Search

In order to acquire satisfying knowledge quickly in current Web and semantic Web, it is very important to find an effective method to present Web resource. We proposed a concept of Ontological Knowledge Node (OKN) as the core of semantic search model and give a design schema to organize knowledge representation from domain perspective. OKN is a kind of Web resource representation method, which consists of a 6-tuple set and can represent the intension and extension of ontologies as well as the relationship between ontologies. This method was explained in detail from the perspective of granular computing. In order to show advantages of the OKN, a Semantic Search Engine was established, which combines the OKN with Nutch to search campus resources. OKN was proved to work well on semantic search.

The overall structure of WSN (wireless sensor networks) node platform was introduced. This platform was suitable for collecting and transmitting data of vibration signals. The initialization, register visit and data read mode of CC2420 were analyzed in detail, which made it possible to realize the wireless transceiver function of node platform. By applying WSN to collect data of mechanical equipment and using neural network to analyze the data, the usability of WSN in mechanical equipment conditions monitoring is proved.

In the process of using neural network to carry out intelligent fault type identification, how to extract sensitive fault features from the original data is quite important for an accurate diagnosis result. An intelligent fault diagnosis method was proposed, which combined time domain analysis and wavelet analysis method to extract features from vibration data of a motor bearing. The resulting vector obtained from the feature extraction was used as samples to train the BP neural network intelligent classifier to enable the classifier to identify fault type. The comparison of experiment results showed that the proposed diagnosis method was effective.

Wireless sensor network is a thriving information collecting and processing technology, which is widely used in military field, industry and environmental monitoring, etc. In a wireless sensor network which is made up of tens of thousands battery-powered sensor nodes, data fusion technology can be used to reduce communication traffic in order to save energy. With respect to large mechanical equipment, traditional wired sensors are commonly used for fault detection and diagnosis. There will be no wiring problem if wireless sensor networks are used, which is favorable to detect potential problems in mechanical equipment without affecting normal production of enterprises. In this paper, the application of wireless sensor networks in machinery fault diagnosis is studied, a data fusion model for machinery fault diagnosis in wireless sensor networks and PCA neural data fusion algorithm are proposed, and the effectiveness of the method is demonstrated in an experiment.