Optimizing gas production with innovative approach to evaporative halite precipitation and liquid loading management

The article discusses two issues that affect the productivity of gas wells: halite precipitation and liquid loading. Halite precipitation occurs when salt crystals form in the well, while liquid loading happens when liquids build up in the well, reducing gas production. Accurate prediction of these issues is essential to mitigate them effectively. While liquid loading can be managing by reducing the size of tubing, halite precipitation is more challenging to predict and manage. The existing approaches have not sufficiently validated, and the current halite management measures are not cost-effective. The article reports on a study that aims to develop a model using halite envelope to predict the onset of halite precipitation in a gas field in Oman. The study simulated halite envelopes for thirty different gas wells known to have experienced halite precipitation problems. The results showed that there is significant variation in the size and shape of halite envelopes in different wells, making it infeasible to construct a single envelope for the entire field. However, grouping halite envelopes based on inflow resistance and constructing a model for each group proved to be a powerful tool for halite prediction. The study also investigated the use of velocity string (VS) technology as a new approach to managing halite precipitation. The study simulated halite envelopes for a few wells with different sizes of VS installed, estimating the timing of halite precipitation reoccurrence. The results showed that the effectiveness of VS as a halite remedial technique depends mainly on the rock quality and the time of installation. Additionally, the study assessed the impact of installing VS on liquid loading and the synergy between halite management and de-liquefaction by using VS. The results showed that the synergy between halite management and de-liquefaction using VS depends mainly on rock quality and the size of the installed VS.


Introduction
Natural gas production has been playing a vital role in securing the energy demand across the world for the last decades.The upsurge in the dependency on natural gas as a source of energy is attributing to many reasons, such as its abundance in nature as well as being more environmentally friendly compared to oil.Consequently, the number of the gas wells around the globe increases pronouncedly [1].Gas wells are susceptible to various production issues, such as condensate banking, corrosion, scaling, etc. that impair their production during their lifetime [2].These challenges usually have significant impacts on the well's deliverability, leading to dramatic implications in the overall economy of the countries around the globe.Scales, which are hard and inorganic particles present in the formation water, are very common in oil and gas wells and they are usually forming when changes in wells' operational conditions, such the temperature, the pressure and water composition, have introduced.The deposition of the scale particles whether in the wellbore or the reservoir near the wellbore region is always associated with substantial financial losses in the oil and gas industry.
Halite scale precipitation is considering as one of the inevitable serious challenges encountered by the gas wells towards the end of their production lifetime [3].This problem has been a subject of concern for the last two decades, and recent increased concern is attributed to the fact that many gas fields are approaching the end of their production lives and the impacts of these problems has become obvious.The halite scaling differs from other types of scale, and therefore considered as unconventional scale, in its high solubility in water and a minimal change in the solubility results in a large amount of deposited halite particles [3].The halite deposition can be easily identify from the well production data.It is always associated with intermittent or accelerated decline in wells' production [4].Halite clogging is find to be a very common issue in gas or HPHT wells producing from depleted reservoirs in the presence of high saline formation water [5].Its severity increases at reservoir pressure less than 150 bar and low wellhead pressure [4], [12] and [14].Additionally, the likelihood of halite scaling is magnifying following the reduction of the wellhead pressure following the introduction of the surface depletion compression [3], [8] and [5].The formation and deposition of the halite in the gas wells are mainly controlling by the pressure and temperature reduction.The former leads to increased gas expansion, leading to evaporating the formation water and leaving behind the salt crystals, which in turn accumulate and block the flow path of the gas.The later, on the other hand, causes a reduction in the solubility of the salt and therefore increases the likelihood of the salt precipitation [6].The pressure-induced halite scaling, sometimes referred to as "evaporative halite scaling" occurs normally in the lower completion where the highest differential pressure exists.By contrast, the temperature-induced halite scale commonly deposits in the upper completion and the surface facilities where the highest change in temperature is usually expected.The halite precipitation have commonly removed and or prevented by freshwater treatment and mechanical cleanout.However, the economic and operational viability of these techniques is always of a high concern.For example, the availability of low saline water is always a challenge.The mechanical cleanout is also associated with high cost, which sometimes makes this remedial technique economically not viable [7].Additionally, the accurate prediction of the onset of the halite scaling plays a vital role in managing the issue and reducing its severity.However, predicting the startup of the halite issue is not straightforward since it depends on specific well conditions [8].
Liquid loading is another well-known challenge faced by gas wells toward the end of field life.This issue occurs when the gas rate declines to a certain value, so-called "minimum flowrate", at which it loses its ability to carry the produced liquid to the surface [9].Liquid loading has a significant impact on the well capacity.When liquid-loading starts, the well firstly produces at unstable rates as a form of cycles of loading and unloading slugs of liquid until the production completely ceases [10] and [9].This phenomenon had usually addressed through the deliquification processes, which is the process of unloading the well and prolonging the wells' lifetime.Many deliquification techniques are available in the industry, each of which has its advantages and disadvantages, such as intermittent production, gas lift and plunger lift velocity string installation.etc. [5].Installing velocity string, which is a small string that had inserted inside the original completion tubing to reduce its internal diameter, has considered as the most popular deliquification technique, and widely used around the globe due to its high efficiency and its low cost [11].
Therefore, the aims of this research are to develop a halite envelope model to forecast when halite will precipitate in gas wells and to verify the model against actual data collected from gas wells in Oman.Additionally, the research will optimize the use of the velocity string method to regulate halite scaling in the gas well completion strings.The research will also evaluate the impact of installing velocity string on liquid loading and establish a synergy between halite management and de-liquefication using velocity string method.It will also predict when liquid loading will occur.

