考虑测量误差的稳健参数设计

Robust Parameter Design Considering Measurement Error

在产品质量设计与制造过程中,通常假定产品或过程的最优参数设计能够准确无误地实施。然而,由于环境温度、湿度以及电压等噪声因素的干扰以及测量误差的作用,过程输入的最优参数设计值往往会呈现出一定的波动,与预定的设计参数值存在一定的偏差。为此,将离线的质量设计技术与在线调整技术相结合,考虑测量误差的情形下,提出一种在线调整的稳健参数设计新方法。首先,通过稳健参数设计方法确定离线控制变量的最优参数设计值;然后,利用传感器技术对生产过程中的可测噪声因子进行在线测量,并借助在线技术调整在线控制变量,对整个过程的期望损失进行调整与补偿,从而降低产品或过程的波动,最终提升产品或过程的设计质量。最后,通过半导体的实际例子验证所提方法的有效性。

It’s usually assumed that the practical process parameters are equal to design parameters during the design and manufacturing phases.Due to the effect of noise variables(temperature of environment,humidity,and voltage inaccuracy of the measurement),the actual process parameters of production may be different from the design parameters.As for process input parameters universal existence deviation from the design value,the new optimization model of robust design combines robust parameter design with online adjustment.Firstly,the optimum parameter values of offline control variables were determined.Then,the online variables according to the measurable noise variables were adjusted to reduce the quality loss.Finally,an example of semiconductor was given to illustrate the efficiency of the proposed approach,compared with the optimization solutions obtained by other methods.