Practical Gas Lift Manual 3r322i

Schlumberger Interventions and Completions

A Practical Gas Lift Manual For Designs in WellFlo Design Guidlines

Practical Gas Lift Manual: Design Guidelines /var/www/apps/conversion/tmp/scratch_2/156060266.doc -1-

1.0 Introduction WellFlo is a Nodal Analysis program. Its function is to analyse the behaviour of petroleum fluids in wells. This behaviour is modelled in of the pressure and temperature of the fluids, as a function of flowrate and fluid properties. The program takes as its input a description of the reservoir, of the well completion (i.e. the hardware of the well), and of the surface hardware (i.e. pipelines etc…). This is combined with fluid properties data (PVT). A well is modelled with a series of nodes from the reservoir to the separator and the program then performs calculations to determine the pressure and temperature of the fluids at each node. Different modes of operation can be employed to either solve for flowrate given controlling pressures (typically done for deliverability calculations), or solving for pressure drops given measured flowrates (typically done for diagnostic calculations). WellFlo calculations are based on Nodal Analysis. There are two main types of Nodal Analysis. The first is the determination of flowrates from pressures , and the second the determination of pressures from flowrates. The determinations of flowrates is concerned with deliverability calculations, whilst the determination of pressure is concerned with monitoring or diagnostic applications. The main purpose of a gas lift design is to optimise the deliverability of the well for a given gas injection rate under the well conditions and profile provided. To do this, a model must be built and then diagnostic analysis must be performed on it to optimise the design. Page 46 and 47 of the WellFlo manual provide a section on deliverability applications. The gas lift engineer must be concerned about how the spacing of each individual valve station and sizing of the unloading and orifice valve ports will effect the deliverability of the well under the envelope of conditions supplied by the customer. Justification for the inclusion of an additional station will hinge on what it will bring to the overall design in the future. Diagnostic analysis will provide on how a gas lift design will operate under the expected conditions.

2.0 The Gas Lift Data Sheet The first step in any design is that of data gathering for the work ahead. The client should complete a data sheet for the well that he would like examined. The data sheet gives the necessary details required to perform a gas lift design at a specific time, such as: ♦ ♦ ♦ ♦ ♦ ♦ ♦

Well Data Fluid Parameters PVT Data Inflow Parameters Gas Lift Data Production Information Test Data

The information is laid out to aid in the Gas Lift Design process as followed in the WellFlo program. It is important to gather as much of the information laid out on the data sheet as possible. Comprehensive Information will produce a more efficient Gas lift design.

A copy of the electronic version of the data sheet is available as part of the program suite accompanying this document or can be obtained e-mailing John Donachie at [email protected] and requesting it.

3.0 Information Entry into WellFlo Refer to the WellFlo manual on how to install and set-up WellFlo on your system. WellFlo will not operate without a relevant dongle. WellFlo is a windows based program. To start the WellFlo program, double click on the WellFlo icon. The program will automatically open the previously saved file if it is available. Alternatively, it will open a new file. For the purposes of this guideline we will assume that a new file is opened or has been created by using the New File option from the File pull down menu located on the Main Menu bar along the top of the screen.

Fig 3.0.1

Fig 3.0.2

Fig 3.0.3

3.1 Beginning a New Design Create a New file and Save As to save your new file with an appropriate name and in the appropriate folder. The CAMCO Aberdeen network has a logical folder system provided for storing gas lift designs: < H:\ TECH \ GASLIFT \ WELLFILE\ Customer \ Asset \ Field \ Well \ Design File (Revision) > (eg.

This filing system allows for easy data retrieval.

Fig 3.1.1

Fig 3.1.2

Fig 3.1.3

to save your file regularly. PC’s do crash and information is lost.

3.2 Data Preparation 3.2.1 General Data Begin by inputting general data regarding the well requiring a gas lift design. Include information specific to the design such as the Well, Field, Designer, Date, etc. Track design information by inputting historic data in the section provided within the window. This can also include additional information that the client has supplied.

Fig 3.2.1.1

Fig 3.2.1.2

3.2.2 Well and Flow Type

Enter the appropriate information, e.g. Annular or Tubing flow, and Production well or pipeline.

