Identification Method of Flow Regime Images Based on Wavelet-Fractal Features and PNN in Gas-solid Fluidized Bed
In this paper, an approach based on wavelet-fractal features and PNN is developed for flow regime images identification in gas-solid fluidized bed. Two-dimensional wavelet transform was adapted to decompose images into four channels; six kinds of fractal dimension were extracted from those channels. Thus the wavelet-fractal feature eigenvectors of flow regime have established. Finally, PNN were trained by using these eigenvectors as flow regime samples. In the end, the flow regime images intelligent identification is realized. The experimental results demonstrated its can effectively identify five typical flow regimes of gas-solid two- phase flow in fluidized bed. The whole identification accuracy is 95％, opening up a new avenue for the flow pattern recognition.
Gas-solid fluidized bed Flow regime identification Wavelet decomposition Fractal dimension Probabilistic neural network
ZHOU Yunlong FAN Zhenru
School of Energy Resource and Mechanical Engineering, Northeast Dianli University, Jilin, 132012, Ch School of Automation Engineering, Northeast Dianli University, Jilin 132012, China
国际会议
第八届国际测试技术研讨会(8th International Symposium on Test and Measurement)
重庆
英文
2889-2892
2009-08-01（万方平台首次上网日期，不代表论文的发表时间）