SPE-185694-MS

Errors Associated with Waterflood Monitoring Using the Hall Plot for
Stacked Reservoirs in the Absence of Profile Surveys

Qianru Qi, Sophie Pepin, Abdula AlJazzaf, and Iraj Ershaghi, University of Southern California

Copyright 2017, Society of Petroleum Engineers

This paper was prepared for presentation at the 2017 SPE Western Regional Meeting held in Bakersfield, California, USA, 23 April 2017.

This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents
of the paper have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect
any position of the Society of Petroleum Engineers, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written
consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may
not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract
It is common to use the slope of Hall plot as a tool for monitoring changes in well injectivity. Loss of
injectivity at a given well head pressure may relate to a build-up of reservoir pressure or a gradual build-
up of skin. In this study, we focus on injectivity losses caused by the wellbore skin development. Based on
actual field data, we had noted errors in detection of permeability losses when later profile surveys were
consulted. Our objective of the study was to examine the problem and then introduce an improved analytical
formulation for layered systems and assess the limitations and meaning of slope changes observed on the
Hall plot for stacked reservoirs. For calibration purposes, we used both an extended analytical model and
a simulation approach to demonstrate the weakness of the Hall Plot for multi-layered cases. In simulating
the cases, we used a cumulative injection dependent function to model the permeability losses for a given
layer in a stacked system, and examined the limitation of estimated slopes in regarding to indication of the
location and extent of the damaged zones in a stacked reservoir for remediation measures. We included a
case study to exemplify the situation.

Introduction
Waterflooding has been one of the most common methods for improved recovery and pressure maintenance
of semi-depleted reservoirs. Effective pressure management and surveillance are essential to maintaining
successful waterflood operations. For waterfloods, one needs to properly manage voidage replacement
ratios (VRR) for pressure maintenance purposes. This will help to avoid the production of excessive
solution gas and prevent reservoir subsidence for poorly consolidated systems. As such, injection wells
need constant monitoring. However, injectivity losses may develop because of reservoir pressure buildup
or skin development caused by water compatibility issues, clay swelling, high TDS of injection fluid and
fines migration.
In this study, we focus on injectivity losses due to skin buildup. Houston et al. (2006) discussed plugging
effects that can occur when the injection water and reservoir water are incompatible, containing ions that
precipitate out of solution and create a skin effect around the well. Clay swelling is another issue that
can cause injectivity losses. Common water sensitive clays may swell if the injection water is not at an
appropriate salinity. Barkman and Davidson (1972) investigated the effects of injection water with high
2 SPE-185694-MS

content of total dissolved solids, where the injection water may deposit the solids in the injected layer
causing the formation of skin. Whichever are the causes for injectivity losses, it is imperative to monitor
injectivity reductions that could otherwise impact the maintenance of the VRR.
The Hall Plot (Hall3, 1963) method, based on the simple Darcy's equation, was originally proposed for
montitoring injectivity losses and gains for a single layer. This plotting system has been used for many
multilayered waterfloods such as the turbidite formations in California. We have examined the limitations
for California sands and a carbonate reservoir under waterflooding in Kuwait, and have noted that the
commonly used Hall Plot by itself is inadequate in capturing the impact of one or more layers which may
be the most affected in permeability losses.
The Hall plot assumes a succession of steady state flows and only requires injection volume and wellhead
pressure data. It also assumes radial flow through a single layer. Over the years, this method has been
revisited and discussed by various authors. For instance, Silin, Holtzman and Patzek4 studied the effect of
reservoir pressure on Hall Plot. Silin et al5 further applied integration of Hall Plot and other tests as well as
InSAR data to supervise water injection operation to avoid subsidence.
An illustration of slope monitoring of the Hall Plot is shown in Figure 1. This figure shows the Hall plots
for a three injection wells, each exhibiting slope changes, which often indicates a gradual loss in injectivity.
Each of the injectors pressurized stacked sands, and merely based on the corresponding Hall plots one
cannot identify which specific layer may be experiencing a problem in injectivity.

