【作者】 郭一楠； 刘擎宇； 巩敦卫； 张振；

【Author】 GUO Yinan;LIU Qingyu;GONG Dunwei;ZHANG Zhen;School of Information & Control Engineering,China University of Mining and Technology;School of Mechanical,Electrical & Information Engineering,China University of Mining and Technology (Beijing);

【机构】 中国矿业大学信息与控制工程学院； 中国矿业大学(北京)机电与信息工程学院；

【摘要】 巷道掘进过程中,有效控制液压锚杆钻机钻臂的水平移动,对实现巷道中锚杆/锚索的精准安装具有重要意义。然而,复杂的巷道环境以及液压锚杆钻机钻臂的固有非线性和不确定干扰,使其高精度位移控制面临挑战。低精度位置控制使钻机钻臂无法及时移动到预先设定的锚杆/锚索安装位置,导致对巷道空顶不能实施有效支护,甚至出现安全事故。因此,针对液压锚杆钻机钻臂位移系统,设计一种高效控制器十分必要。基于此,通过分析液压锚杆钻机钻臂位移系统组成结构,特别是液压油缸和三位四通比例换向阀特性,建立了液压锚杆钻机钻臂位移系统数学模型。进而,考虑时变参数、外部扰动和系统固有非线性,设计了一种新型神经网络自适应滑模控制策略。该控制方法包括两种重要策略:其一是设计了一种改进的滑模趋近律,从而有效消除滑模抖振;其二是设计了一种基于改进径向基网络的不确定干扰观测器,以快速有效辨识系统的不确定干扰。针对某一类实际锚杆钻机钻臂,基于MATLAB和AMESim联合仿真平台,构建其位置模拟系统,并基于该平台,将所提方法与其他三种控制方法进行控制性能对比。对比实验结果表明:在固定干扰与突加干扰两种情况下,所提控制方法均能够无超调、更快速和精确的跟踪期望位置,从而保障巷道支护任务的可靠完成。更多还原

【Abstract】 During the excavating process of a roadway, effectively controlling the horizontal movement of a drilling arm for a hydraulic roofbolter is of great significance to accurately install bolts or anchor cables in a roadway. However, the complex environment of a roadway, as well as the inherent nonlinearity and uncertain disturbances of a hydraulic drilling arm produce challenges for its high-precision displacement control. To the best of our knowledge, low-precision displacement control cannot guarantee a drilling arm timely and accurately moving to the expected displacement for installing bolts or anchor cables, resulting in ineffective support for the roof of a roadway, even safety accidents. Therefore, designing an effective controller for displacement system of a hydraulic anchor-hole drilling arm is necessary. First, by analyzing the components of displacement system, especially, the characteristics of a hydraulic cylinder and three-position four-way proportional directional valve, the mathematical model of a hydraulic anchor-hole drilling arm is established. Second, taking time-varying parameters, external disturbances and inherent nonlinearity into account, an adaptive sliding mode controller based on neural network is presented. It contains two main contributions. One is to design an improved sliding mode reaching law, with the purpose of effectively eliminating chattering. The other is to introduce an identifier based on improved radial basis function to estimate uncertain disturbances quickly and effectively. A simulation system of a practical drilling arm for a hydraulic roofbolter is built by joint simulation platform of MATLAB and AMESim. Following that, the control performances are compared among the proposed controller and three other control methods. Comparative experimental results show that the proposed control method can track the pre-set displacement faster and more accurately without overshoot under the fixed or sudden disturbance, with the purpose of ensuring the reliable support for a roadway.更多还原