Bubble Images Research Articles

Ensemble prediction of mean bubble size in a continuous casting mold using data driven modeling techniques

Gas injection into a slab continuous casting mold is a common practice as it enables long casting sequences by reducing SEN clogging. However, too much gas injection can generate lots of bubbles, and bubbles formed inside the mold are also responsible for the occurrence of various quality defects such as deterioration of steel cleanliness, reoxidation of liquid steel, and occurrence of sliver and blister defects, etc. Mean bubble size is a key parameter and controlling it can resolve/reduce these quality issues. Operating parameters such as gas flow rate and liquid flow rate are the major factors affecting the mean bubble size. Two-phase water modeling​ experiments were performed to generate bubbles in the mold, and the mean bubble diameter was captured using various gas and liquid flow rates. Clear images of bubbles were recorded using a high-speed high-resolution camera along with the application of shadowgraphy. Bubble images were processed using an image processing software, ImageJ to obtain bubble characteristics, and Sauter mean diameter was calculated for each operating condition. Advance machine learning techniques such as Multilinear Regression (MLR), Decision Tree (DT), Support Vector Machine (SVM), and Artificial Neural Network (ANN) were used on the experimental data to predict the combined effect of these operating parameters on the mean bubble diameter. All four ML techniques were compared considering the values of cross-validated adjusted R2, and a performance metric is presented to compare the suitability of ML techniques in this case.

Wave Overtopping Plunging on the Rear Side of Composite Breakwater

Adibhusana, M.N.; Lee, J.-I., and Ryu, Y., 2021. Wave overtopping plunging on the rear side of composite breakwater. In: Lee, J.L.; Suh, K.-S.; Lee, B.; Shin, S., and Lee, J. (eds.), Crisis and Integrated Management for Coastal and Marine Safety. Journal of Coastal Research, Special Issue No. 114, pp. 574–578. Coconut Creek (Florida), ISSN 0749-0208. In this study the behavior of overtopping water falling on the rear side of composite breakwater were investigated through physical experiment. A series of regular wave with various wave period and wave height were used as wave parameters while velocity at the rear edge of breakwater, plunging location, and plunging velocity were used as the parameters of wave overtopping behavior on the rear side of composite breakwater. The geometry of the composite breakwater was kept constant through the experiment. The Bubble Image Velocimetry (BIV) and digital image analysis were applied to measure the velocity in the aerated region of overtopping water falling on the rear side of breakwater and the plunging location. The rear edge velocity was linearly increased as the wave parameter increased. The increment value of plunging location shows a logarithmic trend as wave period and wave height increased. The front velocity of overtopping flows right before plunging the rear water surface shows a linear trend with those wave parameters.

High-speed ultrasound imaging of laser-induced cavitation bubbles

While ultrasound and cavitation based therapies have mushroomed over the years, there is a lack of online monitoring of the cavitation bubble dynamics in biological tissue. Here, we demonstrate that with ultrasonic plane wave imaging, the fast dynamics of single cavitation bubbles can be resolved non-invasively in a tissue mimicking material. Due to the high contrast of bubbles, plane wave compounding is not necessary and the frame rates of up to 74 kHz can been achieved with a research grade ultrasound scanner. Comparison with simultaneous high-speed imaging demonstrates excellent agreement of both measurement modalities in water and in a tissue mimicking material.

A flexible image processing technique for measuring bubble parameters based on a neural network

The main parameters in a liquid-gas system are the size, shape, and kinetic parameters of bubbles, which affect the mass transfer process. To describe the mass transfer process accurately, it is necessary to characterize bubble parameters such as bubble volume and surface area under different experimental conditions. However, these bubble parameters are not accurately assessable by conventional methods, which are limited by poor robustness and flexibility, including numerical methods and conventional neural networks. Based on the neural network technique, we propose a flexible method for measuring bubble parameters in three steps: segmentation of bubble images, segmentation of overlapping bubbles, and extraction of bubble feature parameters. This technique combines geometric, optical, and topological information. To verify the effectiveness of our method, we considered a gas-liquid reactor and microbubble radiosensitizer for brachytherapy as testing scenarios and used several groups of experimental images and synthetic images as testing sets. In the first step, the accuracy of the segmentation of bubble images was greater than 90% in all scenes and reached 99% in some scenes. In the second step, the proposed method was capable of segmenting overlapping bubbles and avoiding excessive segmentation. In the final step, our approach considers the geometric shape and deflection angle of bubbles. Therefore, the bubble parameters of largely deformed and deflected bubbles were obtained with high flexibility. Our method is a powerful tool with robustness and flexibility for the study of gas-liquid interface measurements in chemical and biochemical processes.

