A New Drill Bit Selection Method Based on Machine Learning and Intelligent Optimization Process

Abstract

ABSTRACT In the process of oil and gas drilling, the drill bit is the direct tool to break rock. The selection of drill bit needs to consider formation factors, drilling construction factors, field size requirements and other factors. At present, the optimization methods mainly include mechanical analysis, cost analysis, neural network method and other methods, but there are some limitations. In this paper, three categories and more than ten kinds of input parameters are selected, and the rate of penetration is predicted by BP neural network algorithm. Then, the bit library is established according to the actual bit used in the field, and the parameters of the bit are optimized by intelligent optimization process, and the actual demand is taken as the constraint to select the most suitable one. The error of the ROP prediction model was within 10%, and the ROP was increased by 11.67% on average with the optimized bit. In this paper, the rock breaking rule and complex mechanism of drill bit are stored in the model, and the drill bit optimization is carried out according to the actual requirements, which can provide a certain degree of support for the field drill bit optimization. INTRODUCTION With the development of oil and gas resources to deep and complex formations, the number of horizontal wells drilled continues to increase. At the same time, the horizontal well drilling area continues to expand, and complex geological conditions in some areas are revealed, and hard, highly abrasive formations are prevalent in most blocks. To a certain extent, the drilling operation progress is slow, the single drill bit is less advanced, and frequent drilling and replacement are required, which increases the non-drilling time and consumables consumption, thereby increasing the drilling cost. At the same time, as the length of the horizontal section continues to increase, the construction is becoming more and more difficult. This puts forward higher requirements for drilling construction, especially for the feed and life of a single PDC drill bit, the life of drilling tool combination, the adaptability of drill bit and drilling tool combination and formation. If the drill bit can be reasonably selected according to the formation characteristics, construction requirements, engineering parameters and other factors, so that the ROP and single drill bit advance greatly increase, and even achieve a single drill (only 1 drill bit per spud), it will greatly improve drilling efficiency and reduce drilling costs (Fear et al., 1994). However, as the use of PDC drill bits increases year by year, more new varieties and models of PDC continue to appear, and their scope of use continues to expand. Moreover, there are many parameters of PDC drill bits, which are more sensitive to formations, and slight changes in drill bit parameters may produce large effect differences. Therefore, how to establish the relationship between drill bit parameters, formation anti-drilling characteristics and construction parameters, and refine and optimize drill bit is a problem to be solved.