Objective is to produce hardware and software for detecting eventual quantities of natural gas leaks from storage in caverns under the ocean, besides organizing an experimental acoustical database of natural gas leaks

Developing an underwater acoustical monitoring system that is capable of detecting, quantifying, and locating natural gas leaks in the bottom of the ocean is the goal of one of the projects of the FAPESP Shell Research Centre for Gas Innovation (RCGI). Entitled Passive acoustic monitoring system for CO2 leakage detection, Project No. 33 is a part of the research program focusing on CO2 abatement, and its objective is to monitor underwater salt caves that could possibly be used for carbon storage. The study seeks both to improve the hardware for capturing underwater sounds and to develop software to process those sounds, as well as to create an experimental database of acoustical signals.

The project is a huge challenge. Not only because the salt cave technology for storing CO2 is such a new thing, but also because underwater acoustics is an area that is studied very little in Brazil. Furthermore, the team does not have an acoustical database ready to serve the purpose of the research. “We need experimental data on leakage that can validate the Artificial Intelligence algorithms developed to detect these phenomena. But where is there a lead to be recorded? So, we have to produce these data,” says Research Coordinator and Professor Linilson Rodrigues Padovese, who is also the Coordinator of the Acoustics and Environment Laboratory (LACMAM) of the Department of Mechanical Engineering of the Polytechnic School of the University of São Paulo (Poli-USP).

Sound signature – According to Padovese, this is a database of acoustical sounds recorded both in laboratories and in the field, that is, the ocean. The most important sounds are those made by water bubbles – or streams of bubbles – caused by natural gas being released under the water. “Generally, a leak of this type forms a stream of bubbles, which make noise when passing through the water. Obviously, we will not know what to say, by the sound alone, if the bubbles are of CO2 or of any other gas. But, depending on the characteristics of the sounds recorded, and if there is a cave down there on the bottom, it is possible to relate it to that cave (or not),” he explained.

Padovese said that each acoustical event has a “signature” and that each acoustical event has a “signature”, and it is exactly that specificity that allows isolating the information that is most important. “The production of natural gas resulting from the deterioration of organic material, for example, also forms a stream of bubbles, but the duration of that recording is different from that which comes from leakage. That is, they have different behaviors,” she says. “Therefore, if I know that in a certain region there is a salt cave with natural gas in it and I detect the sound of a stream of bubbles the lasts for several days, it is quite likely that it is related to a CO2 cave, and not to some other phenomenon, like the decomposition of organic material,” he added.

Field trip – The Project team took its first field trip at the end of January to collect experimental data in a simulated leak. According to Padovese, it was a basic experiment, in order to see if the simulation would work properly and to prepare future trips, which will likely take place yet in the first half of this year.

“We did a simulation of a leak near the Santos Basin, at a depth of 15 meters. We went out in two vessels: one had the recording equipment, connected by cables to three hydrophones (underwater microphones), which were placed in the ocean. The other carried the leakage simulator,” he said. A compressed air tank and a hose were used, with instruments for controlling the pressure and rate of leakage of the air, as well as special orifices for the air to escape. “This is because the characteristics of the stream of bubbles depends on pressure, rate of leakage, and the type of openings for the bubbles to escape. The stream can be formed by a number of small bubbles or by larger bubbles, and all of this produces different noises.”

Leakage data were also collected while varying the distance of the simulation equipment from the hydrophones. “We took leakage measurements using the same pressure, the same rate of leakage, and the same point of escape at a variety of distances, in order to teste the equipment’s range of detection,” he added.

These tests will result in articles that will be submitted to the Brazilian Congress on Mechanical Engineering (COBEM), which will be held this year, and another article for an international trade publication. “At this present time, we are focused on data that allow the detection and quantification of natural gas leaks. The location presupposes other types of approach and will be occupied with it later,” Padovese stated.

Special hardware and software – Padovese has a whole arsenal of equipment for this type of research – a good share of which has been developed or improved by him and his team. Besides the hydrophones, the laboratory has recording equipment that can also be used underwater. “The recording equipment can be used in a boat or submerged, inside a hermetically sealed box. The use of one or the other depends on the type and objective of the mission, its duration, the quantity and specific nature of the data that one wishes to record,” he says.

In order to record for long periods of time, or to make simultaneous records at several different points, the equipment generally used has the hydrophone and all of the electronics under water. They are “autonomous” devices that run on batteries and are placed on the sea bottom for the duration of the mission – one day, one week, or several months.

“We have developed recorders with special characteristics for these autonomous devices, which we call ‘ocean pods’. They are cylinders that can be made of PVC, aluminum, or stainless steel, depending on the depth at which they will be placed – containing electronics (recorder and signal conditioners) and batteries – thus allowing an autonomy of up to six months and a recording capacity of 1Tb, or more.”

The research team – consisting of one post-doc, two Ph.D. candidates, one Master’s candidate, and one science undergraduate – also is dedicated to exploring methods of analysis to improve audio capture and processing. “We develop audio detection software, as well; we have created a laboratory technology and a methodology for calibrating the hydrophones. At the same time, we are also developing simulation software for streams of bubbles and the sound that they generate, as well as computer programs for data analysis and visualization. One important aspect of the project is the development of algorithms using Artificial Intelligence techniques to improve the efficiency of detecting and quantifying leakage.”

Various applications – The Professor reminds that the implementation of a monitoring system of this type will also make more detailed inspection procedures possible for those who are responsible for the infrastructure. “If I am monitoring a pipeline or a region of the sea floor and am detecting something that should not be there, that could serve as an alert and generate other procedures that are more to the point: for example, an underwater drone could be put into action to check and see if there is an abnormal situation.

“Actually, we are developing a monitoring technology that will serve to monitor other things, besides natural gas leakage. What happens is that the technology developed for detecting leakage could, eventually, with small adaptations, end up serving other activities,” he says.

Padovese calculates that the acoustics database and all of technology inherent to it will be ready by 2021, which is the time for the duration of the Project. By then, he will have to overcome another big challenge: find researchers interested in working in the area. “I have had a Doctoral scholarship for a researcher in hand for many months and no one has appeared. The formation of human resources is very important in this context,” he states.

According to the Poli/USP Professor, he knows of no other research groups working on this subject in Brazil, except for the team he lead.