Statistical Health Reasoning System of Power Generator Stator Windings against Water Absorption

Abstract

The power generator, as one of the most critical components in a power plant, is typically maintained through use of a time- or usage-based strategy. Either strategy could result in a substantial waste of remaining useful life (RUL), high maintenance costs, and/or low plant availability due to excess, untimely, or missed maintenance. Recently, the field of prognostics and health management has offered new general diagnostic and prognostic techniques to precisely assess health conditions and robustly predict the RUL of engineered systems, with the aim of addressing the aforementioned deficiencies. This paper explores a smart health reasoning system that can be used to assess the health condition of power generator stator windings and their levels of water absorption. The system monitors health based on capacitance measurements of the winding insulations. In particular, a new relative health measure, namely the Directional Mahalanobis Distance (DMD), is proposed to quantify the health condition of stator windings. This paper also proposes an empirical health classification rule, based upon the DMD, which factors in maintenance history. The proposed smart health reasoning system is validated using eight years field data from eight generators, each of which contains forty-two windings.