Please wait a minute...
Chinese Journal of Engineering Design  2026, Vol. 33 Issue (2): 213-222    DOI: 10.3785/j.issn.1006-754X.2026.06.103
Optimization Design     
Structural optimization design method for dry-type distribution transformer based on multi-objective starfish optimization algorithm
Enxin XIANG1(),Hang LI1,Yongjie NIE1,Jingrong GUAN2,Dongyang WANG2
1.Electric Power Science Research Institute, Yunnan Power Grid Co. , Ltd. , Kunming 650032, China
2.School of Electrical Engineering, Southwest Jiaotong University, Chengdu 611756, China
Download: HTML     PDF(2895KB)
Export: BibTeX | EndNote (RIS)      

Abstract  

In order to enhance the service performance of epoxy resin dry-type distribution transformer, taking a 10 kV/0.4 kV epoxy resin dry-type distribution transformer as the research object, a structure optimization design method for dry-type distribution transformer based on the multi-objective starfish optimization algorithm (MOSFOA) was proposed. Firstly, a transformer model coupling electromagnetic-thermal-mechanical fields was established for performance simulation. The cross-sectional areas of the high/low voltage winding conductors, the width of the air ducts, and the radius of the core were determined as the core design variables. Subsequently, high-precision surrogate models linking the design variables to the hotspot temperature rise, loss and short-circuit electromagnetic force per unit volume of the high/low voltage windings were constructed using central composite design and response surface methodology. Finally, collaborative optimization based on the MOSFOA was performed to obtain the optimal combination of structural parameters.The resultsindicate that, while maintaining the insulation safe, compared with the initial plan, the hotspot temperature rise, loss, and short-circuit electromagnetic force per unit volume of the high/low voltage windings were reduced by 22.91%, 33.37%, 9.39%, 42.82%, and 45.52% respectively after optimization. This verified the effectiveness of the proposed optimization method.The research results provide a new method for the optimization design of dry-type transformers, and have significant reference value for improving the design efficiency and operational reliability of dry-type transformers.



Key wordsepoxy resin dry-type distribution transformer      structural optimization      response surface methodology      multi-objective starfish optimization algorithm      collaborative optimization     
Received: 25 September 2025      Published: 28 April 2026
CLC:  TM 741  
Cite this article:

Enxin XIANG,Hang LI,Yongjie NIE,Jingrong GUAN,Dongyang WANG. Structural optimization design method for dry-type distribution transformer based on multi-objective starfish optimization algorithm. Chinese Journal of Engineering Design, 2026, 33(2): 213-222.

URL:

https://www.zjujournals.com/gcsjxb/10.3785/j.issn.1006-754X.2026.06.103     OR     https://www.zjujournals.com/gcsjxb/Y2026/V33/I2/213


基于多目标海星优化算法的干式配电变压器结构优化设计方法

为了提高环氧树脂干式配电变压器的服役性能,以一台10 kV/0.4 kV环氧树脂干式配电变压器为研究对象,提出了一种基于多目标海星优化算法(multi-objective starfish optimization algorithm, MOSFOA)的干式配电变压器结构优化设计方法。首先,建立了电磁-热-力耦合的变压器模型并进行性能仿真,确定以高/低压绕组导体截面积、风道宽度和铁心半径为核心设计变量;其次,运用中心复合设计和响应面法构建了设计变量与高/低压绕组热点温升、损耗及单位体积短路电动力之间的高精度代理模型;最后,基于MOSFOA进行协同寻优,获得最优结构参数组合。优化结果表明:在保持绝缘安全的前提下,相较于初始方案,优化后高、低压绕组热点温升,损耗,高、低压绕组单位体积短路电动力分别降低了22.91%、33.37%、9.39%、42.82%、45.52%,验证了所提出优化方法的有效性。研究结果为干式变压器的优化设计提供了新方法,对提高干式变压器的设计效率和运行可靠性具有重要的参考价值。


