基于深度强化学习的薄煤层刮板输送机功率平衡智能控制方法

Intelligent control method for power balance of scraper conveyor in thin coal seam based on deep reinforcement learning

提出基于深度强化学习的薄煤层刮板输送机功率平衡智能控制方法。基于对薄煤层刮板输送机及传动系统结构分析,将PID控制器待优化参数输入到训练好的Actor-Critic神经网络中,得到PID控制器最优控制参数,并通过优化后的控制参数构建PID优化控制器;将刮板输送机电动机转速作为功率调节的依据,输入期望转速及实际转速的偏差值至PID优化控制器中,调节刮板输送机转速,使其功率平衡控制效果达到最佳。实验结果表明,该方法可以智能控制薄煤层刮板输送机功率,满足薄煤层开采过程中刮板输送机功率平衡的需求。

An intelligent control method for power balance of thin seam scraper conveyor based on deep reinforcement learning is proposed.Based on the structural analysis of thin seam scraper conveyor and transmission system,the parameters to be optimized of PID controller are input into the trained Actor-Critic neural network to obtain the optimal control parameters of PID controller,and the PID optimal controller is constructed through the optimized control parameters;The motor speed of the scraper conveyor is used as the basis for power regulation,and the deviation value of the expected speed and the actual speed is input to the PID optimization controller to adjust the speed of the scraper conveyor,so as to achieve the best power balance control effect.The experimental results show that this method can intelligently control the power of scraper conveyor in thin coal seam to make it balance,and meet the demand of power balance of scraper conveyor in the process of thin coal seam mining.