Noise Testing Data Management And Clustering Argorithms Research Based On WEB

In the field of ship and manufacturing, ship fault diagnosis technology has been playing an important role to ensure run safely. ship fault diagnosis technology for research, which has not only theoretical value, but also a very obvious economic benefits. With the continuous development of science and technology, equipments in a ship are in the increasingly large, complex, high-speed and automation. The equipment progress on the technology, increasing loss due to system downtime and failure followed by a simultaneous. So, it has put forward higher requirements for the fault diagnosis and predicting technique, the traditional diagnosis ways have been unable to meet this demand. In the trend, along with progress of artificial intelligence technology, database technology, ship fault diagnosis have been into the intelligent phrase.The paper is based on Vibration Noise Data System which is carried out by China Ship Scientific Research Center (CSSRC) is a member of China Shipbuilding Industry Corporation (CSIC). To improve self intellectual property and become a information technology platform for ship design performance.Vibration Noise Data System use J2EE development frame. J2EE frame can be used to simplify complicated problems enterprise information development, deployment, management and so on greatly. And it offers transaction, security, naming, persistence and resource management service for enterprise application. So these can enabled developers focus on developing business logic while ignoring the bottom detail, and provide well mechanism for designing and developing the Vibration Noise Data System.The paper research many clustering algorithm and compare all kinds of algorithms application to ship fault diagnosis. A approach is proposed, Theoretical analysis and experimented results testify that the new algorithm has more robust. All clustering algorithms in the paper contains the following contents:Firstly, WRTCM algorithm is new noise-resistant fuzzy clustering approach based probabilistic typicality and cluster repulsion and it use a kind way of metric. The proposed algorithm can solve the identical clustering center problem in noisy environment, and improve the clustering accuracy. This paper presents combines WRTCM algorithm with fuzzy fault diagnoses method applicants to checking equip fault of ship.Secondly, the data in the original space are firstly mapped to a high-dimensional feature space by merce kernel functions, and an additional weighting factor is assigned to each vector in the feature space. Then a modified objective function for fuzzy clustering is introduced in the feature space. Choosing appropriate values for kernel parameters is one of the key problems in many kernel-based algorithm.