Elsevier

Advances in Water Resources

Volume 70, August 2014, Pages 160-171
Advances in Water Resources

Review
Capabilities of Large-scale Particle Image Velocimetry to characterize shallow free-surface flows

https://doi.org/10.1016/j.advwatres.2014.04.004Get rights and content

Highlights

  • The Image Velocimetry (IV) techniques are reviewed by this paper.

  • Free-surface IV is the non-intrusive technique to measure velocities in shallow water.

  • A suite of shallow water flow experiments are conducted with IV techniques.

  • IV technique is the promising means to capture flow hydrodynamics in shallow water.

Abstract

Irrespective of their spatial extent, free-surface shallow flows are challenging measurement environments for most instruments due to the relatively small depths and velocities typically associated with these flows. A promising candidate for enabling measurements in such conditions is Large-scale Particle Image Velocimetry (LSPIV). This technique uses a non-intrusive approach to measure two-dimensional surface velocity fields with high spatial and temporal resolutions. Although there are many publications documenting the successful use of LSPIV in various laboratory and field open-channel flow situations, its performance has not been equally substantiated for measurement in shallow flows. This paper aims at filling in this gap by demonstrating the capabilities of LSPIV to: (a) accurately evaluate complex flow patterns in shallow channel flows; and (b) estimate depth in shallow flows using exclusively LSPIV measurements. The demonstration is provided by LSPIV measurements in three shallow flow laboratory situations with flow depths ranging from 0.05 to 0.31 m. The obtained measurements illustrate the LSPIV flexibility and reliability in measuring velocities in shallow and low-velocity (near-zero) flows. Moreover, the technique is capable to evaluate and map velocity-derived quantities that are difficult to document with alternative measurement techniques (e.g. vorticity and shear stress distributions and mapping of large-scale structure in the body of water).

Introduction

Open channel shallow flows are defined as largely unidirectional, turbulent shear flows occurring in a confined layer [21]. These confined flows lead to the development of strong large scale two-dimensional turbulent structures mutually interacting with the omnipresent small three-dimensional turbulence scales. The main geometrical scale for these flows is their depth. Shallow flows are omnipresent in various areas of the surrounding landscape and situations. They are typically encountered in small streams, ephemeral rivers, ponds, and river floodplains in the natural environment. Ubiquitous shallow flows in urban landscape are the storm water runoff over impervious surfaces (e.g. pavements and compacted soils) and gutter flows. Given their reduced depth, shallow flows are often associated with low-velocities over the cross section.

Documenting flow features in these measurement situations is usually problematic for both traditional methods (e.g. propeller meters) and the newer acoustic-based instruments. The mechanical-based techniques require a minimal water depth to accommodate the instrument measurement volumes (of the order of several centimeters), and they cannot accurately acquire near-zero velocities. The newer acoustic techniques (such as Acoustic Doppler Velocimeter, Acoustic Doppler Current Profiler) are quasi-non-intrusive but they require immersion of the probe in the water body [36]. The minimum measurement volumes of the acoustic instruments are of typically larger than those of mechanical methods and they also encounter difficulties in capturing low-velocity flows [4].

More promising for measuring velocities in natural-scale shallow flows are image-based techniques, collectively labeled herein as Large-scale Particle Image Velocimetry (LSPIV). These techniques stem from the Particle Image Velocimetry (PIV) methodology extensively used in laboratory fluid mechanics studies since the mid-1980s [2]. PIV is essentially a non-contact, non-intrusive velocity measurement technique that quantifies the movement of small and light particles moving within illuminated planes transecting the body of a fluid. The particles, a.k.a tracers, are expected to accurately follow the underlying flow and to uniformly “seed” the area to be measured.

Conventional PIV uses strong light sources (typically lasers) to adequately illuminate the tiny particles visualizing the flow motion. Images of the illuminated surfaces are captured with cameras using various combinations for the timing of the lighting and camera acquisition rate [2]. Velocities are eventually obtained using pattern-recognition algorithms based on statistical methods that allow to “guess with known confidence” where the image patterns are moving in a series of images recorded at known time intervals. Specifically, the recorded images are divided in small “interrogation areas” that are evaluated using statistical means (auto- or cross-correlations) to extract velocities. The process of interrogation is repeated for all areas of the recording leading to a whole-field velocity distribution over the imaged area.

