1 HYDRATE PLUG FORMATION PREDICTION TOOL – AN INCREASING NEED FOR FLOW ASSURANCE IN THE OIL INDUSTRY Keijo Kinnari1, Catherine Labes-Carrier1, Knud Lunde1, Pål Hemmingsen2 , Simon R. Davies3, John A. Boxall3, Carolyn A. Koh3 and E. Dendy Sloan3* 1 StatoilHydro ASA, N-4035 Stavanger, Norway 2 StatoilHydro ASA, N-7005 Trondheim, Norway 3 Colorado School of Mines, Golden, CO 80401, U.S.A. ABSTRACT Hydrate plugging of hydrocarbon production conduits can cause large operational problems resulting in considerable economical losses. Modeling capabilities to predict hydrate plugging occurrences would help to improve facility design and operation in order to reduce the extent of such events. It would also contribute to a more effective and safer remediation process. This paper systematically describes different operational scenarios where hydrate plugging might occur and how a hydrate plug formation prediction tool would be beneficial. The current understanding of the mechanisms for hydrate formation, agglomeration and plugging of a pipeline are also presented. The results from this survey combined with the identified industrial needs are then used as a basis for the assessment of the capabilities of an existing hydrate plug formation model, called CSMHyK (The Colorado School of Mines Hydrate Kinetic Model). This has recently been implemented in the transient multiphase flow simulator OLGA as a separate module. Finally, examples using the current model in several operational scenarios are shown to illustrate some of its important capabilities. The results from these examples and the operational scenarios analysis are then used to discuss the future development needs of the CSMHyK model. Keywords: Flow Assurance, Hydrate Plugging, CSMHyK INTRODUCTION Since 1934 when Hammerschmidt [1] discovered that hydrates were the cause of plugging of natural gas transmission lines, significant research efforts have focused on understanding hydrate formation conditions and on the development of hydrate prevention methods. As the oil and gas industry moved into deeper water and corresponding higher pressures from larger liquid heads, the risk of hydrate formation increased. Recent research has focused on the rate of formation of hydrate plugs in addition to the traditional research on hydrate avoidance and on the new emerging industry practice of hydrate management. The current state- of-the-art model for hydrate plug formation * Corresponding author: E-mail: esloan@mines.edu 2 CSMHyK [2] (the Colorado School of Mines Hydrate Kinetics model) has recently been incorporated into the OLGA multiphase flow simulator and provides an estimate of where and approximately when a hydrate plug may form. In this paper, StatoilHydro presents their view on the industrial requirements for a hydrate plug formation model. The current understanding of the mechanisms for hydrate formation, agglomeration and plugging of a pipeline are then discussed and used as a basis for the assessment of the capabilities of CSMHyK model against industrial needs. Current limitations of the model are identified and future developments are outlined. Finally, example of applying the current model to predict hydrate formation in a fictional subsea tieback is presented. Comparison of model predictions to field data will be presented in a future paper. MOTIVATION StatoilHydro has over the past years done substantial work to reduce the common conservatism inherently present in the standard hydrate control strategies. Successful results have been obtained based on comprehensive R&D work and field experience. A key parameter for success has been the risk assessment process. However, much of this knowledge is still empirical. A model which is able to predict hydrate plugging in a given production system would clearly be a valuable tool. Results would be useful both in design and operations. The key questions to be answered in this context are: 1. How long can one operate safely without plugging within hydrate domain during normal production? 2. How long can one operate safely without plugging within hydrate domain during transient operations (typically start-ups)? 3. Given a hydrate problem, what is the extent of the plugging and where are the plugs located? Given a reliable model for hydrate plug formation, the flow assurance engineer could with confidence identify and implement more cost effective technical solutions both in design and operational procedures. The operator would again with confidence be able to decide on the required counter measures without jeopardizing production from plugging. In order to be able to identify the required characteristics of a plugging model it is necessary first to define the different operating scenarios in various production systems; three categories are identified: 1. Lines with continuous chemical injection 2. Lines with non-continuous chemical injection 3. Gas injection lines 1. Lines with continuous chemical injection In most cases the production systems covered under this category are gas/condensate systems. The lines are typically over 50 km in length with no insulation. In all these cases hydrate inhibitor is continuously injected at the template (examples in StatoilHydro operation or StatoilHydro’s partner operation: Snøhvit, Ormen Lange) or into the export line (Huldra-Heimdal, Troll, Kvitebjørn). Additional injection possibilities might exist at PLEMs. Operational scenarios which can result in hydrate plugging of these lines are discussed below case by case. RFOs (ready for operations) are special types of operations and special caution is normally paid to their design and execution. Plugging risk is practically eliminated. First-time start-up is closer to a normal operational scenario with some peculiarities. Risk assessments are typically performed by simulating the system behavior with OLGA. As a successful first time start-up is of crucial importance, hydrate prevention measures normally provide more than safe operating conditions. Normal operation with failure in the chemical distribution system is by far the most dreaded operational scenario. A key question in this category is how long the production can be maintained before the production must be stopped or other hydrate prevention measures must be implemented. It is also important to know if shut- down of the production would be more risky than continued operation. Clearly a model to predict the consequences to insufficient inhibition will be of great value. 3 Normal operation with limited injection capacity may occur as a result of an underestimation of the water rate in the design phase or as a result of an enhanced production rate. In both cases hydrates could gradually grow and deposit on the pipe wall and gradually increase the pressure drop. Typically, operational experience is used to determine the timescales for plugging. A more confident strategy could be devised if a sound model were available to predict the viable time window for such an operation. Normal production with sufficient chemical injection but with a risk for top-of-the line hydrate formation is a theoretical possibility for large pipes. Order of magnitude estimates can be made using OLGA with knowledge of hydrate thermodynamics. These estimates are necessarily quite conservative. Improved modeling capabilities would provide more accurate estimates and also provide more insight into the phenomenon itself. Shutdown and subsequent restart of production with underinhibited or uninhibited fluid in part of the production conduit is probably the most critical scenario. In most practical situations only a limited section of the production conduit would be in the hydrate domain. The risk will depend on the extent of underinhibition, the fluid, the volume, the time aspects, the geometry and the restart procedure. A model would clearly be of great benefit here. 2. Lines with non-continuous chemical injection This category of production systems covers lines which typically are insulated to keep the fluid temperature above the hydrate equilibrium temperature during normal operation and to provide sufficient cooldown time for implementation of the required hydrate control measures. Insulation is normally applied to pipelines with a maximum length of 50 km. This category also covers short lines and wells (like platform wells) where no insulation is required. Operational scenarios which can result in hydrate plugging in these lines are discussed below case by case. Operation at reduced rates can allow the fluid to cool into hydrate domain. Such a scenario can occur during start-up with limited numbers of wells online, as a result of problems with some wells, or during the tail production. In all of these cases it would be important to know the timescale for hydrate problems or if hydrate problems could be avoided. A mathematical model would clearly be a great advantage in this scenario. Shutdown and start-up operations may bring the system into the hydrate domain. The residence time inside hydrate domain in flowing conditions is normally quite limited (in the order of an hour or two) since the warm reservoir fluid warms up rapidly the production conduit, provided high enough rates are available. In this case it is important to know if sufficient hydrates can form to cause a plug during restart. Due to the consequences of plugging, a lot of conservatism still remains in existing restart procedures. A good predictive model would allow operators to optimize chemical injection rates for a transient operation within hydrate domain. This would result in large savings in chemical usage or in simplified design solutions and through reduced restart time. 3. Gas injection lines This category covers gas injection lines where gas with a relatively high water dew point is injected into a well. Operational scenarios with potential hydrate problems are as follows: Gas injection at a temperature below the water dew point but above the hydrate equilibrium temperature may result in a hydrate problem during a start-up scenario but normal operation will be safe. Water may accumulate in the line during the injection period. This water may form hydrates during restart after a prolonged shutdown and plug the injection conduit. Experimental work and some operational experience have shown that start-up from such a condition can be safe if high enough gas rates are applied. However, a mathematical model to support this kind of decision would be of great value. Savings would arise from simplified design, optimized dew point control, reductions in capacity, and reduced chemical usage in normal operational. Gas injection at low rates with a