Methodology
The main objective of this study is to assess the feasibility of using a generalized halite envelope to predict evaporative scaling in a gas field.Achieving this objective requires three steps done sequentially.The first step is to analyze the MPLT data of several wells known to have this issue.This data will serve as the main input for the following steps.Following the preparation of the input data, halite envelopes for each well will modeled and validated.Then, all modeled halite envelopes will used to define a generic model for the field.This generic model will tested with data point that not used in constructing the generalized model.Another objective of this study is to evaluate the possibility of using velocity string to mitigate or manage the halite scaling issue.This objective will achieved by quantifying the impact of installing VS on halite scaling in term of the time requires for the halite issue to re-occur, post VS.Estimating the time of the halite scaling reoccurrence involves three main procedures.The first step is to model the halite envelope for the case of the VS and analyze the envelope to define the pressure condition at which the halite will re-appear.Then, the outcome of the previous step will used to conduct a nodal analysis to define the inflow and the outflow parameters, which will be used to define the well capacity at a given operating conditions.The last step will be calculating the time of the halite issue with the aid of using the decline curve analysis.
The last objective of this study is to predict the time of the liquid loading issue, evaluating the impact of installing velocity string on liquid loading and assess the synergy between halite management and de-liquification by using velocity string.This objective will achieved by calculating the time it takes the selected wells to load-up and then compare the time to liquid loading with the time of halite re-occurrence when velocity string is installed.The synergy is generating if the halite re-occurrence time exceeds the time to liquid loading.
The flowchart in Figure 1 summarizes the workflow and procedures adapted in this study to achieve the study objectives.

Case Study and Data Gathering
This research will investigate 48 deep gas wells that have affected by halite precipitation.Various datasets from these wells will used to meet the study's goals.These datasets contain the following: 1. Memory Production Logging Tool data (MPLT) of the examined wells.
This data represents the main input to model halite scaling envelopes.Appendix A summarize the MPLT data included in this study and the date of halite occurrences in the individual wells.2. The current production parameters of individual wells including the wellhead pressure (THP) in bar, the flowline pressure (FLP) in bar, the wellhead temperature (THT) in C, condensate gas ratio (CGR), water gas ratio (WGR).See Appendix B 3. Well parameters including the internal diameter of the completion tubing (ID) in inch, the true vertical depth to the top perforation (TVD) in meter, along hole depth to top perforation (AHD) in m, bottomhole temperature (BHT), near wellbore pressure (Pnear) in bar.The gas deviation factor (Z) and the gas specific gravity (SG) are assuming to be constant with default values of 0.95 and 0.65 respectively.See Appendix C 4. Reservoir parameters including the initial reservoir pressure (Pi) in bar and connected gas volume (CGIIP).

Background about the study area.
This research was conducting at a gas field (Field A) in central Oman.The gas wells in this area produce from two sandstone reservoirs hundreds of meters deep.Since the reservoir quality is tight, hydraulic fracturing was using to complete the wells.The wellhead pressure was decreasing from 35-45 bar to 10-20 bar in the field in 2016, because of a surface depletion compression.As a result, after 10 to 20 years of consistent production, several wells began to experience halite precipitation.Freshwater bulheading and mechanical cleanout utilizing coiled tubing equipment are now using to manage halite precipitation in this field.However, these strategies are not always effective since some wells' capacity could not be recovered using these methods.Furthermore, these two procedures are time-consuming and rather expensive.

Predicting halite scaling using a general halite envelope
This section will describe the procedures that will followed to construct a general halite envelope for halite precipitation prediction.The first step is to analyze the MPLT, followed by simulating and validating the halite envelope of different wells, generalizing the halite envelope, and finally validating the general halite envelope.

MPLT data analysis and input preparation
The MPLT data represents the main input to the halite envelope simulation.The MPLT data will analyzed to identify the operating conditions of wells at the time of acquiring MPLT as well as calculating the inflow resistance (A-factor) of each well.This analysis involves conducting the inflow performance of wells.The inflow performance analysis will carried out using an excelbased tool.
This step will conducted through the following steps: -Enter gas production and bottomhole pressure for every MPLT period (Shut-in period, first pass period second pass period etc), for flow unit, reservoir or well.-The tool has its own built-in workflow to provides least square fit of data against Forchheimer equation for flow unit, reservoir or well.
Given the MPLT data analysis is completed, the wells' operating conditions at the time of MPLT collection, representing the no halite cases, will compared with the current operating conditions, representing the halite scaling cases, of the examined wells to define the halite scaling free conditions.

Modeling Halite envelopes and model validation
The halite envelopes of the wells experiencing halite scaling will simulated using Halite Prediction Tool Version 2 (HPT-V2).The purpose of constructing the tool is to be using for screening against the halite precipitation risk for individual wells.The tool has three built-in models, an electrolyte brine model, a gas PVT model and a simple inflow-outflow model.MPLT data is using as a main input to the tool, which has its own add-ins to simulate halite envelope.Figure 2 shows a schematic diagram of the halite envelopes.Koot et.al (2022) [13], provides a comprehensive detail about the development of the tool, its assumptions, and its applications.

