I. Introduction
In the electrified railway industry, the power for locomotives is continuously transmitted through the contact between the overhead catenary system and the pantograph installed on the locomotive roof. More precisely, the sliding contact between the contact wire of the catenary and the slide plate of the pantograph in a high-speed interaction, which is schematically shown in Fig. 1, ensures the normal operation of electric locomotives. The level of stability and of continuity of the sliding contact determines the quality of current collection for the locomotive [1], which is crucial for all pieces of the on-board electrical equipment, especially for the traction drive system. In general, the major measurement that reflects the quality of current collection is the pantograph–catenary contact force (PCCF). At present, statistical indicators such as mean, standard deviation (SD), and maximum and minimum of the contact force, which are time-domain parameters, are adopted to evaluate the contact quality [2], [3]. The fact is that although these indicators can reveal the trend, fluctuation, and abnormal points of a series of contact forces directly, they tend to become observable after the pantograph–catenary interaction has deteriorated. The anomaly that could be reflected by the time-domain indicators already exists irrespective of whether the measurement is performed or not. In a reasonable way, these indicators certainly can be the proof of certain deterioration, but not the prediction of the deterioration.
Schematic of the pantograph–catenary sliding contact.