NATURAL GAS HYDRATE - Determination of Free Water Using Conductivity Measurements

Janne Hauge
Desember 1996

UTVIDA SAMANDRAG

Før ein kunne starte hydratproduksjonen, måtte det finnast ut kva type salt ein skulle tilsetje formasjonsvatnet og kor mykje av det ein skulle bruke. Dette blei gjort ved å delvis fryse ned ulike saltløysningar og deretter separere isen frå væska. Desse vart deretter sett i romtemperatur for smelting og oppvarming. Konduktiviteten til dei ulike løysningane, blei målt både før nedfrysing og etter oppvarming. Av dei fire salta NaCl, KCl, MgCl2 og CaCl2 som var brukte i forsøka, vart det funne at KCl var det beste å bruke i ved hydratproduksjonen. Det vart også funne at ein konsentrasjon på 100 ppm var riktig å bruke. Frå forsøka med is, blei det konkludert med at løyste ioner blir støytte frå krystallstrukturen til is.

Hydrata blei danna på same måten som er forklart i diplomoppgåva til Ofstad (1996). Den einaste forskjellen, var at ei saltløysning vart brukt til formasjonsvatn, og det separerte vatnet blei tatt vare på for konduktivitetsmålingar. Hydrata som blei danna, vart smelta og konduktiviteten til smeltevatnet vart målt. I dette arbeidet måtte smeltevatnet fortynnast då det ikkje blei produsert nok hydrater til å få ca. 25 ml smeltevatn. Dette er mengda som må til for at konduktivitetsmålingane skal kunne utførast. Denne fortynninga førte med seg ei stor usikkerheit i resultatet, men ein kunne likevel sjå at konduktiviteten hadde minka i smeltevatnet til hydratet samanlikna med den løysninga som hadde blitt injisert inn i hydratcella. Konduktiviteten til det separerte vatnet hadde auka.

Ei prosedyre for korleis ein kan finne mengda fritt vatn har blitt utvikla i denne diplomoppgåva, men på grunn av vanskar med utstyret i laboratoriet, blei ikkje resultata så gode som forventa.

SUMMARY

In order to determine what kind of salt and how much of it should be used for the formation of hydrates, experiments in which salt solutions were partly frozen to ice, were conducted. From the salts NaCl, MgCl2, KCl and CaCl2, KCl was chosen as the salt for further use. It was also found that a concentration of about 100 ppm should be used in the hydrate formation experiments. The experiments done with ice, also showed that the dissolved ions were excluded from the crystal structure of ice.

Hydrates were formed according to the procedure described in the diploma thesis of Ofstad (1996). The only difference was that a salt solution was used as formation water, and the separated water was collected for conductivity measurements. The hydrates made were then melted and used for conductivity measurements. The melted hydrate sample had to be diluted before one could do the measurements, because a certain volume was required by the conductivity cell. This dilution included large uncertainties in the result, but it could be seen that the conductivity had decreased in the melted hydrate sample compared to the solution injected to the hydrate cell, and the conductivity of the separated water had increased.

A procedure for how to find the amount of free water was developed in this work, but due to difficulties with the equipment in the laboratory, the results from the experimental work were not as good as expected.

INTRODUCTION

There are several parameters which have to be looked closer into considering the natural gas hydrates, before one can start building the equipment needed in the transport. One of these parameters is the gas content of the hydrate. Some work have earlier been conducted at NTNU (the Norwegian University of Technology and Science) by Nerland (1995) and Ofstad (1996). This work has been based on mass balance, where a sample of hydrate has been melted and the masses before and after the melting have been registered. The temperature and pressure development during the melting have also been recorded. This has given quite satisfactory results, but a main problem has been that one is not able to tell how much of the water comes from the hydrate and how much comes from free water. Free water is in this case water from the ice surrounding the hydrate. This will again lead to an uncertainty considering the gas content of the hydrate, and in terms it will affect the economy of the project. As the free water has been included in the earlier estimates of the gas content, the gas content has been underestimated and thereby also the profit. An economic project is dependent on as high gas content as possible.

