SPE-191185-MS

Comparison of Various Rate-Decline Analysis Models for Horizontal Wells

with Multiple Fractures in Shale gas Reservoirs

Yanan Miao and Xiangfang Li, China University of Petroleum; John Lee, Texas A&M University; Chaojie Zhao and
Yunjian Zhou, China University of Petroleum; Han Li, Texas A&M University; Yucui Chang, China University of
Petroleum; Wenji Lin, Zhihua Xiao, and Nan Wu, PetroChina Coalbed Methane Company Limited

Copyright 2018, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE Trinidad and Tobago Section Energy Resources Conference held in Port of Spain, Trinidad and Tobago, 25-26
June 2018.

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
In recent years, improving the accuracy of production forecast in unconventional reservoirs has been of
growing interest to oil and gas industries. Decline curve analysis (DCA) models have been recognized as the
most efficient and easiest approaches to estimate gas rate. However, fluid flow regime and well rate decline
curves are highly affected by the geological properties of formations. Therefore, the selection of DCA
models based on completion designs and geological properties of formations is important for production
rate prediction.
Traditional DCA methods, particularly Arps' decline model, was originally developed for predicting
boundary dominated hydrocarbon well rate decline, which differs from the dominant long-duration transient
flow regime in shale reservoirs. The Stretched Exponential model, the Duong model, the Arps model with
a minimum terminal decline rate and the scaling method by Patzek were developed to match and forecast
wells with transient flow followed by boundary dominated flow (BDF). In this paper, firstly we developed a
new model to estimate production in shale gas reserviors by considering both Knudsen diffusion of bulk gas
and surface diffusion of adsorbed gas based on the traditional equation of rate versus square-root-of-time.
This proposed model can provide better fits to data in transient linear flow regimes. In addition, a systematic
analysis of numerical simulation cases in CMG were performed to compare with the traditional model.
The results demonstrated that, in most cases, our model which is demonstrated in this paper, provide
more accurate estimation of reserves for numerically simulated cases compared with the traditional decline
methods. Therefore, the work offers critical insights into evaluating production in shale gas reserviors in
a more efficient way.

Introduction
The exploration and development of shale gas reservoirs has become a huge success in recent years, which
is showing that shale will occupy an increasingly important position (Javadpour et al., 2007; Miao et al.,
2018). Compared with conventional gas reservoirs, traditional development means should be replaced by
2 SPE-191185-MS

horizontal wells with multiple fractures, which should be attributed to the lower pemeability and higher
flow resistance. (Semenychev et al., 2014; Figueiredo et al., 2017; Sun et al., 2017) Until now, it has been
widely recognized that the linear flow dominates much of the life of the wells in shale, which in dramatically
different from the boundary domianated flow in conventional gas reservoirs. (Amir et al., 2017; Fan and
Ettehadtavakkol, 2017; Tang et al., 2017) Accordingly, the classical Arps decline model will not be suitable
for shale reservoirs due to its restricted application conditions. Later, a large amount work on production
prediction of shale reservoirs are conducted and new empirical models are construced and applied into shale
industry.
Firstly, Ilk et al. (2008) proposed a new "power law loss-ratio" rate relation that can be validated and
calibrated directly employing rate functions., which assist analyst in obtaining a diagnostic understanding of
the hyperbolic rate decline relation. Secondly, Valko (2009) presented a new decline curve analysis model
that is well suited for evaluating gas production from shale reservours, which was successfully applied into
the production history analysis of more than 10,000 wells in the Barnett Shale. Yu et al. (2013) illustrated
a modified SEPD model by employing a new specialized plot to find all related parameters. This approach
dramatically eliminates the SEPD Method's shortcoming, which is easy to be emplyed and yield reliable
production estimation. Duong (2011) presented a novel decline method where a log-log plot of rate over
cumulative production vs. time and rate vs. time fuction is observed to fit a straight line, which can provide
statistical methods to analyze production forecasts of resource plays and to establish a range of results of
these forecasts. However, Joshi and Lee (2013) put forward a revision to the Duong model to provide better
fits to data in boundary dominated flow regimes.
Although multiple researchers have been focusing on production prediction of shale gas reservoirs,
critical mechanisms of gas transport in nanopores of shale gas reservoirs cannot be well described in
currently reported research, such as, slip flow/Knudsen diffusion of bulk gas and surface diffusion of
adsorbed gas. In this paper, firstly we developed a new model to estimate production in shale gas reserviors
by considering both Knudsen diffusion of bulk gas and surface diffusion of adsorbed gas based on the
traditional equation of rate versus square-root-of-time. This proposed model can provide better fits to data
in transient linear flow regimes. Furthermore, a systematic analysis of numerical simulation cases in CMG
were performed to compare with the traditional model.

