Direct Measurement of Wall Shear Stress in Single

Direct Measurement of Wall Shear Stress in Single- and Multiphase Flows Valery Sheverev, Lenterra Inc. and Bruce Brown, Srdjan Nesic, Ohio University presentation at the Institute for Corrosion and Multiphase Technology Ohio University Athens, OH October 11, 2011 1

Scope We report measurements of wall shear stress taken at five installations of the Institute for Corrosion and Multiphase Technology, using the first of its kind Lenterra Real. Shear. TM sensor 1. Single-phase flows a. 1” pipe flow loop (Britol 50 T oil, flow rate < 9 gpm) b. Thin Channel Flow Cell #1(water, flow rates < 25 gpm) c. Thin Channel Flow Cell #2(water, flow rate < 80 gpm) 2. Multi-phase flows a. Standing slug system b. Moving slug installation 2

Presenter - Lenterra, Inc. • Privately-owned emerging company, a provider of innovative sensor instrumentation based on its proprietary technologies. • Located in the New Jersey Institute of Technology (NJIT) Enterprise Development Center (incubator) in Newark, NJ. • $2. 6 M in SBIR (Small Business Innovation Research) grants from federal agencies (NSF, DOE, NASA) • Sales started in 2011 • 3 patents granted, 4 pending • Founder and current president: Enterprise Development Center, Newark, NJ Valery A Sheverev Industry Professor of Physics, Polytechnic Institute of New York University 3

Flow Force on the Wall y Flow z Pressure, scalar quantity x Wall shear stress (WSS), vector in x, z plane Wall u y x WSS, , is a product of velocity gradient (shear rate) near the wall, and dynamic viscosity of the fluid, : Slope=du/dy|y=0 Wall Sensing area Therefore WSS is an indirect measure of dynamic viscosity of the fluid or shear rate. 4

WSS Measurement – Needs in Oil and Gas Industry The breadth of applications in an industry that utilizes flows is evident when one considers the common use of pressure transducers and the fact that shear stress characterizes flow action much better than pressure • Flow Assurance • Single- and multiphase simulations – WSS is a critical parameter in most models • Characterization of multiphase flows (slug effects etc. ) • Direct detection of high and low viscosity components, especially in harsh environments such as high pressure and low temperature of deep water energy production • Corrosion Analysis • Direct relation: WSS ↑ → Corrosion Rate ↑ , via mass transfer • Corrosion inhibitors testing: WSS ↑ → removes inhibitor from the wall → Corrosion Rate↑ • Important parameter in corrosion models 5

WSS Measurement – History • Indirect measurements: WSS is inferred, through a set of assumptions, from another flow property, such as streamwise velocity or heat transfer rate measured at or near the wall • Require a model of the flow near the wall and knowledge of flow parameters such as temperature and viscosity • Examples: • hot-wire/film-based anemometry – quite rude estimate, no temporal resolution • laser-based near-wall flow velocity measurement • Laser-based or Particle Image Velocimetry (PIV) methods are well developed but they work only in transparent single-phase fluids (water, air) 6

WSS Measurement – Direct Methods u Direct - measure motion of a floating element, positioned flush within the wall. y Floating element displacement measured by : • Electrical techniques x Wall Floating element a. b. Piezoelectric – shear deformation of a PZT element Capacitor-based - floating element is one of the capacitor plates - shift of the floating element changes the capacitance • Drawbacks: - Susceptibility to electromagnetic interference - Narrow temperature range - Difficult to separate WSS from normal force (pressure) a. Optical techniques a. Variety of imaging or resonant methods (such as Fabry-Perot interferometers) to monitor floating element - not durable - require delicate alignment of the resonator - the resonator to be optically clean - difficult to sustain b. field All earlier direct techniques not robust enough for use in 7

Lenterra WSS Measurement Technology • Lenterra’s solution: Combine floating element with mechanical cantilever with micro-optical strain gage that are durable and not affected by the flow Floating element Wall Fiber-optical strain gages - Preferred type of optical strain gage: Fiber Bragg Grating (FBG) Optical fibers • Optical strain gage (FBG) versus: - resistive strain gages – not nearly sensitive Cantilever Sensor enclosure - semiconductor strain gages – sensitivity comparable but narrow temperature range - both types require delicate electronics (preamplifier) embedded in the probe • Robust: Materials used in our sensor: stainless steel + glass 8

Micro Optical Strain Gage - Fiber Bragg Grating FBGs are periodic structures of varying refractive index embedded in optical fibers. • FBG is attached to the cantilever. When the cantilever bends in response to shear stress, the FBG is strained which shifts its optical spectrum. Force Dl ~ Force • By interrogating FBG with a light source, this strain (and therefore WSS) can be measured by tracking the shift in the resonant wavelength. Dl 9

