深海起重机升沉补偿滑模预测控制
Sliding Mode Predictive Control for Heave Compensation of Deep-sea Crane
陈志梅 1卢莹斌 1邵雪卷 1赵志诚1
作者信息
- 1. 太原科技大学电子信息工程学院,太原 030024
- 折叠
摘要
受海浪、风力等因素的干扰,深海起重机升沉补偿系统的响应速度缓慢,系统对于负载位移的控制精度较差.为了提高升沉补偿系统响应速度与系统对负载位移的控制精度,保证起重机在各种海况下正常作业,提出了基于CNN-LSTM深度学习网络的滑模预测控制方法.首先,将CNN网络与LSTM网络结合,建立CNN-LSTM深度学习网络控制系统预测模型.其次,通过参考位移与实际位移的误差建立滑模面,并根据幂次函数设计滑模面参考轨迹;采用粒子群算法(PSO)对性能指标进行寻优,得出控制律,根据控制律控制负载实际位移跟随参考位移.最后,进行了仿真研究.结果表明与传统模型预测控制相比,在该方法的控制作用下,系统的响应速度更快,系统对负载位移的控制精度更高,系统的鲁棒性能更强.
Abstract
Due to the interference of ocean waves,wind and other factors,the response speed of the heave compen-sation system of deep-sea crane is slow,and the control accuracy of the system for load displacement is poor.In or-der to improve the response speed of heave compensation system and the control accuracy of load displacement,and to ensure the normal operation of crane under various sea conditions,a sliding mode predictive control method based on CNN-LSTM depth learning network is proposed.First of all,the prediction model of CNN-LSTM deep learning networked control system is established by combining CNN network with LSTM network.Secondly,the slid-ing surface is established by the error between the reference displacement and the actual displacement,and the ref-erence trajectory of the sliding surface is designed according to the power function.The particle swarm optimization algorithm(PSO)is used to optimize the performance index,and the control law is obtained.According to the control law,the actual displacement of the load is controlled to follow the reference displacement.Finally,the simulation re-search is carried out.The results show that compared with the traditional model predictive control,under the control effect of this method,the response speed of the system is faster,the control precision of the system to the load dis-placement is higher,and the robust performance of the system is stronger.
关键词
升沉补偿/CNN-LSTM/滑模预测控制/粒子群算法Key words
heave compensation/CNN-LSTM/sliding mode predictive control/particle swarm optimization algo-rithm引用本文复制引用
基金项目
山西省自然科学基金(201901D111263)
山西省重点研发计划(201803D121025)
山西省研究生教育改革研究课题(20191G173)
出版年
2024
NETL
NSTL
维普
万方数据