CHAPTER 6 SIMULATION OF RRR MODEL IN OLGA AND PIGGING CASE STUDY

Transcription

1 CHAPTER 6 SIMULATION OF RRR MODEL IN OLGA AND PIGGING CASE STUDY 6.1 INTRODUCTION According to the literature, there are several wax deposition models with different approaches for how to model wax deposition. Unfortunately, none of these models incorporates the physical processes that constitute the wax deposition process, and the process of deposition still remains poorly understood. Most of the focus concerning formation and deposition of wax has been put on single-phase transportation of paraffinic oils, but significant problems with wax deposition may also occur in unstabilized oil and condensate systems (Geankoplis, 2003). An accurate prediction of wax deposition rates and deposited wax distribution in pipelines would represent high value for the industry. This information is valuable for the design stages of the field and also in the scheduling of intervention in the pipeline, in order to assure the oil flow at the desired rate. Four wax deposition models are described here. These are the RRR (Rygg, Rydahl and Rønningsen) model, the Matzain model, the Hydro model and the University of Michigan model. Simulations are to be carried out by using the RRR-and the Matzain model, which are implemented in the transient simulator, OLGA. The most important element is to illustrate how wax deposition models predict wax build-up. 6.2 THE RRR MODEL The RRR (Rygg, Rydahl and Rønningsen) model is a multi-phase flow wax deposition model which predicts wax deposition in wells and pipelines. Note that the RRR model is not applicable for laminar flow (Dirksen and Ring, 1991). A standard steady state multi-phase point model is used to predict pressure drop and liquid hold-up along the pipeline. The effect of deposition on pressure drop and temperature is calculated by integration in time. The multi-component wax model continuously estimates the wax precipitation along the pipeline and the viscosity of the composition. The wax deposition is then estimated from the diffusion of wax from the bulk towards the surface of the pipeline, due to temperature gradients and 99

2 shear dispersion effect. The inner pipe wall friction is varied due to wax deposition (Geankoplis, 2003). The wax deposition model is divided into separate sub-models, where each submodel concerns on specific technical aspect and the components are treated individually (Geankoplis, 2003). Sub-models: Flow model calculates pressure drop and flow regime Thermodynamic Wax Model (TWM) determines the number and properties of the different phases for each section Viscosity model calculates the viscosity Wax deposition model predicts the amount of wax that deposits in a section of the pipeline. Wax deposition build-up is a slower process than flow disturbances in a pipeline. Therefore a semi-stationary model is chosen for wax deposition predictions over longer periods. Semi-stationary is described as changes in boundary conditions like flow rates, pressures and temperature taken into account, as well as varying inner pipe diameter. The model is also compositional, meaning that all components are treated individually in each sub-model. A component mass balance is used for track keeping of all components (Geankoplis, 2003). The pipeline is discretized into a number of sections, where pressure and composition are assumed constant within one section. The temperature gradient reaches from the centre of the pipe to the pipe wall. The energy balance determines the bulk and wall temperatures. All sections are simulated, and independent pressure drops for each section are added up and gives the total pressure drop (Geankoplis, 2003). Wax deposited in one period affects the results in the following time periods. Wax deposits in one section leads to decreased diameter, thus higher pressure drop in the pipeline. Additionally the energy balance is influenced, because the wax has an insulating effect in the pipeline. This results in increased temperature over time for a waxy section (Geankoplis, 2003). 100

