METHOD FOR DETERMINING INJECTOR BOTTOM HOLE PRESSURE IN A HYDROCARBON RECOVERY OPERATION USING LINER-DEPLOYED OUTFLOW CONTROL DEVICES

Abstract

Methods for determining hydrocarbon recovery injection well bottom hole pressures where a downhole liner has liner-deployed outflow control devices, for attempting to ensure operation below a maximum operating pressure. Steam flowrate is varied using a range of valve positions, including shut-in periods (0-5% flowrate) sufficient to approach equilibrium between a surface pressure and bottom hole pressure, generating a determined bottom hole pressure which can be used to calculate an average bottom hole pressure. An average steam flowrate and average surface pressure are determined for a range of valve positions. A pressure drop is calculated for each of the valve positions by subtracting the average bottom hole pressure from the measured surface pressure. Plotting the pressure drop against the average surface pressure for each of the valve positions allows generation of a best fit model defining a polynomial relationship between the pressure drop and the surface pressure. Parameters can then be derived from the best fit model, enabling prediction of the bottom hole pressure for any steam flowrate for the well. The maximum operating pressure can then be more confidently determined for various steam flowrate values. The methods may be automated using a programmable logic controller (PLC) or similar platform located at the well pad.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS(S)

The present disclosure claims priority from U.S. Provisional Patent Appl. No. 63/648,535, filed on May 16, 2024, entitled “METHOD FOR DETERMINING INJECTOR BOTTOM HOLE PRESSURE IN A HYDROCARBON RECOVERY OPERATION USING LINER-DEPLOYED OUTFLOW CONTROL DEVICES,” herein incorporated by reference in its entirety.

FIELD

The present disclosure relates to thermal recovery methods for heavy hydrocarbon reservoirs, and more particularly to determining an optimized steam injection flowrate.

BACKGROUND

It is known in the art of hydrocarbon recovery to employ certain enhanced oil recovery techniques for a hydrocarbon resource that is not recoverable in whole or in part by conventional primary recovery methods. For example, heavy hydrocarbons such as bitumen under normal reservoir conditions do not flow to a well for production to surface, and often require heating which can be achieved in many cases by injection of steam (with or without additives such as solvents).

Cyclic steam stimulation (CSS) and steam-assisted gravity drainage (SAGD) are common types of thermal recovery methods applied to heavy hydrocarbon resources. In the case of CSS, steam is injected downhole through a well, the well is shut-in for some period of time (a “soak” period) to allow the steam to heat the resource and mobilize it, and the now-mobilized hydrocarbon is then pumped to surface through the same well. In SAGD there are typically two wells—an injector well and a producer well—which are drilled down to the reservoir and then commonly extend generally horizontally through the target reservoir area, the injector well drilled near but above the producer well. Steam is injected through the injector well into the reservoir, mobilizing the hydrocarbon resource, and the mobilized hydrocarbon then flows downwardly toward the underlying producer well by gravity where the mobilized hydrocarbon is lifted to surface, using artificial lifting technologies such as electrical submersible pumps (ESPs).

One of the main operational limitations when injecting steam into a reservoir is the maximum operating pressure (MOP) for the reservoir, which cannot be exceeded under applicable industry regulations to avoid cap rock failure and uncontrolled release of reservoir fluids. A higher steam flowrate leads to increased pressure in the reservoir over time; if the pressure approaches MOP, the steam flowrate must be limited (therefore a reduced steam injection with impact on hydrocarbon production). Shallow reservoirs have a lower MOP given their proximity to the surface, and this severely limits the ability to inject steam since such recovery operations are necessarily operated very close to MOP. For perspective, the MOP in a typical heavy hydrocarbon reservoir may be around 9000 kPa, while in shallow reservoirs it may be around 1700 kPa. Although the risk of exceeding MOP may be higher for shallow reservoirs, this risk also exists and is important in deeper wells-every well has an MOP that must not be exceeded, with a potential negative impact on steam injection and hydrocarbon recovery. Note that this is during the SAGD phase of the well's life, usually during the circulation and ramp up phase, and there is a regulatory exception to temporarily increase the MOP to allow for well startup.

Outflow control devices (OCDs) are known for use in SAGD injector wells to control injection of steam into the reservoir and they provide well-known advantages. Two known OCD deployment systems for use in SAGD operations are the liner-deployed outflow control device (LDOCD) in which the OCDs are deployed in the liner (part of the well casing), and the tubing-deployed outflow control device (TDOCD) where a tubing string is run inside the casing/liner and the tubing houses the OCDs. In a prior art system using TDOCD where tubing is used to deliver the steam, there is an annular space between the tubing and the liner with pressure equalized between the annular space and the reservoir (see FIG. 1a); this allows for more accurate injector bottom hole pressure (IBHP) estimation, and the operator can more confidently inject steam to the maximum operating pressure (MOP) of the reservoir. LDOCD systems, however, do not have an annular space (there is no tubing, the steam being injected directly into the liner; see FIG. 1b), so the reservoir pressure is often necessarily overestimated for safety reasons and lack of direct estimation or measurement. The steam injection allowance is thus underestimated, reducing potential hydrocarbon production, leading to an inefficient resource utilization.

The surface measurement of pressure is commonly but not exclusively the wellhead pressure (WHP), which measurement by pressure gauges at the wellhead indicates steam pressure but does not take into account the significant downhole pressure losses along the liner including the OCDs. For TDOCD, a surface blanket gas measurement is equivalent to the pressure in the annular space which is in equilibrium with the injector bottomhole pressure (IBHP), and so the operator can more confidently determine whether the MOP is being reached. For LDOCD, there is no annular space, and therefore the blanket gas measurement and the WHP measurement are equal to the steam pressure; the steam pressure is always equal to or greater than the IBHP. Since the only instrumentation that measures pressure is for WHP or blanket gas pressure, it is always an overestimate of IBHP, and it is a significant overestimate where a correction is essential (especially in shallow reservoirs, where minimal steam can be injected, since even a small steam flowrate will show a WHP>>MOP, and the WHP is larger than the IBHP). The flow of steam through the OCDs causes a pressure drop which must be overcome to allow steam introduction to the reservoir, which biases the IBHP determination upward; any steam flowrate above zero will reduce the operating margin with respect to the MOP, under prior art methods. Any steam flow (or fluid flow in general) will create a pressure drop as it flows through a pipe and/or a restriction, such as the OCDs. Steam flow through the injector well causes a pressure drop due to frictional losses. The WHP incorporates the pressure drop (dP) such that the WHP is less than the MOP. This means that the steam flowrate would be limited in order to reduce the pressure drop (dP) to ensure the WHP<MOP. In general, for well productivity it is desirable to operate at higher pressures because it allows the reservoir to be at a higher temperature and enhance heavy hydrocarbon production. It is also desirable to maintain the reservoir pressure, so higher steam flowrates are needed. Therefore, a correction that accounts for the pressure drop and determines the IBHP would allow one to increase steam injection while still being under the MOP.

