Abstract
In this study, four different versions of the variable metric method (VMM) are investigated in solving standard one-dimensional inverse heat conduction problems in order to evaluate their efficiency and accuracy. These versions include Davidon–Fletcher–Powell (DFP), Broydon–Fletcher–Goldfarb–Shanno (BFGS), Symmetric Rank-one (SR1), and Biggs formula of the VMM. These investigations are carried out using temperature data obtained from numerical simulations.
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs00231-006-0107-9/MediaObjects/231_2006_107_Fig1_HTML.gif)
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs00231-006-0107-9/MediaObjects/231_2006_107_Fig2_HTML.gif)
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs00231-006-0107-9/MediaObjects/231_2006_107_Fig3_HTML.gif)
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs00231-006-0107-9/MediaObjects/231_2006_107_Fig4_HTML.gif)
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs00231-006-0107-9/MediaObjects/231_2006_107_Fig5_HTML.gif)
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs00231-006-0107-9/MediaObjects/231_2006_107_Fig6_HTML.gif)
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs00231-006-0107-9/MediaObjects/231_2006_107_Fig7_HTML.gif)
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs00231-006-0107-9/MediaObjects/231_2006_107_Fig8_HTML.gif)
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs00231-006-0107-9/MediaObjects/231_2006_107_Fig9_HTML.gif)
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs00231-006-0107-9/MediaObjects/231_2006_107_Fig10_HTML.gif)
Similar content being viewed by others
Abbreviations
- e RMS :
-
relative root mean square error, Eq. 23
- f :
-
function of sum of the squares of errors, Eq. 5
- K :
-
number of sensors
- M :
-
total number of time steps (equal to total number of unknowns that should be estimated.)
- q :
-
heat flux (W/m2)
- \(\vec{q}\) :
-
unknowns vector (design vector with dimension M × 1), Eq. 2
- S :
-
vector of descent direction, Eq. 6
- T :
-
temperature calculated by the model (K)
- t :
-
time (s)
- X :
-
pulse sensitivity coefficient (K m2 /W) defined by Eq. 3
- Y :
-
temperature measured by sensors
- (x, y ):
-
Cartesian coordinate system
- α:
-
thermal diffusivity (m2/s)
- ɛ:
-
tolerance error
- λ:
-
search step length
- λ* :
-
optimal step length
- ∇:
-
gradient
- T:
-
matrix transpose symbol
- →:
-
vector
- ′:
-
derivative symbol
- k :
-
sensor index number
- M :
-
total number of time steps
- \(m, \tilde{m}\) :
-
indices of time step, each between 1,2,..., M
- i :
-
iteration number of VMM cycles
References
Beck JV, Blackwell B, Clair CR (1985) Inverse heat conduction. Wiley, New York
Ozicik MN, Orlande HRB (2000) Inverse heat transfer, fundamentals and applications. Taylor & Francis, New York
Alifanov OM (1994) Inverse heat transfer problems. Springer, Berlin Heidelberg New York
Huang CH, Chen WC (2000) A three dimensional inverse forced convection problem in estimating surface heat flux by conjugate gradient method. Int J Heat Mass Transfer 43:3171–3181
Chen UC, Chang WJ, Hsu JC (2001) Two-dimensional inverse problem in estimating heat flux of pin fins. Int Commun Heat Mass Transfer 28(6):793–801
Linhua L, Heping T, Qizheng Y (1999) Inverse radiation problem of temperature field in three-dimensional rectangular furnace. Int Commun Heat Mass Transfer 26(2):239–248
Loulou T, Scott EP (2003) Estimation of 3-dimensional heat flux from surface temperature measurements using an iterative regularization method. Heat Mass Transfer 39:435–443
Rao SS (1995) Optimization; theory and applications, 2nd edn (9th reprint). New Age International (P) Limited Publishers, New Delhi
Lukšan L, Spedicato E (2000) Variable metric methods for unconstrained optimization and nonlinear least squares. J Comput Appl Math 124: 61–95
Zhang ZZ, Cao DH, Zeng JP (2004) Convergence property of a class of variable metric methods. Appl Math Lett 17: 437–442
Biggs MC (1973) A note on minimization algorithms which make use of non-quadratic properties of the objective function. J Inst Maths Appl 12:337–338
Phua KH (1996) Multi-directional parallel algorithms for unconstrained optimization. Optimization 38:107–125
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Pourshaghaghy, A., Kowsary, F. & Behbahaninia, A. Comparison of four different versions of the variable metric method for solving inverse heat conduction problems. Heat Mass Transfer 43, 285–294 (2007). https://doi.org/10.1007/s00231-006-0107-9
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00231-006-0107-9