Real-Time Drilling Parameter Optimization Model Based on the Constrained Bayesian Method

To solve the problems of low energy efficiency and slow penetration rate drilling, we took geological data adjacent wells, real-time logging data, downhole engineering parameters as inputs; mechanical specific unit footage cost multi-objective optimization functions; machine pump equipment limit constraint condition. A constrained Bayesian algorithm model was established for solution, drilling such weight-of-bit, revolutions per minute, flowrate were optimized in real time. Through a comparison with NSGA-II, random search, other algorithms, application results example show that has good effect while maintaining timeliness. It is suitable parameters, can aid driller identifying potential tapping area, provides decision-making basis avoiding low-efficiency rock-breaking working area improving efficiency.