Temperature Field Prediction of Melt-cast Explosives Based on a B-spline Neural Network

doi:10.12382/bgxb.2022.0086

Abstract

Abstract:

The temperature field distribution and evolution inside the mold play a crucial role in determining the casting quality of melt-cast explosive processes. A fast prediction model is developed based on a B-spline neural network for the transient temperature field in a melt-cast explosive process with a water/oil bath. The model is created by first obtaining temperature evolution data samples under different processing conditions through orthogonal numerical experiments. The B-spline neural network is then trained on these data samples to establish a prediction model that represents the relationship between temperature-control parameters and the temperature field inside the grain. This model enables rapid and accurate prediction of the temperature field and solidification front, providing an efficient prediction method for parameter optimization and online control of melt-cast explosive processes. This study serves as a valuable reference for predicting other physical fields in the intelligent development of similar processes in the future.

Key words: melt-cast explosive, B-spline neural network, water/oil bath process, temperature field

TAO Lei, LIU Jianhua, XIA Huanxiong, AO Xiaohui, GAO Feng. Temperature Field Prediction of Melt-cast Explosives Based on a B-spline Neural Network[J]. Acta Armamentarii, 2023, 44(5): 1339-1349.

Figures/Tables 13

References 32

[1]	田勇. 炸药熔铸成型过程监测评价及数值模拟研究[D]. 北京: 中国科学院研究生院, 2009.
	TIAN Y. Process monitoring/evaluation and numerical simulation during casting explosive solidification[D]. Beijing: Graduate University of Chinese Academy of Sciences, 2009. (in Chinese)
[2]	谢志毅, 王冬磊, 刘建波, 等. 炸药熔铸过程数值模拟研究[J]. 太赫兹科学与电子信息学报, 2014, 12(4):617-621.
	XIE Z Y, WANG D L, LIU J B, et al. Numerical simulation of melt-cast explosive solidification process[J]. Journal of Terahertz Science and Electronic Information Technology, 2014, 12(4):617-621. (in Chinese)
[3]	焦云多, 曾晓华, 陈洋, 等. 大当量DNAN基熔铸炸药装药质量控制方法[J]. 兵工自动化, 2019, 38(8):23-26.
	JIAO Y D, ZENG X H, CHEN Y, et al. Method of charge quality control for large equivalent DNAN-base melt cast explosive[J]. Ordnance Industry Automation, 2019, 38(8):23-26. (in Chinese)
[4]	蒙君煚, 周霖, 金大勇, 等. 成型工艺对2,4-二硝基苯甲醚基熔铸炸药装药质量的影响[J]. 兵工学报, 2018, 39(9): 1719-1726. doi: 10.3969/j.issn.1000-1093.2018.09.007
	MENG J J, ZHOU L, JIN D Y, et al. Effect of forming process on casting quality of 2,4-dinitroanisole-based casting explosive[J]. Acta Armamentarii, 2018,39(9):1719-1726.[5] JI C, LIN C S.The solidification process of melt casting explosives in shell[J]. Propellants,Explosives,Pyrotechnics, 1998, 23:137-141. (in Chinese)
[5]	JI C, LIN C S, RAPPAZ M. The solidification process of melt casting explosives in shell[J]. Propellants,Explosives,Pyrotechnics, 1998, 23:137-141. doi: 10.1002/(ISSN)1521-4087 URL
[6]	PEQUET C, GREMAUD M, RAPPAZ M. Modeling of microporosity,macroporosity and pipe-shrinkage formation during the solidification of alloys using a mushy-zone refinement method: applications to aluminum alloys[J]. Metallurgical and Materials Transactions A, 2002, 33(7):2095-2106. doi: 10.1007/s11661-002-0041-5 URL