关键词: 环氧树脂干式配电变压器,  结构优化,  响应面法,  多目标海星优化算法,  协同寻优 
参数高压绕组低压绕组
额定容量/kVA800
额定电压/kV100.4
额定电流/A46.21 154.7
匝数77717
层数373
频率/Hz50
空载损耗,负载损耗/W875,5 840
联结组别Dyn11
Table 1 Basic parameters of dry-type distribution transformer
Fig.1 Grid partitioning of electromagnetic field simulation model for dry-type distribution transformer
Fig.2 B-H curve of core
Fig.3 Grid section of winding area of dry-type distribution transformer
材料

密度/

(kg/m3)

热导率/

[W/(m·K)]

比热容/

[J/(kg·K)]

黏度/

[kg/(m·s)]

铜导体8 933401385
环氧树脂1 2000.461 300
绝缘纸9300.191 340
复合聚酯漆0.23
空气1.2250.024 21 006.431.789 4×10-5
Table 2 Material physical properties for temperature field calculation
Fig.4 Distribution nephogram of electric field of dry-type distribution transformer under rated operating condition
Fig.5 Distribution nephogram of winding loss per unit volume under rated operating condition
Fig.6 Distribution nephogram of winding temperature under rated operating condition
Fig.7 Distribution nephogram of winding short-circuit electromagnetic force per unit volume under rated operating condition
结构参数抽样范围
Slv-30%~30%
Shv-30%~30%
W-30%~30%
R/mm100~156
Table 3 Sampling range of structural parameters of dry-type distribution transformer
组号Shv/mm2Slv/mm2WR/mmThv/KTlv/KS/W
119.68601.750.75132.6774.4073.085 718.0
220.64746.750.97117.7370.4955.755 283.0
317.76616.250.99153.2075.6254.466 361.8
419.04674.251.25112.1375.6154.255 740.2
513.28775.750.93155.07102.9146.476 677.4
612.32877.251.11138.27114.6740.416 636.9
718.08630.750.85102.8082.3278.085 508.0
819.36906.251.17127.0771.4539.915 370.3
915.84645.250.89123.3390.4265.205 996.7
1020.32688.751.19142.0070.0745.935 874.9
1115.52703.251.13128.9391.7750.146 186.0
1212.96558.250.81110.27113.2885.796 443.4
1315.20587.251.15104.6794.0767.706 256.8
1411.36732.250.83130.80129.1959.476 727.2
1518.40514.751.05121.4780.8371.966 280.2
1617.12804.750.95134.5381.6148.635 682.6
1720.00862.750.91147.6070.6744.025 406.9
1814.88790.250.73108.4097.8969.905 498.1
1913.60529.250.87143.87104.3071.587 092.9
2011.68572.751.09125.20125.0863.297 244.7
2112.00761.251.03106.53122.1257.926 330.9
2213.92819.251.29114.00103.2945.006 180.0
2318.72891.750.77119.6077.4856.385 081.7
2416.48543.751.27136.4086.7756.956 794.4
2516.80848.251.21151.3384.8336.656 072.3
2616.16717.750.71145.7390.6461.845 933.1
2717.44833.751.07100.9382.8154.195 347.5
2814.24920.750.79140.13100.8248.615 898.0
2914.56935.251.01115.8799.1146.505 700.0
3012.64659.751.23149.47114.8745.857 301.7
Table 4 Calculation results of hotspot temperature rise and loss of high/low voltage windings
源项平方和自由度均方Fp
R2=0.987 0Radj2=0.974 4
模型8 744.1414624.58169.67<0.000 1
A: Shv8 296.4718 296.472 253.72<0.000 1
B: Slv6.5816.581.790.201 2
C: W6.2716.271.700.211 7
D: R44.27144.2712.030.003 4
AB0.3610.360.100.759 9
AC1.0811.080.290.595 4
AD0.0310.030.010.932 2
BC0.7610.760.210.656 8
BD1.1611.160.310.583 0
CD3.2813.280.890.360 1
A2231.711231.7162.94<0.000 1
B20.0910.090.030.874 6
C29.1019.102.470.136 8
D21.6611.660.450.512 0
残差55.22156.68
Table 5 Variance analysis result for Thv model
源项平方和自由度均方Fp
R2=0.999 2Radj2=0.998 5
模型4 350.7314310.772 691.35<0.000 1
A: Shv2.9512.9525.560.000 1
B: Slv2 111.6112 111.6118 287.34<0.000 1
C: W1 259.6111 259.6110 908.72<0.000 1
D: R784.801784.806 796.67<0.000 1
AB0.0510.050.430.520 2
AC0.4010.403.500.081 0
AD0.0210.020.160.696 3
BC16.97116.97146.93<0.000 1
BD7.6817.6866.53<0.000 1
CD8.8518.8576.61<0.000 1
A20.0010.000.01<0.000 1
B249.37149.37427.520.928 7
C227.53127.53238.40<0.000 1
D210.16110.1687.99<0.000 1
残差1.73150.011 55
Table 6 Variance analysis result for Tlv model
源项平方和自由度均方Fp
R2=0.999 0Radj2=0.998 0
模型10 274 873.5414733 919.542 105.40<0.000 1
A: Shv4 952 479.2614 952 479.261 4207.19<0.000 1
B: Slv2 456 285.4712 456 285.477 046.35<0.000 1
C: W693 143.361693 143.361 988.42<0.000 1
D: R1 399 514.1911 399 514.194014.79<0.000 1
AB561.161561.161.610.223 9
AC4 219.3014 219.3012.100.003 4
AD25 137.49125 137.4972.11<0.000 1
BC2 551.5012 551.507.320.016 3
BD9 200.7119 200.7126.390.000 1
CD148.241148.240.430.524 2
A2108 519.991108 519.99311.31<0.000 1
B264 169.26164 169.26184.08<0.000 1
C2218.471218.470.630.440 9
D22 786.3212 786.327.990.012 7
残差5 228.8515348.59
Table 7 Variance analysis result for S model
组号Shv/mm2Slv/mm2WR/mm