Fig 3.2.2.1

Fig 3.2.2.2

3.2.3 Xmas Tree Data

Double click on the Christmas Tree Icon (located in the graphical representation of the well) and enter the Well Head depths with respect to Mean Sea Level (MSL) and Rotary Kelly Bushing (RKB). Enter the additional information regarding Heat Transfer Coefficients if available, otherwise leave values as default. The Upstream Temperature is defaulted to 60oF, this should remain so unless otherwise stated.

Fig 3.2.3.1 Fig 3.2.3.2

3.2.4 Well Deviation Data Input deviation data from the deviation survey supplied by the client. When entering this use common sense. For example, enter steps in 500 -1000ft or 10-degree increments and increase the number of steps during a build-up. Do not enter the whole deviation survey as this is time consuming and adds little value to the design. Each row of deviation information will add to the calculations that WellFlo performs as part of the nodal analysis, slowing it down.

Fig 3.2.4.1

Fig 3.2.4.2

3.2.5 Well Equipment Data Input the tubing and casing information from the well head to the mid perforation depths. It is advisable to prepare a simple completion schematic in order to visualise the well. Input tubing information first, then casing information. Casing information may already be encomed by the Casing ID column. Check the Tubing ID against regional standards as the default values for the input weight may vary accordingly. Columns have been provided for Roughness and Temperature, these values will remain as default unless otherwise stated (e.g. Roughness changes due to special internal tubing coating)

Fig 3.2.5.1

Fig 3.2.5.2

3.2.6 Surface Equipment Data Input data regarding flow lines and chokes. If this information is not available, the well will develop the model to the tubing head/Christmas tree

Fig 3.2.6.1

Fig 3.2.6.2

3.2.7 Gas Lift Data

Enter the appropriate information. When considering the maximum injection depth take into permanent gauge mandrels, chemical injection valves and PBR’s situated above the packer. The must state the valve differential pressure within this window (the differential between the operating casing pressure and tubing pressure is recommended is to be in the range 150-200psi). A differential must exist or injection gas will not flow though the valve into the production tubing and therefore lighten the specific gravity of the fluid column. The must also specify either Qgi (Gas Injection Rate) or GLRi (Injection Gas/Liquid Ratio). For the purpose of Gas lift Design, select Qgi.

Fig 3.2.7.1

Fig 3.2.7.2

3.2.8 Reservoir Control Select the correct Fluid Type (Normally Black Oil); Set the Entry Model to Manual; Select a Vertical Well Orientation.

Fig 3.2.8.1 Fig 3.2.8.2 Once the above has been selected, within the same window enter the Edit Layer window. Input values for the mid perforation layer (reservoir) pressure and layer temperature, mid perforation depth and the productivity index.

Fig 3.2.8.3

3.2.9 Select the Inflow Performance Relationship (IPR).

Within the same window, enter Choose IPR and select the inflow performance relationship to be used during the gas lift calculations. Use the Straight-line relationship when the reservoir pressure is above the Bubble Point and the Vogel Relationship when the reservoir is below the bubble point, when interstitial gas has evolved. Once all the above parameters are selected press ‘OK’. If Vogel has been selected, it will automatically default to Straight-Line Relationship if the reservoir pressure is above the bubble point.

Fig 3.2.9.1

3.2.10 Fluid Parameters Enter the Oil Specific Gravity or the Oil API Gravity and either the Water Specific Gravity or the Water Salinity. Each pair of values are directly equivalent, only enter one value from each pair, as the other will change automatically (WellFlo highlights this relationship in Blue.)

Fig 3.2.10.1 Once the data preparation has been completed and the gas lift system is being design for specific well conditions, this window allows the Water Cut Per Cent and Production GOR to be modified to suit. Select the highest water cut to design to the worst case.

4.0 Match PVT Correlations The next step in the process is to match the PVT correlations to the measured data. Tuning the PVT model is very important. The influence of fluid properties can have a dramatic influence on pressure drop, particularly the gas/oil ratio. Each correlations is separated into three parts: • • •

Bubble Point (Pb) GOR (Rs) Oil Formation Volume Factor (Bo)

Six correlations are available for the above parameters. In order to select the right one, the must assess how each correlation will affect each parameter. This is achieved by choosing a correlation and selecting Match to enter the Match Oil Properties window. Ensure that the GOR value is correct before performing any calculations in this window. Select match property for all three parameters and enter PVT data at the appropriate temperatures and pressures and press calculate. Note down the match fraction and develop a table to see how each correlation matches up to all three parameters. See table below: Correlation Glaso Lasater Standing Vazquez & Beggs PetroskyFarshad Macary Fig 4.0.1

