文章摘要
铝箔热密封机理和缺陷识别研究
Study on Heat Sealing Mechanism and Defect Recognition of Aluminum Foil
投稿时间:2019-05-24  修订日期:2020-01-15
DOI:
中文关键词: 密封性检测  数值模拟  温度场  Gabor小波  ELM神经网络
英文关键词: feature extraction  numerical simulation  temperature field  Gabor wavelet  ELM neural network
基金项目:
作者单位E-mail
李维军 辽宁石油化工大学 3051379@qq.com 
辛传奇 辽宁石油化工大学 1547332581@qq.com 
张学伟 辽宁石油化工大学  
白璐 辽宁石油化工大学  
摘要点击次数: 29
全文下载次数: 0
中文摘要:
      对铝箔密封受热温度场的机理进行研究并进行密封缺陷识别。首先研究了铝箔密封对电磁加热过程中传热特性进行数值模拟分析,得到不同类型铝箔密封情况的温度场分布特性。然后进行了不同铝箔密封情况的红外热图像缺陷检测实验,得到相应的热像图,与模拟计算得到的表面温度场分布高度吻合,为了加快识别热图像的速度,建立了基于Gabor小波与ELM极限学习机对热图像纹理特征提取并识别的检测铝箔密封性的方法,在相同条件下对比上述检测方法与基于BP神经网络对热像图进行识别方法的训练时间与检测精度。验证了基于Gabor小波和ELM神经网络的算法具有识别速度快、精度高的特点。
英文摘要:
      The mechanism of the heat temperature field of aluminum foil seal is studied and the sealing defect is identified. Firstly, the aluminum foil sealing is studied by numerical simulation analysis of the heat transfer characteristics during electromagnetic heating, and the temperature field distribution character-istics of different types of aluminum foil sealing are obtained. Then the infrared thermal image defect detection experiment of different aluminum foil sealing conditions is carried out, and the corresponding thermal image is obtained, which is highly consistent with the simulated surface temperature field dis-tribution. In order to speed up the recognition of thermal images, a method for detecting the sealing of aluminum foil based on Gabor wavelet and ELM extreme learning machine for thermal image texture feature was established. The above detection method and BP neural network based thermal image were compared under the same conditions. The figure performs the training time and the detection accuracy of the recognition method. The algorithm based on Gabor wavelet and ELM neural network is proved to have fast recognition speed and high precision.
View Fulltext   查看/发表评论  下载PDF阅读器
关闭