Impact of SOC variations on the battery bank sizing of a stand-alone system fed by a passive wind turbine

Abstract

In this paper, the authors compare and analyse two passive wind turbine system models in order to show their equivalence through a storage bank sizing procedures. The main differences between both models reside in the design accuracy and the computational time needed for each model to simulate the wind turbine system behaviour. On the one hand, a first "mixed reduced model" neglects the electrical mode effect and assumes that the DC battery bus voltage is constant (i.e. invariable State Of Charge: SOC). On the other hand, the second "full analytic model" couples SOC fluctuations (i.e. bus voltage variations) in the whole system. When compared to the second model, the "mixed reduced model" allows reducing computational time, which is a major factor in the context of systemic design by optimization. The analysis is performed to put in evidence the correspondence between both sizing approaches with the two corresponding models. The results are finally discussed from the point of view of the compromise design accuracy and computational time reduction. In this paper, the authors compare and analyse two passive wind turbine system models in order to show their equivalence through a storage bank sizing procedures. The main differences between both models reside in the design accuracy and the computational time needed for each model to simulate the wind turbine system behaviour. On the one hand, a first "mixed reduced model" neglects the electrical mode effect and assumes that the DC battery bus voltage is constant (i.e. invariable State Of Charge: SOC). On the other hand, the second "full analytic model" couples SOC fluctuations (i.e. bus voltage variations) in the whole system. When compared to the second model, the "mixed reduced model" allows reducing computational time, which is a major factor in the context of systemic design by optimization. The analysis is performed to put in evidence the correspondence between both sizing approaches with the two corresponding models. The results are finally discussed from the point of view of the compromise design accuracy and computational time reduction. Do not use abbreviations in the title unless they are unavoidable.