3 Deposition in the RRR model is based on molecular diffusion and shear dispersion, both mechanisms which enhance the wax deposition (Geankoplis, 2003). The volume rate of wax deposition by molecular diffusion for a wax forming composition i is found from Where C b i & C w i the molar concentrations of the wax component i dissolved in the oil phase in the bulk and at the wall respectively (mole/m 3 ), wet S is the fraction of the wetted circumference, N wax is the number of wax components, MW i is the molar weight of wax component i (kg/mole), ρ i the density of wax component i (kg/m 3 ), r is the current inner pipe radius (m) and L is the length of the pipe section (m). D is the diffusion coefficient, and the Hayduk-Minhas correlation (m 2 /s), (Hayduk and Minhas, 1982) is used to calculate the diffusion coefficient. δ is the thickness of the laminar sub-layer (m), and Blasius equation is used when calculating the thickness of the sub-layer in turbulent flow (Geankoplis, 2003). Also the volume rate of wax deposited from shear dispersion can be estimated from a correlation by Burger et al. (1981) When accounting for both mechanisms, the total rate of increase in thickness for the wax layer is given as Where, ϕ is the porosity of the deposited wax. The porosity is usually assumed to be in the range from Wax layer porosity is an adjustable parameter in the model. Note that the thickness of the wax layer is averaged around the pipe circumference, even if the inner pipe surface is only partially wetted with liquid. For multi-phase flow the wetted inner surface area depends on the local flow regime predicted and the liquid hold-up (Geankoplis, 2003). The RRR model has been applied to several single-and multi-phase pipeline 101

4 systems. In two cases presented by Rygg et al. (1998), wax-build-up, temperature and pressure drops were simulated over time. Good agreement was shown between calculated and observed pressure losses. Important tuning parameters in this model are the wax layer porosity and the roughness of the wax deposit (Geankoplis, 2003). An assumption made for this model is that the precipitation rate at the wall is not a limiting factor, and that all wax transported to the wall will stick to the surface as long as the temperature is below WAT (Huang et al., 2011). It should be noticed that no removal mechanisms are discussed in this model. Therefore, the main drawback concerning the RRR model is the lack of shear incorporated into the model. Questions should be asked if this is a correct assumption. However, the model may give good enough predictions for low flow rates. Additionally, this model may give reasonable predictions at the time when wax starts to grow at a clean pipe, but after some time the results become less reliable, due to no release-mechanism included in the model. 6.3 FIELD BACKGROUND AND DATA AB field was put on production in April Presently the production from AB field is about 500m 3 /d. The crude from A is pumped to B. At B the entire AB crude is heated to 55 o C and pumped in A-B crude dispatch line. Due to viscous nature of ABcrude high pressure drop is observed during winters in the A-B sector. In past, flow assurance studies wherein blending with lighter crudes, heating at A before pumping and chemical treatment were carried. Asset is resorting to chemical treatment during winters in addition to routine heating before pumping the ABcrude to B to maintain the viscosity within desirable limits for back pressure reduction. Intermittent pumping of crude and water is being done to reduce the back pressure. The lines are being pigged every month for scrapping the wax deposition. The wax deposition propensity and pig stuck up especially during winters is area of concern. 102

5 Pressure kg/cm Study of Dynamic Simulation The dynamic simulation study for wax deposition in trunk lines A-B was carried out in which pressure and temperature trends were analysed Line Pressure Trend (A-B Line) Monthly pressure trend Pressure kg/cm Chemical injection started Chemical injection ended Fig. 6.1 Monthly pressure trend Historical Trend Temperature v/s Pressure Fig. 6.2 Historical trend Temperature v/s pressure 103

6 Historical Trend Volume v/s Pressure Crude Dispatch Schematic Fig. 6.3 Historical trend volume v/s pressure The crude dispatch line from A to B with pumping details at each installation is A 25Km B Fig. 6.4 Crude dispatch line 104

7 6.3.4 Crude Dispatch Line: Coating & Burial Fig. 6.5 Coating & Burial The salient details of line are tabulated below: Table 6.1 Details of Crude Dispatch Line. No. Particulars Unit Value 1 Line size 8 inches 2 OD inches Grade API X-46 4 Wall Thickness Inches 5 ID inches Length A-B Km 25 7 Burial depth 1.2 m 8 Coating Coal tar enamel 4mm 9 Pumping rate m 3 /hr 10 Dispatch temperature 55 o C 11 Dispatch viscosity 20 CP 13 Fluid viscosity 800 cp 105

