Application Of Proper Orthogonal Decomposition Research Articles

Experimental study of flag fluid–structure interaction in a laminar jet and application of POD

A fluttering flag in a uniform laminar flow exhibits complex fluid–structure interaction (FSI) behavior. This investigation experimentally studied aerodynamic load behavior and its connection to the oscillation behavior, flag envelope and fluid flow parameters, and application of Proper Orthogonal Decomposition (POD) to fluid flow data to elucidate changes in flow behavior associated with the changes in oscillation modes and observed drag coefficient values. For this purpose, three flag models with varying lengths were used. The aerodynamic load forces were measured using a high-precision load cell, and two-dimensional Particle Image Velocimetry (PIV) was used to quantify the flow field around the flag. The Reynolds number (Re), mass ratio (R1), and dimensionless rigidity (R2) values were varied between 4.4×104−12.3×104, 1.48−2.77, and 1.3×10−3−15.1×10−3, respectively. The results showed the linear relationship between drag coefficients with normalized amplitude of oscillation and Strouhal number. The results also showed the connection between observed drag and change in flag oscillation modes. The POD analysis showed that the energy content of the POD modes changed with the change in flag oscillation from mode-2 to mode-3 oscillations. The phase portrait of the first four POD modes also showed a unique interplay of POD modes, resulting in a change in the velocity flow field associated with the change in oscillation modes. The low-order reconstruction using select POD modes and control volume analysis of the velocity flow field showed a connection between POD modes and observed drag.

Steady and unsteady coupling in twin weakly underexpanded round jets

We investigate the intermittency of the coupling behaviour in screeching twin round supersonic jets at low Mach numbers across a range of nozzle spacings. Application of proper orthogonal decomposition combined with time-frequency wavelet analysis and spectral proper orthogonal decomposition shows that intermittency can manifest in twin jets as either a competition between the two symmetries, or the jets uncoupling and recoupling. The time scales on which symmetry switching occurs can vary strongly, ranging from $O(10^2)$ to $O(10^3)$ screech cycles. A transition from one symmetry to another is accompanied by a slight change in the screech frequency ranging from 0.30 % to 0.63 %. It was observed that complete uncoupling occurred only at the largest nozzle spacing of $s/D=6$ and at Mach numbers close to modal staging. When the jets are uncoupled they screech at slightly different frequencies, with a disparity of approximately 0.6 %. The coupling is particularly intermittent in the transition from the A1 to A2 branch, where the A2 mode is first observed, and tends toward steady coupling with increasing Mach number.

Application of proper orthogonal decomposition in particle image velocimetry in ocean engineering

Particle image velocimetry (PIV) plays a vital role in flow measurement. Eliminating the effects of overexposure and shadow in particle images is an urgent problem to be solved in the detailed flow measurement of ocean engineering. Seek a robust denoising method for bad test environments, and apply it to the PIV experiments to achieve dual denoising at the image preprocessing and data postprocessing. This study verified the superiority of the Proper Orthogonal Decomposition (POD) filter in pre- and postprocessing compared with traditional mathematical filters. Subsequently, the POD filter is applied to a practical project: the wake field measurement of bulk carrier. Regarding preprocessing, the POD filter is a valuable tool for objectively filtering typical high-frequency spatial noise from the energy perspective based on preserving the original information of particles to the greatest extent, effectively alleviating the impact of overexposure and shadow areas. Regarding postprocessing, the POD filter can fill the data gap by considering the surrounding flow characteristics. At the same time, the flow's spatial distribution and time evolution characteristics can be obtained by POD. The flow reconstructed by selecting the appropriate order model can clearly show the evolution characteristics of large-scale coherent structures, which may provide a deep understanding of complex flow field mechanisms.

