Study of Methane Hydrate Inhibition Mechanisms Using Copolymers

766

Study of Methane Hydrate Inhibition Mechanisms Using Copolymers

B. CINGOTTI,

a

A. SINQUIN,

b

J.P. DURAND, AND T. PALERMO

Institut Français du Pétrole, Division Chimie et Physicochimie Appliquées,1-4 Avenue Bois-Préau, 92 852 Rueil-Malmaison, Cedex, France

INTRODUCTION

In multiphase production and transport of hydrocarbon resources, gas, water, andeventually a liquid hydrocarbon phase (condensate or crude) are in contact under rel-atively high pressure. In cold areas and subsea pipelines, these systems can reachrather low temperatures leading to hydrate crystals that plug production facilities.

To prevent gas hydrate formation, different techniques are already used. Forexample, hydrate formation is avoided by physical actions on the system like insu-lation and/or heating the pipeline, depressurization, water removal, and others.These solutions are often not sufficient, especially in the case of a shut down in sub-sea production. A long residence time in a subsea pipe leads to a prolonged exposureof the system at very low temperatures. Another widespread solution is the injectionof thermodynamic additives such as methanol in large amount (up to 30% withrespect to produced water). These products present the drawbacks of being expen-sive in terms of the large quantities of additives used and also the resulting pollution.

To overcome this crucial industrial problem, a new category of additives has beendeveloped, kinetic inhibitors. These additives are designed to act on the differentsteps of hydrate formation,

1

by delaying nucleation, slowing down their growth, andavoiding agglomeration. Kinetic inhibitors are generally water soluble compoundsthat are effective at low concentrations (less than 1% by weight with respect towater).

Using a semibatch reactor equipped with a turbidimetric sensor to determine theevolution of the particles size distribution (PSD),

2

we apply a new experimental pro-cedure that has a reproducible induction time (

t

i

) even when no solid nucleationagents are present. This experimental procedure is based on the control of hydratenuclei using the turbidimetric sensor. With this new experimental procedure, wehave studied the effect of various hydrosoluble polymers on induction time and onthe PSD.

Telecommunication.

a

Voice: 01 47 52 71 55; fax: 01 47 52 70 58. [email protected]

b

Voice: 01 47 52 71 57; fax: 01 47 52 70 58. [email protected]

767CINGOTTI

et al

.: METHANE HYDRATE INHIBITION

EXPERIMENTAL SETUP

Our experimental setup is shown in F

IGURE

1 and has been described in detailelsewhere.

3,4

The liquid phase (pure water or polymer solution) is placed in a stain-less steel autoclave and pressurized with methane. The reactor is equipped with twosapphire windows to observe hydrate particles. The temperature of the system is con-trolled by circulating a water/glycol mixture in a cooling jacket surrounding theautoclave. Pressure control is achieved by the measurement of the differential pres-sure

∆

P

between the reactor and a ballast pressure reference

P

ref

. A flowmeter placedon the input gas line is directly connected to an integrator and gives the cumulatedgas consumption during the test. The turbidimetric sensor is a UV-visible analyzer

FIGURE 1. Experimental setup.

768 ANNALS NEW YORK ACADEMY OF SCIENCES

that measures,

in situ

, the attenuation of a polychromatic beam in the wavelengthrange 230–750 nm. Using this experimental setup, we can perform experiments witha maximum working pressure of 120 bar and at ambient temperature down to 0

°

C.The data recorded are temperature and pressure in the cell, cumulated gas consump-tion, and turbidity of the media at different wavelengths.

The turbidity measurements can be used as a particle sizing method

2

and an

insitu

PSD of the hydrate suspension can be obtained within the particle diameterrange 10–100

µ

m. Turbidity measurements are possible only if the medium is nottoo concentrated in particles. Consequently, PSD evolutions are determined only atthe early stage of the reaction when the number of solid particles is not too high andat moderate supersaturation.

All the experiments described in this paper were made with ultra-pure water(conductivity, 18.2

Ω⋅

m) or ultra-pure water

+

kinetic inhibitors and methane witha purity of 99.9% provided by Air Liquide.

