Flow Regime Identification of Two Phase Flow based on Image Processing Techniques

Bubble column reactors are common in many industries such as biochemical, energy generation, process industry etc. The performance of the reactor is dependent on the flow regime inside the column. Identification of the flow regimes inside the reactor is significant and a challenging task. In the present paper, non-intrusive, real time approaches for regime identification are developed, based on the image processing technique and Symbolic Dynamic Filtering (SDF) in MATLAB. Flow images are captured using a Photron high speed (1000 fps) camera. The bubble detection algorithm results in the calculation of bubble diameters and the distribution of the bubble diameters for different flow rates, leading to regime identification. The algorithm includes the method of calculation of the optimum threshold luminescence value and converting the input image into binary using that optimum threshold value. From the binary image, boundaries are traced and the average bubble diameter is calculated. The deviation from one regime to another regime is detected by SDF algorithm. Flow measure (μ) is a parameter obtained from SDF, which indicates the deviation of a particular flow from the reference flow. The variation of flow measure is plotted against the air flow rate. It is found that change in the slope of this curve is significant during regime transition. This parameter can be used as a marker of flow regime identification.