The automatic optimisation of drilling performance
The drilling industry, along with many others, is becoming increasingly competitive, demanding greater efforts to improve safety and reduce costs. For this reason, companies are progressively looking towards computerised automation to enhance performance. Unlike most industries however, the drilling industry has been slow to take advantage of the advances in computer and automation technology. Only recently have automatic operations such as tubular handling been placed under computer control. These activities relate to peripheral mechanical handling problems which are relatively easy to solve. The concept of an automatic intelligent drill, capable of making its own or assisted decisions about drilling parameters such as weight on bit or rotational speed, may seem remote and far into the future. Research in drilling automation, at the University of Nottingham, has the ultimate objective of achieving computerized drill control through the the application of an intelligent knowledge induction system. At the University, a laboratory rig has been developed with such a system installed. Decisions for optimal performance are based on either maximum penetration or minimum cost drilling. The system has a self-learning capability, allowing a progressive improvement in performance. The prototype system is currently undergoing trials, using real data collected while the laboratory rig is drilling and artificial data. The results are very encouraging and demonstrate the feasibility and advantages of optimised drill performance. This thesis describes the design and development of this drill optimisation scheme produced by the author. Both the theory behind the optimisation system, and the results of the initial phase of Laboratory testing are included.