Review of popular decline curve models

Arps decline model
The traditional Arps decline model is based on the empirical observation that a well's "loss ratio" is usually
constant with time (Arps 1945; Fetkovich et al. 1996). The constant loss ratio is assumed as "b" by Arps,
shown in Eq. (1):

(1)

Where, qi is initial rate, Mscf/day; b is constant loss ratio, dimensionless; Di is initial decline rate, 1/day.
Several authors have demonstrated rigorously that transient bi-linear and linear flow regimes in
reservoirs, correspond to b values of 4 and 2 respectively (Kupchenko et al. 2008; Maley 1985; Spivey et
al. 2001).

Stretched exponential decline model (SEDM)

The SEDM, proposed by Valkó (2010), is given by Eq. (2).

(2)
SPE-191185-MS 3

Where, τ is characteristic time constant, day; n is exponent, dimensionless; qo is initial production rate,
Mscf/day.
Two advantages of this model over the Arps model are as follows. Firstly, the EUR is bounded; Secondly,
it is designed to model transient flow, unlike the Arps model, which is intended for BDF (Valko and Lee
2010). In previous studies, the SEDM has been tested on grouped data sets and has proved to work well for
historical production periods exceeding 36 months.

Duong model
The Duong model (Duong, 2010) is a new forecasting technique that, as a practical matter, assumes long-
duration linear flow. This method consists of two equations to calculate model parameters a, m and q1. Eq.
(3) is utilized to calculate the parameters a and m by the linear regression analysis.

(3)

(4)
Where,
(5)

Scaling method by Patzek et al.

Defining a dimensionless time as (Patzek et al., 2013):

(6)

The cumulative production of gas mass from a horizontal well with N hydrofractures was assumed as
t. The original mass of gas contained in the reservoir volume drained by this well was defined as M. The
exact solution of the model for cumulative gas production is given by a dimensionless recovery factor:

(7)

The universal solution to the boundary value problem depends on the initial state of the reservoir, the well
flowing pressure, and gas composition, although it is independent of the details of the well geometry and the
hydraulic diffusivity. However, these parameters can be set to typical values within a given shale gas play.
(8)
where is the dimensionless time necessary to extinguish the initial transients in gas flow.

A new production analysis model for shale

The apparent permeability of the bulk-gas phase in nanopores is expressed as:

(9)

Where, ξmb is a correction factor of apparent permeability in nanopores of shale gas reservoirs,
dimensionless; α is the rarefaction coefficient, dimensionless; b is the gas-slip constant, dimensionless; ωm
is the poromechancial-response coefficient of shale matrix, dimensionless; and ωs is the sorption-induced-
swelling-response coefficient of shale matrix, dimensionless. For more information about ωm and ωs, please
see Wu's (2016) study.
4 SPE-191185-MS

Furthermore, depending on the assumption that dynamic equilibrium is achieved between the bulk gas
and the adsorbed gas in the nanopores (Hwang and Kammermeyer, 1966; Medve' and Černý, 2011), the
apparent permeability of adsorbed-gas surface diffusion can be determined in Eq. (10) (Wu et al., 2016).
This is modeled after the Hwang and Kammermeyer model in a low-pressure condition, which incorporates
the adsorbed-gas coverage in a high pressure condition.

