Research on Kernel Feature Construction and Integration Method of Power Transformer under Incomplete Information

The digital transformation strategy puts forward higher requirements on the information support system of the power system, and it also points out the direction for the operation and maintenance of key equipment such as large power transformers. First of all, this paper explains the power transformer holographic data perception architecture, key technologies, and analysis methods from a multi-parameter perspective. Then, based on actual cases, establishing a power transformer information decision table according to equipment components and fault types, and the "kernel" characteristic fingerprint of the power transformer was obtained. Secondly, by analyzing the distribution of the "kernel" feature indicators, a Task Flow Graph-Tasks (TFG-T) data integration architecture and fusion model based on the cloud-pipe-edge-terminal architecture is proposed. The data transmission and integration were described in detail under this model. Finally, the feasibility of the proposed method is verified by simulation analysis of the arithmetic example.