基于LSTM算法的前馈控制系统扰动预测研究
Study on disturbance prediction of forging press feedforward control system based on LSTM algorithm
姚红伟 1尹静文1
作者信息
- 1. 新乡职业技术学院 智能制造学院,河南 新乡 453000
- 折叠
摘要
为了提高机械自动化系统前馈控制精度,设计了一种基于长短期记忆LSTM循环神经网络的前馈控制系统扰动预测模型,并开展实验测试分析.研究结果表明:预测结果中较大间隔下形成的扰动信号序列均方根误差更低,具备更高精度;小间隔时间尺度下形成了更低峭度的扰动信号序列.经过组合预测序列结果更加趋近高占比的序列,获得更小负峭度与序列波动,降低均方根误差,达到更精确的效果.
Abstract
In order to improve the accuracy of feedforward control of mechanical automation system, a distur-bance prediction model of feedforward control system based on LSTM recurrent neural network is designed and tested. The results show that the root-mean-square error of the disturbance signal sequence formed at a larger interval is lower and has higher precision. Lower kurtosis disturbance signal sequences are formed at small interval time scales. By combining the predicted sequences, the result is closer to the sequence with high proportion, smaller negative kurtosis and sequence fluctuation are obtained, and the root-mean-square error is reduced to achieve more accurate results.
关键词
过程控制/前馈系统/扰动预测/LSTM/多时间尺度/峭度Key words
Process control/Feed forward system/Disturbance prediction/LSTM/Multiple time scales/Kurtosis引用本文复制引用
出版年
2024
NETL
NSTL
万方数据