Particle detection and size recognition based on defocused particle images: a comparison of a deterministic algorithm and a deep neural network. - In: Experiments in fluids, ISSN 1432-1114, Bd. 64 (2023), 2, 21, S. 1-16
The systematic manipulation of components of multimodal particle solutions is a key for the design of modern industrial products and pharmaceuticals with highly customized properties. In order to optimize innovative particle separation devices on microfluidic scales, a particle size recognition with simultaneous volumetric position determination is essential. In the present study, the astigmatism particle tracking velocimetry is extended by a deterministic algorithm and a deep neural network (DNN) to include size classification of particles of multimodal size distribution. Without any adaptation of the existing measurement setup, a reliable classification of bimodal particle solutions in the size range of 1.14 μm–5.03 μm is demonstrated with a precision of up to 99.9 %. Concurrently, the high detection rate of the particles, suspended in a laminar fluid flow, is quantified by a recall of 99.0 %. By extracting particle images from the experimentally acquired images and placing them on a synthetic background, semi-synthetic images with consistent ground truth are generated. These contain labeled overlapping particle images that are correctly detected and classified by the DNN. The study is complemented by employing the presented algorithms for simultaneous size recognition of up to four particle species with a particle diameter in between 1.14 μm and 5.03 μm. With the very high precision of up to 99.3 % at a recall of 94.8 %, the applicability to classify multimodal particle mixtures even in dense solutions is confirmed. The present contribution thus paves the way for quantitative evaluation of microfluidic separation and mixing processes.
Experimental study of submerged liquid metal jet in a rectangular duct in a transverse magnetic field. - In: Journal of fluid mechanics, ISSN 1469-7645, Bd. 953 (2022), A10
A liquid metal flow in the form of a submerged round jet entering a square duct in the presence of a transverse magnetic field is studied experimentally. A range of high Reynolds and Hartmann numbers is considered. Flow velocity is measured using electric potential difference probes. A detailed study of the flow in the duct's cross-section about seven jet's diameters downstream of the inlet reveals the dynamics, which is unsteady and dominated by high-amplitude fluctuations resulting from the instability of the jet. The flow structure and fluctuation properties are largely determined by the value of the Stuart number N. At moderate N, the mean velocity profile retains a central jet with three-dimensional perturbations increasingly suppressed by the magnetic field as N grows. At higher values of N, the flow becomes quasi-two-dimensional and acquires the form of an asymmetric macrovortex, with high-amplitude velocity fluctuations reemerging.
Wide field of view stereoscopic PIV measurements in a Rayleigh-Bénard cell. - In: Experimentelle Strömungsmechanik - 29. Fachtagung, 6.-8. September 2022, Ilmenau, (2022), 44
Entwicklung eines magnetohydrodynamischen Pumpsystems für die Mikrofluidik. - In: Experimentelle Strömungsmechanik - 29. Fachtagung, 6.-8. September 2022, Ilmenau, (2022), 38
Kombinierte Geschwindigkeits- und Temperaturmessungen mittels LED und einer Doppelbildkamera. - In: Experimentelle Strömungsmechanik - 29. Fachtagung, 6.-8. September 2022, Ilmenau, (2022), 3
On the benefits and limitations of Echo State Networks for turbulent flow prediction. - In: Measurement science and technology, ISSN 1361-6501, Bd. 34 (2022), 1, 014002, S. 1-18
The prediction of turbulent flow by the application of machine learning (ML) algorithms to big data is a concept currently in its infancy which requires further development. It is of special importance if the aim is a prediction that is good in a statistical sense or if the vector fields should be predicted as good as possible. For this purpose, the statistical and deterministic prediction of the unsteady but periodic flow of the von Kármán Vortex Street (KVS) was examined using an Echo State Network (ESN) which is well suited for learning from time series due to its recurrent connections. The experimental data of the velocity field of the KVS were collected by Particle Image Velocimetry (PIV). Then, the data were reduced by Proper Orthogonal Decomposition (POD) and the flow was reconstructed by the first hundred most energetic modes. An ESN with 3000 neurons was optimized with respect to its three main hyperparameters to predict the time coefficients of the POD modes. For the deterministic prediction, the aim was to maximize the correct direction of the vertical velocities. The results indicate that the ESN can mimic the periodicity and the unsteadiness of the flow. It is also able to predict the sequence of the upward and downward directed velocities for longer time spans. For the statistical prediction, the similarity of the probability density functions of the vertical velocity fields between the predicted and actual flow was achieved. The leaking rate of the ESN played a key role in the transition from deterministic to statistical predictions.
