Based on high order statistics, the non-parametric and parametric methods, which estimate the time delay between the receiving and emissive signal, are presented in wellhead acoustic emissive gas influx detection. By analyzing the signal propaga-tion process, it is seen that the noises, received by two pressure sensors can be modeled as spatially correlated Gaussian noise ap-
proximately. The fourth order cumulant is blind to arbitrary Gaussian noise, so these methods may significantly increase gas influx detection probability compared to the conventional cross-correlation method. Computer simulation confirms their availability.
