These are records of underwater noise made by moving plumes of bubbles, in order to acoustically monitor possible gas leaks rising from the ocean floor

In order to acoustically monitor gas leaks from the ocean floor, at several thousand meters below the surface of the water, researchers from the FAPESP Shell Research Centre for Gas Innovation (RCGI) are establishing a database at the University of São Paulo (USP) – Brazil’s first research institution to have records of this nature. The collection of acoustical data, as well as the hardware and software they are developing for collecting and analyzing the sounds, is linked to another RCGI initiative that analyzes the feasibility of building caverns in the rock layers above the pre-salt layer. These caverns would be used to store CO2 after it is separated from natural gas. The proposal is that the database would provide information for monitoring possible CO2 leaks from this type of reservoir.

The database is gathered from material obtained in the field and in the laboratory. It is constantly growing and, currently, consists of approximately 60 hours of recordings. The researchers are recording the sounds of the bubble plumes originating from simulated underwater gas leaks, at different rates of flow, pressures, and locations. “Leaks usually form a bubble plume that makes noise when it is generated under water. Our goal is to learn the “signature” of the sounds of the bubble plumes, in order to be able to detect, classify, and differentiate them from other type of sounds, explains Professor Linilson Rodrigues Padovese, Coordinator of this RCGI project and also of the Acoustic and Environment Laboratory (LACMAM) of the Department of Mechanical Engineering of USP’s Polytechnic School (Poli-USP).

Therefore, the team began to collect experimental data regarding leaks, for the purpose of validating the Artificial Intelligence (AI) developed for detecting these phenomena. “We performed simulations of leaks, in order to obtain the data. There were three field trips: two to the ocean and one to a reservoir, and we were able to significantly advance the database. We also used the tank in our Poli laboratory, as well as the diving tank at CEPEUSP, USP’s multi-sport club. We already have a good database, with different sizes of bubble plumes and for different rates of flow.”

To simulate leaks, the researchers used compressed and hoses in a reservoir, with instruments for controlling the pressure and flow rate of the air leakage, as well as outlet orifices with different diameters. “For each recorded sound, we know what the flow rate of the air was. When we simulated the leaks, we had already stipulated the flow rate: for example, 1 liter per minute; 10 liters per minute. Therefore, we know the signature sound of the plume for each of the flow rates we had established. That is: we have data for making comparisons.”

New Challenges – According to the Professor from Poli-USP, detecting the noise of the bubble plumes is no longer a problem. “We are now certain that we can detect and differentiate flow rates of 2, 5, and 10 liters per minute at depths of up to 40 meters. Now, our challenge is to create a pilot project for testing the methodology, the software, and the hardware under conditions analogous to those of the desired application. We need to assess whether or not the conclusions we have made, thus far, are applicable at greater depths.”

The engineer explains that, for example, both a leak and the passage of a ship can alter the background sound from the seabed. However, the noise of the ship passes by in just a few minutes, while the leak continues. Therefore, as the time goes by, there is greater confidence that that particular noise could truly be related to a leak.

“I believe that the analysis of noises will be even easier under the conditions of deep salt caverns, because at those depths it is more silent. At a depth of two thousand meters, the acoustical noise is more homogeneous, that is, it is easier to detect something that should not be there, like a leak. When we went to the ocean, we performed experiments near the Port of Santos, where there are noises of all types: dredges, large ships entering, fishing boats going by… so, we believe that if we are able to detect a leak under those conditions, then detection at great depths will be even easier.”

According to Padovese, it is not likely that a leak will cause just one bubble plume. “We will probably have a group of plumes, which makes more noise. I would say that the conditions at great depths would be more favorable, although from the standpoint of the hardware used for capturing the sounds, those conditions will be more demanding. But that is an engineering issue.”

Capture Strategies – The Professor explains that an equipment option the laboratory is developing for monitoring underwater salt caverns is an autonomous recorder, which they call an Ocean Pod: a cylinder (of PVC, aluminum or stainless steel, depending on the depth at which it will be placed) that contains a recorder, signal conditioners, hydrophone, and batteries, providing autonomy of up to six months. He says that the big question now has to do with “seeding” the equipment for capturing the sounds from the seabed.

“We will not work with just one Ocean Pod, but rather we will seed an area with several of them, because the leak could percolate through the geological structure of the cavern at any given point. Therefore, we need to spread the equipment in the area surrounding the cavern. But we are still working on the distance required between each piece of equipment. Each one needs to be located at the edge of the coverage area of the other, so that there will be no blind spots.” He stressed that a variety of monitoring strategies exist, and that will likely be dealt with by the team of the other project, which has the objective of finding the salt caverns and studying their feasibility.