Sectional global sensitivity measures

Pannier, Stephan; Graf, Wolfgang

doi:10.1016/j.ress.2014.09.009

Cited by 10 publications

(11 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Saltelli argues that an appropriate importance measure method should be "global, quantitative and model free." 2 With respect to the proposed EIM indices, from Equations (9), (10), (13), and (14) one can determine that both the individual terms and the interaction terms, including the linear order terms and the higher order terms, have been considered in. Furthermore, the terms are all averaged across the whole distribution space of the error sources.…”

Section: Properties Discussionmentioning

confidence: 99%

“…; (3) The variance‐based methods that use variance to characterize uncertainty of model input factors and model output; (4) The moment independent methods that use the difference between the distribution and the conditional distribution (given one or a group of input factors fixed at an arbitrary position) of model output to characterize the impact of input factors . In addition, there are also other methods have been proposed to tackle specific problems …”

Section: Introductionmentioning

confidence: 99%

“…8 In addition, there are also other methods have been proposed to tackle specific problems. [12][13][14][15] From the descriptions of the GSA methods, one can determine that the GSA methods focus mainly on the "uncertainty", rather than the deviation of random variables from the ideal or target values. We noticed that the ETQL is composed of two components: (a) The quality loss that is caused by variance of QC, and (b) The quality loss that is caused by mean shift (from the target value) of QC distribution.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Importance measure method for error sources of degradation systems based on the present worth of expected Taguchi quality loss

Sun

Cui

et al. 2020

Quality & Reliability Eng

View full text Add to dashboard Cite

Deteriorating of the error sources and the deviation of the quality characteristic (QC) of the product from the target value are ubiquitous in practice. Studying the relative contribution of the error sources to the QC deviation is helpful for improving the product performance effectively. In this study, we propose an error importance measure (EIM) method to quantify the relative contribution of error sources to the Taguchi quality loss that is caused by the QC deviation. The present worth of the expected Taguchi quality loss (ETQL) is used to characterize the deteriorating QC deviation. We use Taylor series expansion to decompose the time‐dependent ETQL and the present worth of ETQL into a series of fractions, representing the individual and the interaction contributions of the error sources. Based on the decomposition, the time‐dependent and the time‐independent EIM indices for one and a group of error sources are defined, respectively. The evaluation procedure for the EIM indices is provided and the properties are discussed. Finally, we provide a realistic application case to an aircraft wheel door mechanism to demonstrate the effectiveness of the proposed EIM method.

show abstract

Section: Properties Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Importance measure method for error sources of degradation systems based on the present worth of expected Taguchi quality loss

Sun

Cui

et al. 2020

Quality & Reliability Eng

View full text Add to dashboard Cite

show abstract

“…For static structures, a mature local sensitivity analysis method has been developed 34‐36 . A global sensitivity analysis determines the effect of arbitrary random input variables on the uncertainty of the output results over the full distribution range 37,38 . After continuous development, many mature global sensitivity analysis methods have been implemented, such as the variance‐based Global sensitivity analysis and moment‐independent global sensitivity analysis, among others 39‐41 .…”

Section: Introductionmentioning

confidence: 99%

Global sensitivity analysis of two‐stage thermoelectric refrigeration system based on response variance

Zhang

Wang

et al. 2020

Int J Energy Res

View full text Add to dashboard Cite

Summary In this paper, a mathematical model of two‐stage thermoelectric refrigeration system is established considering the influence of external heat transfer and its performance is analysed based on finite‐time thermodynamics and Newton's heat transfer law. Taking the cooling capacity and coefficient of performance of the two‐stage thermoelectric refrigeration system as separate objective functions, the general relationship between cooling capacity, coefficient of performance, and working design parameters of the system is determined. The influence of the fluctuation of the input design parameters on the output performance parameters is studied using a global sensitivity analysis based on the variance response. The main and total Global sensitivity indices of input parameters that affect the output performance are calculated, and the related sensitivity ranking are obtained. The results can be used to guide the performance analysis and parameter optimization of two‐stage thermoelectric refrigeration system in application.

show abstract

“…However, as their measures are not functions directly on the input space but on either a quantile value or the boundaries of the considered interval, we find our new sensitivity indices and plots more intuitive and informative. The approach of Pannier and Graf (2015) follows the same general ideas as our method. They subdivide the input space into a finite number of subdomains and calculate derivative based global sensitivity measures on these subdomains.…”

Section: Introductionmentioning

confidence: 99%

Support indices: Measuring the effect of input variables over their supports

Fruth

Roustant

Kuhnt

2019

Reliability Engineering & System Safety

View full text Add to dashboard Cite

Two new sensitivity indices are presented which give an original solution to the question in sensitivity analysis of how to determine regions within the input space for which the model variation is high. The indices, as functions over the input domain, give insight into the local influence of input variables over the whole domain when the other variables lie in the global domain. They can serve as an informative extension to a standard analysis and in addition are especially helpful in the specification of the input domain, a critical, but often vaguely handled issue in sensitivity analysis. In the usual framework of independent continuous input variables, we present theoretical results that show an asymptotic connection between the presented indices and Sobol' indices, valid for general probability distribution functions. Finally, we show how the indices can be successfully applied on analytical examples and on a real application.

show abstract

Sectional global sensitivity measures

Cited by 10 publications

References 21 publications

Importance measure method for error sources of degradation systems based on the present worth of expected Taguchi quality loss

Importance measure method for error sources of degradation systems based on the present worth of expected Taguchi quality loss

Global sensitivity analysis of two‐stage thermoelectric refrigeration system based on response variance

Support indices: Measuring the effect of input variables over their supports

Contact Info

Product

Resources

About