Package Leak Detection Based on Aperiodic Anomaly Extraction
Package Leak Detection Based on Aperiodic Anomaly Extraction
复制标题
DOI：
10.1109/jestie.2021.3056692
复制
发表时间：
2021-07
期刊：
影响因子：
--
通讯作者：
Hisayoshi Muramatsu;Yuta Kakimi;Issei Takeuchi;S. Katsura
中科院分区：
文献类型：
--
作者：
Hisayoshi Muramatsu;Yuta Kakimi;Issei Takeuchi;S. Katsura
食品对我们的社会来说是必不可少的，它们在包装后运输和交付，以避免污染和新鲜度的变质。然而，在自动包装过程中存在包裹泄漏问题，可能会损害人类健康。为了保护人体健康，保证工业自动化食品生产的可靠性，对所有包装进行自动检漏是必要的。本文开发了一种包泄漏测试仪，并提出了一种利用该测试仪进行非周期异常测试的方法，实现了包泄漏的自动检测。所开发的封装泄漏测试仪实现了精确的反作用力测量，从而实现了准确的泄漏检测，而基于泄漏封装模型的利用反作用力的非周期导数的非周期异常测试实现了对扰动的稳健泄漏检测。通过对388个实验包装数据随机产生的10000个包装数据进行测试，验证了包装泄漏测试仪和非周期异常测试的有效性。
Foods are essential for our society, and they are transported and delivered after packaging to avoid contamination and deterioration of freshness. Nevertheless, there is a package leak problem in automatic packaging processes that would risk impairing human health. In order to protect human health and guarantee the reliability of the industrial automatic food production, automatic leak testing is necessary for all packages. This article developed a package leak tester and proposed an aperiodic anomaly test using the tester to detect the package leaks automatically. The developed package leak tester realized accurate reaction force measurement enabling accurate leak detection, and the proposed aperiodic anomaly test using an aperiodic derivative of the reaction force based on a leaking package model achieved robust leak detection against disturbances. The package leak tester and the aperiodic anomaly test were validated by testing 10 000 package data randomly generated by 388 experimental package data.