Investigation of the parameters affecting the winding insulation resistance of synchronous generators and the effects of the improvement processes

The durability of insulation systems employed in electrical machines plays a crucial role in ensuring safe and reliable long-term operation. In this respect, the insulation resistance (IR) of generator windings is widely recognized as a key indicator of the overall health condition of the system. Nevertheless, insulation resistance values are highly sensitive to environmental influences such as humidity, temperature, and surface contamination, and may exhibit significant degradation over time. In this study, two different stator windings used in a synchronous generator were investigated. The winding specimens were subjected to controlled experimental setups designed to simulate humidity and contamination conditions representative of real operating environments. As part of the experimental aging process, insulation resistance and polarization index (PI) measurements were systematically recorded. Furthermore, several modeling approaches were evaluated to characterize the variation in stator winding insulation performance as a function of environmental parameters, including humidity level, exposure duration, and contamination state. Based on the developed models, the study aims to enable early prediction of potential insulation failures in stator windings and to support the planning of condition-based and predictive maintenance activities. To this end, data analysis was conducted using polarization index measurements, and predictive models were developed employing various machine learning algorithms. The results demonstrate that the influence of environmental conditions on insulation behavior can be effectively captured through the proposed modeling framework. This work contributes to insulation condition monitoring practices in the electrical power generation sector and supports improved maintenance planning and continuity of energy supply.