Application of proper orthogonal decomposition for evaluation of coherent structures and energy contents in microbial biofilms

Microbial biofilms are structured communities of surface-associated microbial populations embedded in a matrix of extracellular polysaccharides that provide protection for biofilm cells. Among the wide plethora of microbial species adept at forming biofilms, the fungal pathogen Candida albicans (C. albicans), is one of the most notable. C. albicans biofilm development occurs in a series of sequential steps over a period of 24 h. Various quantitative and microscopic methods are available for the monitoring and evaluation of biofilms, including several innovative real-time methods for the evaluation of the cell-to-cell dynamics occurring during biofilm formation. These methods utilize biosensors which capture electrical, acoustic, and reflectance signals in bacterial populations (Li et al., 2021; Li et al., 2020; Kim et al., 2021; Paredes et al., 2021; Reipa et al., 2021). Additionally, machine learning, deep learning and other computational approaches have progressively been incorporated in the field of microbiology (Qu et al., 2019; Goodswen et al., 2021; Zhang et al., 2021; Ghannam and Techtmann, 2021; Rani et al., 2021; Berg et al., 2019) including some studies in biofilms (Buetti-Dinh et al., 2019; Srivastava et al., 2020; Hartmann et al., 2021; Dimauro et al., 2021) but given the potential of machine and deep learning, this niche is in large need of collaborative work between microbiology and engineering or physics experts to propel machine learning to a higher level. Therefore, whilst promising advances have been made, there is an urgent need for extensive development to take place to study and comprehend the complex interaction of microbial pathogens during biofilm formation. Specifically, there is a lot left to be understood about biofilm energy kinetics, and who the active microbial populations are. We infer that biofilm formation is an extremely diverse phenomenon and that each microorganism exerts different pathways to form a biofilm. Thus, we reasoned on the need for a model that would allow us to study the energy kinetics during C. albicans biofilm development. Modal decomposition techniques (MDTs) commonly used in fluid mechanics are gaining popularity outside their original field and might help decipher some of the dynamically relevant structures of biofilm formation. MDTs permit the identification of coherent structures in fluids and have been used in complex applications of information obtained during a particular time-lapse. A common MDT, Proper Orthogonal Decomposition (POD), can be used in reduced order modelling and machine learning applications. POD allows decomposition of a physical field influenced by different variables that may affect its physical properties. We aimed to evaluate the applicability of this technique in the analysis of energy kinetics during microbial biofilm formation, more specifically C. albicans biofilms. Using POD, we were able to easily distinguish visually distinct modes of growth of C. albicans cells in PBS and RPMI in terms of energy accumulation during the kinetic experiment. Comparing both PBS and RPMI, RPMI contains more energetic and dynamically relevant structures than PBS.

Flow and heat transfer in the wake of a triangular arrangement of spheres

This research work seeks to investigate the influence of spacing and heat transfer on the wake behavior of a triangular arrangement of spheres. Four experimental configurations have been investigated at three Reynolds numbers, Re1 = 350, Re2 = 700, and Re3 = 1050. Two isothermal cases were investigated with spacing between the spheres of zero and one sphere diameter, and two cases were investigated with an applied heat flux at the same spacing conditions. The time resolved particle image velocimetry results revealed various flow phenomena including flow separations, von Karman vortex shedding, and Kelvin–Helmholtz instabilities. The turbulent statistics reveal the effect of proximity and heat transfer on the time averaged values of the wake size, turbulent strengths, and Reynolds shear stress in the wake of each sphere, namely, the laminarization effects from the addition of heat and the suppression of the lead sphere wake from the proximity of the trailing spheres. These results are complemented by the application of proper orthogonal decomposition (POD) to the flow fields, which extracts the coherent structures from the flow. The modes that describe the coherent structures are extracted and described in detail, which provide further insight into effects of the experimental conditions on the temporal behavior of the flow. Many of the low order modes are found to be associated in pairs, corresponding to asymmetric structures or advection of a given structure downstream. The capability of POD to produce reduced order models of the flow is then utilized to facilitate vortex identification analysis. A turbulent kinetic energy based mode truncation criteria, which has been found to enhance vortex identification capability, is applied to select the POD modes and temporal coefficients to be used in the reduced order modeling. The reconstructed velocity fields are then analyzed with vortex identification algorithms to extract the vortex cores and boundaries. The combination of these approaches allows the study of the effect of proximity and heat transfer on the vortex characteristics, such as size, strength, and distribution.