Temperature Compensation • Strain shifts FBG spectrum, but so does temperature • Solution: use two FBGs attached to opposite sides of cantilever Temp Strain FBG 1 (strain due to applied force increases spectral shift due to temperature) Temp Strain FBG 2 (strain decreases temperature shift) Differential signal (shift of FBG 1 spectrum less shift of FBG 2 spectrum) is independent from temperature 10

Spectrum Measurement and WSS Computation • Using laser diode (LD): Monochromatic light from a tunable laser is directed to FBGs and reflected light is recorded by photodiode Scanning laser Photodiode Fibers • As the laser frequency is tuned, reflection spectra (reflected light intensity versus wavelength) are recorded for FBG 1 & FBG 2 FBG • Shift in the resonant wavelengths (Dl = l. FBG 1 - l. FBG 2) calculated • WSS is found from τw = k. Dl where k is the calibration coefficient • Sensors are calibrated by applying a varying mechanical force F to the tip of the cantilever and measuring Dl, τw = F/A (A-area of floating element) • k it is determined by -Properties of FBGs -Area of the floating element - Elastic modulus, length and diameter of the cantilever 11

Lenterra Real. Shear™ Sensor - Probe The Real. Shear ™ sensor • 1/4″-80 threaded housing • FBGs are attached (currently glued) to cantilever • Detailed Specifications are found at www. lenterra. com 12

Lenterra Real. Shear™ Sensor – Complete Measurement System Includes : • A probe with connecting fibers • Controller combining optical components and data acquisition electronics • Computer • Measurement Software 13

Sensitivity to Off-Axis Shear Stress 1. 2 Normalized sensor response 1 0. 8 0. 6 0. 4 0. 2 0 -0. 2 -0. 4 -0. 6 -0. 8 -1 -1. 2 -20 0 20 40 60 80 100 Q, degrees 120 140 160 180 200 Real. Shear™ sensors have a bidirectional response and can be used to find flow direction at the wall. 14

Multi-Phase Detection: Two Disks Apparatus: rotating disk testbed Fluid - glycerin (viscosity 900 c. P) Lower disk rotating at 122 RPM Smaller tooth gap 0. 9 mm Wider tooth gap 1. 2 mm Two full rotations shown Plot on right: Air bubbles are induced behind teeth The Real. Shear ™ sensor 15

Single-Phase Tests at OU: WSS in a 1”Metal Pipe Test section: 1 m long Sensor floating element flush with the inner wall. Laminar Flow (Remax=130) Fluid: Britol 50 T oil 16

Single-Phase Tests at OU: WSS in a 1”Metal Pipe WSS vs. Flow Rate 100 Fluid: Britol 50 T oil Wall shear stress, Pa 80 Viscosity (μ =185 m. Pa-s) measured after completion of tests using a falling ball viscometer. U - averaged velocity (from flow rate), R - pipe radius Estimated data: analytical solution of Navier-Stokes equations for fully developed flow in cylindrical pipe. 60 40 Measured 10% uncertainty in the estimated data is due to uncertainty in viscosity and flow rate Estimated 20 0 0 300 600 900 1200 1500 1800 Pump motor RPM 17

Single-Phase Tests at OU – WSS in a 1”Metal Pipe Instantaneous Signal 18

Single-Phase Tests at OU – WSS in a 1”Metal Pipe Instantaneous Signal-Detail 19

Ohio University - Institute for Corrosion and Multiphase Technology Thin Channel Flow Cell (TCFC) (3 mm x 100 mm x 600 mm) Sensor’s sensitive surface 20

Single-Phase Tests at OU - Thin Channel Flow Cell #1 Instantaneous Signal 70 Random excitation of mechanical oscillations of cantilever by eddies of turbulent flow 65 Wall shear stress, Pa 60 55 50 45 40 35 30 25 20 8. 8 8. 9 9. 0 9. 1 Time, s 9. 1 9. 2 9. 3 21

Single-Phase Tests at OU - Thin Channel Flow Cell #1 WSS vs. Flow Rate 200 Fluid: water 180 U is averaged velocity (from flow rate) ρ is fluid density f is Darcy friction factor Wall shear stress, Pa 160 140 Room Temperature Flow Rate measured by a flowmeter 120 100 80 60 40 Measured Estimated (roughness 0. 1 mm) Estimated (roughness 0. 015 mm) 20 0 0 5 10 15 20 25 30 Estimated: from Darcy– Weisbach equation using Darcy friction factor (found from a Moody diagram, assuming a particular roughness), Flow rate, gpm Measurements arcing downwards – Why? Bruce Brown: top plate of Plexiglas bows up at higher flow rate, increases channel height 22

