Analysis of PCA based compression and denoising of smart grid data under normal and fault conditions

This paper describes application of principal component analysis (PCA) based data compression approaches for smart grid (SG) data storage. PCA methods are applied and compared for steady state operations and various system fault conditions. For signals with Gaussian noise data, a variant of PCA, labelled Iterative PCA (IPCA) is applied to simulated phasor data. The phasor data is simulated for IEEE New England 30-bus system using ETAP software. The simulated data is then stored using the PCA and IPCA, which exploit the sparsity and collinearity among variables in the simulated data. The results indicate that proposed methods compress both the steady state and transient signals effectively while simultaneously removing the Gaussian noise contained in the signals.