Bit bounce detection using neural networks

Real-time monitoring of bottom-hole assembly (BHA) conditions and drill-bit dynamic behavior is a critical factor in improving drilling efficiency. In fact, during rotary perforation, bit and drill-string dynamics can produce anomalous drilling conditions such as bit whirl, stick slip and bit bounce, just to name a few. These dangerous drilling conditions reduce the bit penetration rate and the safety of the operations. Measurement-while-drilling (MWD) may be employed to monitor BHA conditions and prevent unsafe operations, by detecting anomalous drilling conditions directly from the acquired signals, and to support real-time decisions. This paper presents a neural network system for real-time detection of the bit bounce phenomenon. The design of the neural network and its validation test are performed exploiting drill-bit data recorded downhole.