ABSTRACT
In this paper, the quantile impact value (QIV) method was proposed as a weight coefficient distribution method to evaluate the influence of proximate analysis components on the low calorific value (LCV) of blended coals. Through data analysis technology, a new prediction model based on the QIV and the support vector regression (SVR) has been proposed to estimate the LCV. The comparisons between different analysis results reveal that the QIVW-GA-SVR model proposed here is consistently better than those models without weight coefficient distribution.
Highlights
Quantile impact value is proposed as a weight coefficient distribution method.
The QIV can identify the importance of proximate analysis constituents on the LCV.
A reliable model has been established to predict the LCV of blended coals.
The QIVW-GA-SVR model possesses good prediction accuracy and good generation performance.
Additional information
Notes on contributors
Minfang Qi
Minfang Qi received her Ph.D. degree of Thermal Engineering from North China Electric Power University, China, 2016. She is currently a postdoctoral fellow at National Institute of Clean-and-Low-Carbon Energy, Beijing, China. Her major research areas include operation optimization, data analysis and modeling, and energy internet.
Huageng Luo
Huageng Luo received his Ph.D. degree from Georgia Institute of Technology, USA. He had worked for General Electric for 18 years. He is currently a professor in School of Aerospace Engineering, Xiamen University. His major research areas include signal processing, parameter identification and fault diagnosis, structural dynamics, and vibration control.
Zhongguang Fu
Zhongguang Fu received his Ph.D. degree from North China Electric Power University, China, 1999. He is currently a professor in School of Energy, Power and Mechanical Engineering, North China Electric Power University. His major research areas include operation optimization and modeling of complex thermodynamic system.