Uncertainty characterization under measurement errors using maximum likelihood estimation: cantilever beam end-to-end UQ test problem
- 89 Downloads
One goal of uncertainty characterization is to develop a probability distribution that is able to properly characterize uncertainties in observed data. Observed data may vary due to various sources of uncertainty, which include uncertainties in geometry and material properties, and measurement errors. Among them, measurement errors, which are categorized as systematic and random measurement errors, are often disregarded in the uncertainty characterization process, even though they may be responsible for much of the variability in the observed data. This paper proposes an uncertainty characterization method that considers measurement errors. The proposed method separately distinguishes each source of uncertainty by using a specific type of probability distribution for each source. Next, statistical parameters of each assumed probability distribution are estimated by adopting the maximum likelihood estimation. To demonstrate the proposed method, as a case study, the method was implemented to characterize the uncertainties in the observed deflection data from the tip of a cantilever beam. In this case study, the proposed method showed greater accuracy as the amount of available observed data increased. This study provides a general guideline for uncertainty characterization of observed data in the presence of measurement errors.
KeywordsUncertainty characterization Uncertainty modeling Measurement error Systematic measurement error Random measurement error Maximum likelihood estimation
This work was partially supported by the Technology Innovation Program (10048305, Launching Plug-in Digital Analysis Framework for Modular System Design) of the Ministry of Trade, Industry & Energy (MI, Korea). This work was also supported by a grant from the Institute of Advanced Machinery and Design at Seoul National University (SNU-IAMD).
- Easterling RG (2001) Measuring the predictive capability of computational methods: principles and methods, issues and illustrations SAND2001-0243, Sandia National LaboratoriesGoogle Scholar
- Jung BC, Park J, Oh H, Kim J, Youn BD (2015) A framework of model validation and virtual product qualification with limited experimental data based on statistical inference. Struct Multidiscip Optim 51(3):573–583Google Scholar
- Kim T, Lee G, Kim S, Youn BD (2018) Expectation-maximization method for data-based estimation of the cantilever beam end-to-end problem. In: 2018 AIAA Non-Deterministic Approaches Conference. p 1666Google Scholar
- Lee G, Yi G, Youn BD (2018) Special issue: a comprehensive study on enhanced optimization-based model calibration using gradient information. Struct Multidiscip Optim 57(5):2005–2025Google Scholar
- Liang B, Mahadevan S (2011) Error and uncertainty quantification and sensitivity analysis in mechanics computational models. Int J Uncertain Quantif 1(2):147–161. https://doi.org/10.1615/Int.J.UncertaintyQuantification.v1.i2.30
- Lin G, Elizondo M, Lu S, Wan X (2014) Uncertainty quantification in dynamic simulations of large-scale power system models using the high-order probabilistic collocation method on sparse grids. Int J Uncertain Quantif 4(3):185–204Google Scholar
- Romero VJ (2010) Data & model conditioning for multivariate systematic uncertainty in model calibration, validation, and extrapolation. In: 12th AIAA Non-Deterministic Approaches Conference, Orlando, FL, AIAA Paper. p 2010Google Scholar
- Romero VJ, Weirs VG (2018) A class of simple and effective UQ methods for sparse replicate data applied to the cantilever beam end-to-end UQ problem. In: 2018 AIAA Non-Deterministic Approaches Conference. p 1665Google Scholar
- Romero V, Schroeder B, Glickman M (2017) Cantilever beam end-to-end UQ test problem: handling experimental and simulation uncertainties in model calibration, model validation, and risk assessment, Sandia National Laboratories document SAND2017-4689 O, version BeamTestProblem32.docxGoogle Scholar