Generalizing the halite envelope
In this step, the studied wells will divided into three groups based on their inflow resistance (Afactor).Wells with A-factor ranges from 1 to 10 (good quality) are considered as group 1.Wells having A-factor between 10 to 100 (moderate rock) represent group 2 while wells with A-factor more than 100 (tight rock) are categorized as group 3.
Then, the modeled halite envelopes of each group of wells will compiled in a single plot with their respective MPLT point that represents the no-halite points and their corresponding current operating points representing the halite scaling point.Then, by visual inspection, the base case envelope of each group of wells will defined.The base case envelopes will selected for each group of wells based on the following criteria: -The base case envelope should not be neither the smallest nor the largest envelopes of each group of wells.The smallest and the largest envelopes represent the extreme cases.-It should fall in the middle of the other envelopes of each group of wells.
-It should show the best match with the real data.In other words, most of the halite scaling points should fall below the base case envelop and most of the no-halite scaling points should fall above the halite envelop.

Validating the generalized halite envelope and quantifying its accuracy
In this step, a blind test will performed to evaluate the prediction's accuracy of the defined base case envelopes of each group of wells.The wells involved in this test should not use in constructing the base case envelopes described in the previous step.The test will done by simply plotting the scaling and no-scaling points of each well in its respective envelope and compare the model prediction and the field observation.
In order to quantify the accuracy of the model, a confusion matrix statistical representation will used.The 2x2-confusion matrix, as shown in Table 1, correlates the scale prediction and field observation and counts the number of wells for each prediction-observation pair.The true negative (Tn) represents the correctly predicted no halite scaling cases and the true positive represent the correctly predicted halite scaling cases.On the other hand, the false positive (Fp) represents the incorrectly predicted no-halite cases and the false negative (Fn) represents the incorrectly predicted no-halite scaling cases.

Designing Nodal Analysis
The objective of conducting the nodal analysis is to estimate Q1 and Q(halite).This is conducting in two steps as follows: 1-Computing the hydrostatic (B) and friction (C) factors using Equations ( 1) and (2) as shown below. ( (2) 2-Given that the wellhead pressure (THP), the inflow resistance of the well (A-factor), the P1 and P (halite) are known, the Q1 and Q(halite) will be computed using Equation (3).
(3) Nodal analysis will carried out twice, one for the case of 2.44" VS had installed and one for 1.99" VS with the following assumptions: 1-Inflow resistance (A) is constant 2-Tubing head Pressure (THP) is assuming to be constant.

Decline Curve Analysis and Estimating Halite re-occurrence Time
Decline curve analysis for each case is conducted to estimate the time required for the reservoir pressure to decline from the current reservoir pressure (P1) to the reservoir pressure at which halite is expected to occur P(halite).This have done through two steps as follows: 1-The cumulative produced gas, V(gas), for each well will be calculated using Equation ( 4) as shown below.

The synergy between halite management and de-liquefication
To achieve this objective, two main steps are following in this study in sequential manner as described below.

Predicting the onset of liquid loading in the candidates.
In this step, the time of the liquid loading onset for the evaluated wells is calculating through the following procedures: -Carry out the nodal analysis to compute the outflow parameters (B and C factors), and the current gas rate (Q1) at the current reservoir pressure (P1) of the original completion tubing using Equations ( 1), ( 2) and ( 3) respectively.-The Qmin for each well is calculated using modified Turner's expression (Equation (6) (6) x Calculating the Pmin for each well using Equation (7) (7) x The ultimate recovery for each well will be calculated using Equation (8) (8) x Assuming exponential decline, the time required for Q1 to reach Qmin, (TLL), has calculated using Equation 9. (9)

Results and discussion
This section will display the results of the work carried out to achieve the project's objectives.It is dividing into three sections each of which has multiple sub sections.The first section will address the construction and the validation of the general halite envelope model for halite scaling prediction, followed by the results of evaluating the impact of installing velocity string on halite scaling.The last section will display the result of assessing the feasibility of generating a synergy between halite management and de-liquification using velocity string.

Predicting Halite Precipitation Using General Halite Envelope 3.1.1 Analyzing the MPLT data
The inflow performance analysis, using the MPLT inflow performance sheet, was conducting for 49 MPLT data points.The purpose of this analysis, as described earlier, is to estimate the wellhead pressure, the reservoir pressure, and the inflow resistance for each MPLT data set.These data are using for the halite envelope modeling.
The THP and the reservoir pressure from the MPLT analysis are cross-plotted, representing the red dots, as shown in Figure 3, along with the current operating points (yellow dots) of same wells.
All the MPLT data were acquiring before the onset of halite scaling in their corresponding wells and therefore they represent the no-halite scaling points.All the examined wells had experienced halite scaling and thus the current operating points represent the halite scaling points.Figure 3 shows that most of the no-halite scaling points fall at relatively high reservoir pressures and THP above 20 bar.In contrast, the halite scaling points all falls at reservoir pressure less than 150 bar and THP between 12 and 18 bar.
As shown in Table 2, the majority of the analyzed wells have A-factor values between 10-100 (70% of the total population), whereas 15% have low A-factor values and 15% have very high Afactor values.

