王学文, 李素华, 谢嘉成, 等. 机器人运动学与时序预测融合驱动的刮板输送机调直方法[J]. 煤炭学报, 2021, 46(2): 652-666.
引用本文: 王学文, 李素华, 谢嘉成, 等. 机器人运动学与时序预测融合驱动的刮板输送机调直方法[J]. 煤炭学报, 2021, 46(2): 652-666.
WANG Xuewen, LI Suhua, XIE Jiacheng, et al. Straightening method of scraper conveyor driven by robot kinematics and time series prediction[J]. Journal of China Coal Society, 2021, 46(2): 652-666.
Citation: WANG Xuewen, LI Suhua, XIE Jiacheng, et al. Straightening method of scraper conveyor driven by robot kinematics and time series prediction[J]. Journal of China Coal Society, 2021, 46(2): 652-666.

机器人运动学与时序预测融合驱动的刮板输送机调直方法

Straightening method of scraper conveyor driven by robot kinematics and time series prediction

  • 摘要: 随着智能化开采的不断发展,刮板输送机直线度控制对于煤矿安全、高效开采具有重要意义。针对刮板输送机的调直精度不高的问题,提出了一种基于空间运动学与长短时记忆神经网络轨迹预测相融合的调直方法。首先,利用工业机器人的空间运动学知识对液压支架和刮板输送机浮动连接机构的运动规律进行了解析,并以C#语言的形式编入Unity3D仿真系统底层,通过推移机构连接头捕捉刮板输送机中部槽上的关键点,实现液压支架与刮板输送机的连接,实现了液压支架的精准推移,较理想化地解决了销耳间隙的影响;其次,综合考虑到传感器噪声与截割底板轨迹对刮板输送机轨迹检测的影响,在仿真系统中进行相关的补偿后,在MATLAB中利用LSTM(Long Short Time Memory)神经网络对刮板输送机的轨迹进行预测;最后,根据实际工况要求建立了目标调直轨迹的修正模型和轨迹-姿态转换模型,以得到的刮板输送机轨迹为基础,确定及时移架后的液压支架位置与对应中部槽的相对位置差,基于浮动连接机构运动规律液压支架精准推移,实现刮板输送机调直。通过虚拟试验的验证,建立的修正模型和转换模型具有较强的可靠性,在仿真系统与实验室条件下分别在底板存在起伏时进行了调直试验,提出的调直方法的直线度误差分别在±0.2 cm和0.08 cm内,符合调直精度要求,研究结果对刮板输送机的调直研究提供了思路。

     

    Abstract: With the development of intelligent mining,the straightness control of scraper conveyor is of great significance for coal mine safety and efficient mining.Aiming at the low accuracy of scraper conveyor,a method of fusion and straightening based on the prediction of space kinematics and long short time memory neural network trajectory is proposed.Firstly,the movement laws of the floating connection mechanism of the hydraulic support and scraper conveyor are analyzed with the spatial kinematics knowledge of the industrial robot,which is programmed into the Unity3D simulation system through C# language,and the connection between hydraulic support and scraper conveyor is realized by capturing the key points on the middle trough of scraper conveyor through the joint of pushing mechanism,which realizes the precise pushing of hydraulic support and solves the influence of pin ear clearance more ideally.Secondly,considering the influence of sensor noise and cutting bottom trajectory on the track detection of scraper conveyor,after the relevant compensation in the simulation system,LSTM (Long Short Time Memory) neural network is used to forecast the track of scraper conveyor in MATLAB.Finally,according to the actual operating conditions,the correction model of target straightening trajectory and the trajectory attitude transformation model are established,and based on the obtained scraper conveyor track,the position difference between hydraulic support and the corresponding middle trough is determined after timely frame removal,and the straightening of scraper conveyor is realized by precise pushing and advancing of the hydraulic support based on the movement law of the floating connection mechanism.Through the comprehensive verification of virtual experiment and physical experiment,the modified model and conversion model established in this paper have strong reliability,and the straightening tests are carried out in the simulation system and in the laboratory with undulations in the bottom plate,and the straightness errors of the straightening method proposed in this paper are within ±0.2 cm and 0.08 cm respectively,which meets the straightening accuracy requirements and provides ideas for the straightening research of scraper conveyor.

     

/

返回文章
返回