Single-Phase Tests at OU - Thin Channel Flow Cell #2 Time evolution Water Room Temp. Diff. Pressure 76 psi Measurement rate 10 k. S/s 23

Single-Phase Tests at OU - Thin Channel Flow Cell #2 Time evolution - detail Separate turbulences are readily observable 24

Single-Phase Tests at OU - Thin Channel Flow Cell #2 WSS vs. Flow Rate 800 Measured Re= 12350 -73500 Estimated using Darcy-Weisbach equation (0. 1 mm roughness) Estimated using Darcy-Weisbach equation (0. 015 mm roughness) System pressure 40 psig - total pressure is system pressure + differential pressure 700 Wall shear stress, Pa 600 500 Differential pressure between standard ports was directly measured 400 300 No saturation of WSS at higher flow rates (compare with TCFC #1) 200 100 Estimate: same model as with TCFC #1 data 0 0 5 10 15 20 25 30 35 40 45 50 55 Flow rate, gpm 25

Multi-Phase Tests at OU – Standing Slug (Water-Air) 4” Plexiglas pipe Sensor flush with the wall Sensor inserted here Adapter 26

Multi-Phase Tests at OU – Standing Slug (Water-Air) Adapter 27

Multi-Phase Tests at OU – Standing Slug (Water-Air) Clear indication of slug influence on WSS Instantaneous values of WSS in the slug are several times higher those observed in the upstream region ◘ Maximum measured WSS<100 Pa 28

Multi-Phase Tests at OU – Moving Slug (Water-Air) Sensor at the bottom of the pipe, flush with the wall. Water superficial velocity 0. 3 m/s Gas superficial velocity 3. 6 m/s No characteristic features caused by moving slugs observed Thus WSS due to slug < 4 Pa 29

What Did We Learn From the First Tests? • WSS was measured from few Pa to over 1 k. Pa, at pressures <50 • WSS on the pipe wall in laminar flow was measured systematically somewhat higher than predicted • Possibly due to calibration that does not include pressure difference across the floating element • Average values of WSS in two TCFCs (turbulent flows) are in the reasonable range of expected values • Standing water-air slug produced slightly higher instantaneous WSS (under 100 Pa). • Moving water-air slug showed no increase in WSS (<4 Pa) • High measurement rate allows to observe details of the turbulent flow (eddies striking the floating element) that are however masked by mechanical oscillations of the cantilever • Last observation lead to a new combined Wall Shear Stress Corrosion Sensor concept: 30

Advanced Corrosion Sensor Concept Based on Lenterra’s Sensor Design • “When you say the corrosion sensor is flush mounted, make sure it is flush” – words of a professional. • Local fluid turbulence created by a protruding sensor can have a major impact on the damage mechanisms and the rate of damage • Srdjan Nesic: “Lenterra’ WSS sensor can be modified to directly measure corrosion rate!” • V. Sheverev and S. Nesic “Methods And Devices For Monitoring Interaction Between A Fluid And A Wall” – PPA filed June 21, 2011. 31

State of the Art– Tuning Fork Corrosion Sensor • Mechanical oscillator - tuning fork tines attached to a diaphragm that is driven by a piezoceramic or another driver • Resonance frequency f 0=(1/2π)√(k/m) m =system mass, k= stiffness • Corrosion rate is incurred from change in f 0 • Problems: • Tuning forks needs to be immersed in the fluid – how to make it flush with the wall? • Some experiments resulted in decrease in f 0 when loss of material occurred. Why? • Answer: Corrosion affects not only the tip, but also the base – not only m is reduced, but k as well - f 0 may change in any direction 32

Corrosion Sensor – New Concept • Basic design similar to Real. Shear sensor • Oscillator consists of a cantilever and floating element – the oscillator system known as “cantilever with a tip mass” • Only the outer surface of floating element is corrodible (for example coupon inserted) • Due to corrosion, only mass of the floating element reduces changing m, stiffness k not changed since it is determined by cantilever only • The sensor is flush – minimal gaps are needed for detection of oscillation 33

Passive Excitation-Example FFT Spectra Signal from TCFC #1 140 400 120 Amplitude, arb. units 300 250 200 150 100 80 60 40 0 75. 45 75. 5 Time, seconds 75. 55 75. 6 100 110 120 130 140 Frequency, Hz 150 160 1000 140 120 100 80 60 40 20 0 Amplitude, arb. units 0 75. 4 100 20 50 Amplitude, arb. Units Singal, arb. units 350 10 1 0. 01 129 130 131 132 Frequency, Hz 133 100 Frequency, Hz 34

Thank You Any Questions? 35