基于自适应陷波的直拉单晶直径跟踪

Diameter tracking in pulling of crystals based on adaptive notch filtering

为抵消单晶生长过程中晶体直径测量信号存在的周期性干扰,提出一种适合晶体直径测量信号的自适应陷波滤波方法。该方法采用抵消低频分量、保留高频分量的思想建立滤波器优化函数,利用一种改进的粒子群优化算法优化该函数获得周期性干扰的逼近信号,并用此逼近信号抵消周期性干扰。实验结果表明,所提出的滤波方法能有效跟踪直拉单晶直径的变化,精度优于常用的滤波方法,所提出的粒子群优化算法能够使滤波器系数较快收敛。

To offset the periodic interference on crystal diameter measurement in single crystal growth process, an adaptive notch filtering method suitable for crystal diameter measurement has been proposed in this paper. This method builds a filter optimization function through offsetting low frequency signal and maintaining high frequency signal. The function is optimized using an improved PSO （particle swarm optimization） algorithm to obtain an approximation signal close to the periodic interference, which can be used to offset the periodic interference on crystal diameter measurement in single crystal growth process. Experiments show that the filter method based on adaptive notch filtering can efficiently track crystal diameter variations in pulling of crystals and is better than commonly-used methods in calculation precision. These results also show that the proposed PSO algorithm can make the filter coefficient converge to the global optimal point quickly.