The diversification of equipment in power system and the complexity of operating conditions gradually lead to the development of power quality disturbance to the complex power quality disturbance which is released simultaneously by multiple types of disturbance. It is important to identify the type of complex power quality disturbance reliably and effectively for the stable operation of power grid. In this paper, a composite power quality disturbance identification architecture based on GASF-GADF-MTF three-channel multi-dimensional feature fusion plus transformer is proposed. Firstly, the composite power quality disturbance is converted into RGB three-channel data by using Gram summing field(GASF), Gram difference field(GADF) and Markov transfer field(MTF). The two dimensional image of feature fusion is obtained. Then the transformer network is used to classify and discriminate the two-dimensional images containing the complex power quality disturbance characteristics. Simulation results show that the proposed method is reliable and effective, and has the advantages of good noise robustness and strong generalization ability.

complex power quality disturbance(PQDs); Dimension transformation; Multi-feature fusion; transformer; Gram Angle field(GAF); Markov transfer field(MTF)