EXPERIMENTAL PROCEDURE

The aim of our procedure is mainly to obtain reproducible induction time (

t

i

)without adding any impurity, in order to estimate the efficiency of kinetic additives.Our new test procedure is based on the control of the number of nuclei in the solu-tion. The idea is to substitute impurities by hydrate nuclei. The turbidimetric sensoris the essential tool to follow the formation and above all the dissociation ofhydrates. Our procedure comprises three phases:

4

phase 1, first formation; phase 2,dissociation and

over dissociation

; phase 3, second formation.F

IGURE

2 illustrates the experimental procedure and F

IGURE

3 shows the corre-sponding evolution of pressure and turbidity at the wavelength 500 nm versus time.

Phase 1: First Formation

The reactor is filled with ultra pure water at ambient temperature, purged threetimes with methane, the third time the pressure is maintained at 20 bar (point I inF

IG

. 2). The temperature is then decreased to 1.0

°

C under isobaric conditions, toreach the LP point (LP, low pressure in F

IG

. 2). The temperature is kept constant dur-ing the rest of the procedure. Then the pressure is increased to the very high pressurepoint (VHP point, F

IG

. 2) 70 bar and stirring is started at a very high rate. The drivingforce for hydrate formation is very high and hydrates are formed within a few min-utes. Consequently, the turbidity of the medium increases sharply (F

IG

. 3).

Phase 2: Hydrate Dissociation and

Over Dissociation

The dissociation step is achieved by a decompression from the VHP to LP(F

IG

. 2). Since the system is no longer in the hydrate formation zone, hydrates meltslowly and the turbidity of the medium decreases to zero (point V in F

IG

. 3). Whenno more particles are visible in the reactor, a time called the

over dissociation

time(

t

ov

) is measured. During this period, the hydrate melting process is still going onand crystals become nucleï. At the end of

t

ov

, the pressure is increased again to reach

769CINGOTTI

et al


FIGURE 2. Schematic representation of the experimental procedure.

FIGURE 3. Evolution of pressure and turbidity.


the point HP in the hydrate formation zone. This last step represents the end of thedissociation phase.

Phase 3: Second Formation

After a period that corresponds to the induction time

t

i

, at the HP pressure,hydrates appear again. The variations of turbidity and gas consumption are thenrecorded. The end of phase 3 corresponds to a total extinction of the turbidimetricsignal. New cycles of phase 2 and 3 can be repeated throughout the day.

EXPERIMENTAL RESULTS

Various copolymers were studied using this experimental setup and this proce-dure. These copolymers, provided by SNF Floerger (St. Etienne, France), includetwo different monomers, acrylamide (AA), and sodium acrylamidomethylpropane-sulfonate (AMPS). The influence of AMPS molar ratio and the influence of copoly-mer concentration are described below. For each study, the influence of copolymerson

t

i

and on the PSD evolution is discussed, in comparison with a test without addi-tive. All the experiments were performed using the experimental proceduredescribed above with a

t

ov

of 5 min, at 400 rpm, 1

°

C and 45 bar.

Influence of the AMPS Molar Ratio

Copolymers of similar molecular mass with AMPS molar ratios varying from25% to 80% were tested in the reactor at a concentration of 0.15% wt.

Induction Time

The results concerning

t

i

are given in T

ABLE

1. All the additives increase

t

i

butthis increase is not linear with the AMPS molar ratio. An enhanced efficiency isobserved for the copolymer containing 40% of AMPS, and in this case,

t

i

is multi-plied by four regarding the blank test.

Granulometric Measurements

When hydrates start to form, the PSD evolution versus time were studied forall the copolymers and for the blank test. Typical results of mean diameter andD( )

T

ABLE

1. Influence of AMPS molar ratio on induction time (45 bar, 1

C, 400 rpm)

Products

t

i

(min)

a

a

Each value is the mean of at least three experiments.

Standard Deviation

Water 4.5 1.1

25% AMPS 12.3 0.577

40% AMPS 20.5 2.12

50% AMPS 15.0 4.47

60% AMPS 13.0 1.73

771CINGOTTI

et al


particle number (

N

p

) evolution are given F

IGURE

4. These graphs show clearly thatall the copolymers have an effect on the PSD. The first observation is that if anadditive has an efficiency in delaying hydrate crystallization, it will also act on thesize and the number of particles formed. In other words, AA/AMPS copolymersdelay hydrate formation and interact strongly with crystals when they start to form.In F

IGURE

4, we can see two types of behavior. The additives containing 25 and 60%of AMPS have constant and

N

p

rapidly becomes higher than for the blank test.Due to the faster evolution of

N

p

, the upper detection limit is reached earlier and,D

FIGURE 4. Influence of AMPS molar ratio on (A) the mean diameter and (B) thenumber of particles (Np) (45 bar, 1°C, 400 rpm).

