Title :
HVAC Fan Mechinery Fault Diagnosis Based on ANN and D-S Evidence Theory
Author :
Xuemei, Li ; Lixing, Ding ; Yan, Li ; Gang, Xu ; Jibin, Li
Author_Institution :
Sch. of Mech. & Electr. Eng., Zhongkai Univ. of Agric. & Eng., Guangzhou, China
Abstract :
A novel approach to HVAC fan machinery fault diagnosis based on combination of artificial neuron network and D-S evidence theory is presented in this paper. Firstly, the system pre-processes the acquisition data from multi sensor, then makes pattern classifies observations based on BP neural network. Secondly, the output values of BPNN are directly taken as the basic probability assignment of the proposition in the frame of discernment. Therefore realizing the objective treatment of the basic probability assignments and avoiding the complexity of constituting the basic probability distribution function. The experiment results demonstrate that the reliability and accuracy to apply the fusion BPNN results with the D-S evidence theory are higher than the independent BPNN diagnosis.
Keywords :
HVAC; fans; fault diagnosis; mechanical engineering computing; neural nets; pattern classification; statistical distributions; D-S evidence theory; HVAC fan mechinery fault diagnosis; artificial neuron network; data acquisition; probability assignment; probability distribution function; Agricultural engineering; Agriculture; Artificial neural networks; Automatic control; Bayesian methods; Control systems; Fault diagnosis; Machinery; Neurons; Space technology; BPNN; D-S evidence theory; fault diagnosis;
Conference_Titel :
Control, Automation and Systems Engineering, 2009. CASE 2009. IITA International Conference on
Conference_Location :
Zhangjiajie
Print_ISBN :
978-0-7695-3728-3
DOI :
10.1109/CASE.2009.178
