Combining flame monitoring techniques and support vector machine for the online identification of coal blends

Abstract

The combustion behavior of two single coals and three coal blends in a 300 kW coal-fired furnace under variable operating conditions was monitored by a flame monitoring system based on image processing and spectral analysis. A similarity coefficient was defined to analyze the similarity of combustion behavior between two different coal types. A total of 20 flame features, extracted by the flame monitoring system, were ranked by weights of their importance estimated using ReliefF, a feature selection algorithm. The mean of the infrared signal was found to have by far the highest importance weight among the flame features. Support vector machine (SVM) was used to identify the coal types. The number of flame features used to build the SVM model was reduced from 20 to 12 by combining the methods of ReliefF and SVM, and computational precision was guaranteed simultaneously. A threshold was found for the relationship between the error rate and similarity coefficient, which were positively correlated. The success rate decreased with increasing similarity coefficient. The results obtained demonstrate that the system can achieve the online identification of coal blends in industry.

中文概要

目的

混煤在锅炉燃烧中应用广泛。本文利用火焰监测技术提取混煤燃烧的火焰特征量,获取最优的特征量组合,并研究混煤相似度对其辨识错误率和正确率的影响。

创新点

1. 利用ReliefF 算法和支持向量机(SVM)算法定量分析各个火焰特征量在煤质辨识过程中的重要性,获取最优特征量组合;2. 定义混煤的相似度,并分析相似性对其辨识错误率和正确率的影响。

方法

1. 利用火焰监测技术提取火焰图像信号和火焰光强信号,提取20 个火焰特征量(图3 和4、表1);2. 利用ReliefF 算法计算20 个特征量在煤质辨识中的重要性(图7);3. 利用SVM 算法分析特征量个数对煤质辨识正确率的影响,确定最优特征量组合(图8)。

结论

1. 在煤质辨识过程中,结合ReliefF 算法和SVM算法可以将特征量个数由20 降至12,并能保证辨识准确度;2. 混煤与其组分煤种的相似度主要受组分煤种的挥发份含量及掺混比例影响;3. 辨识错误率与相似度之间存在一个阈值,当相似度低于该阈值时,辨识错误率为0,当相似度高于该阈值时辨识错误率与相似度呈正相关;4. 辨识正确率随着相似度的升高而降低。