Investigating Prior Parameter Distributions in the Inverse Modelling of Water Distribution Hydraulic Models

476 Views
161 Downloads
Export citation: ABNT
KOZELJ, Daniel ;KAPELAN, Zoran ;NOVAK, Gorazd ;STEINMAN, Franci .
Investigating Prior Parameter Distributions in the Inverse Modelling of Water Distribution Hydraulic Models. 
Strojniški vestnik - Journal of Mechanical Engineering, [S.l.], v. 60, n.11, p. 725-734, june 2018. 
ISSN 0039-2480.
Available at: <https://www.sv-jme.eu/article/investigating-prior-parameter-distributions-in-the-inverse-modelling-of-water-distribution-hydraulic-models/>. Date accessed: 22 jul. 2019. 
doi:http://dx.doi.org/10.5545/sv-jme.2014.1741.
Kozelj, D., Kapelan, Z., Novak, G., & Steinman, F.
(2014).
Investigating Prior Parameter Distributions in the Inverse Modelling of Water Distribution Hydraulic Models.
Strojniški vestnik - Journal of Mechanical Engineering, 60(11), 725-734.
doi:http://dx.doi.org/10.5545/sv-jme.2014.1741
@article{sv-jmesv-jme.2014.1741,
	author = {Daniel  Kozelj and Zoran  Kapelan and Gorazd  Novak and Franci  Steinman},
	title = {Investigating Prior Parameter Distributions in the Inverse Modelling of Water Distribution Hydraulic Models},
	journal = {Strojniški vestnik - Journal of Mechanical Engineering},
	volume = {60},
	number = {11},
	year = {2014},
	keywords = {Bayesian inference; Calibration; Markov Chain Monte Carlo; Pipe networks; Hydraulics; Water distribution systems},
	abstract = {Inverse modelling concentrates on estimating water distribution system (WDS) model parameters that are not directly measurable, e.g. pipe roughness coefficients, which can, therefore, only be estimated by indirect approaches, i.e. inverse modelling. Estimation of the parameter and predictive uncertainty of WDS models is an essential part of the inverse modelling process. Recently, Markov Chain Monte Carlo (MCMC) simulations have gained in popularity in uncertainty analyses due to their effective and efficient exploration of posterior parameter probability density functions (pdf). A Bayesian framework is used to infer prior parameter information via a likelihood function to plausible ranges of posterior parameter pdf. Improved parameter and predictive uncertainty are achieved through the incorporation of prior pdf of parameter values and the use of a generalized likelihood function. We used three prior information sampling schemes to infer the pipe roughness coefficients of WDS models. A hypothetical case study and a real-world WDS case study were used to illustrate the strengths and weaknesses of a particular selection of a prior information pdf. The results obtained show that the level of parameter identifiability (i.e. sensitivity) is an important property for prior pdf selection.},
	issn = {0039-2480},	pages = {725-734},	doi = {10.5545/sv-jme.2014.1741},
	url = {https://www.sv-jme.eu/article/investigating-prior-parameter-distributions-in-the-inverse-modelling-of-water-distribution-hydraulic-models/}
}
Kozelj, D.,Kapelan, Z.,Novak, G.,Steinman, F.
2014 June 60. Investigating Prior Parameter Distributions in the Inverse Modelling of Water Distribution Hydraulic Models. Strojniški vestnik - Journal of Mechanical Engineering. [Online] 60:11
%A Kozelj, Daniel 
%A Kapelan, Zoran 
%A Novak, Gorazd 
%A Steinman, Franci 
%D 2014
%T Investigating Prior Parameter Distributions in the Inverse Modelling of Water Distribution Hydraulic Models
%B 2014
%9 Bayesian inference; Calibration; Markov Chain Monte Carlo; Pipe networks; Hydraulics; Water distribution systems
%! Investigating Prior Parameter Distributions in the Inverse Modelling of Water Distribution Hydraulic Models
%K Bayesian inference; Calibration; Markov Chain Monte Carlo; Pipe networks; Hydraulics; Water distribution systems
%X Inverse modelling concentrates on estimating water distribution system (WDS) model parameters that are not directly measurable, e.g. pipe roughness coefficients, which can, therefore, only be estimated by indirect approaches, i.e. inverse modelling. Estimation of the parameter and predictive uncertainty of WDS models is an essential part of the inverse modelling process. Recently, Markov Chain Monte Carlo (MCMC) simulations have gained in popularity in uncertainty analyses due to their effective and efficient exploration of posterior parameter probability density functions (pdf). A Bayesian framework is used to infer prior parameter information via a likelihood function to plausible ranges of posterior parameter pdf. Improved parameter and predictive uncertainty are achieved through the incorporation of prior pdf of parameter values and the use of a generalized likelihood function. We used three prior information sampling schemes to infer the pipe roughness coefficients of WDS models. A hypothetical case study and a real-world WDS case study were used to illustrate the strengths and weaknesses of a particular selection of a prior information pdf. The results obtained show that the level of parameter identifiability (i.e. sensitivity) is an important property for prior pdf selection.
%U https://www.sv-jme.eu/article/investigating-prior-parameter-distributions-in-the-inverse-modelling-of-water-distribution-hydraulic-models/
