Journal article
An investigation of the flame detection method based on self-adaptive wavelet conversion
Journal of Engineering for Thermal Energy and Power, Vol.21(6), pp.594-597
11/01/2006
Abstract
In view of the features specific to present-day furnace flame detection systems, such as great difficulty in extracting signal characteristics, low signal-noise ratio and difficulty in making an accurate decision, a judgement method was proposed based on wavelet conversion in combination with a BP neural network. This has been undertaken through an acquisition and analysis of experimental data of pulverized coal combustion in a furnace. Under the above method the collected noise-containing signals will undergo a multi-dimensional wavelet decomposition and characteristics extraction. Thereafter, the signals are subject to a soft threshold value de-noising treatment with the pretreated information serving as a training input to the neural network. A furnace-flame combustion experiment at a power plant has proved that the time-frequency localized analytic method can improve the signal-noise ratio and more effectively identify the combustion state of a flame and is assessed as possessing a wide-ranging practical
Details
- Title: Subtitle
- An investigation of the flame detection method based on self-adaptive wavelet conversion
- Creators
- Li-Min AoJian-Hua LiXuan SongHan Li
- Resource Type
- Journal article
- Publication Details
- Journal of Engineering for Thermal Energy and Power, Vol.21(6), pp.594-597
- ISSN
- 1001-2060
- Language
- Chinese
- Date published
- 11/01/2006
- Academic Unit
- Industrial and Systems Engineering; Injury Prevention Research Center
- Record Identifier
- 9984187045502771
Metrics
17 Record Views