[1]魏周华①,王清华①,何锋军①,等.基于深度学习算法的硝酸铵溶液析晶点检测系统[J].爆破器材,2025,54(01):36-40.[doi:10.3969/j.issn.1001-8352.2025.01.006]
　WEI Zhouhua,WANG Qinghua,HE Fengjun,et al.A Detection System for Crystallization Points of Ammonium Nitrate Solution Based on Deep Learning Algorithm[J].EXPLOSIVE MATERIALS,2025,54(01):36-40.[doi:10.3969/j.issn.1001-8352.2025.01.006]
点击复制

基于深度学习算法的硝酸铵溶液析晶点检测系统()
分享到：

《爆破器材》[ISSN:1001-8352/CN:32-1163/TJ]

卷:
54
期数:
2025年01
页码:
36-40
栏目:
爆炸材料
出版日期:
2025-01-09

文章信息/Info

Title:
A Detection System for Crystallization Points of Ammonium Nitrate Solution Based on Deep Learning Algorithm
文章编号:
5954
作者:
魏周华①; 王清华①; 何锋军①; 党创刚①; 田璐②; 孙伟博②
①陕西北方民爆集团有限公司（陕西西安，715600）
②西安科技大学能源学院（陕西西安，710054）
Author(s):
WEI Zhouhua①;  WANG Qinghua①;  HE Fengjun①;  DANG Chuanggang①;  TIAN Lu②;  SUN Weibo②
①Shaanxi Beifang Civil Explosive Group Co., Ltd. (Shaanxi Xi’an, 715600)
②College of Energy Engineering, Xi’an University of Science and Technology (Shaanxi Xi’an, 710054)
关键词:
硝酸铵水相溶液; 深度学习; 析晶点; EfficientNet算法
Keywords:
aqueous solution of ammonium nitrate;  deep learning;  crystallization point;  EfficientNet algorithm
分类号:
TQ560.7
DOI:
10.3969/j.issn.1001-8352.2025.01.006
文献标志码:
A
摘要:
为解决膨化硝铵生产线水相溶液质量自动检测的问题，根据生产线现场条件，设计了水相溶液质量自动检测系统。通过研究不同深度学习算法对硝酸铵析晶状态判定的准确度发现，EfficientNet算法的准确度最高。对EfficientNet算法进行改良，上调每层特征通道数，在深度上删去了多个MBConv层，减小参数量，降低FLOPs，加速检测，使系统自动测量的析晶点温度与人工测量的平均误差小于0.3 ℃。结果表明：系统可准确测量硝酸铵水相溶液的析晶温度及密度,并自动生成硝酸铵水相溶液检测报告;同时，实现数据的追溯和查询，并对异常数据进行标记，满足生产需要。
Abstract:
An automatic quality detection system for expanded ammonium nitrate aqueous solution was designed based on the on-site conditions of the production line. The accuracy of different deep learning algorithms in determining the crystallization state of ammonium nitrate solution was studied. EfficiencyNet algorithm exhibited the highest accuracy. EfficiencyNet algorithm was improved by increasing the number of feature channels in each layer and removing multiple MBConv layers in depth. The numbers of parameters were reduced, FLOPs were lowered, and the detection were accelerated. The average error between the automatically measured crystallization point temperature by the system and the manually measured temperature is less than 0.3 ℃. The results demonstrate that the system can accurately measure the crystallization temperature and density of ammonium nitrate solution, and automatically generate detection reports for ammonium nitrate solution. At the same time, it can trace and query data, and mark abnormal data, thereby meeting production requirements.

