Measurement of Near-Bed Velocity and Bottom Shear Stress by Ferrofluids
In physical modelling of hydraulic processes, the knowledge of wall shear stress is fundamental to evaluate the energy dissipation of the flow and to assess the flow interaction with the bottom. Furthermore, the interaction between flow and solid walls influences to a great extent the characteristic of the flow itself. For such reasons measurements of wall shear stresses are extremely important.
In the past, several instruments have been developed and used to measure the shear stresses generated at the bottom in the presence of waves and currents.
Simple instruments, such as Preston tubes and Stanton tubes, have been used to measure bed shear stresses . Advantages of such instruments are the robustness. However, because their application is restricted to flows where the normal law of the wall is valid, they are affected by several limits as, for example, the tube dimensions must be smaller than wall layer and the difference between total (dynamic+static) and static pressure must be large.
Nowadays, velocity measurements are often carried out using system based on the acoustic and optic methodologies, such as ADV (Acoustic Doppler Velocimetry), LDA (Laser Doppler Anemometry), PIV (Particle image velocimetry) and PTV (Particle tracking Velocimetry). Some application of such instruments can be found in    end . Bottom shear stresses are derived from such instruments adopting theoretical approaches (e.g. log-fit, momentum integral method, etc). The advantage of such instruments is their reliability. However, such instruments have the following disadvantages: both acoustic and optical techniques cannot measure very close to the bottom, either due to the size of the sampling volume of the acoustic probe or to undesired reflection and disturbances from the bottom itself; the use of optical instruments is limited in large flumes; and measurements with such instruments are extremely difficult in the presence of sediments. Furthermore, for example at the bottom of sea waves, the velocity profiles are not logarithmic which is an underlying assumption of some instrument.
Direct measurements of bed shear stresses can be performed also by means of flush-mounted shear plates, which integrate the force over a relatively large area    . The system allows to resolve shear stress of O(1 mNm-2). However, one of the main problems while using shear plates is the tradeoff between sensor spatial resolution and the sensor capability to measure small forces.
Another class of instruments is the one of thermal sensors, to which hot film anemometers belong  . The system allows to measure turbulent fluctuations. Unfortunately, besides the well-known fragility of the sensors, hot film techniques are traditionally limited by difficulties in obtaining a unique calibration relationship between heat transfer and wall shear-stresses, from the reduction in sensitivity and complications in the dynamic response due to frequency-dependent conductive heat transfer into the substrate, and by measurement errors associated with mean temperature drift.
To overcome some of the above mentioned limits, recent studies     demonstrated the possibility to use ferrofluids sensors to measure near-bed velocity. Another promising method has been presented recently by  using sensor films with arrays of flexible micro-pillars sensing the wall shear stress by their bending in the flow. Ferrofluids sensors can be applied in the presence of sandy bottoms, over which state-of-the-art-instruments usually fail, and are characterized by a high robustness and a low cost.
Ferrofluids are two-state systems made up by small ferromagnetic particles (with size 1 nm-15 nm) dispersed in an organic non-magnetic solvent . A surfactant covers the nano-particles in order to prevent their agglomeration, which would be otherwise induced by the Van Der Waals forces and the magnetic forces. Such fluids behave like Newtonian fluid when the applied magnetic field is null or very small. Whereas if they are exposed to a magnetic field, their viscosity is strongly modified. At the microscopic scale, long chains of particles are formed in the direction of the magnetic field, whereas at the macro-scale a series of spikes aligned along the magnetic field appears. Such a phenomenon is called “the Rosensweig effect” . An example of Rosensweig effect is shown in Figure 1I and in Figure 1II. The use of a relatively large amount of ferrofluid O(10 ml) causes a formation of hexagonal conical patterns which are aligned along the magnetic field lines (see Figure 1b). A single fluidic pattern with a conical shape is generated using a volume of ferrofluid on the order of 2 ml, with a spike height approximately of 1 mm (see Figure 1III and Figure 1IV).
Considering the scheme of Figure 1d, the forces acting in the ferrofluid spike are: i) its weight; ii) the hydrostatic force; iii) the external magnetic field. In addition to the aforementioned forces, when the ferrofluid is subject to water flow, the effects of the dynamics force must be considered. In particular, the ferrofluid is subject to drag force, lift force, the friction generated on the surface of ferrofluid spike and the drag force generated at the bottom of the ferrofluid spike.
