In Oil & Gas drilling, adverse conditions may have to be inferred from real-time data collected from various surface and downhole sensors and the onset of any anomalous event must be predicted well in advance to prevent severe damage to the equipment.  However, noisy sensor measurements and the complex interactions among the measured drilling parameters such as Torque, RPM, Weight-on-Bit, Hook load, Mud Flow make the task of inference formidable and inaccurate.

Our unique PREEMPT framework starts with the denoising of measurement data and feeds the denoised signal or its compact representation to appropriate AI and ML algorithms to comprehend the interactions among multiple parameters with an increased degree of accuracy. Deeply ingrained in the framework is an intuitive ensemble approach to prediction for improved confidence on the prediction of such rare events.