Rare earth resource in fly ashes from coal power plants of China: Based on machine learning model and unit-based estimation

Chang Liu, Yi Yang, Long Chen, Jiayuan Wu, Yuan Sun, Mingzhe Han, Xingpan Guo, Maoyong He, Zhangdong Jin

Research output: Contribution to journalArticlepeer-review

1 Scopus citations

Abstract

Coal fly ashes (CFAs) are an alternative resource of rare earth elements and yttrium (REY). China is the largest producer of CFAs in the world and is likely to hold substantial reserves of CFA-REY resources, while nationwide research on REY resource in Chinese CFAs is lacking. In this work, CFA samples were collected from 118 coal - fired power plants (CFPPs), including eight subjected to long-term monitoring. Based on this, a machine-learned (ML) REY concentration predictive model was developed with a deviation of 16 %, which showed REY concentration and the proportion of air-dried-basis ash yield in coal, and CFPP boiler type were the three governing factors regulating REY concentrations in CFAs. Using this ML model and a unit-based database of Chinese CFPPs, REY concentrations in CFAs from 1062 additional CFPPs were predicted, who accounted for 89.2 % of national coal consumption. Promising CFA-REY resources were defined as those containing ≥300 mg/kg REY in CFA, and were concentrated in North (Inner Mongolia, Shanxi, and Hebei Provinces), East (Shandong, Jiangsu, Zhejiang, and Anhui Provinces), South (Guangdong Province) and Southwest (Guizhou Province) China. Moreover, using a unit-based estimation model, the total amount of rare earth oxides from unutilized CFAs with REY recovery potential in China is estimated to be about 74,300 tons/y, which would meet six months of global demand and have a gross value of US $ 4.2 billion.

Original languageEnglish
Article number104743
JournalInternational Journal of Coal Geology
Volume303
DOIs
StatePublished - 17 Apr 2025

Keywords

  • China
  • Coal fly ash
  • Random-forest machine-learned statistical model
  • Rare earth element
  • Unit-based resource estimation model

Fingerprint

Dive into the research topics of 'Rare earth resource in fly ashes from coal power plants of China: Based on machine learning model and unit-based estimation'. Together they form a unique fingerprint.

Cite this