water production issue in dry gas reservoir

[Smriti]

Hi all,

we have dry gas field.
we are not very sure about the kind of water drive present in the reservoir.
some wells (not all) started producing water after 1.5 years of production and the water rate is continously increasing.
kindly suggest how to find whether its coning problem or localised water body.
we have prepared plots of WGR vs time but with this is it possible to conclude something.


Regards,
Smriti

[anihita]

You can take help of cement visualization logs like CAST-V or USIT. They give a 360 deg image of the well bore behind casing and u can know where the cement is bad water is encroaching. Alternatively, a PLT survey can be carried out to detect the leaky bridge plug or in case of multiple zones punctured, which one is contributing water. But Cement evaluation is must.

[DAH7542]

Smriti, the Cole plot (material balance) may help you understand how strong is your aquifer support.... take a look at this spe paper in case you don't know it: "SPE_75354 - Improvements MB.pdf".

Best regards

[Smriti]

thanks for this valuable information.

[Chee Koh Peh]

Guys,

Basics first.

There are thee possibilities for the water, (a) it may be from an underlying aquifer or (b) if you have a high initial water saturation then you may be producing interstitial water, (c) or if your gas wells that are depleted and BHP is low, then residual water in the gas (there is always residual water) will be condensing in the tubing and hence you will see water at surface.

Did you examine the fields original logs, was there a GWC present? i.e. presence of an underlying aquifer?

(1) A standard P/Z plot will confirm the presence of an aquifer as you will see a deviation from the "volumetric" straight line", an thus eliminates (b) and (c)

(2) If it is an aquifer then which wells are watering out first, are they down dip flank producers? If so, then you may have edge water drive

(3) If it is an aquifer and there is no pattern to the wells watering out is likely that you have bottom drive, where a combination of localized drawdown and/or heterogeneity and/or reservoir depletion controls water breakthrough and the localized GWC.

(4) Coning or localized water bodies are immaterial in your situation since you must have a GWC for coning. Coning is a function of relative permeability and drawdown, and this unlikely unless you have very, very high drawdown, since gas is far more mobile than water.

(5) The fact that you did not initially produce water but only after some time i.e. only after some pressure depletion tends to suggest the upwards movement of a GWC

(6) Was the water breakthrough rapid, leading to high water levels with wells rapidly ceasing to flow? If not then it tends to suggest the upwards movement of a GWC

Rgrds

Chee Koh Peh

[Smriti]

Hi chee,

GWC is present but P/Z plot shows a volumetric straight line for field flowing material balance.
few wells are showing sudden increase in water production ,gas flowrate being the same..
in one well water was increasing at constant rate and then it suddenly increased and then it decreased drastically withing a day and now its producing with that slow rate.
all the wells are showing different behaviour .
is it possible that there is water support in individual wells and not for the whole gas field.
the reservoir is channel sand type.

Rgds,
Smriti

[Chee Koh Peh]

Dear Smriti,

(1) If the reservoir is a channel sand, then there may be large scale heterogeneity and the water break though may be due to high perm streaks that are in communication with the aquifer, rather than a uniform upwards movement of the GWC.

(2) Have you looked at the pressure depletion trend in all the wells of the field? Is the pressure depletion trend similar in all wells? If it is not then you have compartments and this may explain the difference between the wells behaviors you see in the field.

(3) Assuming the flowing material balance P/Z plot is correct, for you to have a GWC contact and not to see any pressure support from an aquifer, would require the aquifer to be very small, and/or compartmentalized.

(4) For noticeable waterdrive the aquifer needs to be at least 20 to 50 > than the size of the reservoir, remember waterdrive is provided by the expansion of the aquifer volume.

(5) Additionally gas reservoirs under waterdrive are rate sensitivity and this can mask the strength of the waterdrive. For example if you are producing the gas reservoir faster than the aquifer can provide pressure support, you may not see any noticeable aquifer support.

Rgrds

Chee Koh Peh

[jsusilo]

I would suggest to first look at all available logs (if available) and see weather or not there's disconnect in the OGWC (if you do then you know right away you may have some type of heterogenity as already mentioned in previous post). If in fact you do have some type of heterogenity, you could confirm it through interference test (depending on quality and characteristics of the reservoir this may take too long to be practical). You may also want to look for structural features, does the well that started producing water happen to be the deepest / nearest to potential aquifer.

The other thing could do is to compare well performance / productivity index (if well test data is available), because the poorer the well the more likely to cone early (requires more drawdown for the same cubic feet of gas). If you see a trend that way perhaps your worst well gives you indication what may come ahead in the rest of (better) well meaning its not localize but rather systemic. You may also want to compare cum gas production, are wells that shows increased WGR happen to be well that have been put on production longer, it is possible that few wells have been produced significantly harder (above the critical cone rate) and hence coning up much earlier then the rest.

[Smriti]

Thank you Chee..
This info is very useful.
Pressure depletion trend is almost similar but in one well water suddendly started coming some 3 months back and now that well started showing somewhat constant pressure (showing compartmentalised water support..)
also in some wells no gas water contact is mentioned but water sands are present.In that case is it possible that water cut can occur?

[Chee Koh Peh]

Dear Simriti,

Okay let's pull this all together now.

(1) You have an initial GWC shown on the logs, is the initial GWC constant across all wells? If yes, then you have little if any "effective" compartmentalization

(2) If pressure depletion is fairly constant across the wells, which indicates "tank like" behavior with pressure equalizing across the entire reservoir, then water break through is most likely due to a rising GWC

(3) Compare all your wells drawdown, firstly by looking at each wells PI, as higher relative drawdown will accelerate water breakthrough. Does this appear to be the case, the wells with higher drawdown suffering earlier water break through

(4) If the P/Z plots show no initial deviation as would be expected when the aquifer begins to energize the system (well at least after 1.5 years of production), then either your gas production rate is greater than the ability of the aquifer to energize the system (i.e. weak aquifer) and your reservoir is rate sensitive, or the aquifer is still acting in transient behavior (i.e. infinite acting) and its affects have not been felt.

(5) If underlying water sands are present and you have no oil zone and/or shale breaks between the gas zone and the water sands, then water will be produced as the gas reservoir is depleted

(6) From all that you shared it would appear you have uniform bottom water drive, with water breakthrough dependent on localized well conditions (draw down and localized reservoir heterogeneity)

(7) However it now becomes tricky as the aquifer is not showing up on your P/Z plot. You need to determine why this is so, to properly manage the fields recovery.


Rgrds

Chee Koh Peh

[Smriti]

Dear Chee ,
Thank you..
i will analyse the field data again to conclude the things..

Rgds,
Smriti

[darcyoil]

Hi Smriti,

One way of identifying aquifer support not so evident on P/z is to run a rate transient analysis. The bump up on the derivative of the Blasingame plot would model very well on the aquifer type curves. Additionally, a flowing material balance will give you the reserves number taking into account the erosion in productivity as well as the effect of the aquifer. RTA is far, far more sensitive to changes than p/z because of the periodicity of the input data sets. Is this an Indian field by the way?