Research on Application of Intelligent Algorithm Based on Multi-Source Data Fusion in Dynamic Assessment of Power Quality

Abstract

With the “double-high” characteristics of a high proportion of new energy and high power electronics in the new power system, the operating conditions of the power grid present strong randomness and time-varying features. Traditional static power quality assessment methods are difficult to adapt to the dynamically fluctuating power grid operating state, with problems such as assessment lag, insufficient accuracy, and single dimension. To address the above pain points, this paper proposes an intelligent dynamic power quality assessment method based on multi-source data fusion. First, a multi-source data preprocessing and feature extraction system for power operation is constructed, and deep mining of heterogeneous data is realized relying on a three-layer fusion mechanism of data level, feature level, and decision level. Second, the temporal intelligent algorithm is improved, and a sliding window dynamic assessment model is built to realize real-time judgment and dynamic update of power quality status. Finally, simulation verification is carried out with the measured power grid data set. The experimental results show that compared with traditional assessment methods, the proposed method improves the assessment accuracy by 9.2% and reduces the dynamic response delay by 47.5%. It has stronger anti-interference and generalization ability, can effectively meet the dynamic assessment needs of power quality under complex power grid conditions, and provides technical support for the refined management and control of power grid power quality.