摘要
针对王村煤业S1204工作面液压支架在自动跟机中存在推移位置控制精度低、跟机速度慢等问题,通过采用理论分析、试验研究等手段,提出基于BP神经网络预测提前动作距离的精准推移控制算法和安装蓄能器的自动跟机策略。采集支架液压系统状态并控制推移液压缸,液压缸控制精度从40~80 mm提高到20 mm内,控制精度大幅提升,表明采用BP神经网络预测控制算法的有效性。支架移架过程分为降柱、拉架、拉架和升柱切换、升柱四个区间,采用95%预测区间上限的方法,依据进液压力预测拉架动作时间。在支架进液端安装蓄能器后,拉架时间由6.54s降为5.39s,降低1.15s;速度均值由0.147m/s提高到0.177m/s,速度提升20.9%,验证快速跟机控制方法的有效性。
In view of the problems of low pushing and movement location control accuracy,slow machine following speed and others in the automatic machine following of the S1204 working face hydraulic support in Wangcun Coal Industry,a precise pushing and movement control algorithm based on BP neural network to predict advance action distance and a strategy of automatic machine following for installing energy accumulator are proposed by using ways of theoretical analysis,experimental research and so on.Collecting the status of the support hydraulic system and controlling the pushing and movement hydraulic cylinder,the control accuracy of the hydraulic cylinder is improved from 40-80 mm to within 20 mm,and the control accuracy is significantly improved,indicating the effectiveness of using the BP neural network predictive control algorithm.The bracket movement process of support is divided into four intervals:lowering column,pulling frame,switching between pulling frame and lifting column,and lifting column.Using 95%to predict the upper limit of the prediction interval method,according to the liquid inlet pressure to predict the pulling frame action time.After installing an energy accumulator at the liquid inlet end of the support,the pulling frame time is reduced from 6.54 s to 5.39 s,a decrease of 1.15 s;The average value of speed improves from 0.147 m/s to 0.177 m/s,with a speed increase of 20.9%,verifying the effectiveness of the rapid machine following control method.
作者
业巧云
Ye Qiaoyun(Jinneng Holding Coal Industry Group Wangcun Coal Industry Company,Shanxi Datong 037000)
出处
《山东煤炭科技》
2024年第8期105-110,共6页
Shandong Coal Science and Technology
关键词
液压支架快速推移
自动跟机
基于BP网络预测
hydraulic support rapid pushing and movement
automatic machine follow
prediction based on BP network
