A MVB network control system was established,including the forward channel delay between the controller node and the actuator node as well as the feedback channel delay between the controller node and the sensor node.The acquired delay data showed the characteristics of random distribution.Autoregressive model and least mean square algorithm were used to predict forward channel delay.Considering the characteristics of the key control system of train,the nonlinear system was described as a number of linear subsystems near different steady-state operating points.A nonlinear network control system model was established based on T-S fuzzy model,and the generalized predictive control method based on the rolling optimization of state space was designed for each subsystem.The control effects with different control methods,sampling periods and load rates were simulated.Results show that the GPC delay control method based on T-S fuzzy model can accurately track reference signals with different frequencies.Compared with the fuzzy adaptive PID method,this method can better suppress the effect of time delay on control performance.It has ideal robust performance and stability performance,and can fully satisfy the requirements for the real-time control of high-speed EMU running process.%建立包括控制器节点与执行器节点之间的前向通道时延,以及它与传感器节点之间的反馈通道时延的多功能车辆总线网络控制系统,获取的时延数据呈现随机分布的特点;采用自动回归模型和最小均方算法进行前向通道时延的预测;考虑列车关键控制系统的特点,将非线性系统描述为不同稳态工作点附近的多个线性子系统,基于T-S模糊模型建立非线性网络控制系统模型,设计各子系统基于状态空间滚动优化的广义预测控制方法.对采用不同控制方法、不同采样周期和负载率时控制效果进行仿真.结果表明:基于T-S模糊模型的GPC时延控制方法可以精确跟踪不同变化频率的参考信号;与模糊自适应PID方法相比较,该方法能较好地抑制时延对控制性能的影响,具有理想的鲁棒性能和稳定性能,完全可以满足高速动车组运行过程的控制实时性要求.
展开▼
