Phase identification is the problem of determining the phase connection of loads in a power distribution system. In modern times, utility operators will generally accomplish this using smart meter data that usually requires some form of feature engineering to achieve practical phase identification using data-driven methods. Feature engineering is essential for voltage magnitude data containing noise, seasonality, and trend. Crucial components of a feature engineering pipeline are presented to perform linear denoising with Singular Value Decomposition, filtering of the denoised data to remove the seasonality and trend, and fuse multiple meter channels. We use the results of the feature engineering to perform phase label correction, a subproblem of phase identification. The authors generate a synthetic dataset from the meshed IEEE 342-Node test feeder circuit with the 2021 Electric Reliability Council of Texas load profiles to evaluate techniques. Our results show that denoising is quite effective for improving phase identification accuracy in the presence of measurement noise. We present some new insight into filtering voltage measurement data to improve accuracy, as well as eliminate the need to determine salient frequencies. We also present the application of a data channel fusion technique that is novel to the phase identification literature. This technique enhances phase identification in cases where there are both wye and delta-connected loads present.

Phase identification is the problem of determining what phase(s) that a load is connected to in a power distribution system. However, real-world sensor measurements used for phase identification have some level of noise that can hamper the ability to identify phase connections using data-driven methods. Knowing the phase connections is important to keep the distribution system balanced so that parts of the system are not overloaded, which can lead to inefficient operations, accelerated component degradation, and system destruction at worst. We present a feature engineering pipeline that combines Singular Value Decomposition (SVD) with an Interquartile Range (IQR)-based nonlinear filter to denoise matrices of voltage magnitude measurements. We empirically show that this pipeline is effective in denoising data that contains either additive Gaussian, Laplacian, Cauchy, or mixed Gaussian/Cauchy noise. This approach reduces Frobenius error and increases the average phase identification accuracy over a year of non-stationary time series data. K-medoids clustering is used on the denoised voltage magnitude measurements to perform phase identification.