[7]	郭朋林, 罗观, 习彦, 等. TNT基熔铸炸药冷却过程温度场变化规律研究[J]. 含能材料, 2010, 18(1):93-96.
	GUO P L, LUO G, XI Y, et al. Study on temperature field variation during cooling process of TNT-based melt-cast explosive[J]. Chinese Journal of Energetic Materials, 2010, 18(1): 93-96. (in Chinese)
[8]	马松, 袁俊明, 刘玉存, 等. DNAN炸药凝固过程的实验与数值模拟[J]. 含能材料, 2014, 22(2):240-244.
	MA S, YUAN J M, LIU Y C, et al. Experiment and numerical simulation of DNAN solidification process[J]. Chinese Journal of Energetic Materials, 2014, 22(2):240-244. (in Chinese)
[9]	胡菲, 刘玉存, 袁俊明, 等. 炸药熔铸装药成型质量监测技术研究进展[J]. 兵器装备工程学报, 2021, 42(9):1-6.
	HU F, LIU Y C, YUAN J M, et al. Review on monitoring technology for quality of casting explosive solidification[J]. Journal of Sichuan Ordnance, 2021, 42(9):1-6. (in Chinese)
[10]	马松, 袁俊明, 刘玉存, 等. 水浴和自然条件下熔铸炸药凝固过程的数值计算[J]. 火炸药学报, 2013, 36(3): 54-58.
	MA S, YUAN J M, LIU Y C, et al. Numerical calculation of solidification process of melt-cast explosive in water bath and natural conditions[J]. Chinese Journal of Explosives and Propellants, 2013, 36(3):54-58. (in Chinese)
[11]	刘威, 曹红松, 张虎成, 等. 分层浇铸工艺对熔铸药柱缩孔的影响研究[J]. 机电技术, 2015(6):122-125.
	LIU W, CAO H S, ZHANG H C, et al. Study of the influence of layer casting process on shrinkage cavity of melt-cast explosive[J]. Mechanical and Electrical Technology, 2015(6):122-125. (in Chinese)
[12]	MENG J J, ZHOU L, ZHANG X R. Effect of pressure of the casting vessel on the solidification characteristics of a DNAN/RDX melt-cast explosive[J]. Journal of Energetic Materials, 2017, 35(4): 385-396.
[13]	岳晓媛, 张会锁, 韩雪莲, 等. 大长径比熔铸装药热芯棒凝固工艺优化仿真[J]. 高压物理学报, 2021, 35(1):1-8.
	YUE X Y, ZHANG H S, HAN X L, et al. Optimization simulation of hot mandrel solidification process of large length-diameter ratio melt-cast explosive[J]. Chinese Journal of High Pressure Physics, 2021, 35(1):1-8. (in Chinese)
[14]	张明明, 张斌宏, 万大奎, 等. 熔铸装药凝固过程温度场变化分析[J]. 弹药先进制造技术, 2021, 40(6): 93-96.
	ZHANG M M, ZHANG B H, WANG D K, et al. Study on temperature field changes during casting explosive solidification process[J]. Ordnance Industry Automation, 2021, 40(6): 93-96. (in Chinese)
[15]	MENG L B, MCWILLIAMS B, JAROSINSKI W, et al. Machine learning in additive manufacturing:a review[J]. The Minerals,Metals and Materials Society, 2020, 72(6): 2363-2377.
[16]	WANG C, TAN X P, TOR S B, et al. Machine learning in additive manufacturing:state-of-the-art and perspectives[J]. Additive Manufacturing, 2020, 36:101538. doi: 10.1016/j.addma.2020.101538 URL
[17]	郝永志, 赵海东, 林嘉华. 基于机器学习的挤压铸造铝合金力学性能预测[J]. 特种铸造及有色合金, 2019, 39(8): 859-862. doi: 10.15980/j.tzzz.2019.08.012
	HAO Y Z, ZHAO H D, LIN J H. Prediction of mechanical properties of squeeze-cast aluminum alloys based on machine learning[J]. Special Casting and Nonferrous Alloys, 2019, 39(8):859-862. (in Chinese)
[18]	SHAHANE S, ALURU N, FERREIRA P, et al. Optimization of solidification in die casting using numerical simulations and machine learning[J]. Journal of Manufacturing Processes, 2020, 51:130-141. doi: 10.1016/j.jmapro.2020.01.016 URL
