With the help of machine learning, toxic materials are detected in water

With the help of machine learning, toxic materials are detected in water ...

Waste materials from oil sands extraction, stored in tailings ponds, may be harmful to the natural environment and neighbouring communities when they leach into groundwater and surface ecosystems.

The challenge for the oil sands industry is that proper analysis of toxic waste materials was difficult to accomplish without lengthy and complex tests. And there is a huge backlog. For example, in Alberta alone, there are an estimated 1.4 billion cubic metres of fluid tailings, according to Nicolas Peleato, an assistant professor at the University of British Columbia''s Okanagan campus.

UBCOsSchool of Engineering''s faculty of researchers has developed a new, faster and more reliable technique of analyzing these samples. This is the first step, according to Dr. Peleato, but the results are promising.

As a way to protect public and aquatic ecosystems, current methods require the use of expensive equipment, and it may take days or weeks to get results. There is a need for a low-cost system to monitor these waters more frequently.

The researchers combined with masters student Maria Claudia Rincon Remolina used fluorescence spectroscopy to quickly detect key contaminants in the water. They also compiled the results through a modelling program that accurately predicted the water''s composition.

The composition may be used as a reference for further testing of other samples, according to Rincon. Researchers are using a convolutional neural network that processes data in a grid-like topology, such as an image. It is similar, according to Rincon, to the type of modelling used to classify hard to identify fingerprints, facial recognition, and even self-driving cars.

According to Rincon, the modeling takes into account the water quality''s background and can distinguish difficult to detect signals, and as a result, it can deliver high-quality results.

The researchers examined a range of organic compounds that are dangerous, including naphthenic acids, which can be found in many petroleum sources. By employing high-dimensional fluorescence, they can identify the most types of organic matter.

Peleato claims that the modelling technique investigates key materials and identifies the samples'' composition. The results of the initial sample analysis are then analyzed in powerful image processing methods to enthuse complete results.

While the results released today are encouraging, both Rincon and Dr. Peleato caution that the technique should be further evaluated at a larger scaleat, which means there may be potential to include further screening of harmful substances.

Peleato declares that this potential screening tool is the first step, but it has a few limitations because not all chemicals or naphthenic acids can be detected only fluorescently. And the technology will have to be expanded for future, more extensive testing.

While it will not replace current analytical methods that are more accurate, Dr. Peleato claims this approach will allow the oil sands industry to accurately screen and treat its waste materials. This is a necessary step to continue to meet the Canadian Council of Ministers of the Environment standards and guidelines.

The research is published in the Journal of Hazardous Materials and is funded by the Natural Science and Engineering Research Council of Canada''s Discovery Grant program.

You may also like: