Flashing liquid jets and two-phase droplet dispersion II. Scaled experiments for derivation of droplet size correlations

Transcription

1 Flashing liquid jets and two-phase droplet dispersion II. Scaled experiments for derivation of droplet size correlations MKOPSC International Symposium, October 28 College Station, TEXAS Peter Kay and Phil Bowen, Cardiff University, UK Henk Witlox, DNV Software, UK

2 Experimental set-up New Dense-spray receiving optics used Experimental conditions: Droplet measurement size range = 2132 microns. Ambient (lab.) temperature circa 15 C Rig pressure = 6-14 barg Data collected 25 mm downstream on nozzle centerline for flashing sprays and 5 mm for subcooled sprays..75, 1, and 2 mm nozzles characterised. L/d ratios from PDA collected continuously data through mechanical break-up, transition and fully flashing regimes Slide 2

3 Test matrix sub-cooled & super-heated experiments Data for the sub-cooled mechanical break-up were recorded: 5 mm down-steam (except for the.75 mm cyclohexane data which was recorded 1 mm down-stream due to large break-up length). 11 radial positions from +1 mm to -1 mm in 2 mm steps. Data for the super-heated sprays were recorded: 25 mm downstream (a compromise between the change in break-up length with superheat) On the centerline only so as to record the transient event. Fuel Break-up Diameter [mm] L/D [-] Pressure [barg] Water Sub-cooled 1, 2 1.1,.5 6, 1, 14 Cyclohexane Sub-cooled.75, 1, 2 1.4, 1.1,.5 6,8,1,12,14 Gasoline Sub-cooled.75, , 3.4 6,8,1,12,14 Water Super-heated.75, 1 3.4, Cyclohexane Super-heated 1, 2 1.1,.5 7.5, 1 Butane Super-heated.75, 1, 2 1.4, 1.1,.5 9.5, 8, 7.5 Propane Super-heated 1, 2 1.1,.5 6.5, 7.5 Gasoline Super-heated Slide 3

4 Flashing results example The transient SMD against temperature of a cyclohexane spray is shown. d = 1 mm L/d = 1.1 Rig temperature, pressure and mass were recorded simultaneously with spray data. Allows the complete trend of droplet diameter with superheat to be studied. SMD [microns] Rig Mass [kg] Nozzle Pressure [bar] Nozzle Superheat Temperature (degc) Temp [degc] Time After SOR [s] Rig Mass Nozzle Pressure Nozzle Temperature Ambient Temperature Slide 4

5 Flashing Cyclohexane Results (1.5barg, 2mm, L/D=.5) Slide 5

6 Flashing Cyclohexane Results (1.5barg, 2mm, L/D=.5) Mass [kg] Pressure [barg] (1) (2) (3) (4) Temperature [degc] 5 2 (1) T 16 C, T sh -65 C Mechanical break-up Time after SOR [seconds] Pressure Mass Ambient Temperature Nozzle Temperature (2) T 4 C, T sh -41 C Mechanical break-up (3) T 85 C, T sh 4 C Transition (4) T 18 C, T sh 27 C Fully Flashing Slide 6

7 Steps in model development Tri-linear approach proposed by Phase II has been verified by transient data. Phase III model based on the Phase II model with the following major changes: New correlation for mechanical break-up SMD (including fluid property ratios). Slight modification of regime transition criteria New SMD proposed for fully flashed condition Mechanical Break-up Transition Fully Flashing A SMD Fully flashed (8 microns) B Cut-off (1 microns) Sub -cooled Superheated Degree of superheat Slide 7

8 Steps in model development Mechanical break-up SMD correlation Regime transition criteria Flashing break-up SMD correlation Mechanical droplet size distribution correlation Flashing droplet size distribution correlation Slide 8

9 Summary Mechanical break-up SMD correlation A three step method was used to derive the mechanical break-up correlation Dimensional form First the correlation was assumed to take a simple dimensional form where the SMD was only proportional to a pressure and nozzle diameter term. Effect of liquid properties Next the effect of the fluid properties on the SMD were examined and included in the correlation. Dimensionless form Finally the correlation was converted into a non-dimensional form and some of the exponents modified to ensure a best fit with the cyclohexane and water data. Phase II Correlation: d d da o We.533 Lo Re.14 Lo L d o.114 Phase III Correlation: d d da o 74We.85 Lo Re.44 Lo L d o.114 Lo water, stp.97 Lo water, stp.37 Lo water, stp.11 Slide 9