Pb 0.844 0.899 0.922 0.845

Rs 0.817 0.878 0.907 0.819

Bo 0.983 0.996 0.997 0.972

0.858

0.842

0.988

0.890

0.831

0.923

Once this table is completed, select the correlation closest to 1.0 (1.0 being a perfect match) for the available parameters. Now the must select this correlation on the fluid parameters and then go to match. Each Parameter is again calculated, however this time directly after calculating the match select Best Fit to tune the correlation to give a perfect match. Please note that you cannot tune both Pb and Rs, as this would cause an inconsistency, both reflect how the solution gas in the oil changes with pressure. In most cases, Tune Pb since data on Rs is usually not extensive, however if information is available then this may be the parameter to tune. Repeat the process separately for the oil and gas viscosity, if information is available.

With all of the above steps completed, a model of the well has now been constructed with the addition of the input data supplied by the operator. Note at this stage that the accuracy of any model and the predictions it generates would only be as good as the input data. It is therefore essential to use accurate PVT data and reservoir information.

5.0 Well Analysis 5.1 Lift Correlations Before carrying out any well performance analysis, the flowing gradient correlation to be used must be decided upon, for both the horizontal and vertical sections of the well bore. Based on the field experience of both CAMCO and EPS, the Hagedorn & Brown (mod) and Beggs & Brill (mod) correlation’s are generally recommended for the vertical and horizontal sections irrespective of the well bore. The other pressure drop correlations and the factors involved are discussed in depth in Appendix 2. In the absence of better information to the contrary, these are considered to offer reasonably accurate predictions of pressure drop for the purpose of the design. Where possible the accuracy of the chosen correlation and the model should be checked against measured data. This can be in the form of flowing gradient surveys or multi-rate tests.

This option allows the designer to import data (flowing gradient surveys or multi-rate tests) for plotting any pressure vs. depth or pressure vs. flowrate graph. Through this, the designer can compare the WellFlo generated values graphically with any measured data available for a more accurate VLP for the design. The Load Measured Data menu option expands to give the options: Depth (MD) vs. Pressure (and Temp)... Depth (TVD) vs. Pressure (and Temp)... Flow rate vs. Pressure… The file extension for Depth (MD & TVD) vs. Pressure (and Temp)… is .DVP, and Flow rate vs. Pressure… is .RVP. The data sets must be ASCII files prepared in a text editor

(Windows NotePad) a maximum of 128 lines of data can be loaded. Title/Legend: The enters a text title on the first line of the file; this will appear as a legend when the data are plotted. If no title line is entered, the default legend will be Measured Data. Comment Lines: Put a hash (#) at the start of a comment line to be certain the program will skip it. Do not put a hash in front of the title line unless you want it to be ignored. Once the measured data is corrected and saved in one of the above three formats using Notepad and loaded onto WellFlo the next step is to calculate the VLP using various correlations and compare to the measured data. Enter the Deepest Injection Operating Point window.

The designer must enter the Sensitivities window for various correlations and then press Calculate. WellFlo now runs through a calculation sequence, once this is completed press OK. Now the results can be viewed through the Plot Results (Select Results, then select Plot thought the Deepest Injection Point window). Once in this window tick the box for Measured Data and select the correlations form Sensitivity. Press Plot again to see which correlation best fits the measured. Input this information into the Correlations section of the Deepest Injection Operating Point window. Lfactors can be used to further calibrate the pressure drop computations in the well to exactly match the measured data through this window. Three factors can be adjusted: • • •

Well & riser flow correlation; used for design Pipeline flow correlation; used in components beyond the well head i.e. flow lines Downcomer flow correlation; used for risers with downflow

Adjusting the L < 1 the computed pressure drops will be reduced, for L > 1 they will be increased. By changing this factor and recalculating, an accurate Vertical Lift Performance(VLP) curve can be selected for the design. Having such data will contribute significantly to the efficiency of a gas lift design. If such tests are not available for the current well design, it is possible to use the correlations selected for another well in the same field. Using such information will give the designer an indication of what correlation is best suited for the field, and can be used for subsequent designs. Do not select a different fields correlations should (Correlations they are known to vary from field to field). Care is needed when manipulating ‘L’ factors, simple changes can have drastic effects on the overall gas lift design. It is better to use established correlations that ‘best fit’ the purposes of the design than to tailor to meet the requirements of the design exactly.

