Design and analysis of a new model predictive current controller for grid connected converters

  • Euan Thomas Andrew

Student thesis: Doctoral Thesis


In recent years, the threat of climate change has led to significant interest in the installation of small-scale, renewable electrical generators due to their ease of installation in remote areas and low capital cost. These distributed energy resources are highly variable in nature and are often connected to the utility grid, mandating the use of a grid-connected converter. The power injected into the power system should be of a high quality to meet local regulations, therefore, the design of these grid-connected systems has received considerable attention in literature. In this thesis, state of the art control algorithms for grid-connected converters are reviewed and classified. Then, key limitations in the existing literature are identified and explained. A new modulated model predictive current controller is proposed, which offers reduced computational burden compared with the prior art, allowing more time to perform additional control functions. The variable switching frequency of existing model predictive controllers is fixed and the power quality is improved. Simulation results are included to prove the equivalent performance of the new approach, whilst the practical implementation in hardware is studied to prove that the computational burden is reduced significantly. The effects of parameter mismatch and grid voltage discretization on the controller are studied. The impact of these phenomena on the calculation of the duty factors is examined and a compensation strategy is derived. The limitations of compensating the effects precisely are discussed and a simplified solution is proposed. Simulation results verify the effectiveness of the proposed compensation. The proposed controller is then extended to unbalanced systems. A Kalman filter estimator is used to extract the positive and negative sequence components and a new calculation time compensation technique is proposed, which offers superior accuracy to existing approaches. The system stability is verified theoretically. Simulation and experimental results are included to prove the superiority of the proposed technique. Despite the improved performance, the execution time is also reduced compared with existing techniques
Date of Award29 Sept 2022
Original languageEnglish
Awarding Institution
  • University Of Strathclyde
SponsorsUniversity of Strathclyde
SupervisorKhaled Ahmed (Supervisor) & Barry Williams (Supervisor)

Cite this