8 6.3.5 Fluid Parameters Table 6.2 Physical Characteristics of crude oil from field AB. No Properties Value 1 Sp. Gravity 60/60oF API Gravity 60 o F 32 3 Pour Point ( o C) Viscosity Dry crude CP o C,40300@ 25 o C 5 Viscosity emulsion (50% w/c) cp o C, o C 6 Viscosity emulsion (90% w/c) cp o C, 4373 o C 7 STO(Wax) WT % SIMULATION STUDIES FOR A B LINE Simulation studies have been carried out in OLGA 7.2 Software in order to find out the pressure and temperature drops in the pipeline under various operating scenarios and different ambient temperatures:- i. Steady state flow (A to B) ii. Wax Deposition and Pigging iii. Skin Heat Tracing System iv. Alternate pumping of Oil and water v. Higher dispatch temperature vi Additional line from A to B Steady State Flow (A to B) The crude from A is diverted to crude oil tanks at B and again pumped into Crude dispatch line with B crude. The steady state flow for two dispatch scenarios has been simulated. SCENARIO-1 Fluid pumped from A to Bshows that node pressure at B end is about 18 bar. 106

9 SCENARIO-2 Fluid pumped from A to B is diverted in to B tanks during winters. The node pressure at B end is about 1 bar. The simulation results are presented below:- Fig. 6.6 Pressure at A v/s Pumping rate Fig. 6.7 Temperature Gradient v/s Pumping rate 107

10 Fig. 6.8 Fluid Viscosity vs. Pipeline length in the A-B line Fig. 6.9 Pressure drop in Ato B segment 108

11 Fig m 3 /d v/s Pumping rate Fig Fluid viscosity at B end 109

12 Fig Fluid shear rate at B end v/s Pumping rate Fig Temperature B end v/s Pumping rate 110

13 Fig Fluid velocity v/s Pumping rate From the graphs above it can be seen that pressure increase is mainly attributable viscosity increase due to heat loss to ambient along the flow path. The simulated pressures for the two scenarios i.e. SCENARIO-1 Fluid pumped from A to B is diverted in to B tanks during winters. Table 6.3 Steady State Flow For Two Scenarios. Pumping rate A Pr.(bar) Scenario-1 Scenario-2 12m3/hr m3/hr m3/hr m3/hr m3/hr It is seen that the model fairly represents the field behaviour for both the scenarios. The model has been used to carry out further studies for wax deposition, skin heat tracing and higher inlet temperatures simulation. 111

14 6.4.2 Wax Deposition and Pigging The AB sector is being pigged for scrapping of wax once in a month. Simulation has been carried out in OLGA 7.2 to predict the wax deposition propensity. The fluid modeling has been carried in PVTSim20 to capture the fluid parameters. The simulation results representing winter conditions i.e. Ambient temperature of 10 o C are presented below:- Fig Wax Mass v/s Days Fig Wax thickness along the line 112

15 A pigging simulation from A to B was carried out after 30 days of wax deposition. Figure 6.17: Wax removal by pigging after 30 days deposition Fig Pig velocity and travel time The pigging case simulation results for winter conditions i.e. ambient 10 o C reveal the following:- Table 6.4 Pigging case Simulation Results. No. Particulars Results 1 Wax deposition Mass in 30 days ~100 kg 2 Distance propensity from A ~ 5 kms 113

16 3 Peak thickness 0.3 mm 4 Average Pig Velocity 0.18 m/s 5 Pig travel time (A - B) ~41 hrs It is seen from above that the wax deposition during winter conditions is not significant. The pigging frequency to once in 30 days is optimum for removal of wax etc Skin Heat Tracing System (SHTS) A simulation model representing the skin heat tracing system similar was built in OLGA7.2. The main components represented are heating system strapped on the line, Polyurethane foam covering, HDPE pipe and buried line. Sensitivity has been carried out on Power v/s Arrival temperature. The SHTS system precludes laying of a new crude despatch line for A to B. Fig Skin heating system The simulation details for winter case are tabulated below:- Fig Fluid arrival Temperature v/s Power 114