temperature below the water dew point and the hydrate equilibrium temperature will eventually result in a hydrate blockage. Conservative estimates can be made based on a worst case assumption of given water 4 fraction converting to hydrates. A mathematical model to allow more precise prediction is needed. THEORY OF HYDRATE GROWTH In the current model, the growth rate of hydrates is proportional to the contact area between water and hydrocarbon and is most likely dominated by the mass and heat transfer between phases and not by intrinsic kinetics. For a given fluid system the type of hydrate formation and growth could be classified according to the prevailing flow regime or pattern. Some fluids may, however, have some indigenous compounds which completely change the hydrate growth characteristics. Such chemical effects are ignored in the following discussions. When the water amount is low, the potential for plugging a conduit will be low in a shutdown situation. Plugging would not be possible if the water content is below a given limit. A lot of experimental work has been carried out to identify such limits in certain geometries and operational conditions [3]. On the other hand during a continuous operation even small amounts of water can over time result in plugging. Stagnant phases During a shutdown situation the phases separate forming a large continuous interface between water and hydrocarbon (except for very stable emulsions). Hydrate growth will occur at this interface which can be either the horizontal oil/water interface or the interface between gas and the water film on the wall. The hydrate volume formed for stagnant interfaces will be relatively small. It is unlikely that they can form a sufficient restriction to impede flow. The situation may change dramatically when the phases are set in motion. During a restart the hydrates formed during shut-in could seed the rest of the system, increasing the hydrate formation rate. A large agglomerate size distribution is expected. Hydrate chunks would be formed with large amounts of remaining free water and gas along with trapped liquid hydrocarbons. In sufficient volumes, these chunks could result in plugging provided that the prevailing hydrodynamic forces are not able to break them down. This kind of behavior would be expected at low flow rates in long enough conduits. Mass and heat transfer will determine the hydrate growth rate. Other forces like capillary forces will also influence the agglomeration rate and the formation of larger chunks. “Stagnant phases” with perturbation The interface between water and hydrocarbon may experience different degrees of perturbation. Perturbations would likely occur during a shutdown when parts of the system are still cooling, creating disturbances to the interfaces along the production conduit. Here no flow is assumed. Another source for perturbations can be movements of the riser. These would result in the largest perturbations on interfaces nearest the riser. The net effect of the perturbations is to increase the contact area and thus the rate of hydrate formation. At even small amplitudes, the initial hydrate growth can be much larger than in the previous case due to increased contact areas and mixing. If the amplitude attains the size of the thickness of the hydrocarbon liquid phase above the water phase, direct contact between water and the gas phase would be obtained. Favourable conditions for hydrate growth would then be present. Hydrate plugs may be formed through the perturbation mechanism during the shutdown. Cooling induced flow When the cooling effects discussed in the previous section are large and lasting, they could induce gas or liquid flows. The initially stagnant or semi- stagnant interfaces would be broken, and fluids rich in hydrate formers will come into contact with the water phase. This type of mechanism is envisaged in a riser base connected to a well-insulated or buried pipeline. The cooling of the pipeline can, in such situations take several days while the riser itself will get into the hydrate domain within few hours. Restart from a state where cooling induced flow has created hydrates can result in plugging. Plugging induced flow In cases where large amounts of hydrates have been formed, the drop in pressure may induce flow through the production conduit. The effect will be largest at lower pressures. For example, 1 m³ 5 hydrates takes typically 150 Sm³ gas, at 100 bar the gas volume would be roughly 1.5 m³, at 200 bar 0.75 m³ etc. This will enhance phase mixing causing increased hydrate formation, thus the process is self accelerating. Restart from this kind of situation would most likely not be possible. Flow with different flow regimes The next sections covers hydrate formation and growth at typical flowing conditions in a multiphase line, assuming that the restart has been possible. There are two different cases: 1) There are no hydrates present initially but these form in the flowing conduit 2) Hydrates are already present in the fluid through some of the mechanisms described above. Initially no hydrates present in the conduit One example of this scenario would be the failure of the chemical distribution system. The flowing multiphase fluids will gradually come into the hydrate domain. Hydrate formation will start at the water/hydrocarbon