Genetic Algorithm Optimized Deep Learning Model to Forecast Ultra-Short-Term Concentration of Fly Ash From Coal Fired Power Plants

 

In the process of advancing carbon peak and carbon neutrality, energy savings, emissions reduction, the transformation and upgrading of coal-fired power plants are of great significance in China.

The team has established an ultra-short-term time series forecasting model, based on deep learning. The model predicts changes in fly ash concentration emitted from a power plant 1 – 2 minutes in advance. This serves as the core of a control system to dynamically adjust key parameters of the electrostatic precipitator, such as power and voltage, to maximise the efficiency of dust removal and save energy. Commercialisation can be carried out through an Energy Performance Contract, and fees will be charged based on the actual electricity cost saved for the power plant.

 

 

Team members

Mr HE Yingjie* (Alumnus, Dept. of Marketing, CityU)
Dr CHU Ying Hao (Shenzhen University)

* Person-in-charge
(Info based on the team's application form)

Achievement

1. Entrepreneurship: HE Yingjie, Forbes China 30 Under 30, 2019
2. Start-up Competitions:
- Outstanding Enterprise Award, the 8th China Innovation & Entrepreneurship Competition
- National Final in Smart Manufacturing, 2019