Halite Envelopes modeling and validation
The Halite Prediction Tool was using to model halite envelopes for 30 wells (HPT).The MPLT data utilized for halite envelopes modeling were collecting prior to the appearance of halite scaling.Examples of modeled halite envelopes and their validation will provided, evaluated, and extensively discussed in this section.
Figure 4 illustrates the halite envelope of well A15 (blue solid line) modeled using MPLT data acquired in 2011 along with the operating point at the time of the MPLT acquisition (grey dot) and the current operating point of the well (orange dot).The X-axis shows the reservoir pressure range at which the precipitation of halite scaling is expecting with the corresponding wellhead pressure (THP) in the Y-axis.The well has completed using 4.7'' completion tubing and still produces from this tubing.The MPLT data estimates the inflow resistance in this well to be 5.4 bar 2 /E3Sm 3 /d, which falls in the range of the good quality rock.The field data shows that the onset of halite scaling has started in 2018 and the halite scaling issue is still ongoing.As illustrated in Figure 4, the MPLT point falls above the envelope indicating that the halite scaling have not anticipated at the time of acquiring the MPLT.Whilst the current operating point falls below the envelope which indicates that the well is currently prone to halite scaling.Given that the MPLT data was collecting prior to the onset of halite scaling and the well is currently experiencing halite scaling, the model matches the field data.Figure 5 shows the modeled halite envelope of well A2.The well was completed using 4.9" tubing ID and it has one MPLT data acquired in 2015.Analyzing the MPLT shows that the inflow resistance of this well (A factor) is 4.8 bar 2 /E3Sm 3 /d (good quality rock).The well has started experiencing halite scaling in 2018.As illustrated in Figure 5, the MPLT data falls in the no-halite scaling region while the current operating point is located inside the halite-scaling region, which indicates that the model matches well with the real data.The maximum reservoir pressure at which halite is predicting to start is 62 bar with maximum THP of 17.8 bar.ONSET THP (BAR)

RESERVOIR PRESSURE (BAR)
Halite envelop for 4.9'' MPLT 2015 OP Figure 6 presents the halite envelope of well A41, which had modeled using MPLT data acquired in 2015.This well has completed with 4.9" tubing ID.Based on the MPLT, the inflow resistance (A) of this well is 24.6 bar 2 /E3Sm 3 /d indicating a moderate rock quality.The production data indicates that the well started experiencing halite scaling in 2018.Figure 6 shows that the MPLT point falls above the halite envelope indicating that the well was halite scaling free at the time of the MPLT collection.On the other hand, the current operating point falls inside the halite-scaling region, which supports the fact that the well is currently scaling.Compared to the previous two envelopes shown in Figure 4Figure 5, the envelope of this well is significantly larger with maximum reservoir pressure of 168 bar and maximum THP of 53 bar.
Figure 6.Shows the modeled halite envelope of well A41 with 4.9" tubing completion ID.
Figure 7 shows two halite envelopes of well A3 modeled using MPLT data collected in 2010.The blue solid line represents the halite envelope of the original completion tubing, while the orange line illustrates the halite envelope of well after reducing the tubing size by installing a velocity string.The well A3 had initially completed by 3.7" tubing ID.In 2016, 2.44" VS string has installed.Analyzing the MPLT data of this well shows that the inflow resistance in this well is 171.2 bar 2 /E3Sm 3 /d indicating a poor rock quality.The field data shows that the onset of halite scaling in this well was in 2016 and the well is still suffering from the issue.As illustrated in Figure 7, the model predicts that the halite scaling had not expected when MPLT data was collecting.It also demonstrates that the present operating point is within the halite-scaling zone, suggesting the existence of a problem, which corresponds to reality.Furthermore, as indicated in Figure 77, installing the velocity string reduces the halite envelope from a maximum reservoir pressure of 277 bar and a maximum wellhead pressure (THP) of 68 bar to a maximum reservoir pressure of 157 bar and a maximum THP of 43 bar.Because reducing the tubing size reduces the region of halite scaling, the risk of halite scaling is also reducing.Figure 8 depicts the predicted halite envelopes of well A29 using MPLT data from 2013.The blue line represents the halite-scaling envelope of the original completion tubing (4.2") and the orange line represents the halite-scaling envelope of the well after installing a velocity string of 2.44" ID.
The inflow resistance of this well based on the MPLT data is estimating to be 14.5 bar 2 /E3Sm 3 /d and indicating a moderate rock quality.The production data shows that the well started experiencing the halite scaling issue in 2016 and the issue still exists.A velocity string (2.44" ID) was installing in 2021.When the MPLT data was gathering in 2013, the model predicted that the well would not experience halite scaling.Furthermore, the model reveals that the well's present operating point is within the halite-scaling region of the velocity string envelope, indicating that the well is currently scaling, which is consistent with reality.The model also shows that installing velocity string reduces the size of the halite scaling envelope significantly, and the therefore the risk of halite scaling also reduced.Nevertheless, in this well, the velocity string installation in 2021 has not yielded to cease or delay the issue.The halite envelopes of well A6 were modeled using MPLT data collected in 2012 as shown in Figure 9.The inflow resistance of this well is 28.6 bar 2 /E3Sm 3 /d (Moderate rock quality).The well was initially completing with 3.7" tubing ID.However, in 2017 the tubing size had reduced by installing a velocity string of 2.44" ID.The production data shows that the onset of halite scaling in this well started in 2016, however, no halite scaling have been reported following the installation of the velocity string.The model outcome predicts that the MPLT point and the current operating point fall outside the region of the halite scaling which matches the real data.Moreover, analyzing the halite envelope of the velocity string shows that the halite scaling will re-occur in this well when the reservoir pressure declines to 54 bar assuming that the well will be operating at the current wellhead pressure (THP=13.5 bar).Increasing the THP will lead to delaying or even terminating the halite scaling in this well.On the other hand, operating the well at a lower THP will accelerate the occurrence of the issue in the well.The modeled halite envelopes of well A18 are illustrating in Figure 10.The modeling was conducting using MPLT acquired in 2012.The inflow resistance of this well is 11.6 bar 2 /E3Sm 3 /d.The well was originally completing with 4.7" tubing ID and the tubing size was reducing in 2021 by installing a 2.44" velocity string.The production data shows that the well started halite scaling in 2016.However, the issue was ceasing following the installation of velocity string.As clearly illustrated in Figure 10, the model predicts no halite when the MPLT data was collecting.It also predicts that the well is currently producing without halite scaling.The model prediction does not contradict with the reality.The model also predicts the re-occurrence of halite scaling issue when the reservoir pressure declines to 28 bar if the operating wellhead pressure (THP) is maintaining at 13 bar.Additionally, the model shows that the re-occurrence of the halite scaling issue can be avoided by increasing the THP by a minimum of 1 bar.Reducing the THP, on the other hand, will accelerate the reoccurrence of the issue.

