包装与食品机械2025，Vol.43Issue(3)：96-104,9.DOI:10.3969/j.issn.1005-1295.2025.03.011
基于长短期记忆网络的仿生锤头压力曲线优化
Optimization of biomimetic hammer head pressure curve based on long short-term memory network
摘要
Abstract
To address the inferior surface slurry-enhancing effect and moisture uniformity of traditional qu-pressing machines compared to manual stepping,this study proposes a biomimetic pressure curve optimization method based on long short-term memory(LSTM)networks.Dynamic mechanical data from manual stepping were collected using flexible plantar pressure sensors.An LSTM model performed time-series prediction to generate standardized optimized curves,applied to multi-station biomimetic hammer heads.Experimental results demonstrated that after applying the optimized pressure curve,the moisture uniformity factor of qu blocks increased from 5.91 to 6.35 and the slurry-coverage area ratio rose from 76.85%to 82.36%,significantly improving both moisture distribution and slurry-enhancing effects.This research provides a data-driven standardized solution for intelligent mechanized qu-making.关键词
长短期记忆网络/仿生压曲/压力曲线优化Key words
long short-term memory network/biomimetic qu-pressing/pressure curve optimization分类
轻工纺织引用本文复制引用
徐永森,徐雪萌,刘权,李颍鹏,吴人鸿,张汉山..基于长短期记忆网络的仿生锤头压力曲线优化[J].包装与食品机械,2025,43(3):96-104,9.基金项目
山东省泰安市科技创新重大专项(2021ZDZX015) （2021ZDZX015）
