摘要
提出基于深度学习的液体装卸车鹤管自动对接方法。首先分析自动鹤管本体结构,采用九点标定法得到鹤管轴线与坐标系之间的夹角,完成鹤管静态中心点坐标获取;然后根据鹤管关节的运动参数,拟合鹤管运动轨迹;最后引入深度学习理论,基于构建的RVFA LSTM模型完成鹤管对接,利用几何特征分析方法进行对接点缩放因子计算,完成罐口对接点的校正,实现液体装卸车鹤管与罐口的自动对接优化。实验结果表明,此方法对接精度高,对接平均绝对偏差低,能够满足工程要求。
An automatic docking method for liquid loading and unloading trunk crane tubes based on deep learning is proposed.Firstly,the structure of the automatic crane tube body is analyzed.The nine point calibration method is adopted to obtain the angle between the crane tube axis and the coordinate system,and the acquisition of the static center point coordinates of the crane tube is completed.Then,based on the motion parameters of the crane joint,the motion trajectory of the crane joint is fitted.Finally,deep learning theory is introduced to complete the docking of crane tubes by using the constructed RVFA LSTM model.Geometric feature analysis is employed to calculate the scale factor of the docking point,and the ground correction of the tank mouth docking point is executed,achieving automatic docking optimization between the crane tube and the tank mouth of the liquid loading and unloading truck.The experimental results show that this method has high docking accuracy and low average absolute deviation,which can meet the engineering requirements.
作者
郭永强
崔春江
王炜
张强
陈宗宇
何坤
GUO Yongqiang;CUI Chunjiang;WANG Wei;ZHANG Qiang;CHEN Zongyu;HE Kun(The Third Gas Production Plant of Changqing Oilfield Branch,Xi’an 710299,China)
出处
《机械与电子》
2025年第6期75-80,共6页
Machinery & Electronics
基金
陕西省工业攻关计划资助项目(2013K746)。
关键词
深度学习
液体装卸车
鹤管自动对接
中心点标定
缩放因子
deep learning
liquid loading and unloading truck
automatic docking of crane tubes
center point calibration
scale factor
