An assessment model for mapping the susceptibility of deposits accumulation on insulators driven by remotely sensed data

Insulators deposits can cause flashovers when the right conditions are met, threatening the normal operation of power grids. Grid staffs performed massive in situ measurements to obtain the ESDD observations and then produced susceptibility level maps, helping guiding cleaning the insulators deposits. Such works are time-consuming and costly, and relies on a limited number of discrete point measurements to infer the susceptibility distribution in the spatial domain. In this work, we proposed a novel assessment model for mapping deposits accumulation on insulators using remotely sensed products, emission inventories and other text data. We have spatialized the text data first and then performed the spatial registration to unify the spatial coordinate systems of multi-sources data. On that basis, we utilized a scoring method to quantitatively evaluate the susceptibility of deposits accumulation on insulators after determining the weights and scoring rule for different attributes using the rough set theory. Results showed that the susceptibility map can very well indicate the major risky regions via comparisons with the present susceptibility level map. Further classification experiments yielded results with an accuracy of 83.6%. We believed that such a novel method can provide great helps for the grid maintenance by reducing costs and improving performances of the insulator deposits susceptibility maps.