D( )


consequently, the PSD analysis is performed in a shorter time than in the blank test(less than 20 min for these copolymers and about 50 min for water). On the otherhand, with the 40 and 50% AMPS copolymers, is similar to that measured in theblank test at the beginning of the crystallization. After 30 min, with the 40%AMPScopolymer, decreases strongly from about 12 to 8

µ

m. The corresponding

N

p

,which is notably inferior than without additive between 0 and 30 min, increasessharply and becomes greater than for the blank test. We conclude that this copolymeris no longer efficient on hydrate crystal growth after 30 min.

The evolution of versus time for the 50%AMPS copolymer and blank test areidentical but this copolymer exhibits a very surprising evolution of

N

p

. The numberof particles seen by the sensor at the onset of the formation is very high (about3

×

10

5

) and remains almost constant throughout the growth stage. Conventionally,hydrate formation in the reactor begins in the thin layer of water at the gas/liquidinterface. In this layer, gas concentration is maximum and primary nucleation starts.The few crystals formed are swept along in the bulk, by the stirring, to seed the bulk.Logically, the number of particles seen by the sensor is almost zero at the beginningand then increases more or less rapidly with the driving force, the stirring rate, andthe additive growth inhibition. In the case of the 50%AMPS copolymer, the fact that

N

p

is very high at the beginning suggests that nucleation does not take place at theinterface but directly in the bulk. Consequently, we emphasize that this additiveblocks efficient primary nucleation at the interface and nucleation occurs in the bulkwhen its efficiency decreases or when the driving force increases. When crystals areformed, the fact that

N

p

remains constant during about 70 min indicates a very stronggrowth inhibiting effect of this copolymer, which avoids an explosive formation ofnew hydrate crystals.

In conclusion, all the AA/AMPS copolymers tested delay hydrate formation andthe 40%AMPS copolymer is the most efficient additive on

t

i

. The copolymers with25 and 60% molar ratio do not have any growth inhibiting effect. On the contrary,when hydrates start to form, the number of particles seen by the sensor is higher thanin the blank test. The 40%AMPS has a growth inhibiting effect during 30 min.Finally, the 50%AMPS has a limited effect on

t

i

but blocks primary nucleation andinhibits growth very efficiently.

Influence of Concentration

The influence of average molecular weight was studied but is not presented in thispaper. The influence of concentration has been studied with a 40%AMPS copolymerof lower molecular mass than that presented above, because of the higher concentra-tions (0.15–1% by weight). The increase of polymer concentration induces a signif-icant increase of viscosity of the solution, but we have verified that there is nosignificant decrease in the mass transfer constant when viscosity increases from1.73 mPa.s (water) to 37.8 mPa.s (1% by weight of copolymer). We emphasizethat the changes in

t

i

and in the PSD are essentially due to the polymer/crystalinteractions.

Induction Time

A first series of

t

i

measurements was made at a HP of 45 bar. The results are givenin T

ABLE

2. At 0.3%,

t

i

is similar to the value obtained in the blank test. Beyond a

D

D

D

773CINGOTTI

et al


threshold concentration of 0.5% by weight, the copolymer dramatically increasesinduction time (35 min at 0.5%). The best

t

i

is obtained at 0.8%,

t

i

is almost ten timesgreater than without polymer. For 1%,

t

i

decreases again. Another series of experi-ments was performed at HP

=

55 bar. Under these conditions, the ti is detected at athreshold concentration of 1% by weight. From a comparison between inductiontimes measured at 45 and 55 bar (see FIGURE 5), we conclude first that ti dependsstrongly on the driving force of the system; and second, that viscosity has no deter-minant effect on induction time, since it is possible to find a threshold concentrationbeyond which the polymer is efficient when the pressure increases from 45 to 55 bar.

TABLE 2. Influence of concentration on induction time (45 bar, 1C, 400 rpm)

Products Viscosity of the Additive Solution at 1°C (mPa.s)

ti (min)a

aEach value is the mean of at least three experiments.

