1.School of Electrical Automation and Information Engineering, Tianjin University, Tianjin300072, China;2.Economic and Technology Research Institute of State Grid Hebei Electric Power Co., Ltd., Shijiazhuang050021, China

When compressed sensing (CS) is applied to advanced metering infrastructure (AMI) systems, it is necessary to solve the problem of low reconstruction accuracy. Therefore, an optimal construction method for AMI based on multi-dimensional CS (AMI-mdCS) is deeply explored in this paper, and the data correlation among multiple smart meters is used to improve the reconstruction accuracy. Firstly, the basic principle of AMI-mdCS is analyzed according to the structural characteristics of AMI. Then, in order to improve the adaptability of the model to AMI, based on Kronecker CS (KCS) and multiple measurement vectors (MMV), a high-dimensional AMI-KCS model and a two-dimensional AMI-MMV model are constructed, and the specific processes of the two models are given. Finally, a joint training for sparse-basis and measurement-matrix based on singular value decomposition (JTSM-SVD) algorithm is proposed for the design of key elements in the model. Compared with one-dimensional model, AMI-KCS model and AMI-MMV model can significantly improve the reconstruction signal-to-noise ratio, and the former model has better improvement effect. Compared with the traditional training algorithm, JTSM-SVD algorithm can further optimize the reconstruction effect.

This work is supported by Tianjin Municipal Natural Science Foundation of China (No. 22JCZDJC00820).