Fhv/

(107N/m3)

Flv/

(107N/m3)

111.68616.250.91145.7310.658.64
214.88877.250.79110.2710.077.37
312.00920.750.93136.409.575.31
419.36790.250.99151.334.704.83
517.76717.750.97100.937.898.34
612.96587.250.89106.5312.7312.14
713.28848.251.07102.809.896.57
818.72659.750.77143.876.517.90
914.24645.251.23104.678.507.91
1018.08833.751.25108.405.625.07
1119.68543.751.11142.004.867.38
1216.80674.250.71114.009.6910.50
1316.48906.251.15147.605.013.80
1415.20514.750.81134.539.3211.91
1517.12688.751.29140.134.895.03
1617.44891.750.75138.276.905.78
1713.92862.751.27125.206.694.51
1816.16630.751.03123.337.438.08
1920.00558.250.87117.736.8110.48
2015.52804.750.95130.807.396.07
2115.84601.751.05153.206.156.84
2220.32819.250.83115.876.316.71
2318.40935.251.01119.605.864.84
2419.04572.751.21112.135.778.06
2513.60529.251.19132.677.898.54
2612.32761.251.17149.477.395.02
2712.64732.250.73128.9312.099.05
2820.64746.751.13127.074.635.35
2914.56775.750.85155.077.496.01
3011.36703.251.09121.4710.837.42
Table 8 Calculation results of short-circuit electromagnetic force per unit volume of high/low voltage windings
源项平方和自由度均方Fp
R2=0.998 9Radj2=0.997 7
模型145.761410.411 814.95<0.000 1
A: Shv88.20188.201 5374.79<0.000 1
B: Slv3.8213.82666.19<0.000 1
C: W33.34133.345 812.32<0.000 1
D: R21.96121.963 827.54<0.000 1
AB0.2710.2746.56<0.000 1
AC1.6211.62282.58<0.000 1
AD0.8810.88153.69<0.000 1
BC0.1110.1119.190.000 5
BD0.0610.0610.910.004 8
CD0.3010.3051.54<0.000 1
A21.7511.75304.43<0.000 1
B20.0010.000.490.495 5
C20.1610.1627.69<0.000 1
D20.0910.0915.350.001 4
残差0.086150.0057
Table 9 Variance analysis result for Fhvmodel
源项平方和自由度均方Fp
R2=0.998 2Radj2=0.996 1
模型135.85149.701 177.90<0.000 1
A: Shv4.2714.27518.63<0.000 1
B: Slv78.33178.339 508.55<0.000 1
C: W35.05135.054 255.40<0.000 1
D: R20.60120.602 500.61<0.000 1
AB0.2410.2429.10<0.000 1
AC0.1410.1417.420.000 8
AD0.0510.056.440.022 7
BC1.8011.80218.78<0.000 1
BD0.7610.7692.17<0.000 1
CD0.3310.3340.41<0.000 1
A20.0110.010.840.374 3
B21.6211.62197.02<0.000 1
C20.1810.1822.440.000 3
D20.0910.0910.550.005 4
残差0.1236150.008 2
Table 10 Variance analysis result for Flv model
Fig.8 MOSFOA flowchart
参数数值
迭代数量/次500
种群规模/个200
平衡参数0.5
Pareto解集规模/个100
Table 11 MOSFOA parameter settings
Fig.9 Pareto frontier
Fig.10 Distribution nephogram of electric field of dry-type distribution transformer under optimized plan
性能指标数值下降幅度/%
初始方案优化方案
Thv/K89.1868.7522.91
Tlv/K55.8037.1833.37
S/W5 869.25 318.39.39
Fhv/(107N/m3)7.594.3442.82
Flv/(107N/m3)7.143.8945.52
Table 12 Comparison of results between optimized plan and initial plan
 
