Data compression and reconstruction of smart grid customers based on compressed sensing theory

• A detailed compressed sensing model for electricity data is proposed.
• The data sparsity is proved with suitable sparse and observation matrices.
• An improved algorithm is adopted to accurately reconstruct the power data.
• Reconstruction results are influenced by compression rate and customer numbers.

In order to improve collection and transmission efficiency of smart electricity information collection system with huge number of data, a compressive sensing model for low-voltage customers was proposed in this paper. This model includes data compression method and reconstruction algorithm. First, we proved that electricity power data satisfy the sparsity condition of compressive sensing in a specific domain. Then, an improved iterative threshold algorithm was adopted to reconstruct the compressed power data, and detail processes of data reconstruction were proposed in succession. Experimental results show that the power data of multiple customers can be accurately reconstructed. Interrelation between reconstruction performance with customer number and compression rate was also analyzed.