面向数控机床运行状态的切削稳定性预测研究

切削加工过程中出现的颤振失稳现象，是限制机床加工质量和加工效率的主要因素。传统切削稳定性预测模型大多基于机床静止状态下的动力学特性，并采用恒定的切削力系数表征不同的切削条件。但在加工过程中，系统动力学特性和切削力系数会随着主轴转速等影响因素而变化，导致预测的切削稳定性叶瓣图在实际工程运用中出现偏差。针对机床运行状态下切削稳定性的准确预测问题，提出一种切削稳定性叶瓣图修正方法。该方法以刀具系统动力学特性与切削力系数为研究对象，首先建立主轴转速样本信息，将考虑转速效应的主轴轴承运行刚度写入机床有限元模型中，获取刀尖频率响应函数及其对应的各阶模态参数，以此结合模态拟合法和插值算法重构任意转速下的刀尖频响函数，同时以各切削参数为变量构建切削力系数响应面预测模型，进而将与转速对应的刀尖频率响应函数和不同切削条件下的切削力系数作为传统切削稳定性预测模型的输入，并通过结合自适应粒子群算法共同求解各转速下的极限切削深度，从而在全转速范围内绘制切削稳定性叶瓣图。将该方法应用于1台3轴立式加工中心的实际工序中，采用多组预测的无颤振切削参数进行切削实验，并通过切削力信号的频谱分析判定切削过程中未出现颤振，验证了稳定性叶瓣图修正方法的有效性，为无颤振切削参数的合理选择奠定了技术支持。

Study on Prediction of Machining Stability for Machine Tool Under Operational State

Chatter in machining process is the main factor limiting the machining quality and machining efficiency. The traditional machining stability prediction models are mostly based on the machine tool dynamics under stationary state, and use the constant cutting force coefficients to represent different cutting conditions. However, in real machining process, system dynamics and the cutting force coefficients will change caused by the factors such as the spindle speed, leading inaccurate chatter prediction. Thus, a method to modify the conventional machining stability lobe diagram (SLD) is proposed to ensure the prediction accuracy based on the relationships among cutting parameters, system dynamics and cutting force coefficients. In this method, the sample information of the spindle speed is initially set up, and then the speed-dependent bearing stiffness is applied into the whole machine tool finite element model (FEM). Then the tool tip frequency response functions (FRFs) are obtained and their related modal parameters are identified. Therefore, with these sample modal parameters, the tool tip FRFs at any speed can be reorganized based on the modal fitting technique and the interpolation algorithm.A response surface method is established for predicting the cutting force coefficients under different machining conditions, which are further combined with the speed-dependent tool tip FRFs to be the inputs of the machining stability mathematical model. This model is combined with the adaptive particle swarm optimization algorithm to achieve the limiting cutting depth at any spindle speed. Then a modified SLD in full spindle speed range is plotted. The proposed method is applied to a three-axis vertical machining center, and different sets of the predicted no chatter cutting parameters are adopted to carry out the chatter tests. And the frequency spectrums of the cutting force signals are analyzed to indicate that no chatter occurred, which verify the feasibility of the proposed method and show that it can lay a foundation for the cutting parameters selection.