Real-time monitoring of the performance of marine seismic sources

The performance of marine seismic sources (air guns) is of great importance to achieve high resolution images of the seabed and subsurface. In challenging offshore conditions, mechanical or electrical failures within the air gun assembly can cause insufficient performance. Air leaks result in pressure drops and are currently recognised as one of the major issues causing downtime and data degradation, often requiring repeating survey lines. As the number of air guns involved in modern surveys increases, it becomes harder to identify potential air leaks in time, increasing the cost of such failures significantly.
To address this issue, this paper presents a monitoring framework for real-time and fast fault detection on individual operating air guns, based on Principal Component Analysis (PCA) and Gaussian Mixture Models (GMMs). For this purpose, we use the output data typically recorded by source controllers employed for air gun synchronisation. The framework exploits the potential of timing sensor measurements not used in seismic post-processing, to develop a robust diagnostic approach. It is successfully benchmarked against data from two types of surveys (dual and triple source). The present approach can be used to identify normal operation and early faults occurring in air guns without requiring a large database of historic data. It can be effectively applied for air gun monitoring during seismic surveying, enhancing online Quality Control (QC).