An efficient inverse approach for reconstructing time- and space-dependent heat flux of participating medium
Shuang-Cheng Sun(孙双成)1,2, †, Guang-Jun Wang(王广军)1,2, and Hong Chen(陈红)1,2$
1 School of Energy and Power Engineering, Chongqing University, Chongqing 400044, China 2 Key Laboratory of Low-grade Energy Utilization Technologies and Systems, Ministry of Education, Chongqing University, Chongqing 400044, China
The decentralized fuzzy inference method (DFIM) is employed as an optimization technique to reconstruct time- and space-dependent heat flux of two-dimensional (2D) participating medium. The forward coupled radiative and conductive heat transfer problem is solved by a combination of finite volume method and discrete ordinate method. The reconstruction task is formulated as an inverse problem, and the DFIM is used to reconstruct the unknown heat flux. No prior information on the heat flux distribution is required for the inverse analysis. All retrieval results illustrate that the time- and space-dependent heat flux of participating medium can be exactly recovered by the DFIM. The present method is proved to be more efficient and accurate than other optimization techniques. The effects of heat flux form, initial guess, medium property, and measurement error on reconstruction results are investigated. Simulated results indicate that the DFIM is robust to reconstruct different kinds of heat fluxes even with noisy data.
Shuang-Cheng Sun(孙双成), Guang-Jun Wang(王广军), and Hong Chen(陈红)$ An efficient inverse approach for reconstructing time- and space-dependent heat flux of participating medium 2020 Chin. Phys. B 29 110202
Fig. 1.
Schematic diagram of CRCHT in 2D participating medium.
Parameter
Lx/m
Ly/m
λ/W⋅m−1⋅K−1
κa/m−1
κs/m−1
n
εw
Φ
Value
1.0
1.0
2.268, 22.68, 226.8
1.0
0.0
1.0
1.0
1.0
Table 1.
Parameters in the verification case.
Fig. 2.
Verification of the FVM–DOM solution for solving CRCHT problem in 2D participating medium.
Fig. 3.
Membership functions of (a) fuzzy input set and (b) fuzzy output set.
em,n
NB
NM
NS
ZO
PS
PM
PB
um,n
PB
PM
PS
ZO
NS
NM
NB
Table 2.
Fuzzy rules of DFIM.
Fig. 4.
Fuzzy inference system of DFIM.
Fig. 5.
Flowchart of DFIM for reconstructing time- and space-dependent heat flux.
Fig. 6.
Reconstruction results of heat flux Q1 using DFIM. (a) Exact distribution, (b) reconstructed results, and (c) relative error distribution.
Optimization technique
Average relative error/%
Maximum relative error/%
Iteration number
Computational time/h
DFIM
0.0002
0.0092
46
0.3
SQP
0.0003
0.0229
306
86.2
CGM
0.0010
0.1341
698
153.3
L-M
0.0015
0.2056
717
172.9
PSO
44.8900
287.7838
10000
23.8
GA
59.2025
392.2087
10000
36.1
Table 3.
Reconstruction results of time- and space-dependent heat flux using different optimization techniques.
Fig. 7.
Retrieval results of heat flux Q2. (a) Exact distribution, and (b) reconstruction result.
Fig. 8.
Retrieval results of heat flux Q3. (a) Exact distribution, and (b) reconstruction result.
Fig. 9.
Retrieval results of heat flux Q4. (a) Exact distribution, and (b) reconstruction result.
Fig. 10.
Retrieval results of heat flux Q5. (a) Exact distribution, and (b) reconstruction result.
Heat flux
Average relative error/%
Maximum relative error/%
Iteration number
Computational time/s
Q1
0.0002
0.0092
46
921.1
Q2
0.0002
0.0099
46
922.7
Q3
0.0001
0.0028
45
919.8
Q4
0.0001
0.0030
46
922.3
Q5
0.0001
0.0075
46
923.9
Table 4.
Reconstruction results of different kinds of heat fluxes.
Initial guess/W⋅m−2
Average relative error/10−6
Computational time/s
Q1
Q2
Q3
Q4
Q5
Q1
Q2
Q3
Q4
Q5
0
1.40
1.97
1.07
0.92
0.79
923.2
923.5
921.0
923.0
921.5
5000
1.38
1.96
1.06
0.93
0.83
921.5
926.4
923.4
921.6
919.4
10000
1.50
1.92
1.05
0.92
0.77
921.1
922.7
919.8
922.3
923.9
30000
1.53
1.90
1.03
0.91
0.83
923.0
921.3
920.1
925.6
921.6
50000
1.99
1.93
1.02
0.91
0.81
922.4
925.4
920.4
922.9
924.1
100000
2.21
2.00
1.16
1.05
0.88
921.6
924.3
922.6
921.7
923.8
Table 5.
Reconstruction results for different initial guesses.
ρ/kg⋅m−3
cp/J⋅kg−1⋅ K−1
λ/W⋅m−1⋅ K−1
κa/m−1
κs/m−1
n
Ceramic
5600
456
1.4
300
12000
1.8
Glass
2219
840
1.025
1244
0
1.4
Table 6.
Parameter of ceramic and glass.[1]
Fig. 11.
Reconstruction results of Q3 in ceramic and glass industries. (a) Exact and reconstructed heat fluxes, (b) relative error distribution of ceramic, and (c) relative error distribution of glass.
Measurement error
Average relative error/%
Maximum relative error/%
Computational time/s
0.01
0.0203
0.1560
923.1
0.05
0.1062
0.6897
921.9
0.1
0.2115
1.3735
922.1
0.5
1.0208
6.9042
919.6
1.0
2.0916
12.2036
921.5
Table 7.
Retrieval relative errors for different measurement errors.
Fig. 12.
Reconstruction results of heat flux Q3 for different measurement errors. (a) σ = 0.01, (b) σ = 0.05, (c) σ = 0.1, (d) σ = 0.5, and (e) σ = 1.0.
Qi H, Sun S C, He Z Z, Ruan S T, Ruan L M, Tan H P 2016 Inverse geometry design of radiative enclosures using particle swarm optimization algorithms Rijeka InTechOpen
[21]
Qi H, Zhao F Z, Ren Y T, Qiao Y B, Wei L Y, Sadaf A I, Ruan L M 2018 J. Quant. Spectrosc. Radiat. Transfer222–223 1
[22]
Li B H, Lu M 2015 J. Univ. Shanghai Sci. Technol.37 225
[23]
Chopade R P, Mohan V, Mayank R, Uppaluri R V S, Mishra S C 2013 Numer. Heat Transfer, Part A63 373 DOI: 10.1080/10407782.2013.733179
Altmetric calculates a score based on the online attention an article receives. Each coloured thread in the circle represents a different type of online attention. The number in the centre is the Altmetric score. Social media and mainstream news media are the main sources that calculate the score. Reference managers such as Mendeley are also tracked but do not contribute to the score. Older articles often score higher because they have had more time to get noticed. To account for this, Altmetric has included the context data for other articles of a similar age.