interfaces. The location of these interfaces depends on the flowrates, fluid characteristics, GOR, diameter, length, geometry etc. At extremely low fluid velocities the phases would be stratified. Some wavy perturbations exist at the oil and water interfaces. Hydrates may start growing at these interfaces and gradually form a slurry phase between the oil and water phases. At low rates no droplets would be present in the gas phase. The gas phase cools down along the pipeline and results in water condensation on the pipe wall. The water will either turn into hydrates or fall back to the liquid phase. Thus in this case hydrates can be formed both on the pipe wall and in the liquid phase. The contact area and mass transfer would be relatively low in the bulk liquid. Contact area in the gas would also be relatively small (water film), but the mass transfer would be fast. The hydrates in the liquid phase may jam further down the line. The growing hydrates on the wall would cause a restriction further upstream. At higher flow rates, waves form and liquid droplets starts to be ejected into the gas phase. Part of the water is dispersed in the oil phase and relatively good mixing in the liquid phase,may enhance the transport of hydrate formers to the water interface. The hydrate formation rate increases considerably as a result of the increased contact area and improved mass transfer characteristics. Hydrate growth in the bulk liquid phase will create a hydrate slurry in which the viscosity increases along the conduit. The particle or agglomerate size distribution is expected to be rather non-uniform with large hydrate chunks. This kind of behavior is often seen experimentally in StatoilHydro’s flow simulator (the flow pattern in these experiments was not stratified but the mixing in the liquid phase could be considered to be close to that discussed here). Hydrates may also deposit on the wall either directly as a result of hydrate growth or as a result of collision of wet hydrate particles/agglomerates from the bulk liquid phase with a water wetted (or bare) wall. Oil wetted walls are not expected to have the right affinity towards water coated hydrate particles. More and more of the water will be dispersed into the oil phase with increasing flow rate. This increases the contact area between water and oil. At the same time more gas is entrained into the liquid phase. The conditions for hydrate formation are thus greatly enhanced. The increased wetting of the pipe wall with oil most likely reduce the possibility for deposition. This kind of behavior is seen in liquid dominated systems. Systems with lower liquid content would approach this behavior at high flow rates in which the wall becomes oil- wet. If there is enough free water present, conditions would promote the formation of large “chunks”. The increasing shear forces from the higher flow rate would counteract the agglomeration of hydrate particles and would reduce hydrate deposition on the wall. Hydrates would in this case be most likely transported in a slurry phase as relatively uniform particles or aggregates. The higher hydrate volume fraction will increase the slurry viscosity. The viscosity increase may result in a blockage (see CSM model). At certain conditions, hydrodynamic slug flow may occur. Slugging would provide the best 6 conditions for hydrate growth in the bulk liquid phase due to the mixing of the phases. Slugging is expected to impede hydrate deposition on the wall both due to shear and wetting of the wall with oil. Most of the hydrate growth would thus occur in the bulk liquid phase. The discussion above is mostly related to flow in horizontal pipes and to pipes with some inclination. Different aspects may apply to vertical pipes such as risers and pipes with large inclinations but the general behaviour is still expected to be valid. Hydrates already present in the fluid The discussion in the previous section is in principle transferable to the case when hydrates already are present in the fluid. This would be the typical situation after a prolonged shutdown situation where hydrates have been formed. The main difference lies in the faster hydrate growth from seeding, and the more non-uniform initial size distribution of the hydrate agglomerates. CMSHYK MODEL DESCRIPTION CSMHyK is a plug-in module for the OLGA (SPT Group) multiphase flow simulator; researchers at CSM have been developing the module in cooperation with SPT Group since 2003. The model predicts the rate of hydrate formation using a first-order rate equation based on the thermal driving force. The rate equation was originally proposed by Vysniauskas and Bishnoi [4] in the absence of mass and heat transfer limitations. In order to accurately simulate hydrate formation in flowloops, it was necessary to reduce the intrinsic kinetic constant by a factor of 500. The lumped parameter accounts for mass and heat transfer limitations in the flowloops. The current model assumes that the hydrate particles convert directly from emulsified water droplets. Nucleation is assumed to occur instantaneously at a subcooling of 3.6°C/6.5°F, a parameter proposed by Matthews [5]. Once formed, the model assumes that these particles remain in the oil phase. The change in relative viscosity of this phase is then found from the Camargo and Palermo [6] correlation for steady state slurry flow. An overview of the current CSMHyK module and its integration into OLGA is shown in Figure 1. Driving Forces Yes No Dissociation Model Pipeline Section n-1 Pipeline Section n+1 Amount of Hydrate Relative Viscosity Growth Rate Formation Model Thermo Properties System Properties Thermo Properties Hydrate forming? CSMHyK Output Fluid Properties Pipeline Section n OLGA® Figure 1: Integration of CSMHyK into OLGA. The CSM plugging model has following characteristics: 1) Water is assumed to be dispersed in the oil phase. 