6684480

Filename

6684480