%0 Journal Article
%R 10.5545/sv-jme.2014.1741
%& 725
%P 10
%J Strojniški vestnik - Journal of Mechanical Engineering
%V 60
%N 11
%@ 0039-2480
%8 2018-06-28
%7 2018-06-28
Kozelj, Daniel, Zoran  Kapelan, Gorazd  Novak, & Franci  Steinman.
"Investigating Prior Parameter Distributions in the Inverse Modelling of Water Distribution Hydraulic Models." Strojniški vestnik - Journal of Mechanical Engineering [Online], 60.11 (2014): 725-734. Web.  22 Jul. 2019
TY  - JOUR
AU  - Kozelj, Daniel 
AU  - Kapelan, Zoran 
AU  - Novak, Gorazd 
AU  - Steinman, Franci 
PY  - 2014
TI  - Investigating Prior Parameter Distributions in the Inverse Modelling of Water Distribution Hydraulic Models
JF  - Strojniški vestnik - Journal of Mechanical Engineering
DO  - 10.5545/sv-jme.2014.1741
KW  - Bayesian inference; Calibration; Markov Chain Monte Carlo; Pipe networks; Hydraulics; Water distribution systems
N2  - Inverse modelling concentrates on estimating water distribution system (WDS) model parameters that are not directly measurable, e.g. pipe roughness coefficients, which can, therefore, only be estimated by indirect approaches, i.e. inverse modelling. Estimation of the parameter and predictive uncertainty of WDS models is an essential part of the inverse modelling process. Recently, Markov Chain Monte Carlo (MCMC) simulations have gained in popularity in uncertainty analyses due to their effective and efficient exploration of posterior parameter probability density functions (pdf). A Bayesian framework is used to infer prior parameter information via a likelihood function to plausible ranges of posterior parameter pdf. Improved parameter and predictive uncertainty are achieved through the incorporation of prior pdf of parameter values and the use of a generalized likelihood function. We used three prior information sampling schemes to infer the pipe roughness coefficients of WDS models. A hypothetical case study and a real-world WDS case study were used to illustrate the strengths and weaknesses of a particular selection of a prior information pdf. The results obtained show that the level of parameter identifiability (i.e. sensitivity) is an important property for prior pdf selection.
UR  - https://www.sv-jme.eu/article/investigating-prior-parameter-distributions-in-the-inverse-modelling-of-water-distribution-hydraulic-models/
@article{{sv-jme}{sv-jme.2014.1741},
	author = {Kozelj, D., Kapelan, Z., Novak, G., Steinman, F.},
	title = {Investigating Prior Parameter Distributions in the Inverse Modelling of Water Distribution Hydraulic Models},
	journal = {Strojniški vestnik - Journal of Mechanical Engineering},
	volume = {60},
	number = {11},
	year = {2014},
	doi = {10.5545/sv-jme.2014.1741},
	url = {https://www.sv-jme.eu/article/investigating-prior-parameter-distributions-in-the-inverse-modelling-of-water-distribution-hydraulic-models/}
}
TY  - JOUR
AU  - Kozelj, Daniel 
AU  - Kapelan, Zoran 
AU  - Novak, Gorazd 
AU  - Steinman, Franci 
PY  - 2018/06/28
TI  - Investigating Prior Parameter Distributions in the Inverse Modelling of Water Distribution Hydraulic Models
JF  - Strojniški vestnik - Journal of Mechanical Engineering; Vol 60, No 11 (2014): Strojniški vestnik - Journal of Mechanical Engineering
DO  - 10.5545/sv-jme.2014.1741
KW  - Bayesian inference, Calibration, Markov Chain Monte Carlo, Pipe networks, Hydraulics, Water distribution systems
N2  - Inverse modelling concentrates on estimating water distribution system (WDS) model parameters that are not directly measurable, e.g. pipe roughness coefficients, which can, therefore, only be estimated by indirect approaches, i.e. inverse modelling. Estimation of the parameter and predictive uncertainty of WDS models is an essential part of the inverse modelling process. Recently, Markov Chain Monte Carlo (MCMC) simulations have gained in popularity in uncertainty analyses due to their effective and efficient exploration of posterior parameter probability density functions (pdf). A Bayesian framework is used to infer prior parameter information via a likelihood function to plausible ranges of posterior parameter pdf. Improved parameter and predictive uncertainty are achieved through the incorporation of prior pdf of parameter values and the use of a generalized likelihood function. We used three prior information sampling schemes to infer the pipe roughness coefficients of WDS models. A hypothetical case study and a real-world WDS case study were used to illustrate the strengths and weaknesses of a particular selection of a prior information pdf. The results obtained show that the level of parameter identifiability (i.e. sensitivity) is an important property for prior pdf selection.
UR  - https://www.sv-jme.eu/article/investigating-prior-parameter-distributions-in-the-inverse-modelling-of-water-distribution-hydraulic-models/
Kozelj, Daniel, Kapelan, Zoran, Novak, Gorazd, AND Steinman, Franci.
"Investigating Prior Parameter Distributions in the Inverse Modelling of Water Distribution Hydraulic Models" Strojniški vestnik - Journal of Mechanical Engineering [Online], Volume 60 Number 11 (28 June 2018)