Generalizing Halite Envelope Model
Having modeled and verified the halite envelopes for all 30 wells, the envelopes with their corresponding MPLT and the current operating points had plotted as shown in Figure 11.The figure shows that the size of the halite envelopes varies significantly indicating that generalizing the halite envelope for the whole field is not feasible.However, it was observing that the size of the envelopes is affecting by the inflow resistance of the well and therefore it has decided to group the halite envelopes based on the inflow resistance.

Grouping the halite envelopes
The wells are dividing into three groups.Group 1 includes the wells with A values range from 1 to 10 bar 2 /E3Sm 3 /d and represents the good quality rock.Group 2 represents the wells with moderate rock quality and A values range from 10 to 100 bar 2 /E3Sm 3 /d while group 3 characterizes the wells with tight rock properties and A values greater than 100 bar 2 /E3Sm 3 /d.The halite envelopes of each group of wells were plotting separately with their corresponding MPLT data and the current operating point.Figure 12 shows the modeled halite envelopes of all group 1 wells.Figure 13 illustrates the halite envelopes of group 2 wells while envelopes of group 3 wells are seeing in Figure 14.These figures were using, as described in the next section, to define the base case halite envelope for each group of wells.

Defining the base case envelopes
The definition of the base case envelope of each group had subjectively done.The smallest and the largest envelopes of each group were discarding as they represent the extreme cases.By visual inspection, the base case envelopes of the three groups were selecting as shown in Figure 15 through Figure 17.

Blind testing the base case envelopes
The base case envelope of each group of wells was blind tested using MPLT data from 19 wells that were not utilizing in generating the base case envelopes.17 MPLT points were collected before the onset of halite scaling issue, and these represent the no-halite points.On the other hand, two MPLT data points were collecting after the start-up of the issue, and these represent the halite scaling points.Additionally, the current operating points of the 19 wells were also using in the blind test and represent no-halite points.
Figure 18 shows the base case envelope of group 1 wells with the no-halite points (red dots) and the halite point (yellow dots).The visual inspection shows that all the no-halite points fall above the halite envelope and all the halite scaling points fall below the envelope.The model prediction of this group of wells matches 100 % with the reality.
Figure 19 shows the base case of halite envelope of group 2 wells with the data points used for the model validation.As shown in the Figure 19, 6 out of 9 of the no-halite points fall above the envelope while only three points fall below the envelope.It also shows that 12 out of 13 halite points fall below the base case envelope while only one point had not accurately predicted.
Figure 20 depicts the blind test on the group 3 envelope.As seen in Figure 20, two of the four nohalite points lie inside the no-halite region, while the other two fall within the halite scaling.Furthermore, all of the halite scaling points have predicted accurately.The accuracy of the models' prediction of group 1, group 2 and group 3 have been measured using confusion matrix as illustrated in Table 3 Table 4 Table 5, respectively.
As shown in Table 3, the prediction of base case envelope of group 1 wells shows 100% accuracy.However, the accuracy of the base case model deteriorate, as the rock quality gets poorer.The base case model for group 2, as shown in Table 4, shows 92.3 % accuracy in predicting the halite scaling (Tp), and 66.6 % accuracy in predicting the no-halite scaling (Tn).It wrongly predicted the halite scaling with the percentage of 7.7% (Fp) and failed to predict 33.3 % of the no-halite cases (Fn).The halite envelope of group 3 showed the least accuracy in predicting the no-halite scaling cases as shown in Table 5.It predicted the halite-free cases correctly in 50 % of the cases (Tn), while it failed to predict in the remaining 50% (Fn).On the other hand, the model predicted the halite scaling all cases (Tp).
Table 3 shows a confusion matrix quantifying the accuracy of the based case halite envelope of Group 1 wells.Analyzing the MPLT data and current operating points of the wells experiencing halite precipitation problems reveals that the halite scaling occurs at depleted reservoir pressures and low THP pressure, which is consistent with the findings of the literature review.The findings show that halite precipitation occurs often at reservoir pressures less than 150 bar, given that the wells are operated at THP ranging from 12 to 18 bar, which is consistent with the findings of Wat et al., 2010;and Goodwin et al., 2016.Additionally, examining the wells' current operating points shows that the significant reduction in the wellhead pressure (THP), caused by introducing a surface depletion compression for example, leads to the decrease in the bottomhole pressure, which in turn increases the drawdown between the reservoir pressure and the bottomhole hole pressure that triggers the halite precipitation.This observation matches well with what was highlighted by Kleinitz et al., 2001;Aquilina (2012) and Veeken et al., 2019 Generally, the prediction of halite envelopes modeled by HPT-V2 matches perfectly with the real data.Different well conditions were examining yet the model prediction matches the field observations.A range of inflow resistance, 4.8 to 210 bar 2 /E3Sm 3 /d, was examining and the results shows a high percentage of correct prediction.In addition to that, there was variation in the tubing size (ID = 3.7" -4.9") of the modeled wells and some of the wells have velocity string installed, yet the prediction was correct.Nonetheless, the high uncertainty of formation water gas ratio (FWGR) can limit the effectiveness of the tool.In this study, the halite envelopes were modeled using the minimum possible value of FWGR and in most cases the model achieved the best match at FWGR=10., In some wells, however, the best match was achieved at very high FWGR, exceeding the reported WGR, and therefore considered unreliable and discarded from further analysis.Additionally, increasing the formation water gas ratio (FWGR) leads to reduce the size of the halite envelope and hence reduces the risk of the halite scaling issue in the well.This had aligned with the literature.
The range of reservoir pressure and THP, at which halite scaling is triggered, differs from one well to another, as indicated by analysis of the halite envelopes of several wells.The halite envelopes of some wells are significantly large while some envelopes are relatively small.The inflow resistance of the well (A) clearly affects the shape of the Halite envelope.The smaller the inflow resistance (good quality rock) the smaller the Halite envelope and hence the less prone to Halite precipitation.
The risk of the halite scaling is decreasing by the installation of velocity string because the size of the halite envelopes had reduced.Installation of velocity strings results in a reduction in tubing size, which reduces drawdown by increasing bottomhole pressure.
It has demonstrated that constructing a universal halite envelope for the entire area to be using for halite prediction is not feasible.This is because the occurrence of halite precipitation is strongly dependent on well conditions, which is completely consistent with the findings of the literature study [8].
Many characteristics can influence the halite-scaling tendency, hence controlling the form and size of the modeled halite envelope.One of the key contributions to the variance in the size of the halite envelope is inflow resistance (A-factor).Wells with a high A-factor have huge halite envelopes, indicating that they are more susceptible to the halite-scaling problem.
However, clustering the wells based on their inflow resistance and picking a representative envelope for each group proved to be a valuable technique for predicting the halite-scaling problem.The blind test reveals a wide difference in model performance in terms of model prediction accuracy.The model prediction of group 1 wells showed very high accuracy in predicting the halite and no halite cases (100%).The high accuracy of this model can explained due to narrow range of the inflow resistance (1-10) compared to the other two groups and thus the variation in the envelopes' shape and size is relatively minimal and therefore the selected base case is representative.However, it can argued that only a handful of wells were used to validate the model and this validation is yet not enough to draw a conclusion about the model performance.The performance of the group 2 halite envelope model was fair.It correctly predicted the no-halite scaling with 66.6 %, while failed to predict the no-halite in 33.3 %.On the other hand, its performance in predicting the halite scaling cases was observing to be reasonably high with 92 % accuracy.The accuracy of the group 3 halite envelope model is poor compared to group 1 and group 2. It correctly predicted 50 % of the no halite scaling cases and failed to predict in 50 % cases.The low performance of this model can be due to the wide range of the A-factor of this group of the wells compared to group 1 and group 2. Additionally, only three wells were using to construct the model, which do not cover the full spectrum of the group-3 wells.Therefore, group 3 model needs to improve by using more MPLT data from more wells.

