基于SSO算法优化神经网络的数控机床热误差建模

针对影响五轴数控机床加工精度的复杂热特性，提出了一种用于摇篮式五轴数控机床热误差建模方法.该方法主要采用鲨鱼嗅觉优化(SSO)算法和神经网络的复合建模方式，有效提高了机床热误差预测模型的精度和建模效率.首先通过使用热成像仪筛选出机床的温度敏感点，然后将温度传感器布置在机床热敏感点的位置，将采集到的热特性数据采用本文所提方法进行热误差建模，结果表明，该方法在建模速度和精度上要优于ABC和PSO神经网络，最后将该热误差预测模型应用于五轴数控机床热误差补偿实验，将试件加工精度提高了32%.

Thermal Error Modeling of Numerical Control Machine Tools Based on Neural Network Neural Network by Optimized SSO Algorithm

In order to explore the complex thermal characteristics of five-axis NC(numerical control) machine tools， a method for thermal error modeling of cradle five-axis NC machine tools was proposed. The principle of shark smell optimization(SSO)algorithm and neural network composite modeling was adopted， which effectively improved the accuracy and modeling efficiency of the machine tool thermal error prediction model. Firstly， the temperature sensitive point was screened by using the thermal imager， and then the temperature sensor was placed at the position of the heat sensitive point of the machine tool. The collected thermal characteristic data were modeled by the above method. The results showed that the method is better than ABC neural network and PSO neural network in terms of modeling speed and accuracy. Finally， the model was applied to the thermal error compensation experiment of the five-axis machine tool， which improves its accuracy by 32%.