In this diploma thesis, the fact that natural gas hydrates contains only water besides the gas molecules will be used (Ngan and Englezos, 1996). This has also been confirmed during the laboratory experiments of this work. If a piece of hydrate is melted, the resulting water will come from both the hydrate and the ice. Considering the hydrates being created from destilled water with low salinity, the conductivity of the melted hydrate would be equal to destilled water if the hydrate contains no free water. If there is free water in the hydrate, the conductivity would be proportional to the concentration of dissolved ions. From this one should be able to come up with a connection to the gas content.

Formation of hydrates from an aqueous electrolyte solution and a natural gas is especially interesting in the hydrate process. A cooling media is needed in order to reach the low temperatures necessary for hydrates to form. This need will be reduced if it is possible to use sea water as the formation water. Sea water contains many different kinds of salt, and as salt is known as a hydrate inhibitor (Tohidi, Burgass, Danesh and Todd, 1993), it could affect the formation process. Salts’ effect on hydrate equilibrium conditions will therefore be looked into in this thesis.

The hydrates needed for conductivity measurements, were made in the laboratory at NTNU, and was part of the work in this diploma thesis. In addition to the formation of hydrates and the conductivity measurements, some PVT- and CSMHYD96-simulations were done in order to investigate the effect that adding salts has to the formation of hydrates.

DISCUSSION

Before the hydrate formation experiments were performed, it had to be decided what kind of salt and how much of it should be used in the experiments. The salts tested in this work were NaCl, KCl, MgCl2 and CaCl2 and the procedure is described in Appendix B.

Freezing of salt solution

From the laboratory experiments, the only conclusion that could be drawn, was that MgCl2 should not be used in the hydrate formation experiments. Figures 12, 13 and 14 show that MgCl2 reaches a lower % reduction in conductivity of ice than the other salts. This would mean that the dissolved ions of MgCl2 to a lesser degree are excluded from the crystal lattice of ice than the dissolved ions of NaCl, KCl and CaCl2. The last mentioned salts seem to be excluded in a similar amount, where KCl has the most consistently high % reduction in conductivity of ice. This could indicate that KCl would be the salt to use for further experiments, but these experiments alone are not enough to make the decision whether to choose one salt instead of another.

The hydrate inhibiting effect of the salt is an important parameter. If the goal is to produce a lot of hydrates, it could be sabotaged if the solution is created of a salt with a large hydrate inhibiting effect. PVTsim-HYD showed that NaCl is preferred to CaCl2, and CSMHYD96-simulations showed that KCl is preferred to NaCl, CaCl2 and MgCl2. The results from the simulations are shown in Figures 2 and 3. The best hydrate inhibiting salt shows the largest shift of the equilibrium line to the left. In this case MgCl2 is the most efficient hydrate inhibitor and KCl is the least efficient. KCl would be preferred to the other salts.

Other parameters that are looked into, are the ion radii and the salting-out capacities shown in Tables 1 and 2. A low salting-out capacity is wanted in the hydrate formation experiments, and according to this, the rank will be the same as it was for the inhibiting effects. MgCl2 has the highest salting-out capacity and KCl has the lowest. The salting-out capacity is a consequence of the ion radius, and therefore the ion with the biggest ion radius has the lowest salting-out capacity while the ion with the smallest ion radius has the highest salting-out capacity. Again, KCl would be preferred to the other salts.

In addition, one has to consider the corrosive effect of the aqueous solutions of chloride salts. Solutions of calcium chloride is much more corrosive than those of sodium chloride, and would therefor not be used in further experiments. This narrows the chose down to two alternatives: KCl and NaCl. KCl is chosen for the experiments done in this thesis; because of the results discussed above but also due to the fact that it is normally used as a reference salt when it comes to aqueous electrolyte solutions.

Three different amounts of salt were tested in this work, 10, 100 and 1000 ppm. 10 ppm was found to be too little as the uncertainties of the measured conductivity values became too large. Simulations showed that the concentrations tested have no significant effect on the equilibrium line for the natural gas hydrates, see Figures 2 and 3, but the experiments showed that 1000 ppm probably was too much. The ice got another consistency than the ice from solutions of 10 and 100 pp. It was more porous and one could see traces of ions in the crystal structure. It was obvious that more ions were included in the crystal lattice of ice. 100 ppm was chosen for the hydrate formation experiments.

