Online flooding monitoring in packed towers using EDPCA method

Traditional flooding monitoring methods have been found insufficient to monitor various types of industrial packed towers. In this work, an enhanced data-driven monitoring method, i.e., enhanced dynamic principal component analysis (EDPCA), is proposed for online flooding monitoring in packed towers. The operation data samples are first clustered into several classes using the fuzzy c-means clustering approach. Then, several single DPCA models are trained with each subset. Furthermore, the Bayesian inference is adopted to integrate these single DPCA models. The obtained results for online flooding monitoring of an air-water packed tower demonstrate that EDPCA can obtain better and more reliable performance, compared with the DPCA method.