Abstract
Automatic under suspended load detection system is a term for an image processing method which is used to identify compliance and safety aspect of crane or lifting & rigging operations in drilling & completion activities especially under suspended load for line of fire prevention which possibly leads to a serious incident and fatality. According to company data statistics, approximately 30% of safety findings are contributed by lifting & rigging operations. As a state-owned company that operates one of the largest fields in Indonesia with an extensive drilling and well intervention programs, Pertamina Hulu Rokan (PHR) commits to protect the safety of their people by reducing lifting & rigging operations safety findings and improving its monitoring process. The company has taken the initiative to explore any digital alternatives that can be applied such as utilization of computer vision and artificial intelligence in online Closed-Circuit Television (CCTV) units to enable prevention of under suspended load case in drilling & completion operations by using deep learning algorithm such as Yolov4 and/or Faster R-CNN.
During development process, the team has several technical challenges to be addressed such as the diversity of lifting and rigging scenarios such as lifting direction, lifting equipment type, and load variety to capture and detect under suspended load zone in 2-dimensional image using straightforward logic. The first step towards building this system is collecting lifting and rigging operations image datasets from various rig areas and with different lifting scenarios and equipment. Deep learning algorithm such as Yolov4, Faster R-CNN are being used to train the model using the dataset which has been labelled on specific objects related to the lifting operation such as crane boom, crane hook, crane loads, crane tires, crane cabin, and crane box to construct under suspended load zone on the given scenarios. The Preliminary results indicate that the method has been useful to identify under suspended load zone and deliver real time automatic notification during lifting and rigging on a drilling or well intervention operations and prevent the safety risk exposure of our personnel.
