Search for: All records

Fault pattern recognition in complex mechanical systems such as gearbox has always been a great challenge. The performance of a classic fault pattern recognition approach heavily depends on domain expertise and the classifier applied. This paper proposes a deep convolutional neural network-based transfer learning approach that not only entertains adaptive feature extractions, but also requires only a small set of training data. The proposed transfer learning architecture essentially consists of two sequentially connected pieces; first is of a pre-trained deep neural network that serves to extract features automatically, the second piece is a neural network aimed for classification which is to be trained using data collected from gearbox experiment. The proposed approach performs gear fault pattern recognition using raw accelerometer data. The achieved accuracy indicates that the proposed approach is not only sensitive and robust in performance, but also has the potential to be applied to other pattern recognition practices. 

Abstract The balloon-borne ANITA [1] experiment is designed to detect ultra-high energy neutrinos via radio emissions produced by in-ice showers. Although initially purposed for interactions within the Antarctic ice sheet, ANITA also demonstrated the ability to self-trigger on radio emissions from ultra-high energy charged cosmic rays [2] (CR) interacting in the Earth's atmosphere. For showers produced above the Antarctic ice sheet, reflection of the down-coming radio signals at the Antarctic surface should result in a polarity inversion prior to subsequent observation at the ∼35–40 km altitude ANITA gondola. Based on data taken during the ANITA-1 and ANITA-3 flights, ANITA published two anomalous instances of upcoming cosmic-rays with measured polarity opposite the remaining sample of ∼50 UHECR signals [3, 4]. The steep observed upwards incidence angles (25–30 degrees relative to the horizontal) require non-Standard Model physics if these events are due to in-ice neutrino interactions, as the Standard Model cross-section would otherwise prohibit neutrinos from penetrating the long required chord of Earth. Shoemaker et al. [5] posit that glaciological effects may explain the steep observed anomalous events. We herein consider the scenarios offered by Shoemaker et al. and find them to be disfavored by extant ANITA and HiCal experimental data. We note that the recent report of four additional near-horizon anomalous ANITA-4 events [6], at >3σ significance, are incompatible with their model, which requires significant signal transmission into the ice.