Research On The Velocimetry Measurements Of Oil-water Two-phase Flow In Horizontal Pipe Based On The PIV Technology

In the processing of oilfield exploitation, it is very significant for increasing therecovery rate to predict and measure the velocity of oil-water two-phase flow in horizontalpipe accurately and precisely. There is often oil droplets overlap and velocity gradient inthe motion, at this time, the particle image velocimetry technology based on the traditionalcross-correlation is already unable to achieve accurate measurements of the velocity field.Therefore, for the high particle concentration velocity measurements, an improvediterative multigrid PIV image deformation algorithm is studied in this paper, and for thelow particle concentrations measurements, a PTV algorithm based on SIFT feature pointmatching is studied.Firstly, due to numerous irrelevant and interferential information existing in theimages of oil-water two-phase flow, this paper studied its corresponding image preprocessmethods, including the image clipping, gray transformation, filtering denoising and imageenhancement, which can remove the irrelevant and interferential information and furtherstress the helpful information to some extent. As a result, it can reduce the influence to thesubsequent matching algorithm, which makes the results of PIV more accurate.Secondly, an improved iterative multigrid PIV image deformation algorithm isstudied in this paper. The method point out that the measurement accuracy depends on thepredicted results of the first iteration largely. Therefore, aiming at this problem,improvements are made on the traditional iterative multi-grid method; namely, a higheraccuracy direct cross-correlation is used in the first iteration with fifty percent overlap ofinterrogation windows.Thirdly, a PTV algorithm based on SIFT feature point matching is studied in thispaper. The method firstly extracts the SIFT key points in images, and then feature pointsmatching is made by slant distance as the similarity measure. When all correctly matchedfeature points are found, particle displacements can be obtained from these points’positions. And further, particle velocity will be achieved.Finally, for different particle concentration of oil-water two-phase flow, experiments are made by the improved iterative multigrid image distortion PIV algorithm and the PTValgorithm based on SIFT feature points respectively, and the results of these experimentsare analyzed in detail.