Evaluations of Simulation Models for Flicker and Harmonics

Abstract

Electric arc furnaces (EAF) are highly dynamic industrial loads that exhibit rapid variations in active and reactive power draw. These dynamics introduce significant power quality disturbances at the point of common coupling (PCC). To mitigate this, a Static Synchronous Compensator (STATCOM) is connected in parallel to dampen the current variations and therefore the power quality disturbances. To size the STATCOM well, preliminary studies with accurate EAF and STATCOM models are vital. This thesis evaluates if a publicly available STATCOM and EAF model can be reliably used to study power quality disturbances. This will be done by comparing the public models with Hitachi’s detailed STATCOM and EAF model based on current measurements. The public EAF model uses a mathematical formula of an arc that emulates flicker behavior with random noise and a non-linear V-I relationship. The public STATCOM has a simple control scheme that works by averaging the current variations from the EAF model and subtracting the variations from the average current. The results indicate that the mathematical EAF model can reproduce the flicker levels observed in the measurement-based model. However, the flicker originates from other frequencies and has a different flicker distribution, overemphasizing the commonly occurring flicker content. Furthermore, the harmonic spectrum of the mathematical EAF model is inaccurate, the second harmonic is 27.6% lower, while the third harmonic is 30.0% higher than in the current measurement model. The public STATCOM model also shows limited accuracy. It only reduced flicker by an average factor of 3.47 compared with the Hitachi’s model that reduced it by an average factor of 7.53. In addition, the public STATCOM increases the current total demand distortion (TDD) by an average of 43.1% compared to a 6.5% decrease for the Hitachi model, and it lacks power factor (PF) control capability. To improve the mathematical EAF model further the second harmonic and interharmonicgroup would need to be increased and more apparent power peaks would need to introduced. The public STATCOM model could be improved by incorporating d-current flicker control, dedicated power factor control and a current rate limiter. In its current form the mathematical EAF model is not suitable for estimating flicker or harmonics. Moreover, due to its poor flicker and harmonic performance the STATCOM cannot be used. Therefore, STATCOM sizing studies are unsuitable to do with the public model.