论文部分内容阅读
研究了以神经网络（ＮＮ）为模型的软件补偿不同机床热误差。提出知识获取是神经网络建模的关键环节。两种数控机床被用来研究分析热变形，通过测量实验获取学习数据，特别是温度测量和传感器数目的选取。介绍了机器学习技术和学习数据组织方法，包括归纳学习和推理学习。给出了预报补偿的结果和精度评价。
The neural network (NN) model was used to compensate the thermal errors of different machine tools. It is proposed that knowledge acquisition is the key link in neural network modeling. Two types of CNC machine tools are used to study the analysis of thermal deformations and to obtain learning data through measurement experiments, especially the selection of temperature measurement and number of sensors. Introduced the machine learning technology and learning data organization methods, including inductive learning and inference learning. The results of the forecast compensation and the precision evaluation are given.