Gas Well Testing

Deliverability testing of gas wells

Deliverability testing refers to the testing of a gas well to measure its production capabilities
under specific conditions of reservoir and bottomhole flowing pressures (BHFPs). A common
productivity indicator obtained from these tests is the absolute open flow (AOF) potential. The
AOF is the maximum rate at which a well could flow against a theoretical atmospheric
backpressure at the sandface. Although in practice the well cannot produce at this rate, regulatory
agencies sometimes use the AOF to allocate allowable production among wells or to set
maximum production rates for individual wells.
Another application of deliverability testing is to generate a reservoir inflow
performance relationship (IPR) or gas backpressure curve. The IPR curve describes the
relationship between surface production rate and BHFP for a specific value of reservoir pressure
(that is, either the original pressure or the current average value). The IPR curve can be used to
evaluate gas-well current deliverability potential under a variety of surface conditions, such as
production against a fixed backpressure. In addition, the IPR can be used to forecast future
production at any stage in the reservoir’s life.
Several deliverability testing methods have been developed for gas wells. Flow-after-flow
tests are conducted by producing the well at a series of different stabilized flow rates and
measuring the stabilized BHP. Each flow rate is established in succession without an
intermediate shut-in period. A single-point test is conducted by flowing the well at a single rate
until the BHFP is stabilized. This type of test was developed to overcome the limitation of long
testing times required to reach stabilization at each rate in the flow-after-flow test.
Isochronal and modified isochronal tests were developed to shorten tests times for wells that
need long times to stabilize. An isochronal test consists of a series of single-point tests usually
conducted by alternately producing at a slowly declining sandface rate without pressure
stabilization and then shutting in and allowing the well to build to the average reservoir pressure
before the next flow period. The modified isochronal test is conducted similarly, except the flow
periods are of equal duration and the shut-in periods are of equal duration (but not necessarily the
same as the flow periods).
Introduction
Contrary to liquids, a gas is highly compressible and much less viscous. In general, gas viscosity
is about a 100 times lower than the least viscous crude oil. It is important, however, to try to
provide the same mathematical treatment to oil and gas hydrocarbons, so interpretation
methodologies can easily be applied in a more practical way. Then, the gas flow equation is
normally linearized to allow the liquid diffusivity solution to satisfy the gas behavior when
analyzing transient test data of gas reservoirs. Depending on the values of reservoir pressure,
viscosity, and gas compressibility factor, the gas flow behavior can be treated as a function of
either pressure to the second power or linear pressure with a region which does not correspond to
any of these and it is better represented by a synthetic function call pseudopressure.
Pseudopressure is a function that integrates pressure, density, and compressibility factor. The gas
system’s total compressibility highly depends on gas compressibility which for ideal gases
changes inversely with the pressure. Then, another artificial function referred as pseudotime is
included to further understand the transient behavior of gas flow in porous media. For instance,
when wellbore storage conditions are insignificant, drawdown tests are best analyzed using the
pseudopressure function.
On the other hand, buildup pressure tests require linearization of both pseudotime and
pseudopressure.
Pseudopressure and Pseudotime
In solving the diffusivity equation certain assumptions are made about the fluid properties, some
of which are not applicable to gas – in particular, that the reservoir fluid is slightly compressible
and that both viscosity and compressibility are independent of pressure:
Above bubble point, most oil properties are fairly constant. Compressibility decreases slightly
with pressure, but it is assumed to be constant over the range of test pressures. Viscosity will
increase slightly with pressure.

For gas, compressibility is inversely proportional to pressure - it is initially very easy to

compress, and becomes increasingly difficult. Gas viscosity increases with pressure, as the
molecules are forced closer together and collide more frequently.
The simplest equation of state that can be used for a gas is the Ideal Gas Law:
PV = nRT
The behavior of real gases is not quite that simple, and a compressibility term is required, the
‘deviation factor’, Z:
PV = nZRT
Assuming this equation of state, it is possible to express the diffusivity equation in the same
form as for a slightly compressible fluid, by introducing the pseudo-pressure (or Real Gas
Potential, defined by Al-Hussainy in 1966):

where p0 is a reference pressure that must be below the lowest measured pressure during the
transient under investigation. It is generally taken as 0.
The diffusivity equation is expressed as:

For gas well test analysis, the pressure terms are replaced by m(p). Analysis based on
pseudopressure is valid in all cases, and today we have no reason not to use it, but historically it
was of interest to make approximations, based upon the variation of µz with pressure:
At low pressures, the gas tends to behave as an ideal gas, and µz is constant. Considering the
integral solution, this means that m(p) is proportional to p2. At high pressures µz is proportional
to p, so m(p) is proportional to p.
So before the advent of computers, the rules of thumb were:
• pmax < 2000 psi: use p2
• pmin > 3000 psi: use p
In between these values no approximation was available, and m(p) had to be used.
In the diffusivity equation on the previous page, the factor µCt is a function of pressure, but for
practical purposes it can often be assumed to be constant during the test. However where
pressure variations are large, such as in a tight formation and/or a very damaged well, this
assumption is no longer valid, as will be seen in the next section.
Assuming that µCt is constant, the dimensionless pseudo-pressure becomes:

where TSC is the temperature at standard conditions, expressed in °Rankine (520), pSC is the
pressure at standard conditions, in psia (14.7), T is the reservoir temperature in °R and qSC is the
gas flowrate at standard conditions, in MSCF/day. The difference between this expression and
the oil equations is that the downhole gas flowrate is computed from the surface rate by way of
an equation of state, rather than a formation volume factor. (B is replaced by T, p and z terms).
The equations for evaluating gas reservoir parameters depend upon which pressure function is
used:
Normalized Pseudopressure
The pressure function can be normalized, and given dimensions of pressure:

The normalized pseudopressure, normalized with reference to gas properties at static conditions,
pi, enables the equations for oil solutions to be used.
Pseudotime
In a tight reservoir, with large drawdowns, the assumption that µCt is constant leads to a
distortion in the early time of the log-log and derivative data:
The raw data shows is re-plotted (as above, solid green line) using the gas pseudo-time function
developed by Agarwal, 1979:

Normalized Pseudotime
The pseudo-time can also be normalized, as suggested by Meunier et al (1987), in such a manner
that at late time the normalized function tpn becomes equivalent to ∆t, ensuring that the standard
superposition calculations can be employed:

Pseudo-Skin, S’
In oil well testing, the skin measured during each transient will be the same, whatever it consists
of: – true formation damage, convergent flow due to partial penetration, inadequate perforations,
etc. In gas wells this is not the case.

In fluid mechanics, the Reynolds number determines whether fluid flow will be laminar, which
in well testing we call Darcy flow, or turbulent, which we call non-Darcy flow. In oil wells, with
the possible exception of some extremely high rate producers, the combination of certain
parameters, such as fluid density and flow velocity, is such that the critical Reynolds number is
never exceeded, and the flow remains laminar. However, in gas wells the Reynolds number
frequently exceeds 2000-3000, and turbulent flow results.
The significance of turbulent flow in well testing is that it creates an additional, rate-dependent
pressure drop. The turbulence occurs where the fluid velocity is highest, which is where the flow
area in the reservoir converges to a minimum – i.e. at the wellbore. This additional pressure drop
very close to the wellbore looks like skin, and can not be separated from the true skin in a single
transient. The total measured skin will be a combination of the true and rate-dependent skins.
To obtain the true skin, the solution is simply to test the well on a number of different flow rates,
as the ‘pseudo-skin’ is related to the true skin by a simple straight-line equation:
S’ = S0 + Dq
where : S’ is the pseudo-skin
S0 is the true skin
D is the non-Darcy flow coefficient q
is the gas flowrate

The pseudo-skin (S’) measured during any transient is such that for a higher flowrate there will be higher
apparent skin. The true ‘damage’ skin is obtained by extrapolating the measured skins to zero flowrate, as
shown overleaf.
(Due to superposition effects it is actually the change in flowrate, ∆q, which determines the apparent skin
value, so that any build-up will have the same S’ as the preceding drawdown.)
It is mainly for this reason that the standard gas test is multiple-rate. The most typical is a
modified isochronal test, in which each drawdown is the same duration (except the last), and
each build-up is the same duration, but build-ups can be longer than drawdowns.
Typical gas test:

By utilizing the superposition principle it would not matter if each transient were a different
duration, but there is another reason for trying to keep the drawdowns and build-ups to fixed
times, and for extending the final flow, as will be seen in the next section. To obtain the true
skin, S’ = S + Dq is the same as ‘y = mx + C’:

If S’ is plotted against q, the slope of the line is D and the intercept is the true skin, S (or S0, the
skin at zero flowrate.) A high pseudo-skin will probably benefit little from an acid job, but may
be significantly reduced by adding more perforations.
Absolute Open Flow
A common measure of the performance of any gas well is its ability to deliver against ‘zero’
bottomhole pressure, or more accurately against atmospheric pressure. How hard could it flow
with a flowing bottomhole pressure of 1 atmosphere? This may sound like a pointless
calculation, as it is certainly unrealistic; a gas well could not even blow out at this rate, the
bottomhole pressure would be too high. However this measurement provides an input value for
the reservoir engineer working with IPR calculations, and is an accepted ‘universal indicator’ for
gas wells. It is called the ‘absolute open flow’ potential of the well, the AOF.
In order to evaluate the AOF, the well needs to be tested at multiple rates – as has already been
done for the skin calculation. The bottomhole pressure must be measured during each
drawdown, after the same time – for example 2 hours into each drawdown, 4 hours, etc.
The final drawdown is extended, as this will provide a single ‘stabilized’ value for the
calculation, pwf5 in the plot below:

Fig. 5-78 Modified Isochronal Test

From this data the following plot is constructed, based upon the empirical relationship that ∆(p2)
vs q, on a log-log scale, will give a straight line, as long as the pwf values are recorded after the
same time in each drawdown. Moreover, the ‘2-hour’, ‘4-hour’, ‘6-hour’ lines, etc., will all be
parallel, converging to the ‘stabilized deliverability curve’ for all of the stabilized points,
obtained if the well had been flowed long enough for each drawdown to stabilize:
 Fig. 5-79 Absolute Open Flow

In a typical test, as shown above, there will be only one stabilized point, and a line is drawn
through this point and parallel to the un-stabilized line. By extrapolating the stabilized
deliverability curve to a ∆( p2) value corresponding to a flowing bottomhole pressure of 14.7 psi,
the AOF is obtained.
Fig. 5-80 Log-Log Responses