Notwithstanding the foregoing, the use of LDOCDs would have numerous benefits over the use of TDOCDs, including a substantial cost savings. Cost reduction is driven by lower material costs and drilling and completion operational costs (cement, labour, rig time, etc.). For example, in addition to savings on the tubing on the LDOCD, in a TDOCD system the tubing may be only 4.5″ in diameter—to provide the same hydraulics as a LDOCD, this might need to be increased to 7″, and the subsequent liner diameter increased even further. The smaller diameter leads to higher frictional losses, which consequently leads to a higher pressure drop across the lateral section of the well. Due to the higher pressure drop, it becomes infeasible at the given pressures to drill potentially longer wells. Therefore, LDOCDs allow for longer wells which are more capital/resource efficient. As part of the sustainability goals is the efficient use of resource reducing the footprint and production intensity (per barrel of bitumen produced). If we upsize a TDOCD to have comparable hydraulics to LDOCD, it would have a larger surface footprint while drilling, making the drilling process less sustainable and more difficult.

SAGD operations in reservoirs must operate below MOP, and while this is a factor in all reservoirs the MOP is even lower in shallow reservoirs. This is a regulatory requirement to avoid fracturing the cap rock and creating an uncontrolled release from the reservoir. The use of LDOCDs would present advantages in such SAGD operations but their use is hampered by the inability under prior art techniques to accurately determine or predict the IBHP to maintain operations below MOP. Being able to operate at higher reservoir pressures and steam flowrates when using LDOCDs would be enabled by a method of estimating the true reservoir pressure (the IBHP), allowing the operation to achieve higher production rates, steam efficiency, capital efficiency, and per well net present values (NPVs).

SUMMARY

According to a first broad aspect of the present disclosure, there is provided a method for determining bottom hole pressure for a steam injection well used in hydrocarbon recovery, the steam injection well configured to introduce steam to a subsurface reservoir containing hydrocarbon, wherein a downhole liner of the well is provided with at least one liner-deployed outflow control device for introducing the steam to the reservoir, the steam injection well further provided with (i) a valve at surface for controlling steam flowrate into the well, (ii) a meter at the surface for measuring the steam flowrate, and (iii) a gauge at the surface for measuring pressure upstream of the at least one liner-deployed outflow control device, the method comprising the steps of:

• • a. opening the valve to allow the steam to pass through the well and the at least one liner-deployed outflow control device into the subsurface reservoir;
  • b. selecting a shut-in period for at least one shut-in sufficient to reduce the pressure to approach the bottom hole pressure;
  • c. operating the valve to implement the at least one shut-in for the shut-in period;
  • d. during the shut-in period, allowing the pressure to approach the bottom hole pressure, resulting in a reduced pressure at the surface; and
  • e. measuring the reduced pressure using the gauge to arrive at a determined bottom hole pressure.

The actual bottom hole pressure can then be more confidently determined based on the determined bottom hole pressure.

In some exemplary embodiments, the pressure measured at the surface is wellhead pressure, although blanket gas pressure could also be measured at the surface. For liner-deployed outflow control devices, the wellhead pressure and the blanket gas pressure are approximately equal.

In some exemplary embodiments of the first broad aspect of the present disclosure, the shut-in period is sufficient to allow the pressure measured by the gauge to reach equilibrium with the bottom hole pressure.

The steam injection well is preferably but not necessarily an injector well of a steam-assisted gravity drainage well pair.

Some exemplary embodiments further comprise operating the valve at step c. at a variety of valve positions to implement a range of steam flowrates including the at least one shut-in, and measuring the pressure for each of the variety of valve positions. The valve positions are preferably but not necessarily maintained for at least 15 minutes. The variety of valve settings preferably ranges from 0% to a selected maximum steam flowrate.

The at least one shut-in is preferably achieved by operating the valve at step c. to between a full shut-in (0%) and a partial shut-in (5%) valve position. In a full shut-in there is minimal steam flowrate at or close to zero, resulting in a negligible pressure drop, while for a partial shut-in (5% or “trickle flow”) there is minimal steam flowrate with a very small pressure drop.

Exemplary methods may further comprise monitoring for changes in the determined bottom hole pressure.

Exemplary methods further comprise the step of, in response to a determined bottom hole pressure exceeding a maximum operating pressure for the reservoir, operating the valve to reduce the steam flowrate.

Step c. preferably comprises at least two shut-ins.

According to a second broad aspect of the present disclosure, there is provided a method for predicting bottom hole pressure for a steam injection well used in hydrocarbon recovery for a range of steam flowrate values, the steam injection well configured to introduce steam to a subsurface reservoir containing hydrocarbon, wherein a downhole liner of the well is provided with at least one liner-deployed outflow control device for introducing the steam to the reservoir, the steam injection well further provided with (i) a valve at surface for controlling steam flowrate into the well, (ii) a meter at the surface for measuring the steam flowrate, and (iii) a gauge at the surface for measuring pressure upstream of the at least one liner-deployed outflow control device, the method comprising the steps of:

• • a. opening the valve to allow the steam to pass through the well and the at least one liner-deployed outflow control device into the subsurface reservoir;
  • b. measuring the steam flowrate using the meter and measuring the pressure using the gauge;
  • c. selecting a shut-in period for at least one shut-in sufficient to reduce the pressure measured by the gauge to approach the bottom hole pressure, resulting in a reduced pressure at the surface, and measuring the reduced pressure to provide a determined bottom hole pressure for the at least one shut-in;
  • d. using the valve, varying the steam flowrate between a plurality of selected values falling within the range of steam flowrate values, including the at least one shut-in;
  • e. measuring the steam flowrate using the meter and measuring the pressure using the gauge for each of the plurality of selected values falling within the range of steam flowrate values;
  • f. determining an average steam flowrate and an average pressure for the measured steam flowrate and the measured pressure for each of the plurality of selected values;
  • g. calculating a pressure drop for each of the plurality of selected values by subtracting an average determined bottom hole pressure for the at least one shut-in from the measured pressure for each of the plurality of selected values;
  • h. plotting the pressure drop against the average steam flowrate for each of the selected values as points, and generating a best fit model for the points defining a polynomial relationship between the pressure drop and the steam flowrate;
  • i. deriving parameters from the best fit model; and
  • j. using the parameters to predict the bottom hole pressure for any steam flowrate along the range of steam flowrate values.

