Design and develop an automatic compensation method of alignment reference network for Tokamak machine assembly

• Novel Automatic Compensation System : Developed an innovative temperature-based compensation method integrating real-time monitoring of biological shield wall supports with displacement trends to address thermal deformation in tokamak alignment networks. • Experimental Validation on EAST Tokamak : Demonstrated 50 % improvement in alignment precision (reducing uncertainty from 0.8 mm to 0.4 mm) through comprehensive testing using 22 temperature sensors and laser tracker verification. • Deformation Database Implementation : Created a predictive compensation database enabling real-time adjustments during assembly, addressing multi-year construction challenges. • Scalable Methodology : Provides a validated framework applicable to next-generation fusion devices, with identified pathways for improvement through expanded datasets and finite element analysis. As the core component of a fusion device, the tokamak comprises multiple complex systems including superconducting magnets, vacuum vessel, thermal shielding, blanket, and divertor systems, all requiring extremely high assembly accuracy over an extended construction period. The alignment reference network serves as the critical datum for component assembly, where its stability directly determines the final device precision. Thermal-induced deformations during the multi-year assembly process present significant challenges to maintaining this network's stability. This paper presents an innovative automatic compensation method to address instability in the alignment and metrology reference network. Our approach integrates real-time temperature monitoring of biological shield wall supports with displacement change trends to predict thermal deformations. A comprehensive deformation database is developed to enable real-time compensation during assembly measurements. Validated on the EAST tokamak using 22 temperature sensors and laser tracker verification, our method successfully reduced alignment uncertainty by 50 %, from 0.8 mm to 0.4 mm. These results demonstrate the effectiveness of temperature-based compensation for large-scale precision assembly in fusion devices.