基于自适应神经模糊推理系统和灰色理论的机床热误差补偿研究

Research on thermal error compensation of machine tool based on adaptive neuro fuzzy inference system and grey theory

提出了自适应神经模糊推理系统(ANFIS)模型,采用灰色理论对机床热误差进行建模,实现机床加工工件定位误差值的最小化。采用自适应模糊推理系统的模糊C均值聚类法,对机床上的温度传感器进行聚类分组和分析,选择出每组最优的温度传感器,将测量温度传感器从76个减少到5个。提出了灰色系统理论,对GM(1,N)公式进行了推导,创建了热误差预测模型。采用实验测量方法对机床运行所产生的误差进行了验证。实验结果显示:补偿前Y轴和Z轴产生热误差的最大值分别为41.5μm和33.8μm,补偿后Y轴和Z轴产生热误差的最大值分别为4.8μm和4.6μm。采用自适应神经模糊推理和灰色系统对机床热误差进行补偿,不仅测量温度传感器数量减少,而且机床主轴运行所产生的误差明显减小,加工精度较高,效果很好。

An adaptive neural fuzzy inference system （ANFIS） model is proposed, which is based on the grey theory to model the thermal error, and to minimize the value of the positioning error of machine tool machining parts. Using the adaptive fuzzy inference system of fuzzy C -means clustering algorithm, the temperature sensor of machine tool for grouping and clustering analysis, select each optimal temperature sensor, measurement temperature sensor will be reduced from 76 to 5. Grey system theory is put forward, and the formula of GM （ 1, N） is deduced, and the thermal error prediction model is established. The error genera- ted by the machine tool is verified by the experimental measurement method. The experimental results show that the maximum value of thermal error of Y axis and Z axis is 41.5μm and 33.8 μm, respectivey, and the maximum value of thermal error of Y axis and Z axis is 4.8 μm and 4.6 μm respectively. Using adaptive neuro fuzzy inference and grey system to compensate the thermal error of machine tool, measure not only reduce the number of temperature sensor and generated by the operation of the machine tool spindle error is significantly reduced, high precision machining, the effect is very good.