The bottom hole pressure can then be more confidently determined based on the determined bottom hole pressure for various steam flowrate values within the range. This method can be used for a plurality of steam injection wells, where each well would have a unique best fit model and resultant parameters.

In some exemplary embodiments, the pressure measured at the surface is wellhead pressure, although blanket gas pressure could also be measured at the surface. For liner-deployed outflow control devices, the wellhead pressure and the blanket gas pressure are approximately equal.

In some exemplary embodiments of the second broad aspect of the present disclosure, the shut-in period is sufficient to allow the pressure to reach equilibrium with the bottom hole pressure.

The at least one shut-in is achieved by operating the valve at step d. to between a full shut-in (0%) and a partial shut-in (5%) valve position.

Exemplary methods preferably further comprise the step of, in response to a determined bottom hole pressure exceeding a maximum operating pressure for the reservoir, operating the valve to reduce the steam flowrate.

Step d. preferably comprises at least two shut-ins. In some such cases, the determined bottom hole pressures for the at least two shut-ins are averaged to arrive at the average determined bottom hole pressure for the pressure drop calculation of step g.

In some exemplary embodiments the plurality of selected values ranges from 0% to a selected maximum steam flowrate.

Each of the steam flowrate values is preferably a valve position.

In some exemplary embodiments, step h. comprises the best fit model being a line of best fit defining a linear relationship between the pressure drop and the steam flowrate. In some such cases with a linear relationship, the parameters could be slope and intercept of the line of best fit.

Some exemplary embodiments comprise identifying a maximum steam flowrate by determining where the polynomial relationship breaks down. In some such embodiments, a maximum steam flowrate can also be determined as an upper boundary for the best fit model. To determine the maximum steam flowrate for a particular steam injection well, the steam flowrate is varied upwardly and the pressure measured until the steam flowrate increases without significant pressure increases indicating failure of the polynomial relationship, in what is called “choked flow”. By this method, the parameters and the maximum steam flowrate can be determined for a given well. In some exemplary embodiments where the relationship is linear, a method to detect the transition from linear domain to non-linear domain may be used to determine the maximum flowrate.

In some exemplary embodiments, the above method for predicting the bottom hole pressure for the steam injection well for a range of steam flowrate values can be automated using a programmable logic controller (PLC) or similar platform located at the well pad. In some embodiments, the PLC receives the measured steam flowrate from the meter and the measured pressure from the gauge, plots the points and generates the best fit model, and derives the parameters for the best fit model, and performs the required verifications and validations (pre- and post-processing of the data).

A detailed description of exemplary embodiments of the present disclosure is given in the following. It is to be understood, however, that the disclosure is not to be construed as being limited to these embodiments. The exemplary embodiments are directed to particular applications of the present disclosure, while it will be clear to those skilled in the art that the present disclosure has applicability beyond the exemplary embodiments set forth herein.

BRIEF DESCRIPTION OF THE DRAWINGS

In the accompanying drawings, which illustrate exemplary embodiments of the present disclosure:

FIG. 1a is a section view of a prior art TDOCD arrangement.

FIG. 1b is a section view of a prior art LDOCD arrangement.

FIG. 2 is a chart illustrating the pressure drop between WHP and IBHP.

FIG. 3 is an exemplary data collection plan.

FIG. 4 is a chart plotting steam flowrate and WHP over time for different valve positions.

FIG. 5a is a chart illustrating plotted points and lines of best fit for a plurality of wells.

FIG. 5b is a table of parameters (slope, intercept, and correlation values) for the plurality of wells plotted in FIG. 5a.

FIGS. 6a and 6b are charts illustrating polynomial (linear) relationships without and with choked flow.

FIG. 7 is a simplified flowchart illustrating an automated method for determining IBHP.

FIG. 8 shows qualitative representations of the three techniques (labeled 1, 2, 3). The best technique is technique 1 (labeled 1) based on the data collected, as technique 1 fits the data in the operational domain better.

FIG. 9 illustrates a steady state response which employs a step detection algorithm; the algorithm detects a step change above a threshold at time t_start, and holds both the steam flowrate (corresponding to the measured flowrate labeled “MV”) and the WHP (corresponding to the measured WHP labeled “CV”) constant for a time period after t_start (i.e., the holding time period between t_start and t_end); this holding time period can be configured to permit a steady state response following the detected step transition at time t_start.

Exemplary embodiments will now be described with reference to the accompanying drawings.

DETAILED DESCRIPTION OF EXEMPLARY EMBODIMENTS

Throughout the following description, specific details are set forth in order to provide a more thorough understanding to persons skilled in the art. However, well known elements may not have been shown or described in detail to avoid unnecessarily obscuring the disclosure. The following description of examples of the present disclosure is not intended to be exhaustive or to limit the disclosure to the precise form of any exemplary embodiment. Accordingly, the description and drawings are to be regarded in an illustrative, rather than a restrictive, sense.

The present disclosure is directed to methods for determining bottom hole pressure for a steam injection well used in hydrocarbon recovery operations, such as for example a SAGD operation, and in particular to such operations where the steam injection well does not house injection tubing but comprises liner-deployed outflow control devices for introducing steam to the reservoir.