Figure 1—Examples of Slope Changes Observed for Stack Reservoirs Injection Data on a Hall Plot.

In summary, if waterflooding operations involve multi-layered systems with a comingled injection, the
practice of the Hall Plot treating a multi-layer reservoir as a single layer may lead to misleading results.
The conventional slope analysis cannot guide the injectivity changes of problematic one or more layers
among the whole stacked sands. Oftentimes, massive acidizing jobs are implemented to treat all the layers.
As such the layers with less or no injectivity loss may have a better flow capacity taking more acids than
the problematic layer(s). This results in unnecessary high-cost acid jobs.
In the paper, we first introduce the analytics and define the slope of Hall Plot for multi-layer cases. We
then show simulated results of a single layer formation experiencing skin development on the Hall plot.
Finally, actual case studies are included to show the importance of incorporating profile surveys when
interpreting the Hall plots.
SPE-185694-MS 3

Analytical Solution to Multilayered Systems for the Hall Plot

According to the Darcy's Equation, for one layer the injection volume I is,

(1)

When injecting into n layers, the total flow injection volume Itot is,
(2)
As such for a number of intervals taking the fluid we obtain:

(3)

Factoring out the change in pressure ΔP = Pbh – Pres,

(4)

leads to:

(5)

Rearranging results we obtain:

(6)

If integrating over the injection duration t, we obtain:

(7)

To address the actual process used in the field in terms of monitoring the Hall plots, now we express the
Hall plot equation in terms of wellhead pressure Pwh.
We know that the pressure difference between wellhead pressure Pwh and Pbh is,
(8)
where ρgDb is the hydrostatic head and ΔPf is the pressure loss due to friction, both in psi.
Substitute (9) into (8), we obtain,

(9)

Rearrange (9), we obtain:

(10)
4 SPE-185694-MS

Assume the reservoir pressure Pres does not build up but maintains at a steady level, (ρgDb+ ΔPf) is also
considered constant, then the summation of the two terms is constant,
(11)
Substituting (11) into (10) and integrating on both sides, we get:

(12)

where Icum is the cumulative injection volume into all layers, Σt Pwh(Δt)·Δt is the cumulative pressure-time
in psi*day. By plotting the Σt Pwh(Δt)·Δt vs. cumulative injection volume, the slope a is,

(13)

As seen, the slope of a Hall plot for a multi-layered reservoir includes a summation terms that considers
the permeability and the skin factor of all the layers.
Example
Assume a three-layer reservoir with same thickness of 20 ft. for each layer, and all the layers start with
100 mD effective permeability before skin develops. If the reservoir has an equivalent radius of 4000 ft.
drainage area and contains the oil of 2 cP viscosity. The initial slope on Hall plot in field units is:

If one of the layers develops a skin of 20, the new expected slope becomes,

This means a slope increase may be caused by just one of the layers; and a major weakness of the
composite Hall plot is that, the damaged layer is not identifiable just from the slope analysis.

Numerical Simulation
In this section, the disproportionate effects of commingled layers can be modeled. To maintain VRR at
1:1, the numerical simulation, we used a two-well reservoir, where one well injects water and the other
produces oil at the same gross fluid rate as injection rate. We modeled a three-layer reservoir to demonstrate
the changes of slopes in the Hall Plot when skin develops in one of the three layers. Reservoir parameters
used are shown in Table 1.
SPE-185694-MS 5

Table 1—Properties and Constraints for The Numerical Experiment.

Property Symbol Value Unit

Reserve OOIP 126.3 MSTB

Reservoir Dimension Lx, Ly, Lz 7200 * 7200 * 60 ft * ft * ft
Gridblock Size ΔLx, ΔLy, ΔLz 24 * 24 * 20 ft * ft * ft
Initial Reservoir Pressure Pres 4000 psi
Depth of Top Layer Dtop 9230 Ft
Porosity Φini 0.3 fractional
Initial Oil Saturation So,ini 0.2 fractional
Absolute Horizontal Permeability kx, ky 80 (layer 1,3); 150 (layer 2) mD
Absolute Vertical Permeability kz 10 mD
Oil API Gravity 30 degree
Bubble Point Pressure Pb 2000 psi
Oil Viscosity μo 2 cP
Water Viscosity μw 0.7 cP
Water Injection Rate Itot 3000 STB/D
Total Gridblock Number 270,000

