| In the project we created the platform for equipment status monitor, with which we research the technology for monitor equipment status. In the paper two parts are included: one concerned with the design of hardware and software of the child node of the platform, and the other concerned with the algorithm for detecting equipment fault.Firstly, we designed the device through which bearing failure is emulated for research. Secondly, we designed the device which uses PC/104 type microcomputer for detecting the vibration signal of bearing. Functions of the detecting device are to acquisition and analyze the vibration signal, and send result of analysis to the parent node of the platform through wireless communication network. Finally we designed the software which runs in the child nodes of that platform for those functions mentioned above.In the paper we research some algorithms for the purpose of detecting bearing fault correctly. It's a popular and efficiently way to detect bearing failure using spectrum analysis, and we descript discrete Fourier transform, power spectrum density analysis which are needed for spectrum analysis. Furthermore we issue the way that use the algorithm of principal components analysis to reconstruct the statistic characteristic value of bearing, then classify the value using the algorithm of support vector machine to identify the type of bearing fault. That is the creative of the paper.Also we give the samples of detecting bearing fault in the paper. Firstly we use the way of spectrum analysis to resolve the vibration signal which is collected from the platform, and judge the fault type through the eigenfrequency of the bearing. Secondly we resolve several statistic characteristic variables using principal component analysis to reconstruct new variable, then we classify those new variables using... |
