电火花加工放电状态的自适应滤波

The adaptive filtering of gap states in electrical discharge machining

电火花加工过程具有非线性、弱稳态的特性,加工过程中放电状态变化剧烈,导致控制变量随之大幅度震动,对系统稳定性极为不利,因此需要对放电状态曲线进行光滑滤波。由于二阶自回归放电状态模型的参数辨识是有偏估计,提出采用中间变量的方法消除参数的有偏估计。采用一个卡尔曼滤波器建立中间变量,另一个卡尔曼滤波器用于二阶自回归放电状态模型的参数辨识,2个卡尔曼滤波器共同组成了一个自适应滤波器。经实验数据验证:该自适应滤波方法不仅能够光滑放电状态曲线,而且能够消除线性滤波本身固有的相移和延迟,同时,自适应滤波方法不会像线性滤波方法造成整个控制系统阶次大幅升高、稳定性下降,能够充分保障实时控制和系统稳定性。

Electrical discharge machining （EDM） has variable non-linear characteristics. The discharging state sometimes changes drastically during machining, which results in violent changes of the control variable and results in process variability. Which indicates that the discharging state curve needs to be filtered. Considering that the pa- rameter identification of a second-order auto-regressive model is a kind of biased estimation, which should be elimi- nated by an instrumental variable method, a Kalman filter was adopted to establish the instrumental variable, and another one was used to identify parameters of the second-order auto-regressive model. The two Kalman filters to- gether formed an adaptive filter. The experiment results show that the state curve can be smoothed without phase shifts or time delays, which are normal defects of linear filters. This adaptive filter is only a second-order system, not raising the order of the whole control systemmuch. This adaptive filtering approach can fully ensure real-time control and system stability.