Digital innovation: Machine learning for optimizing reverse osmosis

New technology developed by the Alberta-based startup, Synauta, is revolutionizing the way desalination and wastewater plants operate. Their patented machine learning algorithm is helping optimize reverse osmosis water treatment systems across the globe, with the potential to save the global industry more than 12 million tonnes in carbon dioxide emissions every year.

Mike Dixon, CEO and founder of Synauta, describes reverse osmosis as the ‘heart of the desalination process’. It uses high pressure pumps to force salt water across a semi-permeable membrane, filtering out salt and other impurities. It’s a technique used in water reuse plants across Alberta, helping the oil and gas, mining and food and beverage industries treat wastewater.

A reverse osmosis plant for water desalination.

Operating a reverse osmosis system, however, is a complex and expensive task. From planning and maintenance to energy and chemical consumption, operators are challenged to meet water production and environmental targets while keeping expenses low. Managing these systems manually is often further complicated by a lack of available personnel, complex calculations and over or under cleaning membranes.

Synauta’s solution is a machine learning algorithm for reverse osmosis that optimizes both energy and chemical use.  It integrates with existing control software so plant operators can easily and reliably manage and optimize variations and trends in water conditions while adapting to fluctuations in feedwater. Ultimately, this technology minimizes energy use, contributes to a lower environmental footprint and lowers operating costs.

Machine learning also has the potential to solve problems unique to many remote and Indigenous communities. Operators in these areas may lack the training required to run desalination or wastewater treatment plants or have limited capacity to do manual optimization work. Synauta’s technology can support remote monitoring and determine optimal energy use for a plant in minutes. Its unique algorithm reduces energy consumption by up to 15 per cent and saves up to 25 per cent in the costly chemicals required for cleaning reverse osmosis membranes.

Synauta’s products and services have a strong market in Canada and abroad. The company, which was recently acquired by Gradiant International Holdings, is doubling their team in Canada this year and their machine learning for reverse osmosis has been deployed with customers in Singapore, the Middle East, Australia, Spain, Chile, Africa and California.

Learn about the supports available through our Digital Innovation in Clean Energy Program.

This article is part of Making Waves, an annual Alberta Innovates publication highlighting water solutions we support. You can read more stories from this issue below.