1D CNN Based Detection and Localisation of Defective Droppers in Railway Catenary

Defective droppers pose a significant threat to the performance of contact between train pantograph and railway catenary. In this paper, impact damaged on pantograph–catenary interaction behaviour is analysed, varying degrees damage each dropper labelled. To improve classification accuracy when both degree position are considered, model integrating multiple 1D CNNs proposed. Approaches including randomly searching optimal hyper-parameters K-fold cross-validation used prevent overfitting ensure regardless training data subset selected. Compared with conventional CNN, integrated method demonstrated using metrics accuracy, F1-score, precision recall. It concluded that, through use can be detected localised based force. Hence, intelligent catenary inspection enhanced.