Using AI to Treat Polluted Water

Baihe River is located in the Sichuan Basin, southwest China. (PHOTO: CFP)

In front of the main building of the Chinese Research Academy of Environmental Sciences (CRAES), there is a big rock bearing the words, "Mountains and rivers green are mountains of silver and gold."

It is a constant reminder for Zhang Lieyu, senior researcher at CRAES, to think about how to use science and technology to treat water.

For the past decade, Zhang has visited innumerable smelly waterways and polluted rivers. He describes himself as a "water doctor" who specializes in treating "sick" rivers and lakes. More than 80 polluted rivers have been returned to full health after his treatment.

Recently, the first home-made artificial intelligence (AI) river basin supervision system that Zhang participated in, has been successfully launched.

This system, which can monitor 96 square kilometers of the Baihe River in Sichuan Province in real time, can realize the functions of water quality traceability, pollution prediction and intelligent decision making.

Zhang Lieyu. (COURTESY PHOTO)

Treating polluted water like a "clinician"

Zhang Lieyu studied in Wuhan University of Technology as an undergraduate, and got his PhD in environmental science at the Institute of Hydrobiology, Chinese Academy of Sciences.

As early as a PhD student, Zhang determined to make protecting clean water his lifelong career.

Seventeen years ago, the ring waterway of Caoyang in Putuo District, Shanghai, was a renowned polluted waterway with black water and blue algae reproduction. It took less than a year for Zhang and his PhD instructor to find the source of pollutant and carry out targeted interventions. After countless practical research and purification experiments, the ring waterway was transformed into a green recreational park.

AI managing water treatment

As the general leader of a demonstration project for water treatment, Zhang designed an AI supervision system. "We use advanced technologies such as big data, AI, block chain, and the Internet of Things to create this intelligent watershed monitoring system," said Zhang. See page 2

According to him, the AI-based system can use fewer sensors to trace back source of emissions exceeding the standard level within 30 minutes, and predict possible pollutant diffusion with an accuracy of over 85 percent.

Besides supporting human management and scientific decision making on river management, this system also has great potential in cultivating environmental protection industries and expanding the scope of environmental protection services.

As a new tool for water management, the system also provides practical solutions to overall environmental supervision, which can significantly improve the efficiency of environmental management.

Global protection for clean water

A recent UN report on global clean water and sanitation suggests that almost two billion people are using water resources contaminated with sewage. And by 2050, it is projected that at least one in four people will be suffering from a shortage of water.

To ensure universal access to safe drinking water for all, effective water treatment need to be adopted. More international cooperation is also needed to encourage water efficiency and support treatment technologies in developing countries.

Zhang told S&T Daily, as soon as the river basin supervision system is complete, they intend to build a smart system for atmospheric environmental monitoring. By all accounts the era of an AI managed water environment has definitely arrived.