The reliable characterization of uncertainties is critical for system evaluations and safety designs. However, insufficient samples are frequently encountered especially in composite materials. Four characterization methods for sparsely sampled quantities (SSQs) were discussed in the current study, including extreme value (EV) interval, tolerance interval (TI), optimized kernel density estimation (KDE), and the Dempster–Shafer (D-S) theory. Two different problems, a Sandia challenge function and a constitutive model for C/SiC composites, are considered for method evaluations. Performances of different methods are evaluated regarding reliabilities and errors in the output quantity derived with the uncertainty propagation. A manufactured solution method is employed. Results show that the TI, KDE, and D-S methods converge to the true distribution with increasing samples, whereas the EV method tends to expand the uncertainties. Under small samples ( 18 ), the TI method exhibits the highest reliability, and the KDE method exhibits the smallest error. Under extremely small samples ( 3 ) the D-S method offers a balance between reliability and error. The TI method can apply well in the constitutive model for the characterization of material uncertainties, where failure is of high consequence and conservatism is emphasized. The obtained results are general and can be employed in the future material characterization with experimental data.


    Zugriff

    Zugriff prüfen

    Verfügbarkeit in meiner Bibliothek prüfen

    Bestellung bei Subito €


    Exportieren, teilen und zitieren



    Titel :

    Uncertainty Characterization Methods for Sparsely Sampled Quantity: A Tradeoff Analysis Considering Propagation


    Beteiligte:
    Yang, Qiang (Autor:in) / Han, Guokai (Autor:in) / Xie, Weihua (Autor:in) / Meng, Songhe (Autor:in) / Jin, Hua (Autor:in)

    Erschienen in:

    AIAA Journal ; 58 , 7 ; 3129-3138


    Erscheinungsdatum :

    2020-04-20


    Format / Umfang :

    10 pages




    Medientyp :

    Aufsatz (Zeitschrift)


    Format :

    Elektronische Ressource


    Sprache :

    Englisch





    Texture recognition from sparsely and irregularly sampled data

    Petrou, M. / Piroddi, R. / Talebpour, A. | British Library Online Contents | 2006



    Multi-Sensor Excision of Sparsely Sampled Nonstationary Jammers for GPS Receivers

    Zhang, Yimin D. / Wang, Ben / Amin, Moeness G. | British Library Conference Proceedings | 2015


    A Comparison of Methods for Representing and Aggregating Uncertainties involving Sparsely Sampled Random Variables - Final Results

    Romero, V.J. / Swiler, L.P. / Urbina, A. et al. | British Library Conference Proceedings | 2013