Effect of cationic collector on the attachment of glass beads to a stationary bubble

Glass beads in contact with a stationary air bubble immersed in dodecylamine hydrochloride (DAH) collector solution were investigated to examine the effects of a cationic collector on bubble–particle attachment. A high-speed video camera was used to record images of bubble–particle interactions. When the bubble diameter was 1.05–1.9 mm, the induction time initially decreased then increased as the bubble size increased in a 10 −5 mol/L DAH solution. Experiments demonstrated that the induction time increased as the collision angle increased. However, in solutions with higher DAH concentrations, the distribution of induction time was scattered, and the jump-in events were observed above the horizontal equator of the bubble; particle aggregation was also observed. When the DAH concentration increased from 10 −5 to 10 −3 mol/L, the induction time initially increased and then decreased. Analysis of the attachment for large numbers of particles verified that the correlation between induction time and attachment efficiency was stronger than the correlation between contact angle and attachment efficiency. At the solution concentration for which the induction time was maximum, the particle attachment efficiency was the lowest. The effects of the collector on bubble–particle attachment were examined in multi–perspectives. This study introduces new insights into the bubble–particle attachment and offers an increased, more comprehensive understanding of the attachment mechanisms. • Induction time initially decreases and then increases as bubble size increases. • Induction time increases as a positive quadratic function of the collision angle. • The distribution of the induction time is random at high solution concentrations. • Induction time increases firstly and decreases afterward with the concentration.

A classifier design for micro bubble generators based on deep learning technique

This paper aims to study the sizes of bubbles generated by the micro-bubble generator device (MBGs) in the water because of its impact on the dissolved oxygen percentage in the water and we find this in aquaculture where oxygen is important for marine life and in many applications. Where we have designed a classifier for micro-bubbles of different sizes and the effectiveness of the device has been proven in generating bubbles whose size ranges from 20 to 50 and calculating their numbers in the image by converting the image from (RGB) to (HSV), it has proven its effectiveness in maintaining lighting and neglecting color information Which is not important in identifying bubbles in the image, as well as increasing the processing speed, and then finding edges in the image, using canny edge detection, which has proven its great effectiveness in finding weak and strong edges. Where we then used the (Circular Hough Transform) algorithm to identify the tiny bubbles in the image, calculate their average diameter and numbers in the image, and find the overlapping bubbles between them by using (Two Thresholds) one for the edge and the other for the center of the bubble and then making (Segmentation) for these bubbles and inserting them on (Gaussian Normal Distribution) Do not neglect the dark bubbles, because in most cases they are unreal bubbles and may be distortions in the image or reflections in the lighting. And then make (zero paddings) for the extracted images to make them equal in size and avoid the distortion that will happen in the image if another method is used to make all bubbles of equal dimensions to prepare them for classification. And then we calculated the (mean square error) of the resulting bubble images to avoid repeating the same bubble in (datasets). The efficiency of the classifier was calculated by using deep learning technology. Where we built a (Convolution Neural Network) consisting of (15) layers and then we trained these layers on the images resulting from the (CHT), which are individual bubbles, each of them classified according to size. The dataset was divided into two parts, 70% of which is for education and the remaining 30% is for (validation), and the test percentage is 100%.

Experimental study on the influence of heating surface inclination angle on heat transfer and CHF performance for pool boiling

Pool boiling heat transfer is widely applied in nuclear engineering fields. The influence of heating surface orientation on the pool boiling heat transfer has received extensive attention. In this study, the heating surface with different roughness was adopted to conduct pool boiling experiments at different inclination angles. Based on the boiling curves and bubble images, the effects of inclination angle on the pool boiling heat transfer and critical heat flux were analyzed. When the inclination angle was bigger than 90°, the bubble size increased with the increase of inclination angle. Both the bubble departure frequency and critical heat flux decreased as the inclination angle increased. The existing theoretical models about pool boiling heat transfer and critical heat flux were compared. From the perspective of bubble agitation model and Hot/Dry spot model, the experimental phenomena could be explained reasonably. The enlargement of bubble not only could enhance the agitation of nearby liquid but also would cause the bubble to stay longer on the heating surface. Consequently, the effect of inclination angle on the pool boiling heat transfer was not conspicuous. With the increase of inclination angle, the rewetting of heating surface became much more difficult. It has negative effect on the critical heat flux. This work provides experimental data basis for heat transfer and CHF performance of pool boiling.

Pump-probe X-ray holographic imaging of laser-induced cavitation bubbles with femtosecond FEL pulses

Cavitation bubbles can be seeded from a plasma following optical breakdown, by focusing an intense laser in water. The fast dynamics are associated with extreme states of gas and liquid, especially in the nascent state. This offers a unique setting to probe water and water vapor far-from equilibrium. However, current optical techniques cannot quantify these early states due to contrast and resolution limitations. X-ray holography with single X-ray free-electron laser pulses has now enabled a quasi-instantaneous high resolution structural probe with contrast proportional to the electron density of the object. In this work, we demonstrate cone-beam holographic flash imaging of laser-induced cavitation bubbles in water with nanofocused X-ray free-electron laser pulses. We quantify the spatial and temporal pressure distribution of the shockwave surrounding the expanding cavitation bubble at time delays shortly after seeding and compare the results to numerical simulations.

Numerical investigation of alternating skimming flow over a stepped spillway

Abstract This study aims to illustrate the influence of stepped spillway width on alternating skimming flow development. A computational fluid dynamics (CFD) model in Ansys Fluent® was established to simulate the flow over stepped spillways, using a volume of fluid model (VOF) and Reynolds Averaged Navier–Stokes (RANS) turbulence model (SST k-ω). The model was first validated by comparisons of velocity profiles at step niches and water depth at step edges with existing measurements acquired by the bubble image velocimetry (BIV) technique and an ultrasonic sensor, in a 0.5-m wide stepped spillway physical model. The SST k-ω model gave good results for velocity and water depth, and the numerical predictions of the vorticity in the skimming and recirculating flows were qualitatively adequate. The model was used to analyse the flow regime for six different stepped spillway widths. The careful examination of flow patterns at the different stepped spillway widths showed that the alternating skimming flow appears for the stepped spillways wider than 0.35 m due to the asymmetrical distribution of vorticity patches that are generated in the step cavity. These vorticity patches are of uniform size and shape when the spillway width is less than 0.35 m, which does not produce an alternating skimming flow. However, for wider stepped spillways, the vorticity increases, and an alternating skimming flow appears closer to the crest.

Oil droplets dispersion under a deep-water plunging breaker: Experimental measurements and numerical modeling

Abstract (1141370) Wave tank experiments were performed to measure the droplets size distribution under the plunging breaking wave. A deep-water plunging breaker of height 20 cm was generated using the dispersive focusing method, and a shadowgraph camera was used to take images of droplets and bubbles of different sizes. For droplets smaller than the 1000 microns, the number-based DSD matched the DS correlation (Delvigne and Sweeney 1988), which gives N(d) ~ d−2.3, but N(d) ~ d−9.7 for diameters larger than 1000 microns. A numerical method was designed to study the oil dispersion under breaking waves by coupling the computational fluid dynamic (CFD) with the Lagrangian particle tracking code (NEMO3D) and population balance model (VDROP). The wave hydrodynamics was reproduced using the Reynolds-averaged Navier Stokes approach within a commercial CFD code ANSYS Fluent. The obtained wave hydrodynamics was then used as inputs for the NEM3D code and VDROP model. The numerical results show reasonable agreement with our experimental observation. The approach adopted to produce the DSD reduces the empiricism of the DS correlation, as the approach uses oil properties and measurable wave properties. The proposed numerical method was ready to be used in other scenarios of oil spills (i.e., oil jets in deep oceans and oil dispersion in riverine systems). It could also be potentially used in large scale forecast and hindcast simulations for oil spill response and research.

Statistical analysis of helium bubbles in transmission electron microscopy images based on machine learning method

Helium bubbles, which are typical radiation microstructures observed in metals or alloys, are usually investigated using transmission electron microscopy (TEM). However, the investigation requires human inputs to locate and mark the bubbles in the acquired TEM images, rendering this task laborious and prone to error. In this paper, a machine learning method capable of automatically identifying and analyzing TEM images of helium bubbles is proposed, thereby improving the efficiency and reliability of the investigation. In the proposed technique, helium bubble clusters are first determined via the density-based spatial clustering of applications with noise algorithm after removing the background and noise pixels. For each helium bubble cluster, the number of helium bubbles is determined based on the cluster size depending on the specific image resolution. Finally, the helium bubble clusters are analyzed using a Gaussian mixture model, yielding the location and size information on the helium bubbles. In contrast to other approaches that require training using numerous annotated images to establish an accurate classifier, the parameters used in the established model are determined using a small number of TEM images. The results of the model formulated according to the proposed approach achieved a higher F1 score validated through some helium bubble images manually marked. Furthermore, the established model can identify bubble-like objects that humans cannot facilely identify. This computationally efficient method achieves object recognition for material structure identification that may be advantageous to scientific work.

