An Approach for Improving the Surge of a Distributed Power Grid Containing the Impact Loads by Forecasting Status Classifiers Based on Big Data

In order to suppress circular current of distributed power grid we have proposed a general approach based on big data theory in this paper. This method includes two stages: the offline identify the status classifiers that described the characteristic of the impact load and the online calculate the drive instruction by comparing sampling data to the classifiers. We utilize big data theory to analyze and forecast the status classifiers of reactive power and harmonic current according to sampled massive data caused by the impact load. The topology of proposed TAPF consists of TSF (Thyristor Switched Filter) and APF (Active Power Filter) is controlled by same DSP2812. Combination rules that include the APF and TSF works individually and these two works together have been given by analyzing status classifiers. The circular current which caused by APF and TSF works together has been suppressed by connecting an extra resistance in series with the channel of TSF. The feasibility and effectiveness of this approach has been illustrated with an injection machine. Experiment and simulation show this approach can suppress the surge caused by the impact loads by which the grid can meet the requirements of national standard GB/T-14549-93.