[19]	QI X B, CHEN G F, LI Y, et al. Applying neural-network-based machine learning to additive manufacturing: current applications,challenges,and future perspectives[J]. Engineering, 2019, 5(4):721-729. doi: 10.1016/j.eng.2019.04.012 URL
[20]	JIN N, LIU D R. Wavelet basis function neural networks for sequential learning[J]. IEEE Transactions on Neural Networks, 2008, 19(3):523-528. doi: 10.1109/TNN.2007.911749 pmid: 18334370
[21]	GÁLVEZ A, IGLESIAS A, PUIG J. Iterative two-step genetic-algorithm-based method for efficient polynomial B-spline surface reconstruction[J]. Information Sciences, 2012, 182: 56-76. doi: 10.1016/j.ins.2010.09.031 URL
[22]	COELHO L, PESSOA M. Nonlinear identification using a B-spline neural network and chaotic immune approaches[J]. Mechanical Systems and Signal Processing, 2009, 23:2418-2434. doi: 10.1016/j.ymssp.2009.01.013 URL
[23]	WANG J J, CHEN B L, YANG C J. Approximation of algebraic and trigonometric polynomials by feedforward neural networks[J]. Neural Computing and Applications, 2012, 21:73-80. doi: 10.1007/s00521-011-0617-3 URL
[24]	DOKUR Z, OLMEZ T. Heartbeat classification by using a convolutional neural network trained with Walsh functions[J]. Neural Computing and Applications, 2020, 32: 12515-12534. doi: 10.1007/s00521-020-04709-w
[25]	HAHM N, HONG B. The capability of localized neural network approximation[J]. Honam Mathematical, 2013, 35(4):729-738.
[26]	徐应祥. 一种新型B-样条拟插值[J]. 仲恺农业工程学院学报, 2021, 34(4):60-65.
	XU Y X. A new B-spline quasi-interpolation[J]. Journal of ZhongKai University of Agriculture and Technology, 2021, 34(4):60-65. (in Chinese)
[27]	HONG X, IPLIKCI S, CHEN S, et al. A model-based PID controller for Hammerstein systems using B-spline neural networks[J]. International Journal of Adaptive Control and Signal Processing, 2014, 28:412-428. doi: 10.1002/acs.v28.3-5 URL
[28]	HONG X, CHEN S. The system identification and control of Hammerstein system using non-uniform rational B-spline neural network and particle swarm optimization[J]. Neurocomputing, 2012, 82:216-223. doi: 10.1016/j.neucom.2011.11.016 URL
[29]	SETIAWAN I, PRIYADI A, MIYAUCHI H, et al. Adaptive B-spline neural network-based vector control for a grid side converter in wind turbine-DFIG systems[J]. IEEE Transactions on Electrical and Electronic Engineering, 2015, 10:674-682.
[30]	WU Z Q, JIA W J, ZHAO L R, et al. Short-term wind speed forecast based on B-spline neural network optimized by PSO[J]. Mathematical Problems in Engineering, 2015:278635.
[31]	谢心和, 赵东明, 吕明昊, 等. B样条神经网络用于区域重力异常曲面拟合[C]∥第十六届国家安全地球物理专题研讨会论文集. 北京:中国地球物理学会, 西安: 陕西省地球物理学会, 2020:39-44.
	XIE X H, ZHAO D M, LÜ M H, et al. Surface fitting of regional gravity anomaly based on B-spline neural network[C]∥Proceedings of the 16th National Security Geophysics Symposium.Beijing:Chinese Geophysical Society, Xi’an: Shaanxi Geophysical Society, 2020:39-44. (in Chinese)
[32]	CHITRA K, VENNILA C. A novel patch selection technique in ANN B-Spline Bayesian hyperprior interpolation VLSI architecture using fuzzy logic for highspeed satellite image processing[J]. Journal of Ambient Intelligence and Humanized Computing, 2020, 12(6):6491-6504. doi: 10.1007/s12652-020-02264-9