2) Hydrate growth rate depends on the interfacial area between water and oil. Surface area is calculated from OLGA. 3) The size of the water droplets is fixed to one single value which can be modified by the user (default 40 μm). 4) The primary hydrate particles agglomerate in the oil phase. The size is determined from the Camargo and Palermo force balance [6] of adhesive and shear forces. In practice this gives single agglomerate size for given hydrate particle concentration in the slurry at a given shear rate. 5) Plugging is caused by the viscosity increase of the hydrate slurry from increasing hydrate volume fractions. Plugging is defined by the attainment of a preset maximum allowable viscosity value. 6) The model does not account for hydrate deposition on the wall. All the hydrate growth occurs in the liquid phase. Deposition can, however, be simulated by imposing a low slip value between hydrate and the oil. In this case the hydrate particles will have a low velocity and accumulate where they initially form. 7) The model accounts for the exothermic heat generated from hydrate formation. 8) The model does not account for the physicochemical effects of the fluid. 7 DISCUSSION The current version of the CSM model is clearly a simplification of the hydrate growth mechanisms as only one growth mechanism is considered. Hydrate growth and eventually the plugging of a production conduit normally occurs in a more complex way including both the pipe wall and the bulk liquid phase as discussed in the previous sections. Hydrate plugging often occurs by the formation of large hydrate “chunks” which subsequently jam the pipeline. Substantial evidence for this type of behavior has been obtained in the numerous visual observations in the StatoilHydro flow simulator at Rotvoll and indirectely in other test systems like the StatoilHydro’s flow assurance pilot at Kårstø. The CSM model does not yet allow for jamming to occur and the current rheological model rarely allows such large hydrate chunks to form. For systems with high shear rates and an oil wetted wall the hydrate particles would be dispersed in the oil phase. In such a case the behavior would be quite close to what already can be predicted by the CSM model. One of the strongest features of the model is the heat balance combined with the information OLGA can provide on the water holdup distribution (but still assuming monodispersed water droplets), flow rates and heat transfer. The positive impact from increased temperature as a result of hydrate formation is most often ignored in practice. The temperature increase may in certain cases be so high that it stops further hydrate growth or at least considerably slows it down. Experimental tests carried out at the StatoilHydro Flow Assurance Pilot show large temperature increases as a result of high hydrate formation. At the current state the model has limited capability to provide answers with respect to all the aspects of plugging. It may, however, provide some valuable insights in risk assessment work when properly used. For example, in certain cases the risk assessment includes identification of water accumulation points and the potential for converting this water into hydrates in critical pipe sections. The model could be used to give an indication of the zone where plugging would occur in the liquid phase. POTENTIAL INDUSTRIAL APPLICATIONS In this section an example is used to highlight potential industrial applications of the model. In this example, a fictional subsea tieback is simulated and the effect of adjusting 4 key parameters is investigated. Subsea Tieback This example demonstrates how the adjustable parameters can be used to evaluate the blockage potential of a fictional tieback. The four main adjustable parameters in the CSMHyK-OLGA model are as follows {OLGA Keyword in block capitals and applicable adjustment range in square brackets}: 1. K1SCALINGFACTOR: Adjustment of the intrinsic kinetic formation rate [0-1] 2. SUBCOOLING: Adjustment of the temperature sub-cooling (hydrate equilibrium T – system T) before hydrate formation initiates [0+] 3. SOIL: Adjustment of the velocity difference (slip) between the oil velocity and the hydrate [0-1] 4. SIZESCALING: Adjustment of the individual hydrate (monomer) particle size used in the aggregation model [0+] The keyword K1SCALINGFACTOR refers to a multiplier for the intrinsic kinetic rate. A value of 0.01 refers to slowing the intrinsic kinetic rate by a factor of 100. The keyword SUBCOOLING is the extent that the system temperature must fall below the hydrate equilibrium temperature for hydrates to start forming (i.e. SUBCOOLING = Teqm – Tsystem). Once hydrates are present, they can continue to form as long as the system temperature is below the hydrate