Title :
Online coal calorific value prediction from mutiband coal/air combustion radiation characteristics
Author :
Cheng, Yanting ; Xu, Lijun ; Li, Xiaolu ; Guo, Zhanshe
Author_Institution :
Sch. of Instrum. Sci. & Opto-Electron. Eng., Beihang Univ., Beijing, China
Abstract :
A SVR modeling based online coal calorific value prediction method was introduced. Through wavelet transform and PCS process, coal flame radiation characteristics were extracted. Through SVR modeling, relationship model between radiation characteristic variables and coal calorific value was established. Proper SVR construction parameters were chosen through grid computing. Experiment results showed good coherence between SVR model prediction and laboratory results of coal calorific value. Proposed SVR based system relies on low cost multiband photoelectric sensors, which is easy to be installed at production spot, while satisfying system performance could meet practical needs of online operation adjusting to enhance production efficiency and diminish pollutant emissions.
Keywords :
boilers; coal; grid computing; mining industry; photoelectric devices; power engineering computing; production engineering computing; pulverised fuels; regression analysis; support vector machines; wavelet transforms; PCS process; SVR modeling; coal flame radiation characteristics; grid computing; multiband air combustion radiation characteristics; multiband coal combustion radiation characteristics; multiband photoelectric sensors; online coal calorific value prediction method; support vector regression; wavelet transform; Computational modeling; Kernel; Laser modes; Measurement by laser beam; Predictive models; Production; Training data; PCA; SVR; coal calorific value; multiband photoelectric sensors;
Conference_Titel :
Instrumentation and Control Technology (ISICT), 2012 8th IEEE International Symposium on
Conference_Location :
London
Print_ISBN :
978-1-4673-2615-5
DOI :
10.1109/ISICT.2012.6291636
