Numerical and experimental analysis of fatigue crack growth under random loading J. Zapatero a , B. Moreno a , A. Gonza ´lez-Herrera a , J. Domı ´nguez b, * a Department of Civil and Materials Engineering, University of Malaga, Malaga 29071, Spain b Department of Mechanical Engineering, University of Seville, Sevilla 41092, Spain Received 21 July 2004; received in revised form 2 November 2004; accepted 6 December 2004 Abstract This article analyses fatigue crack growth under random loading using experimental results taken from literature on the subject and from growth simulations carried out using the Strip Yield Model. The capacity of the Strip Yield Model in representing the retardation effects produced during the growth process is analysed. The effect of different statistical parameters of the random load process and of the representative histories of the same on the crack growth life and on the variability of the results obtained in different representative tests of the same real load process is likewise studied. Some of the parameters considered are the bandwidth of the random loading process, the number of cycles of the load histories employed and the effect of truncating the histories. It can be seen that under certain conditions, the numerical model employed can quite closely predict some of the behaviours of the crack growth. Likewise, it can also be seen that the bandwidth has a significant effect on the fatigue life, and that the length of the histories employed in tests or simulations has a great effect on the variability of the obtained results. It can also be seen that, although the elimination of the overloads tends to give rise to shorter lives, in certain cases, from a statistical point of view, it may be the origin of non-conservative predictions. q 2005 Elsevier Ltd. All rights reserved. Keywords: Strip Yield Model; Fatigue crack growth; Random loading 1. Introduction Fatigue crack growth under random and variable amplitude loading has been analysed by many different authors [1–5]. Some of them pay special attention to the simulation of representative load histories of defined random loading processes [6–8]. Others analyse the effect of overloads and their distribution throughout the load history of the crack growth life [9–11]. There are also many articles that present numerical or analytical models to simulate the crack growth under this type of loading [12–15], or which deal with the effect of many of the other parameters of the loading process on the growth rate [16–20]. Normally, the results of different crack growth tests or simulations under the same random loading process can be different. In the case of tests, this difference in the results is mainly a consequence of the difference in the behaviour of the material employed in each of the specimens and of the randomness of the applied loads. The history applied in each test is different, although they are all representative of the same random loading process. As regards growth simu- lations, the difference in the results is due solely to the randomness of the loads. Quite often, the load histories applied in tests or simulations are of a finite duration, and therefore, must be consecutively repeated up to failure. This finite character of these histories is another source of dispersion in the results. Therefore, in order to estimate the crack growth life of any mechanical system it will be important to have the representative load histories of the employed process available and that several tests be executed to enable the expected dispersion in the real results to be estimated. Thus, it will be possible to estimate the real life of the system from a statistical point of view. One of the objectives of this article is to analyse the effect of some of the statistical parameters of the random loading process on the variability of the results obtained in different crack growth tests under loads that represent the same International Journal of Fatigue 27 (2005) 878–890 www.elsevier.com/locate/ijfatigue 0142-1123/$ - see front matter q 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.ijfatigue.2004.12.008 * Corresponding author. E-mail address: jaime@us.es (J. Domı ´nguez).