Application of SPOD analysis to PIV data obtained in the wake of a circular cylinder undergoing vortex induced vibrations

Vortex induced vibrations (VIV) of a circular cylinder have been investigated experimentally using a cyberphysical apparatus with m∗ = 8, ζ = 0.005, and Re = 4000. This study considers the application of proper orthogonal decomposition (POD) and spectral POD (SPOD) analysis to the wake dynamics of the low-mass-ratio VIV of a circular cylinder in the lower branch at U∗ = 7.5. SPOD has been previously shown to better separate frequency-centered modal dynamics, compared to POD. Coherent POD and SPOD modes were compared and the newly separated third SPOD mode pair was found to have a periodicity characteristic of vortex shedding and a peak in the temporal coefficient spectra at St = f D/U∞ = 0.2248. The literature has identified that the wake dynamics within the lower branch are synchronized to the cylinder motion; however the present study suggests that some hidden dynamics persist at the Strouhal frequency. Low order models based on the first eight POD and SPOD modes were compared, and it was found that the filtering operation in SPOD removes the uncorrelated stochastic energy component of the POD modes while producing a comparable representation of the coherent deterministic part of the wake dynamics. Using SPOD to separate the distinct frequency-centered dynamics into unique, interpretable mode pairs will simplify future efforts to develop sparse dynamical models of the flow.

Application of Proper Orthogonal Decomposition to Study Coherent Flow Structures in a Saccular Aneurysm.

Aneurysms are localized expansions of weakened blood vessels that can be debilitating or fatal upon rupture. Previous studies have shown that flow in an aneurysm exhibits complex flow structures that are correlated with its inflow conditions. Therefore, the objective of this study was to demonstrate the application of proper orthogonal decomposition (POD) to study the impact of different inflow conditions on energetic flow structures and their temporal behavior in an aneurysm. To achieve this objective, experiments were performed on an idealized rigid sidewall aneurysm model. A piston pump system was used for precise inflow control, i.e., peak Reynolds number (Rep) and Womersley number (α) were varied from 50 to 270 and 2 to 5, respectively. The velocity flow field measurements at the midplane location of the idealized aneurysm model were performed using particle image velocimetry (PIV). The results demonstrate the efficacy of POD in decomposing complex data, and POD was able to capture the energetic flow structures unique to each studied inflow condition. Furthermore, the time-varying coefficient results highlighted the interplay between the coefficients and their corresponding POD modes, which in turn helped explain how POD modes impact certain flow features. The low-order reconstruction results were able to capture the flow evolution and provide information on complex flow in an aneurysm. The POD and low-order reconstruction results also indicated that vortex formation, evolution, and convection varied with an increase in α, while vortex strength and formation of secondary structures were correlated with an increase in Rep.

Modal decomposition of the unsteady non-reactive flow field in a swirl-stabilized combustor operated by a Lean Premixed injection system

This work is focused on the numerical study of low-order coherent structures of a high-swirled laboratory-scaled combustor operated by a Lean Premixed (LP) injection system in non-reacting conditions through different flow modal decomposition techniques. This will provide valuable insight into the time-spatial modal structure detecting coherent spatial patterns. Experiments suggest the appearance of a self-excited hydrodynamic instability characterized by a single dominant frequency. On the one hand, the dominant pulsating energy components associated with the Precessing Vortex Core (PVC) are identified through the application of Proper Orthogonal Decomposition (POD) to the instantaneous velocity field. On the other hand, Dynamic Mode Decomposition (DMD) is proven to effectively highlight the relation between the frequency of the most dominant unsteady vortex structures and their spatial distribution within the combustor. Since DMD analysis generates a global frequency spectrum in which each mode corresponds to a specific discrete frequency, its application has been demonstrated to be more efficient than POD when dealing with temporally coherent problems. In this way, the DMD technique has proved to be a robust and systematic method that can give accurate and consistent interpretations of the periodic physics underlying hydrodynamic instabilities in the combustor studied in the present investigation.

Numerical investigation of unsteady cavitation around a twisted hydrofoil

In this paper, the unsteady cavitating flow around a symmetrical twisted hydrofoil is investigated numerically. Cavitating flow characteristics are analyzed in terms of dynamical behaviors and temporal/spatial fluctuations of the cavities along the hydrofoil. At the midsection of the foil, sheet cavity firstly grows with nearly constant speed until the occurrence of the reverse flow at the closure line. During the reverse phase the flow moves upstream, and the sheet cavity keeps growing turbulently with low vapor content at the closure area. Fast Fourier Transform (FFT), Bispecturm and Dynamic Mode Decomposition (DMD) analyses show the existence of harmonics of the shedding frequency, of which the double and triple frequency are captured but only the fundamental frequency dominates the cavitating field. Pressure fluctuations around the hydrofoil and force coefficients are governed by the acceleration of vapor cavity. Application of Proper Orthogonal Decomposition (POD) displays the large-scale coherent structures among cavity shedding evolution. The sheet cavity growth, the main cavity shrinking and developing into a pair of root-like cavity structure and the sector cavity structure are captured by the first mode.