(10)

Where, the gas is assumed as ideal gas; Js is the mass flux of the adsorbed-gas surface diffusion, kg/
(m2·s), Lm is the gas mobility, molηs/kg; u is the chemical potential, J/mol; ξms is a correction factor of surface
diffusion, with the detailed formulas in Wu's study (2015); R is the universal gas constant, J/(mol·K); T is
temperature, K; p is pressure, Pa; μg is the gas viscosity, Pa·s; M is the gas molar mass, kg/mol; and Ds is
the surface-diffusion coefficient, m2/s.
Overall, by combining Eqs. (9) and (10), the apparent permeability for the total gas through shale
nanopores is expressed in Eq. (11), which comprehensively incorporates the effects of rarefaction, wall
roughness, poromechancial response, and sorption-induced swelling response on bulk-gas transfer, as well
as the influence of adsorbed-gas coverage in a high pressure condition.

(11)

The permeability ratio defined as the ratio of apparent permeability to the liquid-equivalent permeability
is introduced in this study. Then, the advanced pseudo-pressure and pseudo-time to account for slip flow,
Knudsen diffusion and surface diffusion are provided here:

(12)

(13)

(14)

Here, is the modified pseudo-pressures at initial pressure to account for slip flow, Knudsen diffusion
and surface diffusion, psia2/cp; is the modified pseudo-pressures at flowing pressure to account for
slip flow, Knudsen diffusion and surface diffusion, psia2/cp; is the modified pseudo-time to account for
slip flow, Knudsen diffusion and surface diffusion, hours; is the permeability ratio defined as the ratio
of modified apparent permeability to the liquid-equivalent permeability, dimensionless; k∞ is the liquid-
equivalent permeability, mD; and is the modified apparent permeability to account for slip flow, Knudsen
diffusion and surface diffusion, mD; μg is the gas viscosity, cp; Ct is the total compressibility, psia-1; Z is the
gas compressibility factor, dimensionless; is the gas viscosity at the average pressure, cp; is the gas
compressibility at the average pressure, psia , and subscript i refers to initial pressure.
-1

The cumulative production at time t of gas wells is:

(15)

Where, q is gas rate, Mscf/day, t is time, hours, and Gp is the cumulative production, Mscf.
The gas-in-place in the region of influence is:
SPE-191185-MS 5

(16)

Here, A is the area of the region of influence, ft2; Sgi is the initial gas saturation; and Bgi is the initial gas
formation volume factor, ft3/scf.
Kang et al. (2013) showed that a generalization of the "radius of investigation" concept for transient
radial flow becomes. Corresponding to the transient linear flow, the "depth of investigation" is expressed
as (Seidle at al., 2016):

(17)

Where, di is called the "depth of investigation", ft, which will be half the distance between adjacent
fractures.
The definition of area of the region of influence is shown in Eq. (18):
(18)
Then substituting Eq. (17) and Eq. (18) into Eq. (16) leads to:

(19)

On the basis of the relationship between the cumulative gas rate and average pressure (Moghadam et al.,
2011), as shown in Eq. (20):

(20)

Here, is average pressure, and and are the modified Z-factors introduced by Moghadam et al.
(2011) at average pressure in the region of influence and initial pressure, respectively.
Substituting Gp and G from Eq. (15) and Eq. (19), respectively, into Eq. (20) results in:

(21)

Based on above equations, the average pressure in the region of influence can be determined. Because
the average pressure in the region of influence is constant (Nobakht and Clarkson, 2012), the corrected
pseudo-time, , becomes:

(22)

Depending on the linear flow theory (El-Banbi and Wattenbarger, 1998; Wattenbarger et al., 1998), a
new fully analytical rate-transient model is proposed as follows, which takes both the slip flow/Knudsen
diffusion of bulk gas and surface diffusion of adsorbed gas into consideration. Here, it is assumed that the
fracture has infinite conductivity, no skin and wells produce at a constant flowing pressure.

