ANALYSIS OF ACCELERATIONS AND STRAINS MEASURED ON A TIE-DOWN SYSTEM OF A HEAVY NUCLEAR TRANSPORT PACKAGE DURING A ROUTINE RAIL

Year
2013
Author(s)
A.D. Cummings - International Nuclear Services (INS) Warrington, United Kingdom
J. Krywonos
Peter C. Purcell - International Nuclear Services, Hinton House, Risley, Warrington, UK
G. Rothwel - Liverpool J. M. University
C. Matthews - Liverpool J. M. University
File Attachment
439.pdf960.7 KB
Abstract
The design and development of nuclear packages is critical for the safe transportation of new fuel and irradiated waste. The renaissance of the nuclear industry in recent years has increased motivation for the development of optimised transport and storage solutions. The design of mechanisms to safely constrain nuclear packages, commonly referred to as tiedown systems, has become more challenging as package masses have increased. This paper focuses on characterising the loading environment that a tie-down system is subjected to using signals processing techniques on previously measured acceleration and strain time histories. The measurements were taken on a tie-down system for a nuclear package, weighing 99.7 tonnes, during a routine rail journey. Similar previous studies on tiedowns have omitted frequency analysis of the measured signals on tie-down systems. A frequency analysis has been used to determine the nature of the loading experienced by a tiedown system and also the extent of vibration transmission into the package. A means for obtaining a suitable filter cut-off frequency is also presented by comparing frequency spectra from different measurement points. To extract quasi-static accelerations from the raw data, several digital filters have been designed to study their effects on the resulting signals. By comparing the low pass and band pass filtered time histories some insightful trends in the accelerations peaks have been found. To demonstrate what constitutes a good or bad filter design, sensitivity studies have been conducted to show how the distributions of peaks and their statistics are altered significantly with poorer filter choices.