A Damage Estimation Method for Packaging Systems Based on Power Spectrum Densities Using Spectral Moments

• Published in: Packaging technology & science 2016-06, Vol.29 (6), p.303-321
• Main Authors: Huart, Victor, Nolot, Jean-baptiste, Candore, Jean-Charles, Pellot, Jérôme, Krajka, Nicolas, Odof, Serge, Erre, Damien
• Format: Article
• Language: English
• Subjects: Constants; Damage; Damage assessment; Density; Engineering Sciences; Materials; Mathematical models; Packaging; PSD; Spectra; spectral moment; Vibration; Wöhler
• ISSN: 0894-3214; 1099-1522
• DOI: 10.1002/pts.2211

Packaging systems always endure some type of damage during transportation, and Wöhler curves are typically used to determine the mechanical wear evolution of a packaging system. Damage estimations can be performed using a Rainflow method, but continuous stress recording is required. However, this recording is not always possible, so global transportation vibrations are represented by the power spectral density (PSD). In this paper, the damage of a system according to the Wöhler curve was studied. The Basquin model was used to determine the mechanical wear evolution of the system, and the mechanical behaviour was established by extracting the Basquin coefficient b and the constant system C from the Wöhler curves. A Wöhler curve was further realized for a packaging system, and its Basquin coefficient b and system constant C were also extracted. Damage is estimated by the accumulation of the stress cycles using the Palmgren–Miner rules. In this paper, the maximum peaks of the cycle were detected in a signal. The density probability of the maximum stress apparition was also constructed from a multivariate Gaussian model. PSDs have the same statistical properties as the corresponding temporal events, so these statistical properties were extracted by the spectral moment method. Furthermore, a damage equation was proposed using the PSD, and the damage induced by different PSDs was determined from the damage equation and compared with the actual duration on a vibration table. This paper focuses on packaging damage estimations using PSD. It describes the conditions of stress cycles apparition and their probability densities. Then, statistical parameters are extracted from PSD using spectral moments to complete the model and estimate the damage. The damage model is experimented on carboardbox.