Sensing and measurement devices are keeping pace with the advancement in the industrial power distribution system. The ability to provide time-synchronized measurements at a fast reporting rate by distribution-level PMUs (D-PMUs) specially with increasing distributed energy resources (DERs) offer great opportunities for monitoring and control. However, unlike the transmission systems, the distribution system waveforms typically have more noise, harmonics and unbalanced phases, posing unique challenges to estimate phasors at the distribution-level. Lack of specific standards for performance requirements of D-PMUs make this further challenging. This work proposes a novel smart synchrophasor device architecture for estimating phasors on polluted signals. The proposed sensor architecture is adaptive to varying system conditions and can adjust reporting rates based on system demands. The proposed approach employs a Sliding Fast Fourier Transform (SFFT) and Signal Estimation by Minimizing Parameter Residuals (SEMPR) technique to simultaneously estimate the harmonic components along with the fundamental phasor. Further, to accommodate the signals generated from varying system conditions in the distribution system, an approach is proposed to update the measurement model for the PMU estimation using adaptive filtering and goodness-of-fit (GoF) measure.