(23)

Where, T is reservoir temperature, °F; h is net pay thickness, ft; ∅ is reservoir porosity, percent; and xf
is fracture half-length, ft.
6 SPE-191185-MS

Therefore, the slip flow/Knudsen diffusion of bulk gas phase and surface diffusion of adsorbed gas phase
are directly incorporated into the new model of rate-transient analysis.

Model Validation and Comparison

To validate the novel model in this paper, a number of numerically simulated test cases are applied. The
common reservoir and fluid parameters among all cases are shown in Table 1. The modeling parameters
used in the calculation of apparent permeability are summarized in Table 2.

Table 1—Reservoir and fluid parameters in six numerically simulated cases.

Parameters(Unit) Case 1 Case 2 Case 3 Case 4 Case 5 Case 6

Initial reservoir pressure(psi) 4615 4615 4615 4615 4615 4615

Reservoir temperature (°F) 250 250 250 250 250 250
Reservoir thickness (ft) 200 200 200 200 200 200
Bottomhole pressure(psi) 500 500 500 500 500 500
Reservoir porosity 0.068 0.068 0.068 0.068 0.068 0.068
Reservoir permeability(mD) 0.000138 0.000138 0.0005 0.0005 0.001 0.001
Fracture permeability(mD) 32 32 32 32 32 32
Fracture spacing(ft) 250 100 250 100 250 100
Fracture half-length(ft) 375 375 375 375 375 375
Fracture width(ft) 0.5 0.5 0.5 0.5 0.5 0.5
Dimensionless fracture conductivity 309 309 309 309 309 309

Table 2—Summary of modeling parameters employed in the calculation.

Parameters (Unit) Values Parameters (Unit) Values

Shale tortuosity (dimensionless) 4.3 Gas-molecule diameter (m) 3.8×1010

Universal gas constant[J/(mol·K)] 8.314 Initial microfracture aperture (m) 5.0×10-6
Gas molar mass (kg/mol) 1.6×10-2 Langmuir pressure (Pa) 6.89×106
Biot coefficient for microfracture
0.8 Avogadro's constant (1/mol) 6.02214×1023
(dimensionless)
Isosteric adsorption heat at the gas coverage of
Biot coefficient for shale matrix (dimensionless) 0.65 14000
zero (J/mol)
Microfracture normal stiffness (Pa/m) 3.58×1012 Shale-matrix Young's modulus (Pa) 4.0×1010

The comparison results between calculated rates and numerically simulated rates for the whole production
life of shale gas wells are presented in Fig. 1. It is noticeable that the model results are in perfect agreement
with the numerically simulated data, especially in linear flow period.
SPE-191185-MS 7

Figure 1—Comparison between calculated rates and numerically simulated rates in log-log
coordinates for six cases: (a) Case 1; (b) Case 2; (c) Case 3; (d) Case 4; (e) Case 5; (f) Case 6.

Additionally, in order to show the advantage of the novel analytical model, comparisons between the
traditional method and novel approach are conducted by six test cases. The traditional equation of gas rate
vs. time (Eq. (24)) to match the same numerical simulation data is applied and the comparison plots are
illustrated in Fig. 2. As expected, the rate-transient obtained from the newly analytical model will be a much
better consistent with the numerical simulated solution than the traditional one.
8 SPE-191185-MS

Figure 2—Comparison among simulated rates, calculated rates from the novel model and calculated rates from the
traditional model in Cartesian coordinates for six cases: (a) Case 1; (b) Case 2; (c) Case 3; (d) Case 4; (e) Case 5; (f) Case 6.

(24)

Where, t is time, hours; Ppi is the pseudo-pressures at initial pressure, psia2/cp; and Ppwf is the pseudo-
pressures at flowing pressure, psia2/cp.
SPE-191185-MS 9

Results
• A new rate-transient analysis model for shale gas reservoirs is proposed by incorporating the slip
flow/ Knudsen diffusion of bulk gas and surface diffusion of adsorbed gas directly into the model.
It is a fully analytical approach rather than an empirical one, which can be widely applicable in
regions whether production is relatively well established or not.
• Perfect matches between the forecast rates generated from the novel rate-transient analysis model
and numerically simulated rates are obtained.
• The accuracy of estimating gas rate using the novel method and the traditional method are
compared by a number of test cases, which indicates that the newly developed model leads to more
confident production forecasts.