Analysis of an unsteady quasi-capillary channel flow with time-resolved PIV and RBF-based super-resolution. - In: Journal of coatings technology and research, ISSN 1935-3804, (2022), insges. 14 S.
We investigate the interface dynamics in an unsteady quasi-capillary channel flow. The configuration consists of a liquid column that moves along a vertical 2D channel, open to the atmosphere and driven by a controlled pressure head. Both advancing and receding contact lines were analyzed to test the validity of classic models for dynamic wetting and to study the flow field near the interface. The operating conditions are characterized by a large acceleration, thus dominated by inertia. The shape of the moving meniscus was retrieved using Laser-Induced Fluorescence-based image processing, while the flow field near was analyzed via Time-Resolved Particle Image Velocimetry (TR-PIV). The TR-PIV measurements were enhanced in the post-processing, using a combination of Proper Orthogonal Decomposition and Radial Basis Functions to achieve super-resolution of the velocity field. Large counter-rotating vortices were observed, and their evolution was monitored in terms of the maximum intensity of the Q-field. The results show that classic contact angle models based on interface velocity cannot describe the evolution of the contact angle at a macroscopic scale. Moreover, the impact of the interface dynamics on the flow field is considerable and extends to several capillary lengths below the interface.
Experimentelle Strömungsmechanik - 29. Fachtagung, 6.-8. September 2022, Ilmenau. - Karlsruhe : GALA, 2022. - verschiedene Seitenzählungen ISBN 978-3-9816764-8-8
A combined velocity and temperature measurement with an LED and a low-speed camera. - In: Measurement science and technology, ISSN 1361-6501, Bd. 33 (2022), 11, 115301, S. 1-12
Microfluidic devices are governed by three-dimensional velocity and temperature fields, and their boundary conditions are often unknown. Therefore, a measurement technique is often desired to measure both fields in a volume. With astigmatism particle tracking velocimetry (APTV) combined with luminescence lifetime imaging, the temperature and all velocity components in a volume can be measured with one optical access. While the three-dimensional particle position is determined by evaluating the shape of the corresponding particle image, the temperature measurement relies on estimating the temperature-dependent luminescence lifetime derived from particle images on two subsequent image captures shortly after the photoexcitation. For this, typically a high-energetic pulsed laser is required to ensure a high signal-to-noise ratio. However, it can also cause additional heating of the fluid. We show that this problem is solved by replacing the pulsed laser with an LED. To compensate for the lower power provided by the LED, we adapted the timing schedule and vastly extended the illumination time and the exposure time for both image captures. In addition, we were able to replace the typically used high-speed camera with an ordinary double-frame camera. In this way, very low measurement uncertainties on all measured quantities can be achieved while keeping the temperature of the fluid unaffected. Random errors dominate within the two focal planes of APTV, yielding a standard deviation of the temperature of individual particles of about 1 only. The measurement error caused by the movement of tracer particles during the much longer illumination and exposure time were found to be acceptable when the measured velocity is low. With the circumvention of light-source induced heating and the lower cost of hardware devices, the adapted approach is a suitable measurement technique for microfluidic related research.
Highly efficient passive Tesla valves for microfluidic applications. - In: Microsystems & nanoengineering, ISSN 2055-7434, Bd. 8 (2022), 1, 97, S. 1-12
A multistage optimization method is developed yielding Tesla valves that are efficient even at low flow rates, characteristic, e.g., for almost all microfluidic systems, where passive valves have intrinsic advantages over active ones. We report on optimized structures that show a diodicity of up to 1.8 already at flow rates of 20 μl s^-1 corresponding to a Reynolds number of 36. Centerpiece of the design is a topological optimization based on the finite element method. It is set-up to yield easy-to-fabricate valve structures with a small footprint that can be directly used in microfluidic systems. Our numerical two-dimensional optimization takes into account the finite height of the channel approximately by means of a so-called shallow-channel approximation. Based on the three-dimensionally extruded optimized designs, various test structures were fabricated using standard, widely available microsystem manufacturing techniques. The manufacturing process is described in detail since it can be used for the production of similar cost-effective microfluidic systems. For the experimentally fabricated chips, the efficiency of the different valve designs, i.e., the diodicity defined as the ratio of the measured pressure drops in backward and forward flow directions, respectively, is measured and compared to theoretical predictions obtained from full 3D calculations of the Tesla valves. Good agreement is found. In addition to the direct measurement of the diodicities, the flow profiles in the fabricated test structures are determined using a two-dimensional microscopic particle image velocimetry (μPIV) method. Again, a reasonable good agreement of the measured flow profiles with simulated predictions is observed.