Organization Name: Xi’an Jiaotong University

Project introduction:

Over 90% of machine tools in China areused in environment without temperature control. To study the key technologies of thermal error modeling and compensation that restrict the accuracy and its stability of machine tools, A standardized modeling and parameter dynamic correction method of machine tool thermal error integrating physical characteristics and deep learning is proposed, which solves the problem of poor robustness of machine tool thermal error model under different structures and working conditions. An automatic modeling method of thermal error model based on big data intelligence and multi factor decoupling drive is proposed to improve the efficiency and model adaptability of machine tool users' thermal error modeling.The thermal error modeling tool and condition adaptive compensation device based on intelligent algorithm are developed, and the domestic high-end CNC system with the functions of unified thermal error modeling and batch compensation is developed. The accuracy control level is equivalent to that of Siemens 840D CNC system. It realizes batch industrial application on more than 20 kinds of domestic ordinary machine tools. T65ms series CNC lathes of Zhejiang Headman developed with the technology of this project have reached the accuracy level of Japanese Mazak LGMQTN150 machine tool, and the dimensional error of parts processed in batch for a long time has been reduced to 5 μm below, the accuracy of domestic ordinary machine tools is improved massively, and it has good accuracy retention despite of seasonal changes. The technology promotes domestic CNC systems and ordinary machine tools to move from medium and low-end to medium and high-end.