Model Predictive Control of Modular Energy Systems

  • Model Predictive Control of Modular Energy Systems

Overview

We developed new control methods for complex systems with an inherent modular structure. Such systems are prevalent in energy systems, particularly in electrical systems. From a mathematical point of view, the associated design problems are inherently non-convex and, thus, practical solutions depend upon judicious use of simplifications. Our focus was on the design of Model Predictive Control algorithms for Modular Multilevel Converters.

Modular Multilevel Converters allow voltage stress of hundreds of kV by dividing it between modules. Their potential to yield high quality waveforms make them excellent candidates for energy distribution.

Model Predictive Control (MPC) has the advantage of dealing with constraints and system failures. It also has the potential to reduce switching losses and harmonic distortion. However, the complexity of Modular Multilevel Converters raises significant challenges to the design of practical MPC algorithms.

Outcomes

A reduced order model and MPC algorithm for Modular Multilevel Converters.

Key features and advantages

RMS of the voltage ripple of each capacitor vs the 2nd harmonic or the circulating current
  • The model developed allows us to
    • characterise resonant frequencies of the converter
    • simplify the control design by reducing the number of actuation signals from one per module to only two in total.

 

  • The proposed MPC algorithm tightly controls
    • the current and voltage at the load
    • the circulating current and capacitor voltages in the converter

Funding

Amount: $265k

Duration: 2013 – 2018

Partners:

  • ABB Corporate Research Switzerland

Team

Further reading:

Quevedo, Aguilera and Geyer, “Model Predictive Control for Power Electronics Applications,” in Handbook of Model Predictive Control, pp. 551-580, Birkhäuser Basel, 2019.
Lopez, Quevedo, Aguilera, Geyer and Oikonomou, “Limitations and Accuracy of a Continuous Reduced Order Model for Modular Multilevel Converters,” IEEE Transactions on Power Electronics, vol. 33, no. 7, pp. 6292-6303, July 2018.
Lopez, Quevedo, Aguilera, Geyer and Oikonomou, “Validation of a Reduced Order Model for Modular Multilevel Converters and Analysis of Circulating Current,” Proceedings of the 17th European Conference on Power Electronics and Applications, EPE’15 ECCE Europe, Geneva, Switzerland, 2015.
Lopez, Quevedo, Aguilera, Gever and Oikonomou, “Reference Design for Predictive Control of Modular Multilevel Converters.” Proceedings of the Australian Control Conference 2011