Identification of structures and mechanisms in a flow field by POD analysis for input data obtained from visualization and PIV

This paper investigates the application of proper orthogonal decomposition (POD) for data obtained from visualizations. Using the POD method, the flow field behind one and two cylinders in a staggered configuration was analyzed. The data processed by this method were obtained from experimental measurements of flow fields using the particle image velocimetry (PIV) method and visualization. The dominant frequencies of the flow pattern from these data were compared using constant temperature anemometry (CTA) measurements. Attention was mainly focused on the flow at three Reynolds numbers: 500, 1200, and 2500. Velocity and vortex fields were created from PIV measurements in the wind tunnel for Re = 500, and video images of flow fields were obtained from dye visualizations in the hydrodynamic tunnel. The components of velocity, vorticity (both of PIV), and change in grayscale (from visualization) were used as input data for POD analysis. A methodology for data processing from visualizations was developed for subsequent analysis using the POD method. A new technique has been found to identify structures in the wake of the cylinders in a staggered configuration by analyzing POD based on various types of input data. The measured fields of dominant frequencies from the CTA and a thorough analysis of the POD modes and their relative energy values for each type of data made it possible to identify the structures and mechanisms that occur in the wake of cylinders. This analysis facilitated a better understanding of the importance of these structures and mechanisms, which can then be used to control the flow behind the cylinders.Graphic abstract

Salient three-dimensional features of the turbulent wake of a simplified square-back vehicle

In this paper, the unsteady wake of a simplified square-back vehicle, with and without wheels, is investigated using large scale Tomographic Particle Image Velocimetry (TPIV), at a Reynolds number of ReH = 5.78 · 105 (based on the model height). In the no-wheel case, the time averaged wake features a balanced toroidal shape, with a good level of symmetry in both vertical and lateral directions. However analysis of the wake dynamics shows this widely accepted result to be a poor model of the wake structure. Application of Proper Orthogonal Decomposition (POD) to the unsteady data reveals the existence of the widely reported bi-stable behaviour, consisting of random switches between twolateral symmetry breaking states. For the first time, the 3D topology of each state is fully characterised and the changes in wake topology during the switches between bistable states are also described. Each symmetry breaking state is shown to feature a characteristic ‘hairpin vortex’ structure that is the result of the merging of two horseshoe vortices, aligned with the vertical edges of the model base. The mutual interactions between these vortices are found to be at the origin of the bi-stable mode. The vertical symmetry is lost when wheels are added to the model, resulting in the formation of an upwash dominated wake. The bi-stable behaviour is removed but considerable mobility in the near-wake remains, in the form of a swinging motion of the rear recirculation.

Three dimensional structure of the unsteady wake of an axisymmetric body

The near-wake of an axisymmetric body has been investigated using base pressure tappings and large scale Tomographic Particle Image Velocimetry (TPIV) at a Reynolds number of ReD = 3.2 × 105, based upon model diameter. Insights into the near-wake dynamics are provided by the application of Proper Orthogonal Decomposition (POD) to the pressure and the TPIV datasets. The first two POD modes show that the axisymmetric topology seen in the time averaged field is the result of the combination of different reflectional symmetry preserving states, each one featuring a hairpin vortex surrounded by an annular structure developing in proximity to the wake closure. The “head” and the “tails” of each hairpin vortex appear to be dynamically linked, as also proven by the existence of a second pair of modes, visible only in the TPIV dataset, featuring a twisted two-lobe structure. The analysis of the temporal evolution of the radial position of the centre of pressure over the model base reveals the existence of two different low-drag scenarios, characterised by the restoration of the axial symmetry or the selection of a single plane of reflectional symmetry. The first state is reported to become the only admissible low-drag configuration when the short-time wake dynamics are removed from the unsteady pressure signal.