项目	炸药	模具	冒口
材料	DNAN基炸药	Steel	铝合金
导热系数λ/ (W·m^-1K^-1)	0.72(20℃)	50	176
	0.70(65℃)
	0.64(110℃)
密度ρ/(kg·m^-3)	1858(20℃)	7940	2700
	1810(65℃)
	1700(110℃)
比热容c_p/(J·kg^-1K^-1)	1130(20℃)	450	900
	1320(65℃)
	1450(110℃)
固相线温度T_sol/℃	85
液相线温度T_liq/℃	90
浇注温度T_j/℃	100
与外界的换热系数h₁/ (W·m^-2K^-1)	20	300	200
与炸药的换热系数h₂/ (W·m^-2K^-1)		200	100
环境温度/℃	T₀

项目	炸药	模具	冒口
材料	DNAN基炸药	Steel	铝合金
导热系数λ/ (W·m^-1K^-1)	0.72(20℃)	50	176
	0.70(65℃)
	0.64(110℃)
密度ρ/(kg·m^-3)	1858(20℃)	7940	2700
	1810(65℃)
	1700(110℃)
比热容c_p/(J·kg^-1K^-1)	1130(20℃)	450	900
	1320(65℃)
	1450(110℃)
固相线温度T_sol/℃	85
液相线温度T_liq/℃	90
浇注温度T_j/℃	100
与外界的换热系数h₁/ (W·m^-2K^-1)	20	300	200
与炸药的换热系数h₂/ (W·m^-2K^-1)		200	100
环境温度/℃	T₀

试验组	水浴初始温度 T₁₀/℃	水浴降温速率v₁/ (℃·min^-1)	油浴初始温度 T₂₀/℃	油浴保温时间 t₂/h	油浴降温速率v₂/ (℃·min^-1)
1	90	0.15	95	4	0.5
2	90	0.2	100	5	0.6
3	90	0.25	105	6	0.7
4	95	0.15	95	5	0.6
5	95	0.2	100	6	0.7
6	95	0.25	105	4	0.5
7	100	0.15	100	4	0.7
8	100	0.2	105	5	0.5
9	100	0.25	95	6	0.6
10	90	0.15	105	6	0.6
11	90	0.2	95	4	0.7
12	90	0.25	100	5	0.5
13	95	0.15	100	6	0.5
14	95	0.2	105	4	0.6
15	95	0.25	95	5	0.7
16	100	0.15	105	5	0.7
17	100	0.2	95	6	0.5
18	100	0.25	100	4	0.6
19	92	0.175	102.5	5.5	0.55
20	95	0.225	100	5	0.6
21	98	0.2	97.5	4.5	0.65

试验组	水浴初始温度 T₁₀/℃	水浴降温速率v₁/ (℃·min^-1)	油浴初始温度 T₂₀/℃	油浴保温时间 t₂/h	油浴降温速率v₂/ (℃·min^-1)
1	90	0.15	95	4	0.5
2	90	0.2	100	5	0.6
3	90	0.25	105	6	0.7
4	95	0.15	95	5	0.6
5	95	0.2	100	6	0.7
6	95	0.25	105	4	0.5
7	100	0.15	100	4	0.7
8	100	0.2	105	5	0.5
9	100	0.25	95	6	0.6
10	90	0.15	105	6	0.6
11	90	0.2	95	4	0.7
12	90	0.25	100	5	0.5
13	95	0.15	100	6	0.5
14	95	0.2	105	4	0.6
15	95	0.25	95	5	0.7
16	100	0.15	105	5	0.7
17	100	0.2	95	6	0.5
18	100	0.25	100	4	0.6
19	92	0.175	102.5	5.5	0.55
20	95	0.225	100	5	0.6
21	98	0.2	97.5	4.5	0.65

测试组	参数	时间/s
测试组	参数	1000	6000	8000	10000	15000	25000
	μ	0.0778	-0.0921	0.0279	0.0578	0.0410	0.0704
19	σ	0.2202	0.4657	0.5285	0.3485	0.3903	0.3197
	μ_abs	0.1852	0.3434	0.3842	0.2260	0.2459	0.2070
	ΔT_max	0.5171	1.4519	1.6954	1.3802	1.3620	1.2470
	μ	0.1737	-0.0025	-0.0703	-0.0063	0.0219	-0.0286
20	σ	0.2219	0.6407	0.4891	0.3089	0.4052	0.1813
	μ_abs	0.2505	0.4450	0.3360	0.2283	0.2699	0.1153
	ΔT_max	0.5154	2.1489	2.6952	0.9250	1.5040	0.6326
	μ	-0.0060	0.0074	0.1287	-0.0456	0.0008	0.0145
21	σ	0.1800	0.3327	0.5125	0.3314	0.3535	0.1936
	μ_abs	0.1297	0.2368	0.3699	0.2586	0.2397	0.1574
	ΔT_max	0.5051	1.1007	1.8330	1.2420	1.2572	0.5140