Identification method for safety hazard behavior in offshore drilling operators

Abstract

During offshore drilling operations, both human inexperience and fatigue can contribute to drilling stagnation, overflow, blowout, and other accidents. Despite significant advancements in the automation of equipment and facilities on offshore drilling platforms, humans continue to play a central role in drilling operations. Utilizing the TPE-LightGBM model, a method for identifying safety hazard behavior among offshore drilling operators based on eye movement data was proposed. The main contributions of the present study are as follows: Firstly, the application of eye-tracking technology to identify safety hazards among offshore drilling operators overcomes the lack of risk perception and monitoring methods for mitigating the safety hazards experienced by operators. Secondly, a method utilizing composite hotspot mapping for eye movement to determine sensitive area divisions during offshore drilling operations was proposed. This method aims to reduce the influence of irrelevant, missing, and incorrect eye movement data on recognition results. Thirdly, in conjunction with qualitative selection and analysis of eye movement laws, a feature selection method for eye movement in the context of safety hazard behavior in offshore drilling was proposed. This method enhances the efficiency of recognizing safety hazard behavior. Lastly, the TPE algorithm was introduced to enhance the LightGBM model, addressing the issue of inaccurate identification of safety hazard behavior in offshore drilling operators. The results indicate that the proposed method effectively enhances the accuracy and execution efficiency for identifying safety hazard behavior among offshore drilling operators.