On the Identification and Decomposition of the Unsteady Losses in a Turbine Cascade

The present paper describes the application of proper orthogonal decomposition (POD) to large eddy simulation (LES) of the T106A low-pressure-turbine profile with unsteady incoming wakes at two different flow conditions. Conventional data analysis applied to time averaged or phase-locked averaged flow fields is not always able to identify and quantify the different sources of losses in the unsteady flow field as they are able to isolate only the deterministic contribution. A newly developed procedure allows such identification of the unsteady loss contribution due to the migration of the incoming wakes, as well as to construct reduced order models that are able to highlight unsteady losses due to larger and/or smaller flow structures carried by the wakes in the different parts of the blade boundary layers. This enables a designer to identify the dominant modes (i.e., phenomena) responsible for loss, the associated generation mechanism, their dynamics, and spatial location. The procedure applied to the two cases shows that losses in the fore part of the blade suction side are basically unaffected by the flow unsteadiness, irrespective of the reduced frequency and the flow coefficient. On the other hand, in the rear part of the suction side, the unsteadiness contributes to losses prevalently due to the finer scale (higher order POD modes) embedded into the bulk of the incoming wake. The main difference between the two cases has been identified by the losses produced in the core flow region, where both the largest scale structures and the finer ones produces turbulence during migration. The decomposition into POD modes allows the quantification of this latter extra losses generated in the core flow region, providing further inputs to the designers for future optimization strategies.

POD-Based Data Mining of Turbulent Flows in Front of and on Top of Smooth and Roughness-Resolved Forward-Facing Steps

A new application of proper orthogonal decomposition (POD) to uncover the relation of the instantaneous energetic large-scale turbulent structures to the dominant POD modes had been reported. Motivated on this method, the data mining from the velocity vector fields measured by the particle image velocimetry and structural analysis on the selected POD-based reconstructed turbulent flows in front of and on top of smooth and roughness-resolved forward-facing steps (FFSs) has been performed. The velocity fields containing the most energetic large-scale structures are conditionally chosen. The conditional criterion is that the chosen velocity fields whose POD temporal coefficients of the first or second mode are correspondingly larger than twice of their root mean square values. Typically, the most energetic instantaneous structures are the large-scale second-quadrant (Q2) or fourth-quadrant (Q4) events and the mostly open separation bubbles in front of these FFSs; while the large-scale structures behave as a strong shear layer, and near or in which are a few spanwise prograde or retrograde vortices; and sometimes the alternating node and saddle points appeared. Similarly, for the flow on top of the FFSs, the most energetic structures are presented as a great many large-scale Q4 or Q2 events and a few secondary vortices at the very near wall; while the large-scale structures are overall exhibited as a strong shear layer, and in which a large number of vertex structures are created. These energetic large-scale structures here are not only sensitive to the surface roughness conditions but also to the spanwise locations.

On the Application of Proper Orthogonal Decomposition (POD) for In-Cylinder Flow Analysis

Proper orthogonal decomposition (POD) is a coherent structure identification technique based on either measured or computed data sets. Recently, POD has been adopted for the analysis of the in-cylinder flows inside internal combustion engines. In this study, stereoscopic particle image velocimetry (Stereo-PIV) measurements were carried out at the central vertical tumble plane inside an engine cylinder to acquire the velocity vector fields for the in-cylinder flow under different experimental conditions. Afterwards, the POD analysis were performed firstly on synthetic velocity vector fields with known characteristics in order to extract some fundamental properties of the POD technique. These data were used to reveal how the physical properties of coherent structures were captured and distributed among the POD modes, in addition to illustrate the difference between subtracting and non-subtracting the ensemble average prior to conducting POD on datasets. Moreover, two case studies for the in-cylinder flow at different valve lifts and different pressure differences across the air intake valves were presented and discussed as the effect of both valve lifts and pressure difference have not been investigated before using phase-invariant POD analysis. The results demonstrated that for repeatable flow pattern, only the first mode was sufficient to reconstruct the physical properties of the flow. Furthermore, POD analysis confirmed the negligible effect of pressure difference and subsequently the effect of engine speed on flow structures.

Scale-adaptive simulation of unsteady flow and dispersion around a model building: spectral and POD analyses

The present paper evaluates the relative performance of scale-adaptive simulation (SAS) in modelling unsteady concentration and flow fields around a model building relative to other transient simulations such as large eddy simulation (LES) and unsteady Reynolds-averaged Navier–Stokes (URANS) models. A novel application of proper orthogonal decomposition (POD) and time–frequency analysis is carried out in order to evaluate the transient behaviour and dominant structures of the flow fields predicted by SAS and LES. Results represent the outstanding performance of SAS in comparison with the URANS computation based on the SST k–ω model. This better performance is related to the accurate reproduction of unsteady fluctuations around the model building by SAS. In addition, the quantitative and qualitative agreements for the shapes and magnitudes of POD modes between SAS and LES confirm the LES-like behaviour of SAS in the wake region. However, in terms of computational performance, SAS imposes an extra CPU cost as compared with LES for the same grid resolution.

