论文部分内容阅读
By researching the data quality problem in the monitoring and diagnosis system (MDS), the method of detecting non-condition data based on the development trend of equipment condition is proposed, and three requirements of criteria for detecting non-condition data: dynamic, syntheses and simplicity are discussed. According to the general mode of data management in MDS, a data quality assurance system (DQAS) comprising data quality monitoring, data quality diagnosis, detection criteria adjusting and artificial confirmation is set up. A route inspection system called MTREE realizes the DQAS. Aiming at vibration data of route inspection, two detecting criteria are made. One is the quality monitoring parameter, which is found through combining and optimizing some fundamental parameters by genetic programming (GP). The other is the quality diagnosis criterion, i.e. pseudo distance of Spectral-Energy-Vector (SEV) named Adjacent J-divergence, which indicates the variation degree of adjacent data′s spectral energy distribution. Results show that DQAS, including these two criteria, is effective to improve the data quality of MDS.
By researching the data quality problem in the monitoring and diagnosis system (MDS), the method of detecting non-condition data based on the development trend of equipment condition is, and three requirements of criteria for detecting non-condition data: dynamic, syntheses and simplicity are discussed. According to the general mode of data management in MDS, a data quality assurance system (DQAS) comprises data quality monitoring, data quality diagnosis, detection criteria adjusting and artificial confirmation is set up. A route inspection system called MTREE realizes the DQAS. Aiming at vibration data of route inspection, two pairs of surveying are made. One is the quality monitoring parameter, which is found through combining and optimizing some fundamental parameters by genetic programming (GP). The other is the quality diagnosis criterion, ie pseudo distance of Spectral-Energy-Vector (SEV) named Adjacent J-divergence, which indicates the variation degree of adjacent da ta’s spectral energy distribution. Results show that DQAS, including both two criteria, is effective to improve the data quality of MDS.