摘要
针对岩石试验机电液伺服控制系统PID控制参数整定复杂、用户使用不友好的问题,提出一种基于系统辨识与智能优化的线性二次型最优控制方法。建立自回归外生变量(ARX)模型对岩石试验机控制对象进行参数辨识,并构建系统状态空间模型。在此基础上设计线性二次型最优控制器(LQR),结合卡尔曼滤波算法实现系统状态实时估计与干扰抑制。采用差分进化-粒子群混合优化算法(DE-PSO)对LQR性能指标中的状态权重矩阵与控制代价矩阵进行协同优化,有效平衡系统响应速度与控制能耗之间的关系。仿真与试验结果表明:所提LQR控制器在阶跃响应中实现了0.48 s的调节时间与0.31%的超调量,能够快速抑制外部扰动;在变速率阶梯加载试验中,系统均实现了无稳态误差的跟踪控制,满足相关岩石力学试验标准的要求;对玄武岩与花岗岩进行单轴强度测试时,在不同加载速率下均实现了试验力超调量小于1%无稳态误差的高精度控制,验证了该方法的鲁棒性与适应性。基于LQR与DE-PSO的岩石试验机控制系统无需手动执行复杂的参数整定过程即可实现试验力无稳态误差控制,为岩石力学试验装备的智能化升级提供了新方法。
In response to the complex tuning of PID control parameters and unfriendly user experience in the hydraulic servo control system of rock testing machines,a linear quadratic optimal control method based on system identification and intelligent optimization was proposed.An auto-regressive with extra inputs(ARX)model was established to perform parameter identification for the control object of rock testing machine,and then a system state-space model was constructed.Based on this model,a linear quadratic regulator(LQR)was designed and Kalman filter algorithm was combined to achieve real-time state estimation and disturbance rejection.A differential evolution-particle swarm optimization(DE-PSO)hybrid algorithm was employed to conduct collaborative optimization of the state weighting matrix and control cost matrix in the LQR performance index and effectively balance the relationship between system response speed and control energy consumption.Simulation and experimental results demonstrate that the proposed LQR controller achieves a regulation time of 0.48 s and an overshoot of 0.31%in the step response,and can quickly suppress external disturbances.In the variable-rate step loading test,the system has achieved steady-state error-free tracking control,meeting the requirements of relevant rock mechanics test standards.When conducting uniaxial strength tests on basalt and granite,high-precision control with a test force overshooting of less than 1%and no steady-state error is achieved at different loading rates,verifying the robustness and adaptability of this method.The rock testing machine control system based on LQR and DE-PSO eliminates the need for manual parameter tuning while achieving steady-state error-free control of the test force,providing a novel method for the intelligent upgrading of rock mechanics testing equipment.
作者
高继开
姚志宾
GAO Jikai;YAO Zhibin(State Key Laboratory of Intelligent Deep Metal Mining and Equipment,Northeastern University,Shenyang Liaoning 110819,China)
出处
《机床与液压》
北大核心
2025年第21期121-126,共6页
Machine Tool & Hydraulics
基金
国家自然科学基金面上项目(52274114)
东北大学研究生揭榜挂帅项目(01220021302007*006)。