Application of proper orthogonal decomposition and radial basis functions for crack size estimation using particle swarm optimization

Complex engineering problems require simulations, which are computationally expensive in cases of inverse identification tasks since they commonly requires hundreds of thousands of simulations. This paper propose a method based on model reduction for crack size estimation, combining the proper orthogonal decomposition method with radial basis functions. The reduced model is validated by comparing the obtained boundary displacements with the corresponding results from a finite element model. This inverse procedure is formulated as the minimization of the difference between the measured and computed values of displacement at selected boundary nodes, called sensor points, using particle swarm optimization algorithm. Convex and a non-convex specimens have been considered for investigations of crack presence, and identification of its size, different crack sizes have been tested to demonstrate the efficiency of the proposed approach.

Faster prediction of wildfire behaviour by physical models through application of proper orthogonal decomposition

Physical models of wildfires are of particular interest in fire behaviour research and have applications in firefighting, rescue and evacuation. However, physical models present a challenge as a result of the large computational resources they often require, especially for the analysis of large areas or when multiple scenarios are investigated. The objective of this paper is to explore the opportunity to reduce the computation time requested by physical wildfire models through application of a model order reduction technique, specifically the proper orthogonal decomposition (POD) technique. POD is here applied to a simple one-dimensional physical model. The full physical model for illustration of the concept is first tested with experimental data to check its ability to simulate wildfire behaviour; it is then reduced using the POD technique. It is shown that the reduced model is able to simulate fire propagation with only small deviations in results in comparison with the physical model (~6.4% deviation in the rate of spread, ROS) and a drastic reduction (~85%) in computational cost. The results demonstrate the advantages of applying effective reduction techniques to create new generations of fire models based on reduced physical approaches. The potential applicability of POD to more complex models is also discussed.

Application of Proper Orthogonal Decomposition in Fast Fourier Transform—Assisted Multivariate Nonstationary Process Simulation

AbstractThe classic spectral representation method (SRM)-based nonstationary process simulation algorithm is used extensively in the engineering community. However, it is less efficient owing to the unavailability of fast Fourier transform (FFT). In this paper, an efficient, almost accurate, and straightforward algorithm is developed for the simulation of the multivariate nonstationary process. In this method, an evolutionary spectral matrix is decomposed via Cholesky method, and then proper orthogonal decomposition (POD) is used to factorize decomposed spectra as the summation of the products of time and frequency functions. Because original time-dependent decomposed spectra are decoupled via factorization, FFT can be used to significantly expedite the simulation efficiency. This POD-based factorization is totally data-driven and optimal, and fewer items are required in matching decomposed spectra. Therefore, the accuracy and efficiency of the factorization can be guaranteed at the same time. Another att...

Application of Proper Orthogonal Decomposition to the morphological analysis of confined co-axial jets of immiscible liquids with comparable densities

The development of a round liquid jet under the influence of a confined coaxial flow of an immiscible liquid of comparable density (central to annular flow density ratio of 8:10) was investigated in the vicinity of the nozzle exit. Two flow regimes were considered; one where the annular flow is faster than the central jet, so the central liquid jet is accelerated and one where the annular flow is slower, so the central liquid jet is decelerated. The central jet was visualised by high speed photography. Three modes of jet development were identified and classified in terms of the Reynolds number, Re, of the central jet which was in the range of 525 < Re < 2725, a modified definition of the Weber number, We, which allows the distinction between accelerating and deceleration flows and was in the range of −22 < We < 67 and the annular to central Momentum Ratio, MR, of the two streams which was in the range of 3.6 < MR < 91. By processing the time resolved jet images using Proper Orthogonal Decomposition (POD), it was possible to reduce the description of jet morphology to a small number of spatial modes, which isolated the most significant morphologies of the jet development. In this way, the temporal and spatial characteristics of the instabilities on the interface were clearly identified which highlights the advantages of POD over direct observation of the images. Relationships between the flow parameters and the interfacial waves were established. The wavelength of the interfacial instability was found to depend on the velocity of the fastest moving stream, which is contrary to findings for fluids with large density differences.