Research On Engineering Vehicle Fault Prediction Based On Clustering And Association Rules Algorithm Under Hadoop Platform

Engineering vehicle has the larger size, components of complex, long service time and so on, how to timely and effectively prevent the happening of the engineering vehicle failure is the enterprise urgent need, also many scholars are studying. Hadoop as the most widely application scope on the data processing of the open source framework has attracted widespread attention. Its main job is to prepare the corresponding Map Reduce programs for large data processing using huge amounts of data generated by the use of storage and management. Extracting the data information contained in the potential for data resource management and its value for further development.Data mining algorithms in big data platform have been better developed.In this paper, after studying the big data framework and traditional clustering and association rules algorithm, according to the characteristics of the excavator on the acquisition of data and data processing characteristics of big data platform, Author puts forward the improved clustering algorithm, combines clustering and association rules algorithm, sets up a Hadoop data platform, the design of the large data programming operation combining algorithm framework, reached under the condition of large amounts of data for excavator failure prediction. This article main research content is as follows:1.The improved algorithm. This paper studies on the maximum and minimum k-means algorithms and based on the shortage that traverse the data continuously clustering algorithm is improved. M+k-means algorithm is proposed. Paper by multiple sets of experiments compared the improved algorithm m+k-means algorithm with the original maximum minimum k-means algorithm, obtained the acceleration ratio, expansion ratio and running time contrast figure.2.Programming Framework Design. Big data platform construction and pseudo-distributed configuration. Combines the improved clustering algorithms and association rule algorithm. In view of the two algorithms designed graphs programming framework. Use and combination Mapper and Reducer class category, and the application interface.3.The realization of fault prediction system. Papers will combine the two algorithms, transplanted into the design of the programming framework, by entering multiple sets of experimental data collected in Shan Tui Construction Machinery co.,Ltd, it is concluded that the abnormal data association rules of engineering vehicles, so as to achieve the purpose of the exception and failure probability prediction.The topic of research is supported by Doctor Start up project‘Big data platform for engineering machinery research and design’,the project number is:20132021.