Adaptive Fault-tolerant Control for Trajectory Tracking and Rectification of Directional Drilling

Motivated by the increasing demands on complex borehole trajectories in oil and gas directional drilling, an adaptive fault-tolerant control (AFTC) method for drilling trajectory tracking and rectification of rotary steerable system (RSS) is proposed by adopting actor-critic reinforcement learning (RL) and integral sliding mode control (ISMC) in the presence of system uncertainties and fault signals. Considering a discrete delay differential equation (DDE) with distance delays, uncertainties and fault signals, first we design an online learning framework via actor-critic RL and radial basis function neural network (RBFNN) in order to make drill bit can track pre-designed trajectory accurately and smoothly. Then in order to handle the fault signals problem, we utilize ISMC to eliminate it as weak as possible and rectify drilling trajectory which may derivate original direction caused by it. The system stability and convergence have been analyzed to ensure uniformly ultimately boundedness of tracking errors and fault-tolerant control signals. The proposed method would have wide application potentials in realizing the trajectory tracking and rectification with automatic operations of directional drilling. The effectiveness and accuracy of it are validated by simulation results with ramp and sine input signals.