The present disclosure uses LDOCDs and employs regular injection well shut-in periods (during which the steam valve position is preferably reduced to 0-5%, with little to no fluid flow through the injection well), which allows pressure to equalize or approach equalization between the wellhead and the reservoir after some period of time required for the blanket gas pressure (measured as WHP) to be in or approach equilibrium steady state with the reservoir pressure. When flow through the OCDs is nearly or fully stopped, since the pressure loss through the OCDs is proportional to flow, then the measured wellhead pressure will eventually become an accurate estimation of reservoir pressure after accounting for static pressure. This process is a first order process and requires time to achieve steady state, on the order of at least 15-45 minutes. This determined IBHP method is preferably also used in order to calibrate and validate a model that can be used for continuous monitoring/prediction. Input data may be continuously collected (not just during shut-in periods) as steam flowrate is varied, and adjusted for pressure loss by using the output of the model so there is a means to estimate the IBHP regularly and monitor for pressure changes during operations. By achieving a better IBHP estimate, LDOCDs can be employed instead of TDOCDs and steam injection can be optimized (allowing the operator to run at higher steam rates and reservoir pressures) and thus enhance well productivity. Using LDOCDs instead of TDOCDs, cost savings, higher steam qualities, longer wells, and other advantages may be achieved.

While full equalization of the WHP and IBHP is most desirable during the shut-in periods, it has been found that it can sometimes take 2.5 hours or more for the WHP to decay to the correct IBHP value. Due to risks in thermal cycling, the exemplary embodiments disclosed herein use 15-minute shut-ins to generate the IBHP ground truth value. This necessarily biases the ground truth values with approximately 50 kPa of positive error, resulting in the model overestimating IBHP by that amount but building in a conservative margin within the model.

In addition to a method for determining IBHP from a data collection process involving shut-in periods and pressure equilibrium, a further desired end is to be able to construct predictive models that may in some embodiments facilitate automated estimation of IBHP at various steam flowrates and thus the allowable IBHP (dependent on steam injection) to remain below MOP. A simpler construct is to have an automated estimation of the pressure drop (ΔP, between the well head and the reservoir), which models can then be used to estimate an IBHP for each well.

ΔP (the pressure drop across the whole liner including OCDs) between WHP and IBHP (essentially the difference between surface and reservoir pressures) is relatively large, as illustrated in FIG. 2, and so an accurate IBHP measurement or determination/prediction is required to help ensure maximum hydrocarbon recovery rate but while not exceeding the MOP-exceeding can overpressure the reservoir rock and overlying cap rock.

In one exemplary method, the valve is adjusted to allow injection of the steam downwardly through the well to the liner, and the steam flowrate and WHP are measured (using the meter and the gauge, respectively). A data collection plan is employed to run the operation through a plurality of valve settings including one or more 0-5% shut-in periods. FIG. 3 illustrates one exemplary data collection plan. The operator (or potentially the process is automated, as described below) uses the valve to systematically vary the steam flowrate according to the data collection plan. In the illustrated data collection plan, each step of the process is greater than 15 minutes. In the steps where steam valve position is reduced to 0-5%, the selected time period should be chosen to allow sufficient time for pressure equilibrium to be achieved between the wellhead and the injector bottom hole (reservoir), which has been found to be 15-45 minutes in some contexts (although in some cases, as noted above, full equalization may require a significantly longer period). The steam flowrate is systematically varied from a selected maximum value to zero and then back to the maximum, while only measuring steam flowrate at surface and WHP. The steam flowrate and WHP are measured continuously at surface for every well, resulting in a unique data set for each well.

In the exemplary data collection plan as shown in FIG. 3, the maximum steam flowrate is 50 t/hr, with a minimum measurable steam flowrate of 10 t/hr; these values are exemplary only, provided solely to illustrate embodiments of the present disclosure. There are two shut-in valve positions that are employed at the noted stages during data collection: 0% steam flowrate (full shut-in) and 5% steam flowrate (so-called “trickle flow” steam flowrate). The 5% steam flowrate may offer greater operational sustainability, as it mitigates the risk of pipe freezing and obviates the necessity for pipe drainage procedures, but the efficacy depends on several factors including line pressure and steam properties (depending on these variables, a 5% valve position may still permit a substantial steam flow, thereby preventing the WHP from equilibrating with the reservoir pressure) as would be clear to those skilled in the art.

Following are the exemplary embodiment plan details:

• • 1 Initial Condition Setup:
    • Begin with the steam rate at its maximum steam flowrate (selected with MOP in mind).
    • Maintain this rate for 15 minutes to establish a baseline.
  • 2. First Reduction:
    • After the initial 15 minutes (period X), reduce the steam rate to 0% (shutin1) valve position.
    • Hold the 0% rate for at least 15 (period X) minutes.
  • 3. Step Reductions:
    • Following the 0% hold, increase the steam rate to the maximum value minus the step size indicated.
    • Maintain this reduced rate for at least 15 minutes. (period X)
  • 4. Second Reduction:
    • Reduce the steam rate to a 5% (shutin2) valve position, ensuring the reduction is done gradually to prevent water hammers due to a sudden change.
    • Hold the 5% rate for at least 15 (period X) minutes.
  • 5. Subsequent Reductions:
    • Reduce the previous steam flowrate by the step size indicated once more.
    • Maintain for at least 15 minutes (period X) before proceeding.
  • 6. Cycle Repetition:
    • Continue to alternate between shutin1 (0%), shutin2 (5%), and the incremental reduction of the previous steam rate by the step size, maintaining each for at least 15 minutes. (period X)
    • Repeat these steps until the minimum measurable steam flowrate is achieved.
  • 7. Incremental Increase:
    • Once the minimum measurable steam flowrate is reached, incrementally increase the steam flowrate by the step size (period X) until back to the initial maximum flowrate. There is no need for 0% (shutin1) and 5% (shutin2) intermediate drops during this phase. The previous step was decrement in flowrate, while this is the opposite (increments). The reason for the incremental increase is to get two data points at each flowrate which increase the confidence in the data collected. There is no need for shut-ins since the purpose is to check the WHP and flowrate accuracy. Lack of shut-ins also makes the test duration shorter which is a more efficient use of time.
  • 8 Restoration:
    • After reaching the targeted maximum reduced rate, return the steam flowrate to the pretrial rates.
    • (In automated methods, set the controller mode back to its original state, whether Remote or Local, as per the standard operating procedure.)
  • 9. Documentation (logging for manual or automated methods):
    • Record the experiment date, start time, and end time.
    • Record any observations, anomalies, or safety issues encountered during the trial.
  • 10. Post-Trial Inspection:
    • After the trial is complete, perform a thorough inspection of the system to ensure no damage has occurred, and verify that the system returns to stable operating conditions.

While the above method may be sufficient to determine an IBHP from the data collection plan including shut-in periods, the data collection plan may also be used as the basis for predicting IBHP for a variety of steam flowrate values in a specific well.