Formation of hydrates

In this work, hydrates were produced in order to find the amount of free water. The procedure described by Ofstad (1996) was followed in detail except from the solution injected into the hydrate cell. In this work a 100 ml 100 ppm solution of KCl was used as hydrate formation water.

The gas content was found according to the procedure also described by Ofstad, and are shown in Table 10. The table shows that the gas content is indeed very low. Values of 40 Sm3/m3 is not as high as expected, and much lower than the values reported in the diploma thesis of Ofstad (1996) and Nerland (1995). This could be due to the salt solution used as hydrate formation water. Ofstad and Nerland used buffer solutions that were well fitted for experiments where a high gas content was searched.

Another reason for the low gas content could be too short separation time. The nitrogen gas injection was turned off when the excessive liquid seemed to be separated out (3-5 min), but more liquid could be trapped in the cell due to hydrate blockage on the filter. The nitrogen injection should perhaps have lasted for 10-15 minutes more.

Amount of free water

The amount of free water was found by conductivity measurements, and the results are shown in Table 10. The table shows an increase in conductivity of the separated liquid. This increase is probably due to the ion exclusion when hydrates are formed. The conductivity of the hydrate sample varies. Theoretically it should decrease due to the ion exclusion, and for most cases it does, but for some experiments an increase is observed. This could be due to the dilution of sample water which includes large uncertainties in the result. The uncertainties are shown in Table 10; calculated as described in Appendix B. Another reason could be pollution of the hydrate formation water. If the reactor was not thoroughly cleaned before an experiment, this could cause large uncertainties in conductivity.

Uncertainties

The hydrate formation and melting of a hydrate sample is done according to the procedures described in the diploma thesis of Ofstad (1996), therefore the uncertainties of this part of the experimental work are also the same. In this work, uncertainties concerning the dilution of the hydrate sample are included. This uncertainty is large and should be avoided. If hydrates can be formed in such an amount that no dilution is needed for the conductivity measurements, the uncertainties would be much smaller.

Recommendations to further work

Experiments should be performed where enough hydrates are made to conduct direct conductivity measurements; that is, avoid the dilution. This could be done by injecting more formation water into the reactor, but it requires a more powerful stirrer than was available in this work.

Some experiments should also be done where a higher gas content is obtained than in this work. This could clearer show the effect of ion exclusion. In order to achieve a higher gas content, longer separation time should be tried.

CONCLUSIONS

· By assuming the hydrates having similar qualities as ice, a procedure for finding the amount of free (additional) water from ice in the hydrate by conductivity measurements was developed.

· PVTsim-HYD showed that NaCl is the least efficient hydrate inhibitor of the salts NaCl and CaCl2. CSMHYD96-simulations showed that KCl is the least efficient hydrate inhibitor of the salts KCl, NaCl, CaCl2 and MgCl2. MgCl2 is most efficient.

· It is recommended to use a 100 ppm KCl solution as formation water for the hydrate formation experiments.

· Laboratory experiments showed the trend that is included in the assumption above, and seemed promising. Due to some difficulties with the equipment more experiments should be done to confirm the method developed in this work.

REFERENCES

· Gudmundsson, J.S.and Børrehaug, A.,«Natural Gas Hydrate an Alternative to Liquefied Natural Gas», Petroleum Review, Vol.50, No 592, May 1996, pp 232-235

· Nerland, B.,«Natural Gas Hydrates, separation and gas content», Diploma thesis, Department of Petroleum Engineering and Applied Geophysics, Norwegian Institute of Technology, University of Trondheim, 1995, 74 pp

· Ofstad, K.,«Hydrate Formation and Separation», Diploma thesis, Department of Petroleum Engineering and Applied Geophysics, Norwegian University of Science and Technology, 1996, 90 pp

· Ngan, Y., Englezos, P,.«Concentration of aqueous solutions by clathrate hydrate formation», 2nd International Conference on Natural Gas Hydrates, Toulouse, France, June 2-6, 1996, pp 467-474

· Berecz, E. and Balla-Achs, M.,«Gas Hydrates», Studies in Inorganic Chemistry 4, Elsevier, Amsterdam-Oxford-New York, 1983