参考文献/References:

［1］陆明, 吕春绪, 刘祖亮. 硝酸铵的膨化机理研究［J］.兵工学报, 2002, 23(1): 30-34.
LU M，Lǔ C X，LIU Z L. A study on the expansion mechanism of ammonium nitrate［J］.Acta Armamentarii, 2002, 23(1): 30-34.
［2］黄麟, 赵旭东, 席韬, 等.现场混装乳化炸药中低温制备及性能表征［J］.工程爆破, 2023, 29(4): 143-148.
HUANG L, ZHAO X D, XI T, et al. Mid-low temperature preparation and properties characterization of onsite mixed emulsion explosive ［J］. Engineering Blasting, 2023, 29(4): 143-148.
［3］徐尊, 吴红波, 缪志军, 等.析晶点对乳化炸药稳定性的影响研究［J］.广东化工, 2015, 42(8): 66, 60.
XU Z, WU H B, MIU Z J, et al. Effect of crystallization point on the stability of emulsion explsoives［J］.Guangdong Chemical Industry, 2015, 42(8): 66, 60.
［4］王枚, 迟平, 吴攀宇, 等.硝酸铵水溶液析晶点在线无人化测量方法的研究［J］.火工品, 2023(4): 71-75.
WANG M, CHI P, WU P Y, et al. Study on online unmanned measurement method of crystallization point of ammonium nitrate aqueous solution ［J］.Initiators & Pyrotechnics, 2023(4): 71-75.
［5］刘万义, 耿图禹, 徐森, 等.硝酸铵溶液析晶点测量方法研究［J］.爆破器材, 2021, 50(3): 60-64.
LIU W Y, GENG T Y, XU S, et al. Study on measurement method of crystallization point of ammonium nitrate solution ［J］.Explosive Materials, 2021, 50(3): 60-64.
［6］崔雪峰, 张日强, 孟祥宇, 等.基于Arduino的硝酸铵溶液析晶点温度测量系统的研究［J］.现代矿业, 2020(12): 165-167, 173.
［7］刘德晶.W/O法在乳化炸药水相析晶点测试中的应用［J］.广东化工, 2018, 45(3): 38-39.
LIU D J.Application of W/O method in the determination of crystallization point of emulsion explosive ［J］.Guangdong Chemical Industry, 2018, 45(3): 38-39.
［8］张东杰.乳化炸药用液体硝酸铵技术指标在线检测系统［J］.火工品, 2018(4): 45-48.
ZHANG D J. Research on online detection system for technical indicators of liquid ammonium nitrate for emulsion explosives ［J］.Initiators & Pyrotechnics, 2018(4): 45-48.
［9］傅杨攀, 刘勇健, 陈贡发, 等.基于CSMR和卷积神经网络的岩质边坡稳定性分析［J］.自然灾害学报, 2023, 32(1): 114-121.
FU Y P, LIU Y J, CHEN G F, et al.Stability analysis of rock slope based on CSMR and convolution neural network ［J］.Journal of Natural Disasters, 2023, 32(1): 114-121.
［10］王潇怡, 罗运华, 喻忠军, 等.基于EfficientNet的多通道雷达目标微动特征分类方法［J］.系统工程与电子技术, 2024,46(9): 1-14.
WANG X Y, LUO Y H, YU Z J, et al.Multi-channel radar target micro-motion feature classification method based on EfficientNet ［J］.Journal of Systems Engineering and Electronics, 2024, 46(9): 1-14.
［11］周子云, 黄洪. 改进EfficientNet图像分类的恶意流量检测模型［J］.四川轻化工大学学报(自然科学版), 2023, 36(6): 49-56.
ZHOU Z Y, HUANG H. Improved malicious traffic detection model for efficientnet image classification［J］.Journal of Sichuan University of Science & Engineering（Natural Science Edition）, 2023, 36(6): 49-56.

备注/Memo

备注/Memo:
收稿日期：2024-05-23
第一作者：魏周华（1990—），男，工程师，从事工业炸药生产及工程爆破方面的研究。E-mail：574366916@qq.com
通信作者：孙伟博（1979—），男，博士，副教授，从事爆破技术及爆破装备方面的研究。E-mail：sunweibo@xust.edu.cn
更新日期/Last Update: 2025-01-10