Abstract
The process of oil and gas drilling is a complicated and uncertain field practice. Moreover, uncertainty and fuzziness of formation is also a threat to drilling safety. Due to inaccurate cognition of objective environment and decision-making errors of subjective consciousness, underground complex conditions and serious accidents are caused.
Stuck pipe is one kind of downhole accident with high probability, complex handling difficulty and cost of manpower and material resources in drilling engineering. To deal with the stuck pipe accident, workers on the well-site are often required to judge the geological conditions of formation drilled, to adjust drilling parameters, drilling fluid performance and other methods to avoid and remove. Without fixed reference basis and determination methods, it is difficult to provide early warnings.
Through the collection, while drilling in formation, such as geological lithology, designed well structure, real-time drilling fluid performance, rock physical properties of backflow cuttings, and drilling engineering parameter et,al. Artificial neural network learning model is used to predict the risk probability of stuck pipe.
In this paper, face to formation collapse stuck pipe, after collected and analyzed the physical and chemical properties of backflow cuttings, drilling fluid system performance, well-site drilling data and other well-site drilling parameters according to the influencing factors of collapse and stuck pipe, combined with computer intelligent technology, and predicted collapse stuck pipe risk probability in advance.
Through well-field data collected, the occurrence of stuck pipe is predicted successfully. Through accumulated database, the accuracy of prediction model is constantly improved and modified. Early warning mechanism, timely adjustment of drilling fluid additives, maintenance of drilling fluid system performance and improvement of drilling parameters for different formations, increase drilling speed and avoid unnecessary waiting time