For this case, the three layers are considered to have equal thicknesses. Consider water injection into the
three layers where among the three one has a higher initial permeability. The more permeable layer will
take more water than the other two layers and as such, exhibits injectivity losses by the formation of a zone
of reduced permeability. The assumption made is that the formation of the skin caused either by TDS or
clay swelling, is influenced by the volume of water injected.
To examine the gradual development of homogeneous skin in the numerical test we used the function
by Liu and Civan6 (1993), as shown below:
(14)
They modeled skin development based on the fraction of pores plugged, where kini and Φini are the initial
permeability and porosity; while k and Φ are current permeability and porosity. Parameter (f) is referred to
as the fraction of the initial area open to flow and kf is a constant for flow allowed through plugged pores.
Figure 2 shows the simulation result from the vertical injector. Here when plotting the cumulative pressure
vs. cumulative injection, the slope gradually increases from 1.58 to 1.74 as shown, which matches with the
analytics. The reference line does not have any skin development which is demonstrated by its constant
slope on the Hall Plot, it then curves upward with the loss of injectivity caused by skin development after
6 months of injection in one of the layers.
6 SPE-185694-MS

Figure 2—Hall Plot of the Three-layer Reservoir with One Layer Experiencing Damage.

In Figure 2, the change in the Hall Plot slope is obvious, but it is only caused by one layer of the injection
stack. This shows the necessity of using an independent source of information to identify the affected layer
before corrective steps are taken to mitigate the injectivity losses.

Case Study
Figure 3 and Figure 5 show two Hall plots in a field in Kuwait with a vertical injector, and a horizontal
injector, respectively. Notice slope changes associated with changes in the injectivity. The plot by itself has
a limited resolution to identify the troubled zones. In these cases, fortunately, profile surveys were available
as shown in Figure 4 and Figure 6.

Figure 3—Hall Plot for the Vertical Injector from Kuwait.

SPE-185694-MS 7

Figure 4—Water Profile Survey for the Vertical Injector from Kuwait.

Figure 5—Hall Plot of the Horizontal Injector from Kuwait.

8 SPE-185694-MS

Figure 6—Water Profile Survey for the Horizontal Injector from Kuwait.

In the case of the vertical injector, based on Table 2 and the survey shown in Figure 4, two sections
received more than 99 percent of the fluid and injectivity losses were truly affected by high input layers.
These layers only represent about 50% of the total interval. It becomes clear that any remedial measures
must have access to the profile survey data. Figure 3 shows the changes of injectivity but fail to deliver
the more detailed information.
SPE-185694-MS 9

Table 2—Water Profile Survey Results for Vertical Injector from Kuwait.

Vertical Well Water Profile Survey showing water injected per zone.

Perforation Zone Thickness Water Injected Water Injected (%)

Section
ft ft ft bbl/d %

1 7930 7956 26 13,676 38.2%

7973 7983 10 204 0.6%

2 7983 8010 27 21,954 61.3%

8010 8019 9 0 0.0%

3 8025 8040 15 0 0.0%

4 8059 8066 7 0 0.0%

Cumulative 94 35,834 100.0%

Similar observations are made for a horizontal injector, as shown in Table 3 and Figure 6. Again, it is
the profile survey that shows less than 23 % of the interval is taking more than 86 % of the water, which
cannot be identified from just the Hall Plot in Figure 5.

Table 3—Water Profile Survey Results for the Horizontal Injector.

Horizontal Well Water Profile Survey showing water injected per zone.

