Abstract:

Objective To investigate the work-related musculoskeletal disorders（WMSDs） associated with poor working postures among underground support workers，and to construct a risk identification and assessment system for WMSDs in workers. Methods Ten support workers from a coal mine of Shendong Group in northern Shaanxi Province from July to September 2023 were selected as the evaluation objects. The BlazePose model within the MediaPipe framework system was integrated with the rapid upper limb assessment（RULA） method to evaluate WMSD risks，and comparing it with expert scores to evaluate the reliability of this method. Results Comparing the RULA scores of three experts with the evaluation results of this assessment method，the intraclass correlation coefficients（ICC） values were all >0.85. According to RULA score，the average angle of neck was 14.2°. The concentrated average angles for the wrist，upper arm，and forearm were 24.8°，and 22.5°-40.5°，respectively，corresponding to RULA scores of 2-3 points. The average wrist angle was 24.8°，with a score distribution of 3 points. Neck postures centered around 14.2°，with most receiving 4 points. In rib and roof support tasks，level 3 risks accounted for 50.0% and 41.7% of postures，respectively，while level 4 risks exceeded 20.0% in both tasks. Conclusion The computer vision-based RULA assessment method constructed in this assessment demonstrates strong stability and accuracy under low-light underground mining conditions. The analysis identifies key tasks contributing to level 4 WMSD risks among support workers，providing reference for reducing the risk of WMSDs in support workers.

Key words: Work-related musculoskeletal disorders, Coal mine support workers, Rapid upper limb assessment method, Influencing factors, Work posture, High-risk areas