A further exemplary method is for predicting bottom hole pressure for a steam injection well, using the same data collection plan described above. As with the above exemplary method, the valve is opened to allow steam injection to the liner, and the steam flowrate and WHP are measured (by the meter and gauge, respectively). The steam flowrate is varied through a plurality of selected values using the valve, including at least one shut-in period (0-5% steam valve position). The steam flowrate and WHP are measured for each of the selected steam flowrate values, with a time period for the shut-ins sufficient to reduce the WHP to approach the IBHP or even allow equilibrium to occur between the WHP and the IBHP, by which the IBHP can be determined for the 0-5% valve position(s), as described above.

In this further exemplary method, an average steam flowrate and an average wellhead pressure are then determined for the measured steam flowrate and the measured wellhead pressure for each of the plurality of selected values (valve positions). The exemplary data collection process generates data that can be presented in a chart such as FIG. 4, which illustrates data for a sample well. The chart shows steam flowrate and WHP as a function of valve position over time. It can be seen that during the shut-in periods the curves drop to zero flowrate, 0% or 5% valve position, and WHP drops to an equilibrium steady state IBHP (determined IBHP). For each of the steps in the data collection process, which are shown as curve maximums and minimums (square steps) in FIG. 4, the steam flowrate and WHP can be averaged to arrive at an average steam flow rate and an average WHP for each step.

Next, a pressure drop is calculated for each of the plurality of selected values. First, all of the determined bottom hole pressure determinations for the plurality of shut-in periods (0-5% valve positions) are averaged to arrive at a single average IBHP value for the specific well. For one non-limiting example, you may have the following: IBHP at 0%=3499 kPa, and at 5%=3501 kPa, resulting in an average IBHP of 3500 kPa. Second, a pressure drop (WHP minus average IBHP value) is calculated for each of the plurality of selected steam flowrate values. For example, if there is a steam flowrate of 50t/h and the WHP is 4000 kPa, then the pressure drop is 4000 kPa-3500 kPa=500 kPa. If there is a steam flowrate of 35 t/h and the WHP is 3700 kPa, then the pressure drop is 3700 kPa-3500 kPa=200 kPa. This calculation would be repeated for all the other combinations of WHP and steam flowrate, resulting in a set of matched pressure drops and average steam flowrates.

The pressure drops and average steam flowrates can then be plotted on a chart, such as the exemplary FIG. 5a. In FIG. 5a, each of the points are plotted on a graph of pressure drop (kPa) against average steam flowrate (t/h). Once the points have been plotted for each of the selected steam flowrate values, it becomes clear that there is a polynomial relationship, and specifically in this embodiment a linear relationship, between pressure drop and average steam flowrate, and a best fit model (in this embodiment a line of best fit) can be generated for the points defining the polynomial relationship between the pressure drop and the average steam flowrate for that well. As can be seen in FIG. 5a, plotted points and a line of best fit are shown for six steam injection wells, and all manifest a polynomial relationship. Other fitting methods were tested such as exponential and logarithmic, but polynomial (and specifically linear) was selected due to its simplicity. For linear fit, three techniques were tested based on how the intercept (c) of the linear model is determined, and the techniques are described below and shown in FIG. 8:

• • Technique 1: Fit to the data without shut-ins. This is practically sound since the steam flowrate meter at the low range has a high error.
  • Technique 2: Force the intercept at (0,0). This is theoretically sound, since we would expect the pressure drop (dp) to be zero at zero flowrate.
  • Technique 3: Fit to the data with shut-ins.

FIG. 8 is a diagram showing qualitative representations of the three techniques (labeled 1, 2, 3). The best technique is technique 1 (labeled 1) based on the data collected, as technique 1 fits the data in the operational domain better.

Given the linear relationship in this embodiment and the ability to then generate a line of best fit for the points derived from the data collection process for a steam injection well, one can further derive a slope and intercept from the line of best fit. FIG. 5b is a table showing the derived slope and intercept (and a coefficient of correlation) for each of the six sample wells.

Now that the line of best fit has been generated to allow definition of the linear relationship between pressure drop and average steam flowrate for a specific steam injection well, one can use the derived slope and intercept to predict or model an IBHP (steady state response) for any steam flowrate and WHP for the specific steam injection well:

$\Delta P=\mathrm{WHP}-\mathrm{IBHP}$ $\Delta P=m\times \mathrm{flowrate}+c\left(\mathrm{due}\mathrm{to}\mathrm{the}\mathrm{observed}\mathrm{linear}\mathrm{relationship}\right)$ $\mathrm{IBHP}=\mathrm{WHP}-\Delta P$ $\mathrm{IBHP}=\mathrm{WHP}-\left(m\times \mathrm{flowrate}+c\right)$ $\mathrm{IBHP}=\mathrm{WHP}-m\times \mathrm{flowrate}-c$ $\mathrm{where}m\mathrm{is}\mathrm{the}\mathrm{derived}\mathrm{slope}\mathrm{and}c\mathrm{is}\mathrm{the}\mathrm{derived}\mathrm{intercept}.$

This allows the operator to more confidently determine a safe steam flowrate and remain below the MOP for the specific well by ensuring that the IBHP<MOP. This method can be used for a plurality of steam injection wells, where each well would have a unique best fit model and resultant parameters (e.g., slope and intercept for a linear relationship).

Due to the dependence on instruments used to measure physical properties such as WHP and steam flowrate, there is an inherent lag in the system between the measured variables. This leads to period of transition which causes incorrect (not real) spikes in the data. To correct for these spikes, two methods can be employed.

• • Method 1: Steady state response which employs a step detection algorithm. FIG. 9 illustrates a steady state response which employs a step detection algorithm; the algorithm detects a step change above a threshold at time t_start, and holds both the steam flowrate (corresponding to the measured flowrate labeled “MV”) and the WHP (corresponding to the measured WHP labeled “CV”) constant for a time period after t_start (i.e., a holding time period between t_start and t_end) as shown in FIG. 9. The holding time period can be configured to permit a steady state response following the detected step transition at time t_start. In embodiments, the holding time period between t_start and t_end can be 5 minutes or some other suitable value.
  • Method 2: In some exemplary embodiments, a dynamic response calculation to reflect real-time can be incorporated in addition to the above steady state calculation of predicted IBHP:

$\mathrm{IBHP}=\mathrm{WHP}-m*\left({\mathrm{transient}}_{\mathrm{response}}\right)*\mathrm{flowrate}-c$ ${\mathrm{transient}}_{\mathrm{response}}=\left(1-e^{-\frac{t-\theta }{\tau }}\right)$

Therefore,

$\mathrm{IBHP}=\mathrm{WHP}-m*\left(1-e^{-\frac{t-\theta }{\tau }}\right)*\mathrm{flowrate}-c$

The parameters of the dynamic response are determined using a step-test, a common procedure using by controls specialists to determine the time constant, and the dead time.

