Two-phase flow circuit characterization by pressure drop mechanistic models, image processing technique and flow pattern mapping

Abstract

Two-phase flow phenomenon is considerably present in many industry sectors and plays an important role in multiple operations, processes and equipment projecting. Literature has shown that modeling these occurrences is not completely well understood and it is still a technological and conceptual challenge. In this way, predictions of flow pattern and friction loss are compared with visual observations and meters’ measurements, considering also an extensive bibliographic review on models for predicting patterns and behavior of multiphase flows. Further, such development is applied aiming to estimate the operational envelop of NEMOG’s multiphase flow circuit. A set of experiments is planned to verify the real operational capacity of the facility and results are further compared to predictions provided by models. In parallel, an image processing algorithm is developed to identify the liquid height through high-frame rate video recordings of a translucent test’s section. Using a selected region of interest, the output of liquid plug height over time is then used to calculate the gas velocity by two different approaches for intermittent flow patterns: geometrical and frequency shift. Furthermore, single-sided amplitude spectrum and power spectral density are studied in order to supply an additional tool to evaluate flow patterns. Capabilities and limitations of the algorithm are discussed, and possible future developments are indeed suggested. This study concluded that, selected mechanistic models are quite capable of providing an estimation of present flow patterns and pressure drop. However, localized pressure drop is not considered in these models, and it may be an issue for certain facilities. The image processing algorithm performed well for definite intermittent pattern and is a valid approach for collecting additional information of a two-phase flow. Situations in which the flow is not intermittent or liquid fraction and Reynolds number is high, represented a challenge to be overcame. Lastly, amplitude spectrum is considered an additional tool to identify two-phase flow pattens based on higher amplitude frequencies.