Predictive control of four-leg converters for photovoltaic energy systems

Venkata Yaramasu, Marco Rivera, Apparao Dekka, Jose Rodriguez

Research output: Chapter in Book/Report/Conference proceedingChapter

2 Scopus citations


Photovoltaic energy systems are one of the most widely adopted distributed generation facilities. This book chapter presents predictive based current and voltage control strategies for four-leg converters employed in grid-connected and standalone photovoltaic energy systems, respectively. The proposed approach employs the novel stationary frame sampled-data models of the four-leg converters with inductive (L) and inductive-capacitive (LC) filters on the output side to predict the control variables such as output currents and load voltages. These predictions are performed using all the possible switching states of four-leg converters. The objective of minimizing the error between reference and predicted variables (load currents or voltages) is fulfilled through a cost function in the predictive control schemes. In addition, the voltage balancing of DC-bus capacitors is considered with the four-leg neutral-point clamped converters. The optimal switching states corresponding to the minimal cost function value are chosen and directly applied to the converter. The predictive control strategies fulfil the control requirements such as load cur-rent/voltage control, DC-bus voltage balancing, and neutral-leg switching frequency minimization. The simulation and experimental studies conducted using unbalanced and nonlinear loads to validate the proposed predictive control strategies.

Original languageEnglish (US)
Title of host publicationPower Systems
Number of pages25
StatePublished - 2019

Publication series

NamePower Systems
ISSN (Print)1612-1287
ISSN (Electronic)1860-4676

ASJC Scopus subject areas

  • Energy Engineering and Power Technology
  • Electrical and Electronic Engineering


Dive into the research topics of 'Predictive control of four-leg converters for photovoltaic energy systems'. Together they form a unique fingerprint.

Cite this