Dataset of two-phase flow in a horizontal pipe: synchronized measurements of acceleration, pressure, void fraction and high-velocity camera

This dataset provides a valuable resource for the detailed study of two-phase flows and flow-induced vibrations in pipes, essential in sectors such as oil, chemical, and nuclear industries. Data collection and analysis were conducted at the Experimental Laboratory of Petroleum (LabPetro) within the Center for Energy and Petroleum Studies (CEPETRO) at the <i>Universidade Estadual de Campinas</i> (UNICAMP). The experimental data comprises 23 points for a gas-water mixture flowing in a horizontal pipe, representing three flow patterns: stratified, slug, and dispersed bubble. To capture the complex characteristics of the flow, synchronized data from void fraction sensors, piezoelectric pressure sensors, and accelerometers were acquired for 600 seconds at a sampling rate of 25.6 kHz. In addition, high-velocity videos were recorded at 800 fps for approximately 15 seconds, also synchronized with the sensors data. A deep learning algorithm was applied to segment the bubbles in the images, allowing accurate identification of their contours. Image processing provided information on the contours, diameters, number, and mean diameter of the smaller bubbles in all recorded images. Additionally, for some frames, a manual image segmentation process was performed to obtain accurate information on the location and contour of each bubble, enabling tracking and calculation of the individual velocity of each bubble. The integration of sensor data with image processing results and bubble tracking allows for a comprehensive analysis of the multiphase flow dynamics. In conclusion, two databases were produced from this experimental data. The first, <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:mi>D</mml:mi> <mml:mi>a</mml:mi> <mml:mi>t</mml:mi> <mml:mi>a</mml:mi> <mml:mi>s</mml:mi> <mml:mi>e</mml:mi> <mml:msub><mml:mi>t</mml:mi> <mml:mrow><mml:mn>25.6</mml:mn> <mml:mi>k</mml:mi> <mml:mi>H</mml:mi> <mml:mi>z</mml:mi></mml:mrow> </mml:msub> </mml:mrow> </mml:math> , contains the raw time series from the sensors (pressure, acceleration, void fraction) and the trigger signal, all at a sampling rate of 25.6 kHz. The second, <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:mi>D</mml:mi> <mml:mi>a</mml:mi> <mml:mi>t</mml:mi> <mml:mi>a</mml:mi> <mml:mi>s</mml:mi> <mml:mi>e</mml:mi> <mml:msub><mml:mi>t</mml:mi> <mml:mrow><mml:mn>800</mml:mn> <mml:mi>H</mml:mi> <mml:mi>z</mml:mi></mml:mrow> </mml:msub> </mml:mrow> </mml:math> , was generated by downsampling the first to 800 Hz to match the camera's frame rate (800 fps). This final database integrates the sensors data with the image processing results, including the corresponding frame, number of bubbles, diameters, contours, segmentation, and bubble tracking information.