Number density of bubbles under ultrasonic horn measured from stroboscopic images

Although image measurement is essential in the analysis of acoustic cavitation bubbles, it is impossible to determine the position of the bubble along the optical axis of the imaging system from the images. Thus, the number density of the bubbles cannot be measured from the image. This paper proposed a method to determine bubbles’ positions along the optical axis using the bubble image brightness. The relationship among the bubble position along the optical axis, the bubble diameter, and the bubble image brightness is clarified using the bubble in the single bubble system. A measurement method of the bubble number density based on this relationship is established. Using this method, the time-averaged spatial distribution of bubble number density under ultrasonic horn is revealed.

Three-Dimensional Reconstruction of Dilute Bubbly Flow Field With Light-Field Images Based on Deep Learning Method

The three-dimensional reconstruction of bubble flow field is of great significance to study the motion of gas-liquid two phase flow. Combined with the abundant information of bubbles in the light field images, a fast and accurate 3D reconstruction method for dilute bubbly flow based on the deep learning algorithm DIF-LeNet (Double Information Fusion with LeNet5 net) is proposed in this work. The calibration method and bubble segmentation algorithm of light field image for data processing is detailed. DIF-LeNet realizes the fusion of different dimension data and regression prediction. By DIF-LeNet, bubble depth could be predicted by fusing the refocused bubble image with the focal distance of the refocused image. Then, the three-dimensional bubble field could be reconstructed with the bubble depth, bubble center and bubble diameter. Comparing with the reconstruction method based on statistics and LeNet5, the reconstruction accuracy and speed of the proposed method based on DIF-LeNet are improved. Especially for the bubble without focused image in the refocused image sequence, the effect of 3D reconstruction based on DIF-LeNet is much prominent.

The effect of liquid co-flow on gas fractions, bubble velocities and chord lengths in bubbly flows. Part II: Asymmetric flow configurations

This paper describes the effects of uniform and non-uniform liquid co-flow on the bubbly flow in a rectangular column (with two inlets) deliberately aerated unevenly. The two vertical bubbly streams, comprising uniform bubbles, started interacting downstream of the trailing edge of a splitter plate. This study quantifies the emergence of buoyancy driven flow patterns as a function of the degree of a-symmetric gas sparging and (non-)uniform liquid co-flow by using Bubble Image Velocimetry (BIV) and dual-tip optical fibre probes. Without liquid co-flow, small differences in the gas fraction of the left and right inlet had a large effect on the mixing pattern, whereas a liquid co-flow stabilized a homogeneous flow regime and the flow pattern was less sensitive to gas fraction differences. Void fractions, bubble velocities and chord lengths were measured at two fixed position in the flow channel, whereas BIV provided a global overview of the flow structures. A correlation was developed to predict (a-symmetric) operating conditions for which the gas fraction of the left and right inlet are balanced, such that the bubble motion is governed by advection and no buoyancy driven flow structures arise. The data obtained is highly valuable for CFD validation and development purposes.

Classification of gas dispersion states via deep learning based on images obtained from a bubble sampler

The gas dispersion state within the bubble columns and reactors greatly affects their performance. A bubble sampler is a useful device to measure gas dispersion in the operating bubble columns and reactors as it has an adjustable sampling part that allows the device to obtain bubble samples from desired regions. However, implementing the bubble sampler as a real-time gas dispersion monitoring tool is difficult owing to the lack of reliable and automated image processing approach.In the present study, we developed a new convolutional neural network model to classify the gas dispersion states in a bubble column based on the images obtained with a bubble sampler. The model was trained with a labeled bubble image dataset comprising five different classes, which corresponded to five different gas dispersion states in the column. The average classification accuracy of the model was 97.5%. It was demonstrated that the trained model can accurately identify the change in gas dispersion state in real-time. The dataset created in this investigation and the pretrained BubbleNet can be found at http://dx.doi.org/10.17632/m3zjf8z286.1.

Research on statistical detection method of micro bubbles in transparent layer of quartz crucible based on image processing

Quartz crucible is a widely used manufacturing equipment in silicon crystal manufacturing industry, and the detection of its quality has a great influence on the purity of subsequent silicon crystals. In this paper, a machine vision-based bubble detection statistical system for transparent layer of quartz crucible is designed, and an intelligent algorithm for measuring the diameter, area and number of bubbles is proposed. Firstly, the bubble image is filtered and enhanced pre-processing operation to enhance the difference between the bubble and the background, and then the binary contour of the bubble is extracted by using a bubble contour extraction algorithm that fuses the morphological algorithm. In order to improve the accuracy of bubble detection, a curvature-based corner detection algorithm is proposed for overlapping bubbles.The experimental results show that the algorithm can accurately and real-time detect the number and diameter of bubbles. Compared with the connected component algorithm(CCA) and edge chain code information (EDCI) algorithm, the proposed method has lower cost and higher detection efficiency.

