[1]许弘焜,张树忠,国钰,等.基于BP-Adaboost的变转速泵控缸系统位移软测量[J].福建理工大学学报,2026,24(01):90-94.[doi:10.3969/j.issn.2097-3853.2026.01.012]
　XU Hongkun,ZHANG Shuzhong,GUO Yu,et al.Position soft-sensing of variable speed pump-controlled differential cylinder system based on BP-Adaboost[J].Journal of Fujian University of Technology;,2026,24(01):90-94.[doi:10.3969/j.issn.2097-3853.2026.01.012]
点击复制

基于BP-Adaboost的变转速泵控缸系统位移软测量()
分享到：

《福建理工大学学报》[ISSN:2097-3853/CN:35-1351/Z]

卷:
第24卷
期数:
2026年01期
页码:
90-94
栏目:
出版日期:
2026-02-25

文章信息/Info

Title:
Position soft-sensing of variable speed pump-controlled differential cylinder system based on BP-Adaboost
作者:
许弘焜; 张树忠; 国钰; 王亚兵; 凌泽懿; 阮玉镇
福建理工大学机械与汽车工程学院
Author(s):
XU Hongkun;  ZHANG Shuzhong;  GUO Yu;  WANG Yabing;  LING Zeyi;  RUAN Yuzhen
School of Mechanical and Automotive Engineering, Fujian University of Technology
关键词:
位移软测量; 变转速泵控系统; 反向传播神经网络; 自适应提升算法; 液压挖掘机
Keywords:
position soft-sensing; variable speed pump-controlled system; back propagation neural network; adaptive boosting algorithm; hydraulic excavator
分类号:
TH137
DOI:
10.3969/j.issn.2097-3853.2026.01.012
文献标志码:
A
摘要:
针对传统位移传感器测量液压缸位移存在的不足以及变转速泵控差动缸系统的非线性特征,采用基于自适应提升算法与反向传播神经网络相结合( back propagationneural network?adaptiveboosting algorithm,BP?Adaboost)的方法进行位移软测量。首先,利用MATLAB/ Simulink 搭建变转速泵控差动缸液压系统模型和机械臂机械模型,并与试验结果对比,证明了模型的准确性。其次,基于软测量原理分别搭建BP 神经网络软测量模型和BP?Adaboost 神经网络软测量模型;将已验证的泵控缸模型与20 t 挖掘机机械模型相结合,以该挖掘机铲斗的装卸不同物料为例,通过批量仿真获取数据集。最后,对所搭建的两个神经网络模型的铲斗位移软测量结果进行对比分析。结果表明:BP?Adaboost 神经网络在约800 mm 的行程中的位移软测量的平均误差和最大误差分别为0.9 mm 和9.6mm,与BP 相比分别降低了47.1%和50.8%,提高了预测的鲁棒性和泛化性。
Abstract:
Aiming at a series of shortcomings of traditional position sensors in measuring the hydraulic cylinders’ position and the nonlinear characteristics of the variable speed pump-controlled differential cylinder system, a position soft-sensing method based on the combination of adaptive boosting algorithm and back propagation neural network (BP-Adaboost) is adopted. Firstly, a model consisting of a variable speed pump-controlled differential cylinder system and a mechanical boom is built with MATLAB/Simulink, and its accuracy is proved by comparison with the test results. Secondly, the BP neural network position soft-sensing model and the BP-Adaboost position neural network soft-sensing model are constructed based on the soft-sensing principle respectively；the verified pump-controlled cylinder model is coupled with the mechanical model of a 20-ton hydraulic excavator, and the loading and unloading of different materials in the bucket of this excavator is taken as an example. The data set is obtained through batch simulation. Finally, the results of the bucket position soft-sensing of the two neural network models are compared and analyzed. Results show that the average and maximum errors of the BP-Adaboost neural network are 0.9 mm and 9.6 mm for the position soft-sensing over a stroke of about 800 mm,which are 47.1% and 50.8% lower than those of the BP, respectively, and the robustness of the prediction and generalization are improved.

参考文献/References:

［1］ 权龙，杨德荣，李风兰. 应用霍耳开关元件检测液压缸位移原理的研究［J］. 电子测量与仪器学报，2002，16（4）：11-15.［2］ 路勇，权龙. 超声波在液压缸活塞位移测量中的应用［J］. 太原理工大学学报，2003，34（5）：550-551，554.［3］ 樊烨男，高飞，廉自生，等. 液压缸用外置式磁感应位移传感器设计［J］. 机电工程，2022，39（6）：775-782. ［4］ 时航，李玥华，周京博，等. 基于STM32的液压缸位移测量系统设计［J］. 工业仪表与自动化装置，2023（1）：9-15. ［5］ AYHAN S，SCHERR S，PAULI M，et al. FMCW radar in oil-filled waveguides for range detection in hydraulic cylinders［C］∥9th European Radar Conference ，2012. Amsterdam: IEEE，2013：63-66.［6］ 王波，郝云晓，权龙，等. 分腔独立变转速泵控电液伺服系统特性研究［J］. 机械工程学报，2020，56（18）：235-243.［7］ 沈伟，崔霞. 变转速泵控马达自抗扰控制［J］. 液压气动与密封，2023，43（4）：39-45. ［8］ MINAV T A, LAURILA L I E, PYRHNEN J J. Relative position control in an electrohydraulic forklift［J］. International Review of Automatic Control, 2013,1(6):54-61.

更新日期/Last Update: 2026-02-25