10 Comparison of mechanical break-up correlation with experimental Summarydata Predicted SMD [microns] 1 1 Cyclohexane - d = 1mm, L/d = 1.1 Cyclohexane - d =.75mm, L/d = 1.5 Cyclohexane - d = 2mm, L/d =.5 Gasoline - d = 1mm, L/d = 3.4 Gasoline - d =.75mm, L/d = 4.53 Water - d = 1mm, L/d = 1.1 Water - d = 2mm 3% deviation Measured SMD [microns] Phase III Correlation Predicted SMD [microns] Cyclohexane - d = 1mm, L/d = 1.1 Cyclohexane - d =.75mm, L/d = 1.5 Cyclohexane - d = 2mm, L/d =.5 Gasoline - d = 1mm, L/d = 3.4 Gasoline - d =.75mm, L/d = 4.53 Water - d = 1mm, L/d = 1.1 Water - d = 2mm 3% deviation Measured SMD [microns] Phase II Correlation Slide 1

11 Summary Regime Transition Criteria Correlation The transition criteria is assumed to take the same form proposed in Phase II: Criteria for Transition A: Ja 48 We 1 7 v, with 1 e v 23 L Criteria for Transition B: Ja 18 Wev 1 7 For fully flashing conditions, the SMD is assumed to be a constant of 8 microns and decreases at a rate of 1 micron per 1 degrees until a final limit of 1 mircons is achieved. Fluid [-] Pressure [barg] d [mm] L/d [-] Measured [ºC] Phase II [ºC] Phase III [ºC] A B A B A B Water Water Cyclohexane Cyclohexane Cyclohexane N/A N/A Gasoline N/A N/A Slide 11

12 Summary Droplet distribution correlation The droplet size distribution adopted is the Rosin-Rammler distribution: ( D) 1 e D a RR D SMD brr The exponents used for mechanical break-up and fully flashing are presented in the table below: Correlation a RR b RR Elkobt Phase III (mechanical).4 2. Phase II (fully flashing) Slide 12

13 Comparison of model with experimental data DISC (JIP-III, 2 mm) Measurement Experimental (Mechanical Break-up) 1 SMD [microns] 6 4 SMD [microns] DISC (JIP-II, 1 mm) 2 Experimental data (1 mm) DISC (JIP-III, 1 mm) Experimental (Mechanical Break Up) Superheat (K) Superheat (K) Cyclohexane, d = 1 mm, L/d = 1.1, P = 7.5 barg Cyclohexane, d = 2 mm, L/d =.5, P = 12 barg Slide 13

14 Comparison of sub-cooled droplet size distribution with experimental data Cyclohexane, d =.75 mm, L/d = Normalised Mass-under-size [-] bar (measured) Droplet 1 Diameter bar (measured) [microns] 14 bar (measured) 6 bar (predicted) 1 bar (predicted) 14 bar (predicted) Slide 14

15 Comparison of flashing droplet size distribution with experimental data Water, d =.75 mm, L/d = 1.4, P = 1 barg Mass-under-size (-) Droplet Diameter (microns) Measured Predicted Slide 15

16 Summary A new model for predicting the SMD and droplet size distributions of a release from a sharp edged orifice from the initial conditions in the nozzle has been developed. The model covers a wide range of liquids, nozzle diameters, superheat temperature and aspect ratios. The new mechanical break-up correlation includes the influence of fluid properties and agrees well with the measured data. The tri-linear modeling approach has been verified, across data sets. Successful simultaneous measurement of rig conditions and droplet data. New correlations for the regime transition criteria have been developed and verified against experimental data. New correlations for the droplet size distributions show relatively good agreement with the experimental data. Slide 16

17 Future work Some data were clipped due to limitations of the laser diagnostic technique (the technique chosen was still the most appropriate), hence the droplet size distribution correlation may need modifying due to advances in measurement technology Bimodal droplet size distributions under (near-)flashing conditions Droplet size correlations for very high superheats Additional large-scale experiments and multi-component releases Slide 17