Method 1 (steady state response) is preferred. The steady state model uses fewer parameters making it simpler and more straightforward to apply and interpret than the dynamic model. Due to its simplicity, the steady-state model is easier to generalize across different wells or conditions without the need for recalibrating the dynamic aspects, which might vary significantly with operational conditions. Steady state modeling requires less data for calibration as it primarily depends on polynomial relationships and does not require data on the transient behavior of the system.

The prediction presents a well-defined and bounded problem, offering distinct advantages when contrasted with the direct prediction of reservoir injector bottom hole pressure. This bounded nature of the problem is underpinned by specific constraints that inherently govern it: pressure drop must always be greater than zero, aligning with physical principles; and the predicted pressure drop should remain within the bounds established by the determined bottom hole pressure:

$0\le \Delta P<\mathrm{WHP}$

These constraints serve as pivotal guiding principles, contributing to the robustness and relevance of the pressure drop prediction model.

In some embodiments, a maximum steam flowrate can also be determined as an upper boundary for the best fit model, using determination of “choked flow” as illustrated in FIGS. 6a and 6b. Choked flow behaviour may be seen in the relationship between pressure, steam flowrate, and valve position. At choked flow the polynomial correlation between steam flowrate and WHP changes, with the flowrate staying the same while the pressure rapidly increases, leading to an overestimation of the IBHP based on the calculation. FIG. 6a illustrates a non-choked behaviour where the pressure increases as the flowrate increases. FIG. 6b, however, illustrates a rapid increase in pressure while the flowrate remains just over 30 t/h. In the illustrated embodiment, there is a linear correlation between pressure drop and steam flowrate, but the choked flow demonstrates non-linearity, and thus maximum flowrate for the steam injection well. The detection algorithm is based on detecting when the non-linearity starts. An efficient approach is using a binary search algorithm.

Models according to the present disclosure are calibrated using data from routine shut-ins and periodic data collection. It is desirable in some exemplary embodiments to perform recalibrations of the model on a periodic basis, for example monthly, and preferably also on the occurrence of certain trigger events. A recalibration is a retest of the original pressure drop and shut-in values on certain flowrates representative of the operating range of the well. This retest would be completed and then the results analyzed to produce new estimates for the parameters from the polynomial relationship defined by the best fit model. These values can then be used to update the model.

Recalibration is expected to be completed periodically to update the model for time dependent factors, but it can also be triggered due to events that could invalidate a model. Such trigger events include the following:

Environmental/Operational:

• • Well shut-in (Change in OCD status, well completions)
  • Switch to NCG/Solvent
  • Large change in reservoir pressure (threshold, based on safety range to MOP)

Measurement and Data Quality:

• • Instrument failure (failure signal due to sudden loss of data, or out of range data)
  • New measurement device introduction
  • Significant deviation in pressure data

Model Performance and Validation:

• • High Error (residual) when doing a spot check
  • Drift in model predictions detected
  • Failed recalibration
  • Large change in the parameters

Regulatory and External Factors:

• • Manual trigger by operations
  • Manually turning determination on after a threshold period
  • Changes in MOP value (temporary changes)

In addition to the model recalibration techniques, spot checks may be implemented for verification. A spot check is a single point verification of the model to help ensure that the current prediction is reflective of the actual IBHP pressure. These spot checks are preferably done periodically (for example, every two weeks) but more frequently than recalibration.

The spot checks may be either aggressive or less intrusive. In an aggressive spot check, the steam flowrate is reduced by one step (via a drop in the steam valve position) to obtain a physical measurement related to the IBHP; a failed aggressive spot check preferably also triggers a recalibration.

Additional less intrusive spot checks are preferably scheduled to help ensure the operation is under the reservoir MOP, and they are not intended to trigger recalibration. Instead of reducing the valve position to 0% or 5% to help determine the true IBHP, the valve position need only be adjusted until the WHP is below the MOP. For example, in one hypothetical case under an aggressive spot check:

$\mathrm{IBHP}=3000\mathrm{kPa}$ $\mathrm{MOP}=5000\mathrm{kPa}$ $\mathrm{WHP}=5500\mathrm{kPa}$

The pressure drop (ΔP) is calculated as: ΔP=WHP−IBHP=5500 kPa−3000 kPa=2500 kPa.

For a less intrusive spot check, at 0% or 5% valve position, WHP is assumed to be approximately equal to IBHP (3000 kPa). However, to ensure WHP is below MOP, all we need is for WHP to be less than 5000 kPa, and not equal to IBHP (3000 kPa). To achieve this, we can reverse calculate the ΔP needed (ΔPmop used to differentiate from the ΔP in the aggressive spot check technique):

$\mathrm{WHP}=\mathrm{MOP}+\Delta P_{\mathrm{mop}}$ $5500\mathrm{kPa}=5000\mathrm{kPa}+\Delta P_{\mathrm{mop}}$ $\Delta P_{\mathrm{mop}}=500\mathrm{kPa}\left(\mathrm{thus}\mathrm{less}\mathrm{aggressive}\mathrm{and}\mathrm{less}\mathrm{intrusive}\right)$

While any polynomial relationship can be used, following are equations for a linear relationship. We can reverse calculate the flowrate required to achieve this ΔP_mop of 500 kPa using the equation:

$\Delta P_{\mathrm{mop}}=m\times \left(\mathrm{steam\_flowrate}\right)+c$

• • where (m) is the slope and (c) is the intercept derived from the line of best fit for the linear relationship.

Thus, the steam flowrate is steam_flowrate=(ΔP_mop−c)/m, where (m) and (c) are pre-determined from the previous calibration.

In this case, ΔP_mop is 500 kPa. Therefore, we only need to reduce the flowrate to achieve a pressure drop equivalent to 500 kPa (less intrusive) instead of 2500 kPa (as with the aggressive spot check). In this case, it might be a valve position of 40% (for example) instead of 0 or 5%.

