The recent development of feature extraction algorithms with multiple layers in machine learning and pattern recognition has inspired many applications in multivariate statistical process monitoring. In this work, two...The recent development of feature extraction algorithms with multiple layers in machine learning and pattern recognition has inspired many applications in multivariate statistical process monitoring. In this work, two existing multi-layer linear approaches in fault detection are reviewed and a new one with extra layer is proposed in analogy. To provide a general framework for fault diagnosis in succession, this work also proposes the contribution propagation analysis which extends the original definition of contribution of variables in multivariate statistical process monitoring. In fault diagnosis stage, the proposed contribution propagation analysis for multilayer linear feature extraction algorithms is compared with the fault diagnosis results of original contribution plots associated with single layer feature extraction approach. Plots of variable contributions obtained by the aforementioned approaches on the data sets collected from a simulated benchmark case study(Tennessee Eastman process) as well as an industrial scale multiphase flow facility are presented as a demonstration of the usage and performance of the contribution propagation analysis on multi-layer linear algorithms.展开更多
D-statistic contribution analysis has been frequently used in practice for fault diagnosis.Existing algorithms for computing contributions to D-statistic tend to distribute cross-term contribution equally between two ...D-statistic contribution analysis has been frequently used in practice for fault diagnosis.Existing algorithms for computing contributions to D-statistic tend to distribute cross-term contribution equally between two correlated variables.This leads to increased variance in contribution estimation and hence poor separability of faulty and normal variables.A new method for contribution calculation to D-statistic is proposed here which introduces a weighting scheme capable of distinguishing the contributions of two correlated variables.Simulation examples show that the proposed approach achieves improved resolution for distinguishing faulty and normal conditions.展开更多
基金Supported by National Basic Research Program of China （973 Program） （2009CB320602）, National Natural Science Foundation of China （60721003, 60736026）, and Changjiang Professorship by Ministry of Education of P. R. China
基金supported by the funding from the European Union's Horizon 2020 research and innovation programme (No. 675215-PRONTO-H2020-MSCA-ITN2015)
文摘The recent development of feature extraction algorithms with multiple layers in machine learning and pattern recognition has inspired many applications in multivariate statistical process monitoring. In this work, two existing multi-layer linear approaches in fault detection are reviewed and a new one with extra layer is proposed in analogy. To provide a general framework for fault diagnosis in succession, this work also proposes the contribution propagation analysis which extends the original definition of contribution of variables in multivariate statistical process monitoring. In fault diagnosis stage, the proposed contribution propagation analysis for multilayer linear feature extraction algorithms is compared with the fault diagnosis results of original contribution plots associated with single layer feature extraction approach. Plots of variable contributions obtained by the aforementioned approaches on the data sets collected from a simulated benchmark case study(Tennessee Eastman process) as well as an industrial scale multiphase flow facility are presented as a demonstration of the usage and performance of the contribution propagation analysis on multi-layer linear algorithms.
基金the National Basic Research Program (973) of China(No.2010CB731800)the National Natural Science Foundation of China(Nos.60974059, 60736026 and 61021063)
文摘D-statistic contribution analysis has been frequently used in practice for fault diagnosis.Existing algorithms for computing contributions to D-statistic tend to distribute cross-term contribution equally between two correlated variables.This leads to increased variance in contribution estimation and hence poor separability of faulty and normal variables.A new method for contribution calculation to D-statistic is proposed here which introduces a weighting scheme capable of distinguishing the contributions of two correlated variables.Simulation examples show that the proposed approach achieves improved resolution for distinguishing faulty and normal conditions.