[[1]]   曹辰, 黄贺伟, 朱稀驰, 等. 多次短路冲击下变压器绕组累积效应及抗短路能力研究[J]. 电机与控制学报, 2025, 29(11): 44-55. doi:10.15938/j.emc.2025.11.005
CAO C, HUANG H W, ZHU X C, et al. Research on winding cumulative effect and short-circuit resistance of transformer under multiple short-circuit impacts[J]. Electric Machines and Control, 2025, 29(11): 44-55.
doi: 10.15938/j.emc.2025.11.005
[[2]]   马文嘉, 王丰华, 党晓婧. 基于稀疏自适应S变换的变压器短路冲击绕组状态声信号检测[J]. 电网技术, 2021, 45(9): 3755-3762. doi:10.13335/j.1000-3673.pst.2020.1545
MA W J, WANG F H, DANG X J. Acoustic signal detection of transformer winding under short-circuit impact based on sparse adaptive S-transform[J]. Power System Technology, 2021, 45(9): 3755-3762.
doi: 10.13335/j.1000-3673.pst.2020.1545
[[3]]   MENG T N, LI Y, LI P, et al. Research on distribution of winding leakage magnetic field of three-phase dry type transformer under short-circuit condition[J]. Energy Reports, 2023, 9: 1108-1115.
[[4]]   WANG Z H, LIU S L, ZHANG L. Analysis of vibration characteristics of dry-type transformer iron core and windings based on multi physical field[C]//2022 12th International Conference on Power and Energy Systems, Guangzhou, China, Nov. 23, 2022.
[[5]]   朱福寿. 树脂浇注干式变压器散热气道布置与成本优化探讨[J]. 变压器, 2010, 47(12): 24-25.
ZHU F S. Discussion on layout and cost optimization of heat dissipation air duct of resin cast dry-type transformer[J]. Transformer, 2010, 47(12): 24-25.
[[6]]   SMOLKA J, NOWAK A J. Shape optimization of coils and cooling ducts in dry-type transformers using computational fluid dynamics and genetic algorithm[J]. IEEE Transactions on Magnetics, 2011, 47(6): 1726-1731.
[[7]]   薛杨, 刘森, 吕杰, 等. 基于耦合场快速计算的核电厂主变压器数字孪生体搭建[J]. 工程设计学报, 2024, 31(6): 716-724.
XUE Y, LIU S, LÜ J, et al. Construction of digital twin of nuclear power plant main transformer based on fast calculation of coupling field[J]. Chinese Journal of Engineering Design, 2024, 31(6): 716-724.
[[8]]   YUAN S, ZHOU L J, CAI F L, et al. A lightweight optimization study for large capacity dry-type on-board traction transformer of EMU based on CFD and RSM[J]. IEEE Transactions on Transportation Electrification, 2023, 9(2): 2941-2954.
[[9]]   MA Y K, ZHOU X, WANG X K, et al. Study on the influence of winding height on the short-circuit withstand capability of 110 kV transformers[J]. Sensors, 2025, 25(21): 6528.
[[10]]   DAWOOD K, TURSUN S, KÖMÜRGÖZ G. Optimized gap positions for improved electromagnetic forces in dry-type transformer design[C]//2024 6th Global Power, Energy and Communication Conference, Budapest, Hungary, Jun. 4-7, 2024: 211-215.
[[11]]   LI W Y, WANG D Y, XIA Y Y, et al. Improvement of mechanical performance of large capacity dry-type on-board traction transformer under short-circuit electromagnetic force[J]. High Voltage, 2025, 10(6): 1558-1570.