As a further non-limiting example, in a shallow reservoir, a limiting case provides further clarification:

$\mathrm{IBHP}=3000\mathrm{kPa}$ $\mathrm{MOP}=3050\mathrm{kPa}\left(\mathrm{note}\mathrm{that}\mathrm{the}\mathrm{MOP}\mathrm{is}\mathrm{only}\mathrm{slightly}\mathrm{higher}\mathrm{than}\mathrm{the}\mathrm{IBHP}\right)$ $\mathrm{WHP}=4000\mathrm{kPa}$ $\Delta P=\mathrm{WHP}-\mathrm{IBHP}=1000\mathrm{kPa},\mathrm{and}$ $\Delta P_{\mathrm{mop}}=\mathrm{WHP}-\mathrm{MOP}=950\mathrm{kPa}$

Note that for a shallow reservoir, the ΔP and ΔP_mop are very similar whether we look at the difference of WHP to IBHP or to MOP. Thus, for a shallow reservoir, we would always go to 0% or 5% valve position (near zero steam flowrate).

The computational simplicity of the polynomial model allows for sequential updates and integration with existing control systems. The system was designed to continually recalibrate the models through periodic data collection, ensuring accurate and reliable IBHP measurements. Recalibration events will occur monthly, with bi-weekly spot checks to ensure that the model has not drifted and is still representative of current conditions.

It has been found that the average steam flowrate is a significant predictor of differential pressure (pressure drop), with changes in average steam flowrate consistently influencing differential pressure across a wide range of conditions, with a strong polynomial relationship between steam flowrate and pressure drop. This relationship is always polynomial for a given well, but the parameters (such as slope/intercept for a linear relationship) change over time (not stationary) and from well to well. For the solution to be commercially reasonable on a large scale for a large number of wells, it should be proceduralized and automated so that a well can self-identify its conditions and periodically re-calibrate/validate itself against shut-in measurements.

In terms of the physical system components, there is a steam flowmeter, and a pressure meter with transmitter at surface, downstream of the steam flowrate control valve—the control valve will be manipulated automatically in the exemplary embodiment.

In some exemplary embodiments, the above method for predicting the bottom hole pressure for the steam injection well for a range of steam flowrate values can be automated using a programmable logic controller (PLC) or similar platform computational platform. The skilled person would be aware of other computing systems useful for embodiments of the present disclosure, but PLCs are desirable as they are conventionally already present at a well pad and suitable for the calculations such as those described herein. In such embodiments, the PLC receives the measured steam flowrate from the meter and the measured wellhead pressure from the gauge, plots the points and generates the best fit model, and derives the parameters (such as slope and intercept for a linear relationship) for the best fit model. It also performs the required verifications and validations at pre-processing and post-processing steps on the data. This is a fully automated end-to-end system.

An exemplary PLC comprises a CPU (to interpret inputs and execute control program stored in memory), a memory unit (for storing inputs and program to be executed), a power supply unit, input and output interfaces, and a communications interface. Specifically, the input interfaces would communicate through the communications interface with the meter and the gauge to receive the measured steam flowrates and WHP values, which measurements are then stored in the memory unit. The program (described below and illustrated in FIG. 7) is stored on the memory unit and executed on the CPU. The CPU is configured to determine an optimized steam flowrate based on the within methods described above, determining a corresponding valve position and communicating same by the communications interface to the valve to control same and adjust the valve to the desired position. Another advantage of using PLCs at well pads is that you can enable shutdown at the individual well pad, and if the specific PLC crashes only that one well pad is shut down (other well pads can continue operation), and in this case presents no incremental reliability risk since the PLCs are critical infrastructure for the well pad operation. Further, reliability is problematic if using remote third-party computer service, which can be subject to crashing, communication loss, etc.

In a first step, the logic (built in this case as structured text code, but it could be ladder logic, or functional charts) is built in Logix 5000 (RunMode and BuildMode as shown in FIG. 7). The exemplary system continually and automatically calibrates the models in production using periodic data collection as per the data collection plan described above.

Turning to FIG. 7, “BuildMode” is the testing mode of the system; it has setpoints indicating the need to change valve operation when achieved. Parameters (e.g., slope, intercept) and maximum flowrate go into the “RunMode” (where the pressure drop calculation is conducted). RunMode calculates the determined IBHP so steam injection can be determined. This is used by the safety system in place to not exceed MOP. It will be clear to those skilled in the art that actual operation may incorporate a reduction from true MOP to better ensure there is no exceedance.

An algorithm is preferably used in this exemplary embodiment to detect the step changes in the data collection process. This algorithm may be built to identify and analyze steps within a dataset, particularly focusing on a variable of interest, which would be the valve position. An example of a step detection method used is described here, however, any well known step (change point) detection techniques can be used. The method employed here is used for its simplicity, therefore, it can be implemented on the PLC. It begins by initializing necessary data structures, including a list to store identified steps and a dictionary to keep track of the current step being analyzed, with fields such as start and end indices, sum of values, and counts of data points within each step. As it iterates through the dataset, it checks for significant changes in the variable of interest compared to the average value within the current step, determined by a user-defined threshold. When a change surpasses this threshold, the current step is concluded, and a new one begins. Once all data points are processed, the function refines step statistics and assigns statuses based on the average values. Steps with average values within predefined ranges for valve positions are labeled as ‘shutin1’ (0%) or ‘shutin2’ (5%), while others are marked as ‘operational’. Overall, this function may then provide an automated approach to detecting and characterizing steps in a dataset, offering insights into temporal variations or events within the monitored system.

In some other embodiments, machine learning and sequential learning techniques may be employed. In some exemplary embodiments with a linear relationship, the method predicts the value of Y using a linear equation: Y=(predicted slope)*X+(predicted intercept). However, the data is not static, and new data points (X,Y) keep arriving over time. To ensure the model is up-to-date, various techniques can be used:

• • Technique 1: Full data collection test (BuildMode) on a fixed periodic interval
  • Technique 2: Full data collection test (BuildMode) on a dynamic interval, the interval is automatically determined based on the confidence in the estimate slope and intercept.
  • Technique 3: Sequential learning using one point at a time, thus, not requiring the full data collection test previously described. This is described below.