5.2 Optimum Gas Injection Rate The next stage in the design procedure is to calculate the pressure traverse with no gas lift. The must ensure that gas lift injection rate is equal to zero by going to Gas Lift Design window and check that the value for gas injection is zero. The next step involves Analysis → Deepest Injection Point → Operating Point, the must enter calculation nodes i.e. top node = Christmas tree/separator pressure, bottom node = layer and the solution node = casing. If temperatures are known select the calibrated option and enter the appropriate figures. Select liquid rate ranges for the calculation, ensure Iterate to exact operating point is on, and press Calculate. The program then calculates a liquid rate for the well. For viewing calculations in graph format, go to Results and then Plot for various graph formats. This gives the wells vertical lift performance and temperature with no gas injection.

Select Iterate to exact operating point to enable regression to refine the intersection point of the inflow and outflow curves. Once this point is found, WellFlo generates a pressure versus depth curve for the exact flow rate. Without this option selected, only a estimate of the inflow/outflow intersection point is made. It is possible to plot a graph with a range of gas linjection rates. This enables a designer to have a visual representation of how the gas injection rate is affecting the production performance. Using Sensitivities the can calculate various performance curves for this well with regard to increasing water cut, decreasing PI and reservoir pressure etc. When sensitivities are selected the must state the selection for up to 2 variables to see

how changing these factors would affect the well performance. The first task for the design engineer is to determine the optimum gas injection rate by comparing varying gas injection rates with sensitivity to water cut, PI and reservoir pressure. The optimum gas lift rate is at the point where the performance curve changes from a positive slope to zero or a negative slope. An example of these curves with respect to water cut, PI and reservoir pressure sensitivities is presented in Appendix 3 for better understanding. However, most companies operate with a threshold that is calculated by the slope of stb/d/mmscf. This threshold is in place for economic reasons, e.g. for one operator an increase of 1mmscf injection can only be justified if an extra 50-100 stbo/D is produced. Therefore, for each step of 1mmscf a prediction in the increase of oil produced is made to justify the subsequent step up in injection rate. Now the designer has selected the optimum injection rate. If the gas injection rate increases to a level where the gradient becomes negative, the gas injected is making the well inefficient by injecting too much gas that takes up the well fluid volume.

Insert this value into Gas Lift Data window (through the Data Preparation pull down menu) and will be the injection rate throughout the design process. If the client gives a kick off pressure, add this to the model later. The kick off pressure, a higher pressure that is achieved by ing a smaller amount of gas, allows the first mandrel depth to be located further down in depth. For a conservative gas lift design, even if a kick-off pressure is theoretically available, it would be sensible not to use this kick off pressure and design to the maximum allowable casing pressure that is available on a day-to-day basis. Had the designer stipulated a kick-off pressure for the design and this pressure could never be achieved the gas lift design would fail at the first unloading valve rendering the design unusable.

Print off an Inflow/Outflow plot for each water cut percentage.

Reservoir Factors Having built the well model the analysis can begin. The most difficult part of a gas lift design is deciding what factors are likely to vary over the life of the well and thus require a more detailed investigation. Here, CAMCO’s considerable Gas Lift experience helps identify the areas that will need consideration.

The gas lift designer must consider: •

•

Maximise Oil Production The objective of a good gas lift design is to maximise oil production over the life of the well, achieved by ensuring a mandrel spacing that will always allow the deepest injection point to be accessed whatever the well conditions. As stated previously, this may entail running initially redundant mandrels to allow transfer deeper at a later time or visa versa. Minimise Well Intervention

By careful design considerations, mandrel spacing and the correct port size selection it is possible to design gas lift systems that will automatically transfer the depth of injection to match changes in well conditions. However, there are drawbacks to this design technique and it is generally reserved for subsea wells where minimum wireline intervention is intended. •