· Sloan Jr., E.D.,«Natural Gas Hydrates», Journal of Petroleum Technology, Dec. 1991, pp 1414-1417

· Tohidi, B., Danesh, A., Todd A.C. and Burgass, R.W.,«Measurement and Prediction of Hydrate-Phase Equilibria for Reservoir Fluids», SPE Production & Facilities, May, 1996, pp 69-76

· Bishnoi, P.R. and Dholabhai, P.D.,«Experimental study on propane hydrate equilibrium conditions in aqueous electrolyte solutions», Fluid Phase Equilibria, Vol. 83, 1993, pp 455-462

· Englezos P. and Ngan, Y.T.,«Incipient Equilibrium Data for Prpane Hydrate Formation in Aquroud Solutions of NaCl, KCl and CaCl2», Journal of Chemical Engineering Data, Vol. 38, 1993, pp 250-253

· Ussler III, W., Paull, C.K.,«Effects of ion exclusion and isotopic fractionation of pore water geochemistry during gas hydrate formation and decomposition», Geo-Marine Letters, Vol. 15, 1995, pp 37-44

· Sassen, R. and MacDonald, I.R.,«Evidence of structure H hydrate, Gulf of Mexico continental slope», Org. Geochem., Vol. 22, No. 6, 1994, pp 1029-1032

· Mehta, A.P. and Sloan Jr., E.D.,«Structure H Hydrate Phase Equilibria of Methane + Liquid Hydrocarbon Mixtures», J. Chem. Eng. Data, Vol. 38, No. 4,1993, pp 580-582

· Lederhos, J.P., Mehta, A.P., Nyberg, G.B., Warn, K.J. and Sloan, E.D.,«Structure H Clathrate Hydrate Equilibria of Methane and Adamantane», AIChE Journal, Vol. 38, No. 7, July 1992, pp 1045-1048

· Mehta, A.P. and Sloan, E.D.,«Structure H hydrates: the state-of-the-art», Colorado School of Mines, Golden (USA), 2nd International Conference on Natural Gas Hydrates, Toulouse, France, June 2-6, 1996, pp 1-8

· Mehta, A.P. and Sloan Jr., E.D.,«Structure H Hydrate Phase Equilibria of Paraffins, Naphthenes and Olefins with Methane», J. Chem. Eng. Data, Vol. 39, No. 4,1994, pp 887-890

· Weast, R.C. and Astle, M.J.,«Handbook of Chemistry and Physics», 60th edition, 1979-80, CRC Press, Inc., Boca Raton, Florida

· Barrow, G.M.,«Physical Chemistry», Third edition, McGraw-Hill Book Company, 1973, 787 pp

· Cole Parmer, «Teknisk, kjemisk», a/s WaMo, 1995-96

· Laider, K.J., Meiser, J.H.,«Physical Chemistry», The Benjamin/Cummings Publishing Company, Inc., 1982, 919 pp

· Horvath, A.L.,«Handbook of aqueous electrolyte solutions: physical properties, estimation and correlation methods», Ellis Horwood Limited Publishers, Chichester, England,1985, 631 pp, (pp 249-299)

· Robinson, R.A. and Stokes, R.H.,«Electrolyte solutions», London, Butterworths Scientific Publications, 1955, 512 pp

· Barduhn, A.J., Towlson, H.E. and Hu, Y.Chien,«The Properties of Some New Gas Hydrates and Their Use in Demineralizing Sea Water», AIChE Journal, Vol. 8, No. 2, May 1962, pp 176-183

· Tohidi, B., Burgass, R.W., Danesh, A. and Todd, A.C.,«Hydrate Inhibition Effect of Produced Water: Part 1-Ethane and Propane Simple Gas Hydrates», Heriot-Watt U., SPE, Offshore European Conference, Aberdeen, 7-10 September, 1993, pp 255-264

· Sloan, E.D.,«Clathrate Hydrates of Natural Gases», Marcel Dekker, Inc., New York, 1990, 641 pp

· Calsep A/S, «PVT and phase behaviour simulation system», Version 6.1
· Colorado School of Mines, «The Colorado School of Mines Hydrate Prediction Program-CSMHYD», August 5, 1996