17 The above graph in Fig depicts the flowing fluid temperature under various heating conditions i.e 0,2,4,6,8,10 W/M Pumping pressure v/s Power Fig. The above graph in Fig depicts the starting pressure requirement at A under various heating conditions i.e. 0,2,4,6,8,10 W/m. Fig Fluid Viscosity v/s Power The above graph in Fig depicts the fluid viscosity at B end under various heating conditions i.e. 0,2,4,6,8,10 W/m. The various parameters like viscosity, temperature, pressure drop etc. have been plotted below against sensitivity variant for of current provided for heating the line. 115

18 Fig Power v/s Total Pressure drop The variation of pressure drop with respect to power can be observed in Fig above where pressure is in bar and the magnitude of total pressure drop at 0,2,4,6,8,10 W/m can be observed. Fig Power v/s Frictional Pressure Drop 116

19 The salient findings are tabulated below:- Table 6.5 Power Requirement for Ensuring Pressure. No Line Power Arrival Starting Flowing Fluid Pressure. [W/m] Temp Pressure, Viscosity at B end, drop in line, 1 Existing line 2 Skin heat tracing system From the above table it is can be inferred that a power of 10 W/m is required to ensure pumping pressure below 30 bar during winter conditions Alternate Pumping Of Oil And Water During discussions Asset informed that in PQ area oil and water are being pumped alternately to mitigate the high back pressure and potential for line blockage. Simulations have been carried out for P-Q sector considering pumping of oil for 4 hrs followed by water for 4 hrs. The simulation results are graphically represented in Fig below: Fig Oil Water pumping cycle at P 117

20 Fig Pressure at P v/s Conditions Fig Oil-Water holdup in line The simulation results from above Fig & 6.27 reveal that an alternate oil and water would lead to reduced back pressure at AB end. However the model does not account for higher fluid viscosity due to likely formation of oil & water into a emulsion. The predicted flow regime is stratified hence mixing of oil and water may not occur. This can be tried as pilot during winters if sufficient processing and water handling capacity is available at P end Higher Dispatch Temperatures It was informed by Asset that one additional heater treater is being installed at A which will result in better temperature management at A end before pumping the crude. Simulation study has been carried out to ascertain the impact of higher despatch temperatures on pressure drop during winters. The simulation results for 118

21 fluid viscosity with temperature and pressure drop along the length of pipeline are graphically represented in Fig & 6.29 below: Fig Fluid viscosity at B end Fig Pressure drop in line 119

22 Fig Temperature gradient The variation of temperature gradient and overall heat transfer coefficient with respect to the length of the pipeline is shown in Fig & 6.31 respectively. Fig Overall heat transfer coefficient 120

23 Fig Pressure v/s despatch temperature Fig Temperature drop in line From the studies a marginal impact on pumping pressure is observed on account of higher starting temperature is observed for flow rate of 16m 3 /hr in Fig and temperature drop in line can be seen in Fig Additional Line from A to B Option to analyse another option considering a loop line from A to B to reduce back pressure during winters. The flow hydraulics was studied considering one additional line. 121

24 The simulation results are graphically represented in Fig & 6.35below respectively. Fig Pressure Drop in line In both of these graphs two pipes represents loop line for which pressure drop and shear rate can be observed to be increasing with increase in mass flow rate. Fig Shear rate at B end 122

25 Fig Fluid velocity: Flow rate v/s pipe Fig Temperature drop in line From the studies it is seen that additional line would reduce the pumping pressure requirement significantly. However it may not be of help in mitigating gel formation if the pumping has to be shut down for long period. The graph below in Fig depicts the cool down period of the line from a stabilised flow rate. The fluid temperature in entire line is expected to reach below pour point temperature in about 48 hrs. 123

