No abstract

. Predicting component reliability and level of degradation with complex-valued neural networks. Reliability Engineering and System Safety, Elsevier, 2014, 121, pp.198-206 AbstractIn this paper, multilayer feedforward neural networks based on multi-valued neurons (MLMVN), a specific type of complex valued neural networks, are proposed to be applied to reliability and degradation prediction problems, formulated as time series. MLMVN have demonstrated their ability to extract complex dynamic patterns from time series data for mid-and long-term predictions in several applications and benchmark studies. To the authors' knowledge, it is the first time that MLMVN are applied for reliability and degradation prediction. MLMVN is applied to a case study of predicting the level of degradation of railway track turnouts using real data. The performance of the algorithms is first evaluated using benchmark study data. The results obtained in the reliability prediction study of the benchmark data show that MLMVN outperforms other machine learning algorithms in terms of prediction precision and is also able to perform multi-step ahead predictions, as opposed to the previously best performing benchmark studies which only performed up to two-step ahead predictions. For the railway turnout application, MLMVN confirm the good performance in the long-term prediction of degradation and does not show accumulating errors for multi-step ahead predictions.

Accurately estimating station dwell time is critical for timetable planning. Its importance has increased as railways seek to improve timetable stability and network efficiency, while serving more passengers and different types of transport services. This research consisted of developing a station dwell time model in cooperation with the Swiss Federal Railways (SBB).The proposed model estimates dwell times based on the input parameters: vehicle type (number, position, width and level of doorways), infrastructure (platform level) and demand (number and distribution of passengers). The research divides dwell time into five sub-processes: door-unblocking, opening doors, passenger boarding/alighting, closing doors and train dispatching. Each sub-process was evaluated separately to understand its influence on dwell time. The SBB's automatic passenger counting system was used to record the number of passengers boarding and alighting at each door and the beginning/ending time of each sub-process. During eight months over three million measurements were made on four different vehicle types operating on 20 different routes. These data were analyzed and used to develop the dwell time model. This paper describes the research methodology, the structure of the dwell time model, the data collection system and presents a summary of results including statistical distribution and influence factors of sub-process times.

No abstract

No abstract