Parameter identification of DC arc models using chaotic quantum cuckoo search

Abstract

DC arc fault is difficult to be detected, which brings great challenge to the safety of DC distribution system. The investigation of arc fault detection based on simulation method is able to reduce the cost and contributes to the research in the system where it is not suitable to carry out experiments. Accurate arc model is the foundation of simulation research. To address the gaps existing in arc models and corresponding parameter identification methods, two novel arc models and a new optimization algorithm are presented in this paper. First of all, hook static model was proposed to describe the nonlinearity near the transition point in static curve of arc. Then, segmented noise model was developed to flexibly fit the diversiform shape of arc noise’s frequency spectrum (FS). Finally, in order to improve the accuracy of parameter identification of arc models, this paper introduced chaos mechanism into quantum cuckoo search (CQCS): using chaotic map to initialize the population and generate the random parameter p; using lifespan based chaotic local search to further update the position of best solution. Based on the data obtained from established experimental platform, the effectiveness of the proposed models and optimization algorithm is verified. Meanwhile, compared with the existing methodologies, the proposed models and optimization algorithm are of better performance.