Monitoring Multiphase Flow via Earth s Field Nuclear Magnetic Resonance

Transcription

1 Monitoring Multiphase Flow via Earth s Field Nuclear Magnetic Resonance Keelan O Neill University of Western Australia School of Mechanical and Chemical Engineering SUT Subsea Controls Down Under 2 th October 26 Fluid Science & Resources

2 Total shipments ($US million) Motivation for research The future of oil and gas production Development of deeper offshore fields and marginal fields Increased produced water Increased subsea processing Subsea production arrangement Objectives of flow metering Measurement of phase fractions Measurement of phase velocities Flow regime identification 5 World Multiphase flow meter market Subsea manifolds 4 3 4% growth 2 Production hub Satellite field wells Satellite field wells Atkinson, I., et al., A New Horizon in Multiphase Flow Measurement. Oilfield Review, 24. 6(4): p Year J. Yoder, The many phases of multiphase flowmeters, Pipeline & Gas Journal, 24 (23) 4-4 2

3 Superficial Liquid Velocity, U SL [m/s] Multiphase flow meter technologies Common measurement principles Gamma ray attenuation Electrical impedance measurement Ultrasonic measurement systems Advantages of nuclear magnetic resonance Non-invasive measurement Non-radioactive technology Flow regime independent Gamma-ray source Collimator Flow Detector Flow direction Downstream transducer Upstream transducer Ultrasonic emission Commercial magnetic resonance flowmeter: M-Phase 5 Superficial Gas Velocity, U SG [m/s] 3

4 Nuclear magnetic resonance Summary of nuclear magnetic resonance (NMR) magnetic field strengths Earth s field NMR Well-logging tool Magnetic resonance imaging Chemical spectroscopy Ultra-low field Low field High field 5 µt 5 mt 7 T 4

5 Nuclear magnetic resonance Basic classical mechanics description. Align nuclei ( H atoms) with magnetic field (polarisation) 2. Apply radio frequency pulse B Radio frequency pulse Pulse and Collect Experiment Free induction decay Z 9 o RF pulse t delay t a 3. Observe the magnetisation relax X Y M 5

6 The Earth s field NMR flow meter Key system features Detected in the Earth s magnetic field Time of flight measurement Remote detection system Flow direction Pre-polarising magnet Faraday cage EFNMR detection coil 2. Excitation. Polarisation Radio frequency pulse B 3. Detection Z X Independent flowrate measurement EFNMR: Earth s field nuclear magnetic resonance Y 6 M

7 Model for NMR signal Polarisation Intermediate decay Detection Pulse and Collect Sequence 9 o RF pulse Free induction decay Flow Halbach array EFNMR Detection L P τ PD = L PD v S(v, t a ) = S e τ P T L PD τ PD = L PD v e τ PD T L D t delay + t a τ D τ D = L D v e t delay +t a T 2 t delay Parameters L: length τ: residence time v: velocity S : initial magnetization T : spin-lattice relaxation T 2* : observed spin-spin relaxation t a 7

8 Tikhonov Regularisation The inverse problem Model transfer matrix A P v = S P(v) = A S Velocity probability distribution NMR Signal r Pipe velocity distribution z Tikhonov regularisation Least squares fitting method Allows complex models to be fit to experimental data Real image Image with noise Reconstructed image M. Bertero and P. Boccacci, Introduction to inverse problems in imaging. 998, CRC Press. 8

9 NMR Signal [μv] P(v) Single phase velocity analysis Experimental conditions Single phase water flows at 4 52 L/min (.8.5 m/s) Experimental NMR signals fit using regularisation.8 m/s.35 m/s.53 m/s.7 m/s.88 m/s.6 m/s Expt. data Model fit Corresponding velocity probability distributions.8 m/s.35 m/s.53 m/s.7 m/s.88 m/s.6 m/s Time since pulse [s] Velocity [m/s] 9

10 Predicted mean velocity from NMR analysis [m/s] NMR predicted velocity deviation Velocity comparison NMR measurement section Independent rotameter measurement (Uncertainty of ± 5%).2 Comparison of measured mean velocities 5.% Deviation plot.8 2.5%.6.% % -5.% Measured mean velocity from in-line rotameter [m/s] Mean absolute error =.9 % Measured mean velocity from in-line rotameter [m/s]