Sequential learning may therefore be applied by adding one point at a time as described below:

• • 1. Initial Model Training: Start by training an initial model to predict the slope and intercept based on an initial dataset. This initial model captures the relationship between X and Y based on the data available at that time.
  • 2. Model Update: As new data points (X,Y) become available, instead of retraining the entire model from scratch, update the model incrementally. When a new data point arrives, use it to adjust the predicted slope and intercept based on the existing model's parameters.
  • 3. Adaptation Over Time: The model gradually adapts to changes in the data distribution by incorporating new observations. This allows it to reflect the most recent trends and relationships between X and Y.
  • 4. Continuous Learning: Continue to collect data and update the model periodically to ensure that it remains accurate and relevant to the current data distribution. This process is repeated as new data points are received.

Sequential learning may be advantageous in scenarios where the data is non-stationary or evolving, and retraining the entire model from scratch with each new data batch is practical but not preferred. Instead, one may be able to maintain and fine-tune the model's parameters over time, helping to ensure that it stays up-to-date and capable of making accurate predictions as new data arrives.

The foregoing is considered as illustrative only of the principles of the present disclosure. The scope of the claims should not be limited by the exemplary embodiments set forth in the foregoing, but should be given the broadest interpretation consistent with the specification as a whole.

Claims

1. A method for determining bottom hole pressure for a steam injection well used in hydrocarbon recovery, the steam injection well configured to introduce steam to a subsurface reservoir containing hydrocarbon, wherein a downhole liner of the well is provided with at least one liner-deployed outflow control device for introducing the steam to the reservoir, the steam injection well further provided with (i) a valve at surface for controlling steam flowrate into the well, (ii) a meter at the surface for measuring the steam flowrate, and (iii) a gauge at the surface for measuring pressure upstream of the at least one liner-deployed outflow control device, the method comprising the steps of:

a. opening the valve to allow the steam to pass through the well and the at least one liner-deployed outflow control device into the subsurface reservoir;

b. selecting a shut-in period for at least one shut-in sufficient to reduce the pressure to approach the bottom hole pressure;

c. operating the valve to implement the at least one shut-in for the shut-in period;

d. during the shut-in period, allowing the pressure to approach the bottom hole pressure, resulting in a reduced pressure at the surface; and

e. measuring the reduced pressure using the gauge to arrive at a determined bottom hole pressure.

2. The method of claim 1, wherein:

the shut-in period is sufficient to allow the pressure to reach equilibrium with the bottom hole pressure.

3. The method of claim 1, wherein:

the steam injection well is an injector well of a steam-assisted gravity drainage well pair.

4. The method of claim 1, further comprising:

operating the valve at step c. at a variety of valve positions to implement a range of steam flowrates including the at least one shut-in, and measuring the pressure for each of the variety of valve positions.

5. The method of claim 1, wherein:

the at least one shut-in is achieved by operating the valve at step c. to between a full shut-in (0%) and a partial shut-in (5%) valve position.

6. The method of claim 1, further comprising:

monitoring for changes in the determined bottom hole pressure.

7. The method of claim 1, further comprising:

the step of, in response to a determined bottom hole pressure exceeding a maximum operating pressure for the reservoir, operating the valve to reduce the steam flowrate.

8. The method of claim 1, wherein:

step c. comprises at least two shut-ins.

9. The method of claim 4, wherein:

each of the valve positions is maintained for at least 15 minutes.

10. The method of claim 4, wherein:

the variety of valve settings ranges from 0% to a selected maximum steam flowrate.

11. The method of claim 1, wherein:

the pressure at the surface is a wellhead pressure.

12. A method for predicting a bottom hole pressure for a steam injection well used in hydrocarbon recovery for a range of steam flowrate values, the steam injection well configured to introduce steam to a subsurface reservoir containing hydrocarbon, wherein a downhole liner of the well is provided with at least one liner-deployed outflow control device for introducing the steam to the reservoir, the steam injection well further provided with (i) a valve at surface for controlling steam flowrate into the well, (ii) a meter at the surface for measuring the steam flowrate, and (iii) a gauge at the surface for measuring the pressure upstream of the at least one liner-deployed outflow control device, the method comprising the steps of:

a. opening the valve to allow the steam to pass through the well and the at least one liner-deployed outflow control device into the subsurface reservoir;

b. measuring the steam flowrate using the meter and measuring the pressure using the gauge;

c. selecting a shut-in period for at least one shut-in sufficient to reduce the pressure to approach the bottom hole pressure, resulting in a reduced pressure at the surface, and measuring the reduced pressure to provide a determined bottom hole pressure for the at least one shut-in;

d. using the valve, varying the steam flowrate between a plurality of selected values falling within the range of steam flowrate values, including the at least one shut-in;

e. measuring the steam flowrate using the meter and measuring the pressure using the gauge for each of the plurality of selected values falling within the range of steam flowrate values;

f. determining an average steam flowrate and an average pressure for the measured steam flowrate and the measured pressure for each of the plurality of selected values;

g. calculating a pressure drop for each of the plurality of selected values by subtracting an average determined bottom hole pressure for the at least one shut-in from the measured pressure for each of the plurality of selected values;

h. plotting the pressure drop against the average steam flowrate for each of the selected values as points, and generating a best fit model for the points defining a polynomial relationship between the pressure drop and the steam flowrate;

i. deriving a set of parameters from the best fit model; and

j. using the set of parameters to predict the bottom hole pressure for any steam flowrate along the range of steam flowrate values.

13. The method of claim 12, wherein:

the shut-in period is sufficient to allow the pressure to reach equilibrium with the bottom hole pressure.

14. The method of claim 12, wherein:

the at least one shut-in is achieved by operating the valve at step d. to between a full shut-in (0%) and a partial shut-in (5%) valve position.

15. The method of claim 12, further comprising the step of:

in response to a determined bottom hole pressure exceeding a maximum operating pressure for the reservoir, operating the valve to reduce the steam flowrate.

16. The method of claim 12, wherein:

step d. comprises at least two shut-ins.

17. The method of claim 12, wherein:

the plurality of selected values ranges from 0% to a selected maximum steam flowrate.

18. The method of claim 12 wherein each of the steam flowrate values is a valve position.

19. The method of claim 16, wherein the determined bottom hole pressures for the at least two shut-ins are averaged to arrive at the average determined red bottom hole pressure for the pressure drop calculation of step g.

20. The method of claim 12, comprising:

identifying a maximum steam flowrate by determining where the polynomial relationship changes.

21. The method of claim 12, wherein:

the pressure at the surface is a wellhead pressure.