26 Fig. 6.38Gel strength along the pipeline The gel strength of the fluid can be scaled up to the required restart pressure of full size pipelines using the simplified equation below (which assumes the gel strength is constant throughout the gel plug and the entire plug will yield simultaneously): ΔP = 4ƮL/D Where: Ʈ: Yield Strength (Pa) L: flow line length (m) D: flow line ID (m) ΔP: restart pressure (Pa) Of the above back pressure reduction options diversion to B tanks is already being successfully resorted to in the field. However, Skin heat tracing system would allow flow under all operating condition including restart after shutdown. It has been evaluated vis-a-vis Chemical treatment. 6.5 CASE STUDY The present case study uses OLGA to simulate the pigging operations & PVTsim at its backend. Continuous monitoring of pressure & temperature was done through SCADA system. Brief about OLGA: The OLGA dynamic multiphase flow simulator models timedependent behaviours, or transient flow, to maximize production potential. Transient 124

27 modeling is an essential component for feasibility studies and field development design. Dynamic simulation is essential in deep water and is used extensively in both offshore and onshore developments to investigate transient behaviour in pipelines and wellbores. The OLGA Wax module calculates the deposition and transport of wax components along the pipeline. It models the effects of increases in pipeline roughness, decreases in pipeline diameter, and the increased apparent viscosity of the oil phase with precipitated solid wax particles. Wax deposition occurs on the inside surface of a flow line due to molecular diffusion when the pipe wall temperature falls below the wax appearance temperature (WAT). Wax precipitation occurs in the oil bulk flow when the bulk temperature is below WAT. The Wax module supports tuning fluid properties related to molecular diffusion, dissolution, shear related wax transport, and effective viscosity of an oil/wax mixture to dynamically model wax deposition, dissolution, and transport effects. The OLGA simulator also simulates pigging operations for wax layer removal and transport. Brief about PVTsim: PVTsim is a versatile equation of state (EOS) modeling software that allows the user to simulate fluid properties and experimental PVT data. The wax module evaluates wax formation conditions from an ordinary compositional analysis, quantify the amount of wax precipitate, run flash calculations, and plot wax formation conditions through PT curves. If data is available, it is also possible to tune the wax model to an experimental cloud point or to experimental wax content in the stock tank oil. The amount of wax precipitated may be calculated as a function of P for constant T or as a function of T for constant P and quantitative flash calculations will consider gas, oil and wax. Additionally, there is an option to account for the influence of wax inhibitors. The trunk line between A&B is being pigged for scrapping of wax once in 30 days irrespective of weather. Simulation has been carried out in OLGA 7.2 to predict the wax deposition propensity. The fluid modeling has been carried in PVTSim18 to capture the fluid parameters. 125

28 Table 6.6 Trunk line details. No. Particulars Unit Value 1. Line size inches 8 2. OD inches Grade API X Wall Thickness Inches ID inches Length km Burial depth m Coating Coal tar enamel mm 4 9. Pumping rate m 3 /hr Dispatch temperature Deg C Dispatch viscosity cp Pressure at end ( TPL / Tank) Bar 18 / Fluid Temperature at Ramol end Deg C Fluid viscosity at end cp ~ 800 cp 15. Ambient temp (Max/Min) Deg C 38 / Crude dispatch pumps Rated Pressure 60 kg/cm2 Rated flow 2 x 12 m3 hr, 1x 16 m3/hr Table 6.7 Physical Characteristics of crude oil. No. Properties Value 1. Sp. Gravity 60/60 o F API Gravity 60 o F Pour Point ( o C) Viscosity Dry crude cp o C, 40300@25 o C 5. Viscosity emulsion (50% w/c) cp o C, 22500@25 o C 6. Viscosity emulsion (90% w/c) cp o C, 4373@25 o C 7. STO(Wax) WT %

29 Fig Wax Mass vs Days Form the above graph it can be seen nearly 110 kg of wax mass has been predicted. At the pig receiving station nearly 40 kg of un-dissolved wax is observed. So it is very rightly & modest prediction that 70 kg of wax will be in semi dissolved state. Fig Wax thickness along the trunkline It can be seen that a maximum wax thickness predicted is 0.32mm at nearly 4.5 kms from the pig launching station i.e. A. The reasons for this are change in trunkline orientation & temperature gradient. These factors greatly influence the wax 127