11 P(v) NMR predicted velocities [m/s] Two pipe analysis Experimental conditions Overall flow rates: 5 5 L/min Separation distance (L PD ):.68, and.98 m Flow Pipe A Pipe B Halbach array L PD EFNMR detection coil Velocity (pipe A) ~ 33% of velocity (pipe B) 4 Two pipe system velocity probability distributions.2 Comparison of measured mean velocities 3 3 l/min 4 l/min 5 l/min Velocity [m/s] L PD.68 m.78 m.88 m.98 m Rotameter measured velocities [m/s] A Pipe B

12 Liquid Holdup NMR Signal [µv] Gas/liquid analysis Flow regime identification Stratified flow Slug flow Time [s] Video analysis region Video analysis ID: 3 mm EFNMR Detection Coil Halbach Array U SL =.3 m/s U SG =.88 m/s Time [s] 2

13 Velocity [m/s] P(v) Liquid Holdup NMR signal [µv] Velocity and holdup determination Velocity probability distribution Processed signal & model fit Experimental data Model fit Liquid holdup determination S,2P S,Full Velocity [m/s] Time since pulse [s] x L = S,2P S,Full Stratified flow Slug flow Overall time [s] Overall time [s] 3

14 Liquid Holdup Liquid Holdup Liquid Holdup Holdup analysis comparison Stratified flow..75 Slug unit Liquid: 4 L/min Gas: 4 L/min Video holdup NMR holdup Background stratified component Low frequency slug flow Liquid: 8 L/min Gas: 4 L/min Video holdup NMR holdup Background stratified under-prediction Gas bubbles Meniscus Increased relaxation Higher frequency slug flow Liquid: 2 L/min Gas: 2 L/min Experimental Time [s] Video holdup NMR holdup Partial slug capture Slug residence time ~.2 s Scan time ~.7 s Slug not fully captured 4

15 NMR predicted superficial velocity [m/s] NMR velocity standard deviation [m/s] Liquid Holdup Velocity [m/s] Two phase velocity analysis Overall time [s] Overall time [s].4.3 U SL = N i= N x i v i Gas Flow rate L/min 2 L/min 4 L/min Slug Flow.2.. Gas Flow rate L/min 2 L/min 4 L/min Rotameter measured superficial velocity [m/s]. Stratified Flow..2.3 Rotameter measured superficial velocity [m/s] 5

16 Velocity [m/s] P(v) Conclusions Can accurately measure and model the FID signal of a moving fluid through the flow metering system Can determine the velocity probability distribution of liquid moving in the system via Tikhonov regularisation m/s.66 m/s. m/s Velocity [m/s] Able estimate the liquid velocity and holdup over time for stratified and slug flow Overall time [s] Stratified flow Slug flow 6

17 Superficial Liquid Velocity, U SL [m/s] Future work Analyse fluid flow in further air/water flow regimes Incorporate oil into flow metering system Apply dynamic nuclear polarisation for signal enhancement Superficial Gas Velocity, U SG [m/s] Fully develop analysis techniques to be able to interpret three phase flow (oil/gas/water) Gas Oil Water Hogendoorn, J., et al., Magnetic resonance multiphase flowmeter: measuring principle and multiple test results, in Upstream Production Measurement. 25: Houston. 7

18 Acknowledgements Supervisors Michael Johns, Einar Fridjonsson and Paul Stanwix Final Year Project Students Adeline Klotz and Jason Collis Thank you for your attention! Questions? 8

19 P(v) Tikhonov regularisation The inverse problem; Model transfer matrix A P v = S P(v) = A S Velocity probability distribution NMR Signal Apply Tikhonov regularisation; min ( A P v S 2 + λ P v 2 ) 8 6 Pipe velocity distribution r z λ = -3 λ = 75 (optimal) λ = 4 Residual norm Second moment of P(v) 4 2 Generalised cross validation method is used to optimise the smoothing parameter (λ).5.5 Velocity [m/s] 9

20 P(v) Comparison to a theoretical model Theoretical turbulent power law distributions U r = V M n + n 8 6 r R n n = f Re Experimental and theoretical distributions at v =.44 m/s n.44 m s n = 4 n = 5.37 n = 7 Experimental = 5.37 (Zagarola et al. 997) Velocity [m/s] 2