Co-operative sensor fusion using time warping in multimodal tomometry for process con

Abstract

Recent studies in process tomography have mostly focused on image processing and enhancement of images obtained from Electrical Capacitance, Electrical Resistance, Electrical Impedance, Gamma- and X-ray- tomographic modalities. In the process industries, there is a growing need for fast acting measurement and control procedures, architectures and their implementation within existing sensor and control suites. With this industrial relevance, a multimodal tomometric sensor fusion is studied involving pressure and capacitance measurements in a two phase flow rig. Pressure is measured using standard pressure sensors, whereas the capacitance are measured using arrays of capacitance electrodes placed on the periphery of a section of the pipe in the multiphase flow rig. These capacitance measurements are performed by using ECT-modules. In this paper, the focus is on early slug detection and timely characterization of these slugs formed in the two phase flow. To associate simultaneous occurrence of slug based phenomena in the time series of the pressure and ECT-signals, dynamic time warping (DTW) algorithms are used. The usage of DTW and the multimodal tomometric sensor suite with pressure and ECT-modules leads to a co-operative sensor data fusion, which can help to identify characteristic features of slugs in two phase flows, which can be significant inputs to the process control unit in the assessment/implementation of necessary actions, such as activating choke valves, reducing pump outputs of one or more of the phases flowing in the rig. Time series of pressure and ECT-signal are studied using DTW and results are compared and fused with some discussions on how these could be used in process control.