JD iCity, a business unit of JD Digits (the digital technology arm of e-commerce firm JD.com) focusing on empowering intelligent cities with AI and big data, won a bid of digitalization project for a large thermal power plant in China on December 14, 2020. The project aims to apply cutting-edge AI technologies to transform the plant’s power generation management systems to improve boilers’ efficiency and reduce coal costs and pollution. The plant where its being rolled out belongs to China Energy Investment Corporation and is based in Langfang, Hebei province, which borders Beijing. As part of a grand plan, the AI control system will be implemented on more than 70 power boilers in Langfang, the steel heartland of pollution-prone Hebei province. It has beeen in use in Nanning city in the southwestern Zhuang Autonomous Region of Guangxi since last year.
JD won the bid thanks to its boiler combustion optimizing control system. The AI control system, automatically adjusts a range of variables for thermal boilers in real time, including the coal feeding process, air distribution and water vapour levels, according to JD Digits. At the same time, the system can generate operational plans based on self-learning, and big data analysis provides helpful guidance for plant workers, reducing their reliance on high-level experts.
The system has already been implemented in several power plants in China so far, and has proven effective in solving the complex problem of dealing with high-dimensional continuous variables when the boiler is burning in real time. Through intensive self-learning and the adaptive nature of AI, the system is capable of monitoring multiple core measuring points and over 70 control parameters of the boiler, such as the temperature, coal feed rate, air intake damper positions and more，and also implements dynamic tuning for the best combustion effect.
The combustion efficiency of boilers fitted with the new AI-control system increased to 93.9 per cent from 92.75% for boilers without it in a March test, according to the company, which added that an increase in boiler thermal efficiency of only 0.5% would save the country 7 billion yuan (US$1.07 billion) a year in coal consumption and pollution control costs.
Thermal power generation will remain the main source of electricity in China for years to come. This industry has high demand to adopt new technologies to upgrade and modernize itself. China remains heavily dependent on fossil fuels and accounts for nearly 30 per cent of global greenhouse gas emissions. However, in recent years the country has taken steps to rein in the pace of new coal power plant construction to reduce overcapacity and pollution levels. In September, the country vowed to become carbon neutral by 2060.
“The algorithms used in the thermal power optimisation project could also be used in similar industries and related fields in the future … to bring efficiency gains to more industries and create greater value to society,” said Xianyuan Zhan, a data scientist at JD Digits. The team built the system and the algorithms on which it is based by studying an array of internal data on the boilers, and maintain the best combustion conditions, and reduce coal consumption and nitrogen oxide emissions.