The order of magnitude of such forces are as follows : i)ferrofluid weight O(10-5 N); ii) drag force due the pressure in flow direction O(10-7-10-4 N); iii) lift force O(10-7-10-4 N); iv) friction generated on the surface of ferrofluid spike O(10-9-10--6 N); v) drag force generated at the bottom of the ferrofluid spike O(104 N);
As it can be seen from the above list, drag force generated at the bottom of the ferrofluid spike is the largest force in the balance (greater of several order of magnitude). Therefore, the other forces can be neglected. The magnetic force exerted by permanent magnets, although not directly measured, should be large enough to maintain the ferrofluid shape. Thus, the magnetic force must compensate the drag force generated at the bottom of the ferrofluid spike. Consequently, it can be concluded that, under flow action, the dynamic equilibrium lead to deformations and displacements of the ferrofluid spike which are both proportional to the magnitude of the bottom shear stress. Figure 2II and Figure 2III qualitatively show how the displacement of the ferrofluid spike increases as the bed shear stress increases. It follows that through an appropriate calibration of the system, shear stress measurements can be obtained by recovering the deformation of the spike.
State of art on the ferrofluid sensor calibration
With the aim to study the ferrofluid sensor under different flow conditions, in the last years, several tests were carried out at the Hydraulic Laboratory of the Department of Civil Engineering and Architecture, University of Catania. A preliminary analysis on the ferrofluid sensor was carried out in . In such study the tests were performed using a circular channel, named “annular cell” (Taylor Couette System), placed on a rotating platform that is driven to an alternate rotation motion with given amplitude and period to obtain a controlled fluid motion. The ferrofluid was placed on the bottom surface where a magnetic force was applied through the permanent magnet. The ferrofluid shape was monitored through a web-cam. The images of web-cam were used to estimate the displacement of the top apex of ferrofluid spike due to water flow. In the experimental campaign of  the ferrofluid shape was monitored through an inductive readout strategy which makes use of two planar coils located at the bottom of the flume (see Figure 3). The ferrofluid spike was placed initially at the center of the two coils. The movement of the ferrofluid spike produced a variation in the magnetic permeability and consequently a variation of the inductance in the coils. In this way, the perturbation of the magnetic field generated by the displacement of the spike is transduced into a voltage variation by a suitable conditioning circuit.
The conditioning circuit is made up by several components: an initial alternating current (AC) bridge and a differential amplifier (Circuit 1), a full wave rectifier and a filter, a differential amplifier which acts as a subtractor and filter (Circuit 2). The analysis was performed on the output both of the Circuit 1 and of the Circuit 2. The tests were carried out under steady currents in presence of smooth fixed bed (1.36%).
The analysis of the tests allowed to understand that the Circuit 2 is characterized by a larger sensitivity than Circuit 1, approximately by a factor of 30, and to observe a parabolic relationship between the output voltage and the near-bottom velocity and bottom shear stress.
In  the ferrofluid sensor was studied under regular surface waves on a rough bottom covered with medium coarse sand. The tests were carried out both in a flume which is 6 m long, 0.50 m wide and 0.70 m high (see Figure 4) and in a tank which is 18 m long, 3.6 wide and 1.0 high (see Figure 5). Experimental tests were executed both in hydrostatic conditions and in presence of regular linear waves. The sand used during the tests was characterized by a D50 equal to 0.56 mm. Obtained results indicate the applicability of the proposed strategy also in the presence of suspended impurities and a moderate bed sediments transport.
 extend the analysis by both increasing of the number of the tested flow conditions and studying the effects of the possible disturbances of the impact of moving sediment on the ferrofluid sensor. The goals of this study are (i) a quantitative assessment of the effect of the external controlling magnetic field and of the gain of the conditioning circuit on the calibration curves and on the error estimate is performed; (ii) a comparison of the performance of the system between steady current and wave conditions is carried out; and (iii) the effects of the impact of the sediments on the ferrofluid sensor are investigated in weak mobile bed conditions. The steady current tests were executed in a small scale flume which is 3.60 m long, 0.15 m wide and 0.36 m high (see Figure 6), the regular wave tests were executed in canal which is 18 m long, 3.6 wide and 1.0 high. During the experimental campaign, 12 steady current conditions and 83 tests under regular wave tests were carried out. The tests were performed changing the magnet type and the number of magnets.
The results of the experimental campaign can be resumed as follows:
- the lower limit of system working range is equal to 0.08 Nm-2 for both currents and regular waves;
- the upper limit is equal to 0.2 Nm-2 for steady currents and 0.4 Nm-2 in the presence of waves;
- in the presence of steady flow larger gains can be used without introducing significant noise in the measurements (gain of the conditioning circuit equal to 277 is considered in all the steady current tests); in the experimental wave conditions, the optimal value of the gain of the conditioning circuit is equal to 11.6;
- in steady current conditions, sensitivity of the system increases as the intensity of the magnetic field increases and the measurement errors are larger if the magnetic field is smaller;
- in steady current conditions, sensitivity of the system increases as the intensity of the magnetic field increases and the measurement errors are smaller if the magnetic field is smaller;
- the experimental campaign carried out in the presence of sediments confirms that the proposed measurement methodology is robust enough to be used in the presence of a weak bedload transport without costly damages to the sensor.
