Enhanced estimation of seasonal irradiance patterns: The adapted time synchronous average method

Accurately separating the cyclic pattern from background noise in irradiance data is critical for various applications, including monitoring photovoltaic panels. However, seasonality, evident as amplitude modulation in the irradiance waveform, poses a significant challenge to existing methods like Time Synchronous Averaging (TSA) and Cepstral Editing (CE). These methods struggle to effectively remove the influence of seasonality on the estimated cyclic pattern. This work proposes a novel method called the Adapted Time Synchronous Average (ATSA) to address these limitations. The new approach is inspired by the application of cyclostationary tools used in mechanical signals of rotating machinery. ATSA employs a three-step approach: signal normalization to mitigate seasonality, TSA estimation using the normalized data, and adaptive filtering for signal reconstruction. Our analysis demonstrates that ATSA outperforms TSA and CE in separating the cyclic pattern from the stochastic noise component in irradiance data.