The Impact of Velocity String on Halite Scaling
In this section the impact of installing velocity string on halite scaling, on term of halite delay time T(Halite), will be evaluated, quantified, analyzed, and discussed.

Candidates' selection
To achieve this objective, 5 wells were carefully selected.The main basis of selecting these wells as candidates for the full analysis is to have a representative well at least from each group, i.e. good rock (group-1), moderate rock (group 2) and tight rock (group-3).Table 6 summarizes the selected wells with their corresponding inflow resistance and the group each well belongs to.Calculating the delay time of halite scaling after installing the velocity string can conducted in three steps as described below.

Halite envelopes modeling and analysis
In this step, two halite envelopes, one for 2.44" velocity string and one for 1.99" velocity, for each well were modeled and analyzed to estimate the pressure at which halite scaling re-occur (P(halite)) for each type of velocity string.
Figure 21 illustrates the modeled halite envelopes for well A45 modeled based on MPLT data acquired in 2016.The model shows clearly that installing either velocity string leads to cease halite scaling at current operating conditions.The model also predicts that the halite scaling issue will re-occur when the reservoir pressure declines from the current reservoir pressure, 110 bar, to 83 bar in case of 2.44" VS is installing and when reservoir pressure declines to 52.3 bar for the case of 1.99" VS is installing, given that the wellhead pressure (THP) is maintained at 13.5 bar.
The modeled halite envelopes of well A15 are illustrating in Figure 22.These envelopes were modeled using MPLT data collected in 2011.The production data indicates that the onset of halite scaling in this well has started in 2018 and the well is still under water treatment.The MPLT point falls outside the region of halite scaling, indicating that the well was not expecting to experience halite scaling at the time of collecting the MPLT data.The current operating point falls below the envelope of the 3.8" tubing indicating that the well is currently scaling.The model predictions match well with the real data and thus the model is considering reliable.Installing velocity string, either 2.44" or 1.99", will result in ceasing the halite scaling as per the model prediction.However, the halite scaling issue will not terminated permanently, and will reoccur in the future according to the model.The re-occurrence of the issue will happen in this well when the declining reservoir pressure reaches 34 bar, if 2.44" velocity string is installing, or reaches 23 bar in the case of installing 1.99" velocity string.
Figure 23 presents the halite envelopes modeling of well A13 based on MPLT data acquired in 2013.Well A13 was completing with 4.7" tubing ID and no further re-sizing for the tubing was performing.The well started suffering from halite scaling issue in 2013.As illustrated in Figure 23, the model does not anticipate halite scaling prior to the acquisition of the MPLT data but predicts the issue at the current operating conditions, which matches perfectly with the real data.
According to the modeling outcome, installing velocity string will delay the halite scaling issue in this well.In the case of installing the 2.44" VS, the halite scaling issue will re-appear when the reservoir pressure declines to 46 bar, and is predicted to re-occur when the reservoir pressure reaches 23 bar with 1.99" VS.
Figure 24 illustrates the modeled halite envelopes of well A13 based on MPLT data acquired in 2010.According to production data, the well has started experiencing halite-scaling precipitation in 2016.The model predicts that the MPLT data falls in the halite-scaling free area and therefore halite scaling was not predicting in 2010.It also predicts halite scaling at the current operating condition.As shown in Figure 24, the current operating point also falls inside the halite-scaling region for both velocity strings scenarios, indicating that reducing the drawdown by velocity string installation will not lead to any delay in halite scaling.
Figure 25 also provides another example of the ineffectiveness of installing velocity string to manage halite precipitation.It shows the modeled halite envelopes of well A34 using the MPLT data acquired in 2010.The model does not predict halite scaling at the time of acquiring MPLT data and predicts halite at current conditions.The model predicts that installing a velocity string (2.44" and 1.99") will not terminate the halite scaling issue at all.Modeling and analyzing the halite envelopes in the previous step provides an insight into the potential pressure at which halite will re-occur, P(halite), which will be used as an input for the next step as summarized in Table 7.