References
Arps J J. Analysis of decline curves. Transactions of the AIME 1945; 160(01): 228–247.
Amir S Z, Sun S. An efficient two-scale hybrid embedded fracture model for shale gas simulation. Journal of Petroleum
Science and Engineering 2017; 152:683–714.
Duong A N. Rate-decline analysis for fracture-dominated shale reservoirs. SPE Reservoir Evaluation & Engineering
2011. 14(03): 377–387.
El-Banbi A H, Wattenbarger R A. Analysis of linear flow in gas well production. In: Paper presented at the Society for
Petroleum Engineers Gas Technology Symposium, Calgary, Alberta, Canada, March 15-18, 1998; 1998.
Fan D, Ettehadtavakkol A. Semi-analytical modeling of shale gas flow through fractal induced fracture networks with
microseismic data. Fuel 2017; 193: 444–459.
Fetkovich, M.J., Fetkovich, E.J., and Fetkovich, M.D. 1996. Useful Concepts for Decline Curve Forecasting, Reserve
Estimation, and Analysis. SPE Reservoir Engineering 11 (1): 13–22.
Figueiredo B, Tsang CF, Rutqvist J, Niemi A. Study of hydraulic fracturing processes in shale formations with complex
geological settings. Journal of Petroleum Science and Engineering 2017; 152:361–374.
Hwang S T, Kammermeyer K. Surface diffusion in microporous media. The Canadian Journal of Chemical Engineering
1966; 44(2): 82–89.
Ilk D, Rushing JA, Perego A D, Blasingame T A. Exponential vs. hyperbolic decline in tight gas sands: understanding the
origin and implications for reserve estimates using Arps' decline curves. In: SPE Annual Technical Conference and
Exhibition. Society of Petroleum Engineers; 2008.
Javadpour, F., Fisher, D., Unsworth, M., 2007. Nanoscale gas flow in shale gas sediment. J. Can. Petrol. Technol. 46
(10), 55e61.
Joshi K, Lee W J. Comparison of various deterministic forecasting techniques in shale gas reservoirs. In: SPE Hydraulic
Fracturing Technology Conference. Society of Petroleum Engineers; 2013.
Kang S S, Datta-Gupta A, Lee W J. Impact of natural fractures in drainage volume calculations and optimal well placement
in tight gas reservoirs. Journal of Petroleum Science and Engineering 2013; 109: 206–216.
Kupchenko, C.L., Gault, B.W., and Mattar, L. 2008. Tight Gas Production Performance Using Decline Curves. Paper
presented at the CIPC/SPE Gas Technology Symposium 2008 Joint Conference, Calgary, Alberta, Canada. Society of
Petroleum Engineers SPE-114991-MS.
Lee W J, Sidle R. Gas-reserves estimation in resource plays. SPE Economics & Management 2010. 2(02): 86–91.
Maley, S. 1985. The Use of Conventional Decline Curve Analysis in Tight Gas Well Applications. Paper presented at
the SPE/DOE Low Permeability Gas Reservoirs Symposium, Denver, Colorado. 1985 Copyright 1985, Society of
Petroleum Engineers 00013898.
Medved' I, Èerný R. Surface diffusion in porous media: A critical review. Microporous and Mesoporous Materials 2011;
142(2): 405–422.
Miao, Y.; Li, X.; Lee, J.; Zhou, Y.; Wu, K.; Sun, Z.; Liu, S. A new rate-decline analysis of shale gas reservoirs: Coupling the
self-diffusion and surface diffusion characteristics. Journal of Petroleum Science and Engineering 2018, 163, 166–176.
Miao, Y.; Li, X.; Zhou, Y.; Lee, J.; Sun, Z.; Chang, Y.; Wang, S.; Hou, C. A new rate-transient analysis model for shale
gas reservoirs coupled the effect of slip flow and surface diffusion. International Journal of Heat and Mass Transfer
2018, 124, 1–10.
Moghadam S, Jeje O, Mattar L. Advanced gas material balance in simplified format. Journal of Canadian Petroleum
Technology 2011; 50(01): 90–98.
Nobakht M, Clarkson C R. A new analytical method for analyzing linear flow in tight/shale gas reservoirs: constant-
flowing-pressure boundary condition. SPE Reservoir Evaluation & Engineering 2012; 15(03): 370–384.
10 SPE-191185-MS