LSPIV is a newer sub-family of the PIV technique and it is applied to moving fluid bodies with the goal to quantify velocities at the free surface. The first attempts to extend the PIV methodology for quantifying the velocity at the free surface was made in Japanese rivers in early 1990s [11]. Given that the areas subjected to velocity measurements are considerable larger that the laboratory-size PIV measured areas this variation of the technique was labeled “Large-scale” PIV. The technique was subsequently applied in laboratory conditions for quantifying velocity at the free surface in hydraulic models for riverine environment [30]. Ambient light is typically used for visualizing tracers on the free surface in field or laboratory measurements. In this paper we maintain the LSPIV labeling for the technique as the first community where the technique was applied is the hydraulic and hydrologic communities. This community continues to extensively use the LSPIV term even if some of the technique components are widely different that the original ones.

The main distinctions between PIV and LSPIV are the size of the measurement volume and the size and characteristics of the tracers used for visualization of the flow. Since the targeted area for LSPIV velocity measurements is order of magnitudes larger than those measured with conventional PIV, the image recording is usually made from an oblique angle (see Fig. 1). The flow tracers are also considerable larger for LSPIV compared with the micron-size particle used in the laboratory PIV. Practically, the tracers are not anymore individual particles as in PIV but clusters of particles of larger sizes that collectively form floating patterns at the free surface by the moving water body. The oblique angles used for recording inherently introduce distortion of the physical space images due to geometric perspective. The distortion is removed using a topo survey of few selected points contained in the images in conjunction with a geometrical transformation that relates the camera with real-world coordinates. The result of this transformation is that the LSPIV estimated velocities are provided as vectors with actual magnitudes and orientations everywhere over the imaged area.

Since its inception, a plethora of studies demonstrates the capabilities of LSPIV to efficiently measure spatial distributions of velocities that subsequently lead to quantitative and qualitative information for typical large-scale laboratory and field flows [7], [12], [18], [23], [30]. Some of the measurement environments subjected to LSPIV measurements were challenging, considerably more difficult to measure with alternative instruments. For example, acquisition of free-surface velocities in channels covered by ice floes (e.g. [9]). Another area of wide interest for LSPIV is the estimation of discharges in streams. For this purpose, the LSPIV surface velocity is used along with the bathymetry of the cross section within a velocity area method [3].

Although LSPIV has been successfully employed in a variety of field and laboratory conditions, there are substantially few reported measurements acquired with the technique in shallow water bodies. Often time these flows are also slow (at times close to zero velocities) and confined within changing and difficult to access boundaries such as in the floodplain flow situations. The non-intrusive and two-dimensional (2D) nature of the measurements coupled with its capability to instantaneously provide velocity over large areas, makes LSPIV a good candidate for the situation and at times the only one possible measurement solution (e.g. [32]). The goals for this review paper are to illustrate that: (a) LSPIV can measure correctly the free-surface velocity in complex shallow flow situations, and, (b) one can detect the water depth (or bottom surface) in shallow flows using a non-intrusive approach with LSPIV measurements as input. The illustration of the LSPIV use cases consists of results of three laboratory experiments. The first and second experiments demonstrate the first goal while the second one was tested in the third experiment. Before describing the three experimental conditions and sample results to substantiate the LSPIV merits, essential implementation aspects of the technique are presented that currently are scattered in various Refs. [18], [30].

Section snippets

Overview of Large-scale Particle Image Velocimetry

The most distinctive feature of LSPIV is that it is the only available technique that provides instantaneous velocity measurements on a plane. LSPIV is capable to simultaneously measure velocity over considerable area sizes (up to hundreds of m2) with high repetition rates (typically 60 Hz) without deploying a probe in the water body. This is a considerable advantage when flow conditions are hazardous for instruments or operation boats (e.g. during floods and flow with floating debris).

Experimental arrangement and procedure

In the 1960s, many streams and creeks in urban cities in Japan were reconstructed as concrete canals on all side for flood protection. Unfortunately, the reconstruction eliminated the natural flow features of rivers, making them unpleasant esthetically, especially at low flow when the water depth is only several centimeters with monotone water surface over a concrete flat bed, thereby people’s attention to river has been lost gradually. To enrich the stream landscape, installing small

Results

Fig. 3a shows an instantaneous surface velocity distribution for a single arrangement of the largest structure, having a base length of 16 cm and a height of 2 cm. The water depth was kept constant at 2 cm by a downstream tail gate, to keep the structure fully submerged. The distribution of the transverse velocity component shows that the submerged structure generated periodic large-scale von-Karman vortices in the downstream region. Moreover, a time series at a location downstream of the

