Hybrid variable monitoring: An unsupervised process monitoring framework with binary and continuous variables

Traditional process monitoring methods, such as PCA, PLS, ICA, MD et al., are strongly dependent on continuous variables because most of them inevitably involve Euclidean or Mahalanobis distance. With industrial processes becoming more and complex integrated, binary also appear in besides variables, which makes challenging. The aforementioned traditional approaches incompetent to mine the information so that useful contained is usually discarded during data preprocessing. To solve problem, this paper focuses issue hybrid variable (HVM) proposes a novel unsupervised framework with including variables. HVM addressed probabilistic framework, can effectively exploit implicit both at same time. In HVM, statistics strategy suitable for only healthy state defined physical explanation behind elaborated. addition, estimation parameters required derived detail detectable condition proposed method analyzed. Finally, superiority fully demonstrated first numerical simulation then an actual case thermal power plant.