Nodal Analysis for the Candidate Wells
Following the definition of P(halite) of the selected wells, the nodal analysis was carried out for each well as described in this section.
In this step, the nodal analysis was carried out for the 3 wells in which installing velocity string is predicted to delay the halite scaling.Two wells (A34 and A32) were not taking for further analysis as the model predicts halite scaling even if velocity string is installing.
The aim of this step is to computing the outflow parameters (B and C).The current gas rate if velocity string is installed, denoted as Q1, at the current reservoir pressure (P1) and the gas rate when halite scaling occurs, denoted as Q(halite), when reservoir pressure declines to the pressure at which halite scaling is expected to re-occur (P(halite).All these parameters will used in the next step to estimate the delay time.
The nodal analysis was performing for two scenarios, the first one when 2.44" velocity string is installing and the second one for 1.99" velocity string.
Table 8 and Table 9 summarizes the outcomes of the nodal analysis done for scenario 1 and scenario 2 respectively.

Decline curve analysis and halite delay time estimation
As mentioned earlier, the halite scaling will re-occur when Q1 declines to Q(halite) as the reservoir pressure depletes from P1 to P(Halite).By assuming exponential decline and using E.Q ( 12) and E.Q ( 14), the outcome of this step is summarizing in Table 10 and Table 11.Generally, installing velocity string leads to reduce the size of the halite envelopes and therefore the likelihood of experiencing halite precipitation is reducing as clearly demonstrated in the previous examples.The installation of velocity string reduces the completion tubing size, and consequently increases the bottomhole pressure and reduces the drawdown.As clearly covered in the literature review, the evaporative scaling is highly dependent on the gas expansion, which in turn depends mainly on the drawdown.The increase in the drawdown leads to the increase in the gas expansion and accelerates the evaporative scaling.On the other hand, reducing the drawdown tends to limit the gas expansion and hence deaccelerate the occurrence of the issue.
The results also demonstrate that the effectiveness of VS to manage halite scaling is highly controlled by the inflow resistance, and hence the rock quality, of well.Installing velocity string in wells with small value of inflow resistance, good quality rock, leads to cease the halite scaling for longer time compared to installing it in wells with higher value of A.
Using smaller VS leads to further delay in the reoccurrence of the halite scaling issue.However, the well capacity is reducing significantly and leads to production deficiency.
The timing of installing velocity string dictates its effectiveness in managing halite precipitation.A late installation might not cease the issue as demonstrated in well 32 and well A34.On the other hand, a very early installation of velocity string can compromise on the well capacity.The optimum timing of installing velocity string is the time when the production loses due to production loses because of reducing the well capacity is offset by the deferment caused by halite precipitation and the cost incurred to remove the salt.Therefore, a very detailed economic study needs to define the optimum timing of velocity string installation.

3.3
Evaluating the synergy between halite management and de-liquification using velocity string This objective has achieved by following two steps, the first step was to calculate the time of liquid loading and the second step is to compare the delay time caused by installing the velocity string and the time to liquid loading.

Estimating the time of the liquid loading onset
Estimating the time of the liquid loading issue requires three steps.Firstly, the nodal analysis is conducting using original well diameter and assuming no velocity string is installed.The results of this step are summarizing in Table 12.
Table 12 shows the outcome of the nodal analysis of the examined wells using the original tubing diameter and assuming no velocity string is installing.Following the nodal analysis, the second step is to calculate the Qmin and the Pmin for each well as shown in Table 13.The third step is to perform decline analysis, assuming exponential decline, to estimate the ultimate recovery and the time required the well to load up.Table 14 presents the time of the liquid loading onset in the examined wells.The impact of installing velocity string on the onset of liquid loading was evaluating by repeating the above-mentioned three steps assuming the examined wells are operating through 2.44" and 1.99" VS instead of the original completion tubing.The minimum flow rate (Qmin) and the minimum pressure (Pmin) of each well were estimated for the case of 2.44" and 1.99" VS as summarized in Table 15.This was following by estimating the ultimate recovery and the time of onset of liquid loading given 2.44" or 1.99" velocity string is installing as summarized in Table 16.The impact of velocity string on the minimum flow rate (Qmin), the minimum reservoir pressure (Pmin), the ultimate recovery (UR) and the hence the time of liquid loading occurrence are illustrated in Figure 26 through 29 respectively.Figure 26 and Figure 27 show that the minimum flowrate and the minimum pressure of all the three wells will decrease significantly if velocity string is installing.Additionally, the wells' ultimate recovery will enhanced upon the installation of the velocity string as shown in Figure 28 and the time of the liquid loading onset will delayed dramatically as explicitly illustrated Figure 29.