•

•

Maximise Depth of Injection Gas injection is designed to be as deep as mechanically possible by firstly taking into the well’s completion components (e.g. annular packer, gauge mandrels, PBR’s, chemical injection valves, etc.) on this basis the maximum injection depth is decided. Secondly, the designer must bear in mind the valve differential pressure. This value is the theoretical minimum pressure drop through the gas lift valve to allow gas injection. The valve differential is normally set to 100 - 200 psi per the standard gas lift valve design practice. The final, and most obvious, constraints are the casing and tubing pressure ratings; if a casing pressure was to exceed the casing or tubing pressure rating this could lead to either burst or collapse failure of the tubulars. Stability The designer must ensure the well is stable by inspection of the inflow/outflow curves for each design. If the inflow (straight line/curved) curve is intersected by the outflow curve with a positive gradient then the well is stable. If the inflow curve is dissected by an outflow curve with a gradient equal to zero or negative the system is said to be ‘Unstable’. Examples of stable and unstable flow are provided in Appendix 4 for closer analysis. Uncertainties in Reservoir Performance Over the life of the well, reservoir performance is probably the hardest item to predict with any certainty. It is therefore essential that any well analysis covers the likely range in pressures and PI’s to be encountered in the producing life of the well. These will have a big impact on the production and thus the flowing gradient of the well. All of these factors will influence the mandrel spacing and maximum depth of injection. Worst Case Design With the above considerations in mind for the particular well the next step is to initially design the well at the worst case conditions for gas lift. The worst case conditions for gas lift design are: • High reservoir pressure • High productivity index • High water cut percentage With these factors in mind, the will enter the values as described earlier and calculate a pressure traverse. The results are then plotted on a Pressure vs. Depth plot, the scales adjusted appropriately, and printed for a hard copy. See Appendix 5 for an example. The will now manually design the gas lift mandrels for this worst case condition. Positioning the mandrels for the worst case design allows for well unloading and for the well to be gas lifted albeit not necessarily optimised as well conditions change. The unloading procedure for continuous flow gas lift is presented in Appendix 6 for further reading.

Manual Mandrel Design: Surface Close Design Method (This process is automated in WellFlo)

1. Draw static gradient (kill fluid) either from a TVD of 0 to 1000 ft or 0 psi to 450/460 psi; the kill fluids are commonly of a gradient 0.45 or 0.46 psi/ft, see data sheet provided by client. 2. With this gradient, using set squares, draw a line from the wellhead/separator pressure to the maximum allowable casing pressure, where this line intersects the casing pressure draw a horizontal line through the vertical lift performance curve to the depth on the Y axis. The first mandrel is now positioned at this depth. 3. For positioning the second mandrel the designer must firstly decide upon a transfer pressure margin to ensure the second valve will be uncovered if the well’s vertical lift curve over performs. For conservative designs this is 10% of the pressure differential between the tubing pressure and the casing pressure at the depth of mandrel one. For known reservoirs, where vertical lift performance (VLP) is known not to over-perform, the safety margin added should be in the range of 50 - 100 psi. Secondly, the designer must draw in a casing pressure drop of 15-psig safety margin to ensure valve number one closes. Starting at mandrel #1 draw the reduced casing pressure using a parallel line to the maximum allowable casing gradient. 4. Space mandrel #2 using the static gradient starting at the transfer point of 10% of the tubing performance curve for an unknown reservoir, or 50 – 100 psi for known fields with existing working designs, until the line intersects the maximum casing pressure with a drop of 15 psi. Where this line intersects the casing pressure draw a horizontal line through the vertical lift performance curve to the depth on the Y axis. The second mandrel is now positioned at this depth with safety margins in place. 5. The above steps are repeated until there is not enough differential pressure between the vertical lift performance curve and casing pressure at mandrel depth. 6. If the designer is given a kick-off pressure and the client stipulates this is to be used in the design, only one factor is changed. Instead of the depth of the first mandrel being decided upon by the maximum allowable casing pressure the greater kick-off pressure is used. Therefore, this increased pressure allows mandrel one to be placed deeper in the well. The kick-off pressure is only used for the first mandrel depth. After that, the maximum allowable casing pressure is used because as soon as injection gas is vented into the tubing, through the first mandrel, this higher pressure can longer be attained. Once the above is complete enter the mandrel depths into Gas Lift Data window in the Data Preparation pull down menu and go to Analysis and Operating Point to calculate the liquid flowrate. Check that the gas lift design will operate through the Gas Lift Design window in the Analysis pull down menu. Enter the calculated flow rate and check the appropriate values for static gradient etc…

Within the Gas Lift Margins window, check the kick-off pressure (if given) and the casing closing pressure margin is equal to 15 psi. Now Recalculate and WellFlo will let the know if the design will operate and unload for those particular conditions and safety factors incorporated. If the design does not unload the must adjust the mandrel spacing and hence the ideal operating point may not be reached. Once adjusted and fully functional the must justify the placement of each valve, in essence calculate what lift the specific mandrel will provide. If the lower mandrel of the initial design does go to a deeper depth but the oil increment gained by this mandrel is low then this valve is not incorporated into the design.

