A study of a two phases medium in pipe vessel using gaussian filter method and k-NN classifier

Abstract

In this paper we present the existence and uniqueness of solutions to the stochastic problem to determine two phase level of gas and water in the industrial pipe line. In order to distinguish the level phases in a vessel such as high, medium, and low a mechanism of Gaussian Filter Method was proposed in this study. The k-weights of the statistical features were adjusting by the k-NN Classifier and separate the categories for training from the various parameters. Within this study, 6 statistical used features were employed for training. In addition, a time-frequency parameter obtained by continuos signal was also used to identify the pulse during the recognition process. For the classification of all phase level, the average accuracy was higher than 90.0% for the whole data set. Finally, in order to provides the opportunity of obtaining quantitative and real time data on two phase media within a full scale industrial process, such as filtration, without the need of process interruption, tomography methods answer it all.