Rough Controlling TSC for Reactive Current Compensation in Traction Substations

Abstract

In process of reactive current compensation in electrified railway, due to applying fixed parallel compensation while not adapt to dynamic change of traction loads, redundant compensation or deficient compensation therefore frequently happen. Actually, while returning reactive power to power supply but suffering a positive measure, average power factor is lower than the primary one, thus, penalty is inflicted for poor power factor. To rescue the situation, in the paper we propose a novel capacitor switching strategy based on rough functions concept. The method firstly applies artificial neural networks to on-line monitor reactive current in traction load arms, then according to rough function, yields the discrete series space of reactive current for the intensity and speed of capacitor grouping switching, and generates decision-making rules for capacitor grouping switching. Meanwhile, considering the existence of three-time harmonics in traction loads, filter is designed in each grouping capacitor to filter a part of harmonics. The investigation indicates that the proposed method can overcome the deficiencies of traditional methods and duly trace dynamic change of loads, and power factor is also dramatically improved. In the end, an application example shows that the method is an effective dynamic compensation strategy for reactive power in traction substation