Comparison between the time to liquid loading and the halite reoccurrence time
As described earlier, the synergy between halite management and de-liquification have guaranteed when there is an overlap between the time to liquid loading and the halite precipitation reappearance time.Table 17 and Table 18 show the comparison between the time to liquid loading the halite scaling re-occurrence time given 2.44" VS and 1.99" VS installed respectively.
Given that installing velocity in well A32 and A34 (with A-factor being 97.7 and 118.9 respectively) has no impact on halite scaling, the synergy cannot be assured for these two wells for both 2.44" and 1.99" VS.Installing 2.44" VS in well A45 (A-factor =56.3) will delay the halite precipitation for 3.1 year, while liquid loading will take place after 5.5 year given the current production conditions are maintained and no modification on the tubing size is done.Therefore, the synergy between halite management and deliquification using 2.44" velocity string is not applicable in this well.However, the halite delay time caused by using the 1.99" VS exceeds the liquid loading timing and therefore the synergy is this case is possible.Moreover, the synergy can be generated in wells A15 and A13 (with A-factor of 5.4 and 14.2) by either using 2.44" or 1.99" VS as the halite delay time resulted from the VS installation exceeds the liquid loading timing.Additionally, the 1.99" VS tends to delay the halite reoccurrence approximately 3 times longer than the delay time caused by the 2.44" VS as clearly shown in the tables.The feasibility of synergizing halite management and deliquification using the velocity string technology is highly dependent on the inflow resistance (A-factor) and the internal diameter of the installed velocity string.Evaluating wells with different A-factor values revealed that the higher the A-factor the less the potential of the generating the synergy.For wells with tight rock, installing velocity string has no impact on the halite scaling and only beneficial for deliquification purposes.
On the other hand, installing velocity in wells with good rocks (with A-factor values less than 10) leads to delay the halite re-occurrence significantly that exceeds the time to liquid loading given that the well is operated via the original completion tubing.
The internal diameter of the installed velocity string can also play a significant role in dictating the synergy between the halite management and deliquification.The smaller velocity string's diameter, the longer the halite re-occurrence time and the more likelihood the synergy is plausible.
Installing smaller velocity string leads to increase the BHP significantly and reduces the drawdown.As covered in the literature, the halite precipitation is dominating by the drawdown, and the higher the drawdown caused by installing the velocity string, the longer it takes the well to re-experience halite scaling and thus the more synergy can generated.However, this tends to

Figure 1 .
Figure 1.Shows the flowchart of the method used in this study.

Figure 2 .
Figure 2. shows a schematic diagram of the halite envelope.

Figure 3 .
Figure 3. Depicts MPLT and current operating conditions for well analysis: no halite scaling vs. halite scaling.

Figure 4 .
Figure 4. Shows the halite envelope of well A15 with 4.7-inch completion ID (blue solid line), the MPLT point (grey dot) and the well current operating point (orange dot).

Figure 8 .
Figure 8. Displays the modeled halite envelopes for well A29 with 4.2" tubing completion ID (blue) and 2.44" ID velocity string (orange), alongside the current and MPLT operating points.

Figure 11 .
Figure 11.Illustrates modeled halite envelopes for all wells with MPLT and current operating points: no halite scaling (red) vs. halite scaling (yellow).

Figure 14 .Figure 15 .
Figure 14.Reveals halite envelopes for Group 3 wells with MPLT (red dots) for no-halite points and current operating points (yellow dots) for halite scaling.

Figure 16 .Figure 17 .
Figure16.Shows the selected base case halite envelope for group 2 well (moderate quality) with the MPLT and the current operating points of all the wells used to construct the model.

Figure 18 .Figure 19 .Figure 20 .
Figure 18.Shows the blind test performed to validate the base case envelope for group 1 rocks (good quality rock)

Figure 25 .
Figure 25.Shows the modeled halite envelopes of well A34.

Figure 26 .
Figure 26.Shows the impact of installing velocity string on the minimum flowrate in the examined wells.

Figure 27 .Figure 28 .Figure 29 .
Figure 27.Shows the impact of installing velocity string on the abandonment pressure (Pmin) in the examined wells

Table 1
shows an example of a 2 x 2-confusion matrix used to quantify the model accuracy.

Table 2
shows the outcome of the inflow performance analysis of the MPLT data

Table 6
shows the selected candidates for the evaluation

Table 7
provides information about the selected wells and their corresponding P(halite) with

Table 8
Table10summarizes the outcomes of calculating the time of halite re-occurrence for scenario 1

Table 15
summarizes the calculated minimum flow rate and minimum pressure of the examined wells if velocity string is installing.
Table17shows the comparison between the time to liquid loading and the halite delay time given 2.44" VS is installedTable18shows the comparison between the time to liquid loading and the halite delay time given 1.99" VS is installed