- M. R. Head and I. Rechenberg, “The Preston tube as a means of measuring skin friction,” J. Fluid Mech., vol. 14, no. 1, pp. 1–17, 1962.
- D. T. Cox, N. Kobayashi, and A. Okayasu, “Bottom shear stress in the surf zone,” J. Geophys. Res. Ocean., vol. 101, no. C6, pp. 14337–14348, 1996.
- W. N. Smith, P. Atsavapranee, J. Katz, and T. Osborn, “PIV measurements in the bottom boundary layer of the coastal ocean,” Exp. Fluids, vol. 33, no. 6, pp. 962–971, 2002.
- R. E. Musumeci, L. Cavallaro, E. Foti, P. Scandura, and P. Blondeaux, “Waves plus currents crossing at a right angle: Experimental investigation,” J. Geophys. Res. Ocean., vol. 111, no. C7, 2006.
- S.-J. Lee and H.-B. Kim, “Laboratory measurements of velocity and turbulence field behind porous fences,” J. Wind Eng. Ind. Aerodyn., vol. 80, no. 3, pp. 311–326, 1999.
- M. P. Barnes and T. E. Baldock, “Direct bed shear stress measurements in laboratory swash,” in Journal of Coastal Research, 2007, vol. 50, no. Special issue, pp. 641–645.
- K. L. Rankin and R. I. Hires, “Laboratory measurement of bottom shear stress on a movable bed,” J. Geophys. Res. Ocean., vol. 105, no. C7, pp. 17011–17019, 2000.
- Z. You, B. Yin, and G. Huo, “Direct Measurement of Wave-Induced Bottom Shear Stress Under Irregular Waves,” in Advances in Water Resources and Hydraulic Engineering, Springer, 2009, pp. 1213–1218.
- B. M. Sumer, M. M. Arnskov, N. Christiansen, and F. E. Jørgensen, “Two-component hot-film probe for measurements of wall shear stress,” Exp. Fluids, vol. 15, no. 6, pp. 380–384, 1993.
- G. Gust, “Skin friction probes for field applications,” J. Geophys. Res. Ocean., vol. 93, no. C11, pp. 14121–14132, 1988.
- B. Andò, S. Baglio, V. Marletta, E. Foti, and R. E. Musumeci, “Measurement of bottom velocities and shear stresses by ferrofluids at the sea bottom,” in Instrumentation and Measurement Technology Conference (I2MTC) Proceedings, 2014 IEEE International, 2014, pp. 728–731.
- R. E. Musumeci, V. MARLETTA, A. Bruno, S. BAGLIO, and F. Enrico, “Ferrofluid measurements of bottom velocities and shear stresses,” J. Hydrodyn. Ser. B, vol. 27, no. 1, pp. 150–158, 2015.
- R. E. Musumeci, V. Marletta, B. Andò, S. Baglio, and E. Foti, “Measurement of wave near-bed velocity and bottom shear stress by ferrofluids,” IEEE Trans. Instrum. Meas., vol. 64, no. 5, pp. 1224–1231, 2015.
- R. E. Musumeci, V. Marletta, A. Sanchez-Arcilla, and E. Foti, “A ferrofluid-based sensor to measure bottom shear stresses under currents and waves,” J. Hydraul. Res., pp. 1–18, 2018.
- C. Brücker, J. Spatz, and W. Schröder, “Feasability study of wall shear stress imaging using microstructured surfaces with flexible micropillars,” Exp. Fluids, vol. 39, no. 2, pp. 464–474, 2005.
- S. Odenbach, “Ferrofluids: Magnetically Controllable Fluids and Their Applications,” Appl. Rheol., vol. 14, no. 4, p. 179, 2004.
- M. D. Cowley and R. E. Rosensweig, “The interfacial stability of a ferromagnetic fluid,” J. Fluid Mech., vol. 30, no. 4, pp. 671–688, 1967.
- B. Andò, S. Baglio, C. Trigona, and C. Faraci, “Ferrofluids for a novel approach to the measurement of velocity profiles and shear stresses in boundary layers,” in Sensors, 2009 IEEE, 2009, pp. 1069–1071.