Seidle J, Erdle J, Hale B, Houze O, Ilk D, Jenkins C, Lee J, Ritter J, Wilson S, and Weatherford D. Estimating ultimate
recovery of developed wells in low permeability reservoirs. SPEE Monograph 4 2016.
Semenychev V K, Kurkin E I, Semenychev E V. Modelling and forecasting the trends of life cycle curves in the production
of non-renewable resources. Energy 2014; 75:244–51.
Sharma A, Lee W J. Improved workflow for EUR prediction in unconventional reservoirs. In: Unconventional Resources
Technology Conference. Society of Petroleum Engineers; 2016.
Spivey, J.P., Jr., J.H.F., Williamson, J.R. et al. 2001. Applications of the Transient Hyperbolic Exponent. Paper presented
at the SPE Rocky Mountain Petroleum Technology Conference, Keystone, Colorado. Society of Petroleum Engineers
00071038.
Sun, F. R., Yao, Y. D., Li, X. F., Li H., Chen G., Sun, Z. 2017. A numerical study on the non-isothermal flow characteristics
of superheated steam in ground pipelines and vertical wellbores. Journal of Petroleum Science and Engineering. 159:
68–75.
Sun, F. R., Yao, Y. D., Li, X., F., Zhao, L. 2017. Type curve analysis of superheated steam flow in offshore horizontal
wells. International Journal of Heat and Mass Transfer, 113: 850–860.
Tang, H.; Hasan, R.; Killough, J. A Fully-coupled Wellbore-Reservoir Model for Transient Liquid Loading in Horizontal
Gas Wells. Presented at ATCE, San Antonio, TX, 2017.
Tang, H.; Killough, E.; Heidari, Z.et al. A New Technique To Characterize Fracture Density by Use of Neutron Porosity
Logs Enhanced by Electrically Transported Contrast Agents. SPE Journal, 2017.
Valko P P. Assigning value to stimulation in the Barnett Shale: a simultaneous analysis of 7000 plus production histories
and well completion records. In: SPE Hydraulic Fracturing Technology Conference. Society of Petroleum Engineers;
2009.
Wattenbarger R A, El-Banbi A H, Villegas M E, Maggard J B. Production analysis of linear flow into fractured tight
gas wells. In: Paper presented at the Society for Petroleum Engineers Rockey Mountain Regional/Low-Permeability
Reservoirs Symposium and Exhibition, Denver, Colorado, USA, April 5-8, 1998; 1998.
Wu K, Li X, Guo C, Wang C, Chen Z. A unified model for gas transfer in nanopores of shale-gas reservoirs: coupling
pore diffusion and surface diffusion. SPE Journal 2016.
Wu K, Li X, Guo C, Chen Z. Adsorbed gas surface diffusion and bulk gas transport in nanopores of shale reservoirs with
real gas effect-adsorption-mechanical coupling. In: Paper presented at the Society of Petroleum Engineers Reservoir
Simulation Symposium in Houston, Texas, USA, February 23-25, 2015; 2015.
Yu S, Miocevic D J. An improved method to obtain reliable production and EUR prediction for wells with short
production history in tight/shale reservoirs. In: Unconventional Resources Technology Conference. Society of
Petroleum Engineers; 2013.