Perforation Zone Thickness Water Injected Water Injected (%)

Section
ft ft ft bbl/d %

1 8858 9100 242 3,498 27.2%

2 9100 11300 2200 140 1.1%

3 11300 11450 150 1,610 12.5%

4 11450 11914 464 7,618 59.2%

Cumulative 3056 12,866 100.0%

Conclusions
In the absence of a reservoir pressure buildup becasuse of accumulation of fluids around an injector
necessitating an increase in injection pressures, our derivation of a new Hall plot equation and simulation
work highlights the limitations when the diagnostic functionality of a single layer formulation is used for
layered reservoirs.
The important conclusion we have reached is the absolute necessity of incorporating profile surveys
before decisions are made to address injectivity losses based on the Hall Plot alone. The Hall Plot diagnostic
capability, when related to stacked reservoirs, is limited. A higher degree of reliability in diagnostics of
injectivity loss of a particular layer can be detected only when injection surveys are consulted. This will
avoid employing corrective measures such as large scale acidizing that will unnecessarily treat all the layers.

SI Conversion Factors

bbl × 1.5899 E-01 = m3

cP × 1.0 E-03 = Pa·s
day × 1.1574 E-05 =s
10 SPE-185694-MS

ft × 3.048 E-01 =m
mD × 9.869 E-16 = m2
psi × 6.8948 E+00 = kPa

Nomenclature
a = slope in Hall plot, psi*day/STB
acurrent = current slope in Hall plot with skin, psi*day/STB
aini = slope in Hall plot without skin, psi*day/STB
B = formation volume factor of injected water, rb/STB
C = constant in Hall plot, ΔPf + ρgDb + Pres, psi
Db = depth of the bottom layer in the stacked reservoir, ft
Dtop = depth of top layer in the stacked reservoir, ft
f = fraction of initial area open to fluid flow in porous media, dimensionless
h = layer thickness, ft
i = subscript index for layer i
l = water injection rate, STB/day
Icum = cumulative water injection volume into all layers, STB
Itot = total water injection rate into all layers, STB/day
k = permeability, mD
kf = constant of flow allowed through plugged pores, dimensionless
kini = initial permeability, mD
kx, ky = horizontal permeability, mD
kz = vertical permeability, mD
n = total number of layers in the stacked system, count
Pb = bubble point pressure, psi
Pbh = injection bottomhole pressure, psi
Pres = reservoir pressure, psi
Pwh = wellhead pressure, psi
re = reservoir radius, ft
rw = wellbore radius, ft
S = skin factor, dimensionless
So,ini = initial oil saturation, dimensionless
t = injection time, day
ΔP = pressure difference, Pinj – Pres, psi
ΔPf = frictional pressure loss in tubing, psi
Δt = time interval within which a certain Pwh is maintained, day
Φ = instantaneous porosity, dimensionless
Φini = initial porosity, dimensionless
ρg = hydrostatic gradient, psi/ft
μ = viscosity of injected water, cP
μo = viscosity of oil, cP
μw = viscosity of formation water, cP
SPE-185694-MS 11

References
1. Houston, S. J., Yardley, B., Smalley, P. C., & Collins, I. (2006, January 1). Precipitation And
Dissolution Of Minerals During A Waterflood - The Evidence Of Produced Water Chemistry
From Miller. Society of Petroleum Engineers. doi:10.2118/100603-MS.
2. Barkman, J. H., & Davidson, D. H. (1972, July 1). Measuring Water Quality and Predicting Well
Impairment. Society of Petroleum Engineers. doi:10.2118/3543-PA.
3. Hall, H.N. How to Analyze Waterflood Injection Well Performance. World Oil (1963, Oct.)
128–30.
4. Silin, D. B., Holtzman, R., & Patzek, T. W. (2005, January 1). Monitoring Waterflood
Operations: Hall Method Revisited. Society of Petroleum Engineers. doi:10.2118/93879-MS.
5. Silin, D. B., Holtzman, M. J., Patzek, T. W., Brink, J. L., & Minner, M. (2005, January 1).
Waterflood Surveillance and Control: Incorporating Hall Plot and Slope Analysis. Society of
Petroleum Engineers. doi:10.2118/95685-MS.
6. Liu, X., & Civan, F. (1993, January 1). Characterization and Prediction of Formation Damage in
Two-Phase Flow Systems. Society of Petroleum Engineers. doi:10.2118/25429-MS.