30 deposition. It can be seen that later wax deposition is constant as equilibrium is achieved & irrespective of 10, 20 or 30 days its remains nearly the same. Fig Wax removal by pigging after 30 days deposition From the prediction it can be seen that distance propensity of wax deposition is nearly 4.8 kms from A. An anti-exponential nature is predicted further due to equilibrium. A near correct estimate of nearly 41.5 hours has been predicted whereas 42.5 hours had been taken for the pig to travel across the trunkline. Fig Pig velocity and travel time The Figure shows pig velocity, average pig velocity & the pig distance travelled. The average pig velocity is 0.182m/s. 6.6 PRESSURE AND TEMPERATURE VARIATION IN TRUNK LINE Temperature and pressure profile for A - Bfor6 months from November to April. 128

31 O Temperature X Pressure [ 2 P e n ] H i s t o r i c a l T r e n d A M D L I M G G S C O D - T I M D a i l y A v g [ o C ] A M D L I M G G S D P M P - P I D a i l y A v g [ k g / c m 2 ] [ E ] [ E ] N o v D e c J a n F e b M a r A p r Fig Temperature and pressure profile for 6 months from SCADA The above Figures are the flow rate-pressure and pressure temperature variation in A -B trunk line. During winters November to February, the drop in temperature leads to increased wax deposition in the pipeline. This results in high pressure inside the pipeline. While in April due to increase in temperature, wax deposition decreases so pressure also decreases. Figure also shows increase in backpressure (February 4 pig launched) due to wax deposition & it is further increased due to launching of pig. After the pig is received at B receiving station, immediate pressure drop is observed. A gradual increase in trunk line temperature can also be seen with decrease in pressure. One important parameter which plays a deciding factor in wax deposition in trunk line is the Stock Tank Oil s water cut. Higher water cut makes wax inter molecular wax bonding stronger which further increases pressure. An increase in water cut can also cause to trunk line to corrode thereby providing active sites for nucleation for paraffin s. The current simulation gives good results in a 30 day pigging cycle, but as the quantity of liquid to be transported via trunk line will increase this pigging frequency needs to be accordingly modified. Results: The pigging case simulation results for winter conditions i.e. ambient 11 O C reveal the following:- 129

32 Table 6.8 Pigging Case Simulation Results For Winter Conditions. No. Particulars Results 1 Wax deposition Mass in 30 days ~110 kg 2 Distance propensity from A ~ 4.8 kms 3 Peak thickness 0.32 mm 4 Average Pig Velocity m/s 5 Pig travel time (A - B) ~41.5 hrs 130

33 REFERENCES 1. Burger, E.D., Perkins, T.K., and Striegler, J.H., 1981, Studies of Wax Deposition in the Trans-Alaska Pipeline, Journal of Petroleum Technology Vol. 33 (No. 6), PP Dirksen, J., and Ring, T., 1991, Fundamentals of Crystallization: Kinetic Effects on Particle Size Distribution and Morphology. Chemical Engineering Science Vol. 46, PP Geankoplis, C., 2003, Transport Processes and Separation Process Principles, Vol Hayduk, W. and Minhas, B.S., 1982, Correlations for Prediction of Molecular Diffusivities in Liquids. The Canadian Journal of Chemical Engineering Vol. 60 (No. 2), PP Huang, Z., Senra, M., Kapoor, R. and Fogler, H.S., 2011, Wax Deposition Modeling of Oil/Water Stratified Channel Flow. 6. Rygg, O.B., Rydahl, A.K. and Rønningsen, H.P., 1998, Wax deposition in offshore pipeline systems, Proceedings 1st North American Conference on Multiphase technology, Banff, Canada, June, Singh, A., Lee, H., Singh, P., and Sarica, C., 2011, SS: Flow Assurance: Validation of Wax Deposition Models Using Field Data from a Subsea Pipeline, OTC 21641, the Offshore Technology Conference,