The next step of the design process is to consult with the client to determine how the three parameters (Reservoir Pressure, Productivity Index and Water Cut) will vary with time. The worst case design, as detailed above, may only be applicable in the infancy of the well and as the reservoir characteristics change over time the designer must incorporate this into his design. The original design will not inject deep into the well and as the well changes over time, this injection depth will not be optimal. The designer must consider future circumstances to provide the client with optimum lift later in the wells life. If the PI reduces over time, with reservoir pressure and water cut staying constant, this would

increase the drawdown at the perforations, and hence move the VLP curve to the left allowing greater valve differential pressures. More and deeper mandrel depths are required for this change in reservoir conditions. If the PI and water cut remain constant but the Reservoir Pressure (Pr) reduces, there will be no effect on the drawdown, but a lower flowing bottom hole pressure (Pwf) would be seen. This would shift the VLP curve to the left again where deeper mandrels could be placed. If the well’s water break-through did not occur as expected and water-cut stayed relatively low (at say, 10%) this would result in a lighter fluid column than predicted, as water is heavier than oil. Consequently, the VLP would remain stationary for longer making an extra mandrel at a deeper depth more attractive. The above examples are describing the effect of one parameter changing while the others remain constant. This rarely occurs in reservoirs. For most depleting reservoirs the Pr will reduce with an increasing water-cut over time. Also, PI can increase (shift VLP to right) in a well by fracturing or alternatively decrease (shift VLP to left) by the reservoir’s permeability reducing over time. Bearing these factors in mind, the designer needs to decide whether gas lift is required and if it is economical, with respect to the client’s threshold, by way of performance curves and sensitivities.

All of the expected gas lift scenarios generate a host of different designs. The mandrels are then overlain and spacing selected by inspection to give the optimum design to cover the expected range discussed by the designer and client. Mandrel depths are then inputted into WellFlo and the design is checked to ensure that it will work for each of the expected conditions. At this stage, safety factors are built into the mandrel spacing. This is an area where the experience of the gas lift designer is essential.

Valve Port Sizing

Once the mandrel spacing has been calculated, the valve type and port size must be decided. The must calculate the gas required for the unloading production rate to achieve the necessary transfer point that will uncover the valve below. For each mandrel being examined the designer must change the mandrels below the one being inspected to ‘Inactive’ in the Gas Lift Data window (found in the Data Preparation pull down menu). Once this has been done, the well performance is calculated using injection rate sensitivities. When the vertical lift performance curves exceed the necessary transfer pressure, adequate gas injection uncovers the valve below. Insert this injection rate into the model to calculate the Fluid Unloading Rate and Temperature. Calculations are then performed using the Thornhill-Craver method to size the appropriate orifice to the required injection gas rate through the valve. that unloading valves and Orifice valves have different coefficients of discharge and this must be specified in the gas age program. As a rule of thumb, the coefficient discharge for unloading valves is 0.76 and for orifice valves 0.86.

Once the designer has opened the program, he must input the following: ♦ ♦ ♦ ♦ ♦

Gas injection rate required Upstream pressure at depth Downstream pressure at depth Temperature of valve at depth Coefficient of discharge for the valve

Once the above is entered the must press Calculate; the program will then give a port size. The appropriate port sizing is selected with the CAMCO product guide. To double-check, port sizing is entered into the program and the gas injection rate is left blank, the then presses ‘Calculate’ to ensure the appropriate amount of gas will through the port size. Repeat the above procedure for each valve. Please note that changes in reservoir characteristics (vertical lift performance) alter the amount of gas injection required for deeper valves when unloading; therefore, the gas lift age calculations are carried for the worst well performance (deepest injection) to ensure the design will operate at these conditions. All of the valve port sizing is then checked to ensure they will operate throughout the predicted well performance conditions. In short a cross-section of well conditions are taken to represent the range of well parameters in which these valves must perform. The bottom valve, the orifice, must the optimum injection rate decided on earlier.

