In order to characterize cutting mechanics during high-speed milling and micromilling applications, high-end piezoelectric dynamometers with a wide frequency bandwidth are necessary. Nevertheless, when installed into the machine tool their signal bandwidth is limited by the dynamic behaviour of the machining system. Thus, special filters have to be adopted for dynamics compensation. State of the art filters are based on a simplistic 3 × 3 dynamic model of device transmissibility without taking into account the influence of input force location with respect to the centre of the sensing platform. The Upgraded Augmented Kalman Filter has been recently proposed for solving this problem. Although it outperformed the other state of the art filters, it was based on the preliminary identification of a parametric mathematical model that is generally a difficult and non-automatic task. Here a novel non-parametric filter is introduced, that was based on a more general and abstract model of dynamometer dynamics considering both input force direction and location. By so doing, impressive results were found both from modal analysis and from real cutting tests, showing the potential of the new method for an effective and almost completely automatic cutting force dynamic compensation.

为了表征高速铣削和微铣削应用中的切削力学，需要具有宽频带宽的高端压电测功机。然而，当安装到机床中时，它们的信号带宽受到加工系统动态行为的限制。因此，必须采用特殊滤波器进行动态补偿。最先进的滤波器基于一个简单的 3 × 3 设备传输率动态模型，没有考虑输入力位置相对于传感平台中心的影响。最近提出了升级的增强卡尔曼滤波器来解决这个问题。虽然它的性能优于其他最先进的过滤器，它基于参数数学模型的初步识别，该模型通常是一项困难且非自动的任务。这里引入了一种新颖的非参数滤波器，它基于考虑输入力方向和位置的更一般和抽象的测功机动力学模型。通过这样做，从模态分析和实际切削测试中都获得了令人印象深刻的结果，显示了新方法在有效且几乎完全自动的切削力动态补偿方面的潜力。