Stochastic sensitivity analysis

Irving, A.D.

doi:10.1016/0307-904x(92)90110-o

Cited by 23 publications

(11 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Sensitivity analysis aims at studying the relationships between the output and input variables. Differential sensitivity analysis is considered to be the most commonly employed method in sensitivity analysis (Irving, 1992). This method deals with local sensitivity analysis by focusing on the evaluation of the partial derivatives ߲݂ ‫ݕ߲‬ ⁄ of the function f. Many approximation methods are used to calculate the partial derivatives of f .…”

Section: Sensitivity Analysismentioning

confidence: 99%

Estimating the variability of tillage forces on a chisel plough shank by modeling the variability of tillage system parameters

Abo-Alkheer

Kharmanda

Hami

et al. 2011

Computers and Electronics in Agriculture

View full text Add to dashboard Cite

In this paper, a probabilistic approach is proposed for quantifying the variability of the tillage forces for the shank of a chisel plough with narrow tines and to estimate the failure probability. An existing threedimensional analytical model of tool forces from McKyes was used to model the interaction between the tillage tools and the soil. The variability of tillage forces was modeled, taking into account the variability of soil engineering properties, tool design parameters and operational conditions. The variability of the soil engineering properties was modeled by means of experimental observations. The dispersion effect of each tillage system parameter on the tillage forces was determined by a sensitivity analysis. The results show that the variability of the horizontal and vertical forces follows a lognormal distribution ((ߤ ൌ ͲǤ ͺ ʹǡߦ ൌ ͲǤ ͶͶͻ); (ߤ ൌ ͲǤ ͲͲͶǡߦ ൌ ͲǤ ͶͶ)) and the relationship between these forces is positive and quasi-linear ሺߩ(ܲ ு ǡܲ) = 0.93). This lognormal variability was integrated into the estimation of the failure probability for the shank by using Monte Carlo simulation (MCS) and the first-order reliability method (FORM). The results obtained by these two methods, with the assumption of non-correlation between the horizontal and vertical forces, were almost identical. However, the FORM method was faster and simpler, compared to the MCS technique. Furthermore, the correlation between the horizontal and vertical forces has no significant effect on the failure probability, regardless of the correlation strength. Therefore, it is concluded that the FORM method can be used to estimate the failure probability without considering the correlation between horizontal and vertical forces.

show abstract

Section: Sensitivity Analysismentioning

confidence: 99%

Estimating the variability of tillage forces on a chisel plough shank by modeling the variability of tillage system parameters

Abo-Alkheer

Kharmanda

Hami

et al. 2011

Computers and Electronics in Agriculture

View full text Add to dashboard Cite

show abstract

“…Screening methods are more specifically used to reduce large sets of parameters to smaller sets before applying global sensitivity analysis methods, which are more computationally demanding ( [24], [29]). In the simple case of «a time invariant linear system, whose input time series boundary condition experiences a time continuous invariant perturbation» [21], the first-order sensitivity functions are expressed by the partial derivative of the output y with respect to each of the input parameter . In the case of building energy simulations tools, the relationship between the outputs and the inputs may be strongly non-linear and input parameters may be correlated to each other.…”

Section: Brief Introduction To Sensitivity Analysismentioning

confidence: 99%

Empirical validation and local sensitivity analysis of a lumped-parameter thermal model of an outdoor test cell

Cattarin

Pagliano

Causone

et al. 2018

Building and Environment

View full text Add to dashboard Cite

This paper presents the experimental validation of a thermal model describing the ZEB Test Cells Laboratory, located at the Gløshaugen campus of NTNU and SINTEF in Trondheim, Norway. Besides, a local sensitivity analysis identifies the parameters and inputs that are most influential on the thermal behaviour of the test cell, in terms of temperature profiles of the internal air and internal surfaces. The analysis shows that, in free-running conditions, the most important parameters and inputs, out of the 49 tested ones, are: the air temperature in the guard zone, the initial temperature(s) of the test cell envelope, the linear dimension of the square window, the solar irradiance on the vertical plane of the window, the depth of the test cell, the thermal conductivity and the thickness of the polyurethane layer in the envelope, the solar direct transmittance of the window, the internal height and width of the test cell, the external air temperature and the electrical power input to the mixing fan. Based on the outcome of the local sensitivity analysis and on in-field observations, some practical measures to improve the quality of the input data provided to a dynamic energy simulation tool, and thus the accuracy of prediction of the temperature evolution of the test cell. For example, we suggest monitoring accurately the environmental conditions in the guard zone, which are particularly influential under free-running conditions, and installing a global irradiance pyranometer next to the window in order to reduce the uncertainty related to the entering solar load.

show abstract

“…such that the oberserved heat flux at the surface, q(t), is composed of components due to the conductive, radiative and convective processes that interact at the surface. This equation can be written in terms of the physical observable variables and the linear response factors or functions of each process as, process is the heat transfer coefficient or gain of that process [8].…”

Section: Application Of Techniquementioning

confidence: 99%

“…flTt ( k, T), this function will also enable the 'driven' process, if any, to be identified. The area under the estimated first order/linear response function is the linear steady state gain of that process [8]. The steady state gain between the surface heat flux and the temperature difference driving force of a process is the surface heat transfer coefficient of that process [5].…”

Section: Application Of Techniquementioning

confidence: 99%

“…) = L hqt::.T,(T)6.Ts(t-T) + L hqt::.T~(T)a-6.Tr 4 (t-T) + L hqt::.T 1 (T)6.TJ(t-T)(8) where {6.Ts(t), 6.Tr 4 (t), 6.T,(t)} are, respectively, the temperature gradient within the ceiling, the difference between the fourth powers of the surface temperatures of the ceiling and the average of the remaining surfaces within the test cell, and the temperature dif-ference across the air-surface boundary layer, and where hqt::.T. ( T), hqt::.T!…”

mentioning

confidence: 99%

See 1 more Smart Citation