equilibrium temperature. This meta- stability is almost always seen in hydrate nucleation. The default value for the SUBCOOLING is 3.6°C/6.5°F, which is taken from Matthews et al. [5]. The keyword SOIL is the term given to the slip factor between the oil velocity and the hydrate velocity. Alternatively, SWAT can be used which adjusts the slip between the water velocity and the hydrate velocity. The default is 1 for SOIL and 0 8 for SWAT. The calculation of the hydrate velocity is based on the following equation, where V refers to the superficial velocity of the phase (in the subscript): The keyword SIZESCALING refers to the multiplication factor of the monomer hydrate particle size; the default value is 40 µm. The fictitious deepwater tie-back for the simulation is a 48 km/30 mile long pipeline in 1500 m/5000 ft of water. The flow line barely enters the hydrate domain. Figure 2 details some of the basic inputs into the OLGA model generated for this example. Wellhead Separator Riser Flow Line 5000 ft 30 miles, 10” diameter P = 15,000 psia T = 220 °F Water cut = 35% T = 39.2 °F U = 4 BTU/Ft2 F Hr P = 300 psia T = 80 °F Well Area of Hydrate Risk Figure 2: Schematic of the tieback in the simulation. Figure 3 shows the simulated temperature profile and hydrate formed using the default values, i.e. 3.6°C/6.5°F sub-cooling, intrinsic kinetics, oil slip factor of 1 (no deposition), and 40 µm droplet/hydrate particle size. Hydrate formation in this case is heat transfer limited, i.e. limited by the ability of the pipeline to remove heat due to the exothermic heat of hydrate formation. This is seen by the pipeline temperature quickly rising to the hydrate equilibrium temperature. Figures 4 to 7 show the effect of changing each of the parameters, the most sensitive of which is the oil slip. Pipeline T Equilibrium T Hydrate fraction Figure 3: Simulation results using base case values. Pipeline T Equilibrium T K1SCALING = 1.0 K1SCALING = 0.1 Pipeline T K1SCALING = 1.0 K1SCALING = 0.01 Equilibrium T Figure 4: The K1 scaling factor must be reduced by 100 to move the system from heat transfer control. SUBCOOLING = 0.5 F SUBCOOLING = 4 F SUBCOOLING = 6.5 F SUBCOOLING = 9 F Figure 5: Changing the subcooling changes the point of onset of hydrate formation but has little effect on the amount of hydrate formed. 9 SOIL = 0.01 plugged SOIL = 0.01 plugged Figure 6: The oil slip is the most sensitive parameter; A low value for SOIL leads to hydrate accumulation. SIZESCALING = 2 SIZESCALING = 0.5 SIZESCALING = 2 SIZESCALING = 0.5 Figure 7: The droplet size has little effect on the amount of hydrate formed or on the pressure drop; Increasing the droplet size reduces the relative viscosity. CONCLUSIONS The industry is currently lacking predictive tools for hydrate plugging. More accurate knowledge of hydrate plugging risk will benefit system design and operational support. The new CSM model has been developed in order to fill this gap. The main application domains and features of this model have been presented. The current version of the CSM model is the first step towards describing the complex mechanisms in multiphase systems. Model improvements are still required along with further qualification in well defined test systems. Important model extensions include the possibility of having both deposition on the wall and hydrate growth and agglomeration in the liquid bulk. In addition the droplet size distribution needs revisiting and more realistic size distributions should be incorporated in the model. The hydrate growth is most likely controlled by the mass and heat transfer in the close vicinity of the growing particles. It is therefore mandatory to expand the model to include these aspects. The current plugging model is based on the gradual increase in the viscosity of the hydrate slurry. The further model development will include the possibility of forming large agglomerates, depositions on the wall and jamming the pipe by detached deposits. . ACKNOWLEDGEMENTS The authors wish to thank StatoilHydro ASA for allowing publishing of this paper. The authors further acknowledge the financial support received from the Deepstar Consortium of energy companies for the CMSHyK model development. REFERENCES [1] Hammerschmidt E.G. Formation of Gas Hydrates in Natural Gas Transmission Lines. Ind. Eng. Chem. 1934;26:851. [2] Turner D. et al. Development of a Hydrate Kinetic Model and It’s Incorporation into the OLGA 2000® Transient Multiphase Flow Simulator. 5th International 10 Conference on Gas Hydrates, 4018, Trondheim, Norway, 2005, p1231-1240. [3] Kinnari K., Labes-Carrier C., Habetinova E., Straume E. and Hjarbo K. Reduced chemical injection strategy for hydrate control of subsea templates and spools. 5th North American Conference on Multiphase Technology, Banff, Canada, 2006, p. 21-35. [4] Vysniauskas A. and Bishnoi P.R. A Kinetic Study of Methane Hydrate Formation. Chemical Engineering Science, Vol. 38, No 7, pp1061-1972, 1983. [5] Matthews P.N., Notz P.K., Widener M.W. and Prukop G. Flow Loop Experiments Determine Hydrate Plugging Tendencies in the Field. Gas Hydrates: Challenges for the Future. NYAS Vol. 912, p330-338. [6] Camargo R. and Palermo T. Rheological Properties of Hydrate Suspensions in an Asphaltenic Crude Oil. Proc. 4th Int. Conf. Gas Hydrates, 2002, pp880-885, Yokahama.