基于自适应子群协作QPSO算法的机车黏着智能模糊优化控制

Intelligent Fuzzy Optimal Control of Locomotive Adhesion Based on Adaptive Multiple Subgroup Collaboration QPSO Algorithm

为了使机车在运行过程中保持轮轨最佳黏着力,需要对牵引电机转矩进行实时调整。以车轮角加速度和黏着特性曲线斜率的估计值作为输入变量,以牵引电机转矩的调节量作为输出变量,提出机车黏着智能模糊优化控制策略;以考虑轮轨黏着力变化指标和牵引电机转矩波动指标的加权目标函数作为优化设计模糊控制器应满足的适应度函数;基于双输入单输出的二维模糊控制器结构,量化并变换得到模糊控制器的尺度变换函数和隶属度函数,并提出自适应子群协作QPSO算法对这些函数中的参数进行优化。仿真结果表明:在轨面条件发生多种极端突变的情况下,该控制策略都能使机车轮对在很短的时间内调整到最佳黏着状态下,从而发挥最佳牵引力。

In order to keep the best wheel-rail adhesion of locomotive during operation,the real-time adjustment of traction motor torque is needed.An intelligent fuzzy optimal control strategy for the wheel-rail adhesion of locomotive is proposed with the value of wheel angular acceleration and the estimated value of adhesion characteristics curve slope as the input variables,and with the adjustment amount of traction motor torque as the output variable.The weighted objective function considering the change index of wheelrail adhesion and the fluctuation index of traction motor torque is taken as the fitness function for the optimization design of fuzzy controller.The scale transformation functions and membership functions of the fuzzy controller are quantified and transformed based on the structure of two dimensional double-input and single-output fuzzy controllers,and an adaptive subgroup collaboration QPSO algorithm is proposed to optimize the parameters of these functions.Simulation results show that the control strategy can make the wheelset of locomotive be adjusted to the optimal adhesion state in a very short time under a variety of extreme conditions of rail surface,and thus achieve the optimal traction force.