Authors

Affiliations

  • University of Ljubljana, Faculty of Civil and Geodetic Engineering, Slovenia
  • University of Exeter, School of Engineering and Computer Science, UK
  • Institute for Hydraulics Research, Ljubljana, Slovenia
  • University of Ljubljana, Faculty of Civil and Geodetic Engineering, Slovenia

Paper's information

Strojniški vestnik - Journal of Mechanical Engineering 60(2014)11, 725-734

10.5545/sv-jme.2014.1741

Inverse modelling concentrates on estimating water distribution system (WDS) model parameters that are not directly measurable, e.g. pipe roughness coefficients, which can, therefore, only be estimated by indirect approaches, i.e. inverse modelling. Estimation of the parameter and predictive uncertainty of WDS models is an essential part of the inverse modelling process. Recently, Markov Chain Monte Carlo (MCMC) simulations have gained in popularity in uncertainty analyses due to their effective and efficient exploration of posterior parameter probability density functions (pdf). A Bayesian framework is used to infer prior parameter information via a likelihood function to plausible ranges of posterior parameter pdf. Improved parameter and predictive uncertainty are achieved through the incorporation of prior pdf of parameter values and the use of a generalized likelihood function. We used three prior information sampling schemes to infer the pipe roughness coefficients of WDS models. A hypothetical case study and a real-world WDS case study were used to illustrate the strengths and weaknesses of a particular selection of a prior information pdf. The results obtained show that the level of parameter identifiability (i.e. sensitivity) is an important property for prior pdf selection.

Bayesian inference; Calibration; Markov Chain Monte Carlo; Pipe networks; Hydraulics; Water distribution systems