[[12]]   DING Y W, YANG C X, XIONG B. Multi-objective optimal design of traction transformer using improved NSGA-II[C]//2021 24th International Conference on Electrical Machines and Systems, Gyeongju, Korea, 2021: 1470-1474.
[[13]]   ZHOU C D, ZHAO J H, WANG H M, et al. Multi-objective optimization design of linear phase shift transformer based on improved differential evolutionary algorithm[J]. IEEJ Transactions on Electrical and Electronic Engineering, 2023, 18(9): 1439-1450.
[[14]]   KALANTARI KHANDANI M, ASKARZADEH A. Optimal MV/LV transformer allocation in distribution network for power losses reduction and cost minimization: a new multi-objective framework[J]. International Transactions on Electrical Energy Systems, 2020, 30(6): e12361.
[[15]]   许桐, 钟昌廷, 李斯嘉, 等. 基于成功历史参数自适应海星优化算法的多目标桁架结构优化设计[J]. 计算力学学报, 2026, 43(1): 1-8.
XU T, ZHONG C T, LI S J, et al. Success history parameter self-adaptive starfish optimization algorithm for multi-objective truss structural optimization design[J]. Chinese Journal of Computational Mechanics, 2026, 43(1): 1-8.
[[16]]   王巧媛, 贾少锋, 梁得亮, 等. 基于多物理场仿真的低频变压器温度特性分析及导向结构优化[J]. 高电压技术, 2024, 50(5): 2009-2019.
WANG Q Y, JIA S F, LIANG D L, et al. Temperature characteristics analysis and guide structure optimization of low-frequency transformer based on multi-physics field simulation[J]. High Voltage Engineering, 2024, 50(5): 2009-2019.
[[17]]   AGHAJANZADEH I, RAMEZANIANPOUR A M, AMANI A, et al. Mixture optimization of alkali activated slag concrete containing recycled concrete aggregates and silica fume using response surface method[J]. Construction and Building Materials, 2024, 425: 135928.
[[18]]   中国轻工业联合会. 500 kV变压器匝间绝缘纸: [S]. 北京: 中国轻工业出版社有限公司, 2012.
China National Light Industry Council. 500 kV transformer coil winding paper: [S]. Beijing: China Light Industry Press Ltd., 2012.
[[19]]   SOMADASAN S, SUBRAMANIYAN G, ATHISAYARAJ M S, et al. Central composite design: an optimization tool for developing pharmaceutical formulations[J]. Journal of Young Pharmacists, 2024, 16(3): 400-409.
[[20]]   ZHONG C T, LI G, MENG Z, et al. Starfish optimization algorithm (SFOA): a bio-inspired metaheuristic algorithm for global optimization compared with 100 optimizers[J]. Neural Computing and Applications, 2025, 37(5): 3641-3683.
[1] Guang'ao LIU,Yinglong CHEN,Changmin LUO,Bo YAN,Fei GAO. Optimization design and experimental study of gas control valve with low torque[J]. Chinese Journal of Engineering Design, 2026, 33(1): 117-129.
[2] Jinyun CAI,Zhong LIU,Gang WANG,Qingbin ZHAO,Ning AN,Xuwei DU,Dongliang LI,Yuanzhou LI. Optimization design of auxiliary tail rope pulling device for winch mill based on response surface methodology[J]. Chinese Journal of Engineering Design, 2024, 31(2): 178-187.
