Contents lists available at ScienceDirect Engineering Failure Analysis journal homepage: www.elsevier.com/locate/engfailanal Experimental and FEM numerical assessment of multiaxial fatigue failure criteria for a rolling Stock's seats structure R. Sepe a , A. Greco b , A. De Luca b, ⁎ , E. Armentani a , F. Berto c a Dept. of Chemical, Material and Production Engineering, University of Naples Federico II, P.le V. Tecchio 80, 80125 Naples, Italy b Dept. of Engineering, University of Campania Luigi Vanvitelli, Via Roma 29, 81031 Aversa, Italy c Dept. of Mechanical and Industrial Engineering, Norwegian University of Science and Technology, 7034 Trondheim, Norway ARTICLE INFO Keywords: Fatigue testing High cycle fatigue Fatigue assessment Finite element analysis Railway engineering ABSTRACT This paper deals with the experimental and numerical investigations of the structural behaviour of a rolling stock's seats system under static and fatigue loading conditions, aimed to assess the reliability of some multiaxial fatigue failure criteria implemented in the FE model. According to the NF F31–119 standard, a specific full-scale multiaxial testing machine has been used for the fatigue test. During the test, a visual inspection approach with a detection interval of 3·10 5 cycles has been adopted. After 1.2·10 6 cycles the test has been stopped and a Liquid Penetrant Inspection (LPI) has been carried out in order to detect accurately cracks nucleation and pro- pagation. Moreover, numerical analyses, based on the Finite Elements (FE) Method, have been performed on the EN-AW 6060 T5 aluminium alloy structure of a rolling stock's seats to predict the most affected zones where the cracks nucleation may occur under multiaxial fatigue loads. Since the in-service applied loads belong to Multiaxial High-Cycle Fatigue (MHCF) load class, Sines, Crossland and Dang Van criteria have been used to post-process the predicted results achieved by the FE analyses. For validation purpose, numerical and experimental results have been compared. According to the numerical-experimental results comparison, Dang Van criterion provided the best level of accuracy. 1. Introduction The continuous need for the weight reduction, the improvement of the design optimization of components and the need to increase the lifespan of equipment fostered the understanding of the fatigue behaviour of materials, either under multiaxial loading cycles or under increased number of loading cycles [1]. As mentioned by Armentani et al. [2], it is well known that there is an increasing interest in the use of new materials such as composites and new aluminium alloys which induced the development and improvement of the design criteria able to take into account phenomena such as fatigue, fracture, buckling and many others. Consequently, many design models based on complex analytical, empiric/semi-empiric formulations or numerical methods have been developed. Moreover, experimental tests are required to support these theoretical analyses. These tests should take into account the real specimen shape and size and the loading conditions. For this purpose, with reference to the fatigue behaviour, it is necessary the realization of multiaxial testing machines too, with the aim to investigate simple or complex structures under loading conditions similar to the real ones. It is generally recognized that multiaxial stresses occur in many full-scale structures, being rare the oc- currence of pure uniaxial stress states. https://doi.org/10.1016/j.engfailanal.2019.04.065 Received 15 February 2019; Received in revised form 3 April 2019; Accepted 25 April 2019 ⁎ Corresponding author. E-mail address: alessandro.deluca@unicampania.it (A. De Luca). Engineering Failure Analysis 102 (2019) 303–317 Available online 26 April 2019 1350-6307/ © 2019 Published by Elsevier Ltd. T