Dome Charges In order to calculate the Dome charges for all of the unloading valves in the workshop, a temperature correction must be implemented. This is aided by a program called NITPROP, where the inputs the valve depth and temperature and then the workshop temperature of 60oF is entered with a pressure of 0psig, NITPROP then calculates the temperature coefficient. This temperature coefficient is noted for each unloading valve.

Predicting accurate temperature profiles in flowing and unloading wells can greatly improve the artificial lift design, enabling the designer to set the valve’s dome pressure more accurately and thereby improving the predictability of valve throughput. The temperatures of the valves at depth using WellFlo are calculated on the basis the well is flowing. However, the reservoir is not performing when the well is initially unloading and the annulus is filled with relatively cold kill fluid making these temperatures somewhat inaccurate. This temperature difference has a significant effect on the operability of the Nitrogen charged bellow valve. Underestimating this valve in the design process might mean the valve would not open. Over estimating the

valve will remain open and interfere with lower valves. The accurate temperature during unloading therefore lies between the geothermal gradient of the crust of the earth and the well flowing temperatures. CAMCO uses the Calibrated option for temperature traverse prediction, the experience gained from CAMCO gas lift engineers has shown that this is the most reliable. This method requires a flowing tubing head temperature and separator temperature for a given flowrate. This is the minimum information required for an accurate design with respect to temperature. With this information collated from various wells on a specific field, CAMCO gas lift engineers are building databases to give a historical background on how temperature relationships vary well to well to improve future designs. If accurate temperature data has been provided by the client, by way of a flowing gradient survey, the temperature traverse created by WellFlo can be corrected by manually calculating a heat transfer coefficient in segments from the reservoir to separator. If there is a significant temperature drop through the seawater effecting the temperature traverse below the seabed, the designer has two options. Firstly, an extrapolated flowing wellhead temperature can be calculated or secondly, the seabed can be modelled as the well head, with the fluids from the seabed to the platform modelled as a flow line.

Calculation Worksheet When sections 5 and 5.1 are completed, fill in the calculation worksheet found in the following directory in CAMCO’s network:

This worksheet allows the designer to input all valve depths, temperatures, port sizes, tubing pressures, casing pressures and temperature coefficients. The worksheet is set-up with calculations to calculate dome charges and so on, and this is the final part of the design.

The next stage is to show it to a colleague for a quality check before discussing with the client in more detail.

Conclusion To summarise, continuous gas lift is a very complex system composed of many interactive components. In deg or optimising a system, all components have to be considered simultaneously. By changing a component from either the inflow subsystem or the outflow subsystem, a drastic effect can be observed on the resulting flow and gas consumption. The total amount of computation required for sensitivity analysis is extensive. Therefore, having the assistance of an easy to use computer program, such as WellFlo, is indispensable in generating a good gas lift design.

Using the Word Macro for Reporting It is possible to modify the format in which the WELLFLO.RPT will portray its information. A Macro for Microsoft Word is available that will partially convert the Automatically generated report from a Word Pad document to a Microsoft Word document. Instructions are available on how to set up the WellFlo Report Macro in the EPS manual provided with the WellFlo program. The Macro EPS provides (Located in WellFlo.txt) is incorrect and should not be used, replace the macro with the CAMCO Macro and follow the same procedure. This is available as part of the suite of programs provided with this document or alternatively it can be sent via e-mail by ing a CAMCO representative. ( [email protected])

Flosystem Product Electronic Help System On-Line Help is available in Flow Sytem either by: • Using the Help option on the main menu bar. • Hitting the F1 key when a menu item is highlighted, or when you are in a dialogue box. • Shift+F1 gives you the option to get help on graphic screen regions by clicking the area of interest (menu bar, toolbar, plot legend, axes, etc.) A large question mark appears beside the mouse pointer when this is active. The Manual The next place to try is the FieldFlo manual. This has provided an index and figures, many of which show specific screens or windows to help point you to the right place. You might find an explanation of the problem here. In any case, please look here first, before calling EPS. Internet At http://www.epsedin.com you will find the EPS web-site. development, providing a base for their clients.

This is in continual

Internal to use the internal provided by your Schlumberger colleagues. External As a last resort it is possible to EPS or a local client directly to get help with problems with the software. The below e-mail addresses should only be used if all of the above options have been exhausted: UK: [email protected] Americas Office: [email protected] Far East Office: [email protected] Middle East : [email protected]