[3] YANG Shi-xiang, LI Wen-qiang. Innovation design of sealing structure of incineration ash treatment equipment[J]. Chinese Journal of Engineering Design, 2021, 28(6): 679-686.
[4] ZHANG Ze, CHEN Yong, LI Guang-xin, LEI Yong-gan, RUAN Ou, WANG Zai-zhou. Design and structure optimization of electro-hydraulic control system assembly of electric vehicle transmission[J]. Chinese Journal of Engineering Design, 2021, 28(3): 335-343.
[5] ZHANG Qing-Chun, QING Qi-Xiang, LIU Jie. The cylindrical teeth's topology optimization design of reverse circulation DTH bit based on gas-solid two-phase flow[J]. Chinese Journal of Engineering Design, 2014, 21(4): 340-347.
[6] ZHANG Can, LIU Feng, LI Li-Juan. Improved bacterial foraging algorithm and its application to design of frame structures[J]. Chinese Journal of Engineering Design, 2012, 19(6): 422-427.
[7] LIU Shun-Yao, LI Jin, HE Hao. Topological optimization of the main girder web based on ESO[J]. Chinese Journal of Engineering Design, 2011, 18(3): 174-177.
[8] LIU Feng, QIN Guang, LI Li-Juan. A quick group search optimizer with passive congregation and its application in spatial structures[J]. Chinese Journal of Engineering Design, 2010, 17(6): 420-425.
[9] WANG Xiao-yan, YANG Zhi-qiang, WENG Meng-chao. Optimization design of nonlinear steel frame based on improved adaptive genetic algorithm[J]. Chinese Journal of Engineering Design, 2009, 16(2): 103-107.
[10] LUO Ming-jun,WANG Wen-lin. Finite element analysis and correlative structural optimization on framework of double-wishbone independent suspension with twin torsion bar[J]. Chinese Journal of Engineering Design, 2009, 16(2): 108-111.
[11] SUN Zhi-Juan,MENG Qing-Xin,WU Jian-Rong. Check ring joint structure design in underwater skirt pile gripper based on body contact nonlinear finite element analysis[J]. Chinese Journal of Engineering Design, 2007, 14(2): 117-120.
[12] HU Ying-Chun, LI Shang-Ping, PEI Shao-Shuai, CHEN Shu-Xun. Research on new and highly-effective methodology for structure optimization[J]. Chinese Journal of Engineering Design, 2007, 14(1): 1-5.
[13] LIAO Xing-Tao, ZHANG Wei-Gang, LI Qing, ZHONG Zhi-Hua. Application research of response surface methodology in thin-walled structural optimization design in crashworthiness[J]. Chinese Journal of Engineering Design, 2006, 13(5): 298-302.
[14] CHEN Shu-Xun, WANG Su-Nuan. New linear iterative algorithm and its application in seeking solution of criterion equations of structural optimization[J]. Chinese Journal of Engineering Design, 2005, 12(5): 270-272.
[15] XUE Cai-Jun, QIU Qing-Ying, YANG Wei. Research on collaborative analysis m ethod of static and dynam ic perform ance of m echanical structure[J]. Chinese Journal of Engineering Design, 2002, 9(4): 183-186.