Standard Deviation

water 1.73 4.5 1.1

0.15% 10.9 3.0 2.0

0.3% 16.5 4.0 1.3

0.5% 23.2 35 1.41

0.8% 30.6 39 1.41

1% 37.8 23.5 0.707

FIGURE 5. Induction time versus viscosity of the additive solutions at 45 and 55bar (1°C, 400 rpm).


Granulometric Measurements

FIGURE 6 shows the evolution of and Np versus time for the experiments per-formed at 45 bar. The PSD evolution at 0.15 and 0.3%, which corresponds to a ti thatis inferior or equivalent to the blank test, exhibits a decrease of and a sharpincrease in Np, with Np evolution similar to that determined with water. As with the25 and 60% AMPS copolymers in the previous section, the acquisition times aremuch shorter than for the blank test. At higher concentrations (0.5 and 0.8%), theevolution of is again similar to that for the blank test. The experiments can be

D

D

D

FIGURE 6. Influence of concentration on (A) the mean diameter and (B) thenumber of particles (Np) (45 bar, 1°C, 400 rpm).

D( )

775CINGOTTI et al.: METHANE HYDRATE INHIBITION

prolonged longer than with water, because of the much lower rate of increase in Np.Logically, the increase in Np is slower at 0.8% than at 0.5%.

In conclusion, we have evidence for a threshold concentration below which thecopolymer does not delay hydrate formation. This threshold concentration increaseswith the driving force applied to the system. At 45 bar, the PSD study shows thatbelow the threshold concentration, the hydrates formed have a greater and Npincreases exponentially. At concentrations beyond the threshold, the evolution ofis similar to that for the blank test and the increasing rate of Np versus time is muchlower than with water. The higher the polymer concentration, the slower the increas-ing rate.

CONCLUSION

A new experimental procedure based on formation/dissociation cycles has beendeveloped. This procedure, based on the control of the number of hydrates nuclei,allows us to attain reproducible induction times (ti) without adding impurities to themedium.

Various AA/AMPS copolymers were studied using this procedure. Influenceof the AMPS molar ratio and polymer concentration have been explored; first, on ti,and second, on the PSD evolution versus time. At a given molecular mass, all theAA/AMPS copolymers delay hydrate formation and there seems to be an efficiencypeak around 40–50% AMPS. The best inhibition performances, with respect to themean diameter and the quantity of hydrates formed, are obtained with a copolymercontaining 50% molar AMPS, which blocks primary nucleation. For a given copol-ymer composition, beyond a threshold concentration, the copolymer increases in-duction time, independently of viscosity. The higher the concentration, the higherthe inhibiting effect on the PSD, as far as increasing concentration prevents the meandiameter from diminishing and the number of particles from increasing very rapidly.

ACKNOWLEDGMENTS

The authors express their thanks to Total, Elf, and SNF Floerger.

DD

TABLE 3. Influence of concentration on induction time (55 bar, 1C, 400 rpm)

Products Viscosity of the Additive Solution at 1°C (mPa.s)

ti (min)a

aEach value is the mean of at least three experiments.

Standard Deviation

water 1.73 4.3 0.577

0.15% 10.9 3.7 1.53

0.3% 16.5 1.8 1.33

0.5% 23.2 2.3 0.577

0.8% 30.6 1 0

1% 37.8 19.3 8.74


REFERENCES

1. LONG, J.P. et al. 1994. Kinetic inhibitors of natural gas hydrates. Proceedings of theSeventy-third GPA Annual Convention. New Orleans, LA.

2. HERRI, J.M. 1996. Etude de la Formation de l’Hydrate de Méthane par TurbidimétrieIn Situ. Ph.D. Thesis, University of Paris VI, France.

3. HERRI, J.M. et al. 2000. Interest of in situ turbidimetry for the characterization ofmethane hydrate formation: application to the study of kinetic inhibitors. Chem.Eng. Sci. Accepted for publication.

4. CINGOTTI, B. et al. 1999. A new method to investigate kinetic hydrate inhibitors. PaperSPE 50757 presented at the International Symposium on Oilfield Chemistry. Hous-ton, Texas, February 16–19.

Documents

Study of Methane Hydrate Inhibition Mechanisms Using Copolymers