Global sensitivity analysis of riveting parameters based on a random sampling-high dimensional model representation

2021 ◽  
Vol 113 (1-2) ◽  
pp. 465-472
Author(s):  
Junqing Yin ◽  
Jinyu Gu ◽  
Yongdang Chen ◽  
Feng Zhang
Keyword(s):  
Sensitivity Analysis ◽  
Random Sampling ◽  
Global Sensitivity Analysis ◽  
Model Representation ◽  
High Dimensional ◽  
High Dimensional Model Representation ◽  
Dimensional Model ◽  
Global Sensitivity

Analytical global sensitivity analysis with Gaussian processes

2017 ◽  
Vol 31 (3) ◽  
pp. 235-250 ◽  
Author(s):  
Ankur Srivastava ◽  
Arun K. Subramaniyan ◽  
Liping Wang
Keyword(s):  
Sensitivity Analysis ◽  
Gaussian Processes ◽  
Global Sensitivity Analysis ◽  
Decomposition Methods ◽  
Variance Decomposition ◽  
Model Representation ◽  
High Dimensional ◽  
High Dimensional Model Representation ◽  
Sobol Indices ◽  
Global Sensitivity

AbstractMethods for efficient variance-based global sensitivity analysis of complex high-dimensional problems are presented and compared. Variance decomposition methods rank inputs according to Sobol indices that can be computationally expensive to evaluate. Main and interaction effect Sobol indices can be computed analytically in the Kennedy and O'Hagan framework with Gaussian processes. These methods use the high-dimensional model representation concept for variance decomposition that presents a unique model representation when inputs are uncorrelated. However, when the inputs are correlated, multiple model representations may be possible leading to ambiguous sensitivity ranking with Sobol indices. In this work, we present the effect of input correlation on sensitivity analysis and discuss the methods presented by Li and Rabitz in the context of Kennedy and O'Hagan's framework with Gaussian processes. Results are demonstrated on simulated and real problems for correlated and uncorrelated inputs and demonstrate the utility of variance decomposition methods for sensitivity analysis.

Variance Based Global Sensitivity Analysis for Uncorrelated and Correlated Inputs With Gaussian Processes

2015 ◽  
Author(s):  
Ankur Srivastava ◽  
Arun K. Subramaniyan ◽  
Liping Wang
Keyword(s):  
Sensitivity Analysis ◽  
Gaussian Processes ◽  
Global Sensitivity Analysis ◽  
Decomposition Methods ◽  
Variance Decomposition ◽  
Model Representation ◽  
High Dimensional ◽  
High Dimensional Model Representation ◽  
Sobol Indices ◽  
Global Sensitivity

Methods for efficient variance based global sensitivity analysis of complex high-dimensional problems are presented and compared. Variance decomposition methods rank inputs according to Sobol indices which can be computationally expensive to evaluate. Main and interaction effect Sobol indices can be computed efficiently in the Kennedy & O’Hagan framework with Gaussian Processes (GPs). These methods use the High Dimensional Model Representation (HDMR) concept for variance decomposition which presents a unique model representation when inputs are uncorrelated. However, when the inputs are correlated, multiple model representations may be possible leading to ambiguous sensitivity ranking with Sobol indices. In this work we present the effect of input correlation on sensitivity analysis and discuss the methods presented by Li & Rabitz in the context of Kennedy & O’ Hagan framework with GPs. Results are demonstrated on simulated and real problems for correlated and uncorrelated inputs and demonstrate the utility of variance decomposition methods for sensitivity analysis.

scholarly journals Global sensitivity analysis for the Rothermel model based on high-dimensional model representation

2015 ◽  
Vol 45 (11) ◽  
pp. 1474-1479 ◽  
Author(s):  
Yaning Liu ◽  
M. Yousuff Hussaini ◽  
Giray Ökten
Keyword(s):  
Sensitivity Analysis ◽  
Global Sensitivity Analysis ◽  
High Dimensional ◽  
High Dimensional Model Representation ◽  
Dimensional Model ◽  
Global Sensitivity ◽  
Sensitivity Indices ◽  
Fuel Model ◽  
Input Parameters ◽  
Fuel Models

Rothermel’s wildland surface fire spread model is widely used in North America. The model outputs depend on a number of input parameters, which can be broadly categorized as fuel model, fuel moisture, terrain, and wind parameters. Due to the inevitable presence of uncertainty in the input parameters, knowing the sensitivity of the model output to a given input parameter can be very useful for understanding and controlling the sources of parametric uncertainty. Instead of obtaining the local sensitivity indices, we perform a global sensitivity analysis that considers the synchronous changes of parameters in their respective ranges. The global sensitivity indices corresponding to different parameter groups are computed by constructing the truncated ANOVA – high dimensional model representation for the model outputs with a polynomial expansion approach. We apply global sensitivity analysis to six standard fuel models, namely short grass, tall grass, chaparral, hardwood litter, timber, and light logging slash. Our sensitivity results show similarities, as well as differences, between fuel models. For example, the sensitivities of the input parameters, i.e., fuel depth, low heat content, and wind, are large in all fuel models and as high as 85% of the total model variance in the fuel model light logging slash. On the other hand, the fuel depth explains around 40% of the total variance in the fuel model light logging slash but only 12% of the total variance in the fuel model short grass. The quantification of the importance of parameters across fuel models helps identify the parameters for which additional resources should be used to lower their uncertainty, leading to effective fire management.

Sign in / Sign up

Export Citation Format

Share Document