Wall shear stress modified by bubbles in a horizontal channel flow of silicone oil in the transition region

We performed laboratory experiments on bubbly channel flows using silicone oil, which has a low surface tension and clean interface to bubbles, as a test fluid to evaluate the wall shear stress modification for different regimes of bubble migration status. The channel Reynolds numbers of the flow ranged from 1000 to 5000, covering laminar, transition and turbulent flow regimes. The bubble deformation and swarms were classified as packing, film, foam, dispersed, and stretched states based on visualization of bubbles as a bulk void fraction changed. In the dispersed and film states, the wall shear stress reduced by 9% from that in the single-phase condition; by contrast, the wall shear stress increased in the stretched, packing, and foam states. We carried out statistical analysis of the time-series of the wall shear stress in the transition and turbulent-flow regimes. Variations of the PDF of the shear stress and the higher order moments in the statistic indicated that the injection of bubbles generated pseudo-turbulence in the transition regime and suppressed drag-inducing events in the turbulent regime. Bubble images and measurements of shear stress revealed a correlated wave with a time lag, for which we discuss associated to the bubble dynamics and effective viscosity of the bubble mixture in wall proximity.

Measurement of the air bubble size and velocity from micro air bubble generation (MBG) in diesel using optical methods

In this paper, we determine the bubble size and velocity from air bubble generation (MBG) in a diesel using optical methods. A KTM Series Pump was used to generate micro air bubbles in diesel. The air bubble radius and velocity measurements can be useful parameters to optimize the bubble generation process. Two optical systems were used for measurement air bubble sizes and their velocities in diesel. First, the optical system without an objective lens was used to determine the velocity of air bubbles in diesel. Another optical system with a 10× objective lens was used to obtain the size distribution of air bubbles generated in diesel. An available optical system with a 10× objective lens can detect a bubble diameter greater than 3.3 µm that air bubble images were processed using the ImageJ program. We measured the size distribution of air bubbles generated using the ImageJ program. The micro air bubble radius measured in diesel was found to be 6.26 µm in the sample after a month from air bubble generation. In addition, the particle image velocimetry (PIV) technique was used to measure the velocity field. Then, we used the OpenPIV program for PIV image processing. The highest velocity distribution was determined to be 90 mm/s for diesel without air bubbles and 20 mm/s for diesel with air bubbles after a month of the bubble generation.

Machine learning shadowgraph for particle size and shape characterization

Conventional image processing for a particle shadow image is usually time-consuming and suffers degraded image segmentation when dealing with images consisting of complex-shaped and clustered particles with varying backgrounds. In this paper, we introduce a robust learning-based method using a single convolution neural network for analyzing particle shadow images. Our approach employs a two-channel-output U-net model to generate a binary particle image and a particle centroid image. The binary particle image is subsequently segmented through a marker-controlled watershed approach with the particle centroid image as the marker image. The assessment of this method on both synthetic and experimental bubble images has exhibited a better performance compared to the state-of-art non-machine-learning method. The proposed machine learning shadow image processing approach provides a promising tool for real-time particle image analysis in industrial applications.

The effect of liquid co-flow on gas fractions, bubble velocities and chord lengths in bubbly flows. Part I: Uniform gas sparging and liquid co-flow

Unique experiments were performed in a homogeneously sparged rectangular 400×200×2630 mm (W×D×H) bubble column with and without liquid co-flow. Bubbles in the range 4–7 mm were produced by needle spargers, which resulted in a very uniform bubble size. Dual-tip optical fibre probes were used to measure horizontal profiles of gas fractions, bubble velocities and bubble chord lengths for superficial gas velocities Usg in the range 0.63–6.25 cm/s and superficial liquid velocities Usl up to 20 cm/s. Images of the bubble column were captured and a Bubble Image Velocimetry technique was adopted to calculate bubble (parcel) velocities. For low gas fractions, when a homogeneous flow regime occurred, both methods agreed very well and the optical fibre probes were found to be rather accurate for our bubbles. A liquid co-flow was found to have a calming effect and to stabilize a homogeneous bubbly flow regime, with less spatial variation in gas fractions and bubble velocities. Bubble chord lengths were almost normally distributed and do not exhibit the theoretical triangular probability density functions. The mean cord lengths were in the range 1.9–3.5 mm and found to increase with Usg and to decrease slightly with increasing Usl, while a liquid co-flow significantly reduced the standard deviation of the chord length distribution.