Results

LSPIV allowed measuring 2D instantaneous surface velocities for the wetted parts of the surfaces. Fig. 7 shows an example of LSPIV velocities averaged on 40 image pairs for the experiment on the sinusoidal surface. LSPIV was able to measure the flow on each corrugation (channel effect) and the flow above the corrugation (in the direction of the main slope) in a satisfactory manner. Flow depth ranged from 15 mm in the corrugations to 1 mm or less above the corrugations, as described by Legout et

Results

Free-surface velocities were measured for the 20 experimental cases. Fig. 9 illustrates a sample of time averaged surface velocity distribution for the case with water depth, 0.254 m, bulk velocity, 0.1 m/s, and the big bump. The accuracy of the measured velocities in the near-wall channel area is lower than in the central part of the flume because of the particles sticking to the flume walls. To avoid this effect, bathymetry estimation was computed only in the area around the flume centerline,

Conclusions

The paper presented a suite of experiments conducted with free-surface IV techniques in shallow flows for documenting various hydro- and morpho-dynamic aspects of shallow flows. The non-intrusive, non-contact features of these techniques enable them to acquire velocity measurements where other techniques are difficult or impossible to employ. The techniques complement numerical simulations and other analytical tools that have provided most of the knowledge of shallow flows. The good performance

Acknowledgments

The first author was partially funded by the Iowa Flood Center, IIHR-Hydroscience & Engineering. The support is greatly appreciated. The corresponding author is supported by the Fundamental Research Funds for the Central Universities. The authors are thankful to three anonymous reviewers for their helpful comments on the manuscript.

References (38)

  • M. Church et al.

    Gravel-bed rivers: processes, tools, environments

    (2012)
  • J. Costa et al.

    Use of radars to monitor stream discharge by noncontact methods

    Water Resour Res

    (2006)
  • J.D. Creutin et al.

    Traceless quantitative imaging alternatives for free-surface measurements in natural streams

    Hydraul Meas Exp Methods

    (2002)
  • Eco-Foam. Four star plastics; 2013. http://www.fourstarplastics.com/pack8.htm Accessed:...
  • A.M. Fincham et al.

    Low-cost, high resolution DPIV for measurement in turbulent fluid flows

    Exp Fluids

    (1997)
  • I. Fujita et al.

    Application of video image analysis for measurements of river-surface flows

    Ann J Hydraulic Eng Japan Soc Civil Eng

    (1994)
  • I. Fujita et al.

    Large-scale particle image velocimetry for flow analysis in hydraulic applications

    J Hydraul Res

    (1998)
  • I. Fujita et al.

    Unseeded and seeded PIV measurements of river flows videotaped from a helicopter

    J Visual

    (2003)
  • I. Fujita

    Development of high-accurate particle image velocimetry based on improved optical flow method

    Ann J Hydraulic Eng JSCE

    (2004)
  • Cited by (56)

    • Adaptive large-scale particle image velocimetry method for physical model experiments of flood propagation with complex flow patterns

      2022, Measurement: Journal of the International Measurement Confederation
      Citation Excerpt :

      Although the principle of LSPIV measurement is similar to that of the traditional PIV method, compared with PIV measurements which use lasers to illuminate particles seeded in a fluid under laboratory conditions, LSPIV can tracks water surface patterns, such as floating debris and foams, under natural light illumination conditions to measure the velocity [35]. Especially, it is more effective for shallow flow measurement [36,37] and high flow measurement [14], even can be applied to evaluate the surface turbulence of a natural river [35,38]. Remarkably, river discharges can be determined as well through LSPIV measurement when river bathymetry and the vertical velocity profile are known [39].

    • Development of a three-axis accelerometer and large-scale particle image velocimetry (LSPIV) to enhance surface velocity measurements in rivers

      2021, Computers and Geosciences
      Citation Excerpt :

      To enhance the available measurement techniques, state-of-the-art approaches to measure the water surface velocity and river discharge are needed. Among these techniques, large-scale particle image velocimetry (LSPIV) using fixed stations is a remote surface velocity measurement system that displays promising potential for measuring surface velocity in river segments via the nonintrusive and continuous capture of images of floating objects (Muste et al., 2014; Tauro et al., 2015; Ran et al., 2016; Li et al., 2019). LSPIV provides a safe and easy option for the measurement of surface velocities and flows at large scales during flash flood events.

    View all citing articles on Scopus
    View full text