基于卡尔曼滤波的动平衡测量过程调节方法

Measurement process regulation of dynamic balancing machines based on Kalman filtering

针对目前所采用的消除动平衡机测量偏倚的定期重新标定方法将造成生产线不必要的停工,从而提高制造企业的生产成本问题,提出基于卡尔曼滤波的动平衡测量过程在线调节方法.利用泰勒展开方法对振动响应测量值进行坐标变换以消除观测方程的非线性.结合多变量统计过程控制(MSPC)对测量过程进行监控以检测测量偏倚,提出通过追踪状态参数变化对测量偏倚进行补偿的方法.通过对样机长期的状态监控数据对该方法进行验证.结果证明,该方法能够快速诊断出测量偏倚并对其进行准确的补偿,在保证测量系统精度的同时,最大限度地缩短停工时间.

The recalibration method, which is usually adopteu to, will lead to the unavoidable stoppage of the whole assembly line and then increase the operation cost of manufacturing enterprises. An online dynamic balancing measurement process regulation method based on Kalman filtering was proposed aimed at the problem. Taylor expansion was used to conduct coordinates transformation of vibration response values and then eliminate the nonlinearity of the observation equation. Multivariate statistical process control （MSPC） was integrated to monitor the measurement process and detect measurement bias. The method to compensate measurement bias through tracing the change of state parameters was proposed. The method was validated by the long-term monitoring data of a prototype machine. Results indicate that the method can rapidly and accurately detect and compensate measurement bias. Then the accuracy of measurement system can be guaranteed and the stoppage was greatly shortened.