Agony aunt for: PTA, Gas MBA, DCA and RTA
Moderator: moderators
Agony aunt for: PTA, Gas MBA, DCA and RTA
Hi. Started a new forum topic covering:
Classical well test analysis
Gas material balance
Decline curve analysis
Rate transient analysis (integration of DCA and PTA physics)
Post the problems you face at work here, and I will try to answer them (when I'm not travelling, and if the problems are not too tough !!!)
Cheers.
Classical well test analysis
Gas material balance
Decline curve analysis
Rate transient analysis (integration of DCA and PTA physics)
Post the problems you face at work here, and I will try to answer them (when I'm not travelling, and if the problems are not too tough !!!)
Cheers.
Awright !
I'll answer Q1 first:
As you know, there are two principal types of storage: one in which the fluid interphase level changes (meaning the liquid level falls or rises depending upon conditions). In the second, the liquid level stays static (more or less), but the dissolved gas within the liquid comes out and establishes its own equilibrium at the top of the well column. Both affect the transient (build up/fall off...), and makes our type curve matching (and our lives!) more difficult. And life gets a lot worse when it is a water injector.
So, when you are doing regression analysis/type curve matching in PTA, remember:
1. Selects rough value of CD (dimensionless storage constant) from your first unit slope of derivative curve
2. Input that into your model and see how the fit is.
3. Then start inputting very small numbers for CaD (Dimensionless apparent wellbore storage coeff) and CpD (Dimensionless storage pressure parameter), and see how the curve moves.
4. If your PTA package is high end, do an APE (automated parameter estimation), but remember to keep a low range on the CD number otherwise it'll keep matching till new year's day !
Try this out and let me know
Second question: DCA and RTA are both "indirect" means since they treat one phase at a time. You could be able to identify banking by productivity shifts on your flowing material balance and Blasingame plots.
I'll answer Q1 first:
As you know, there are two principal types of storage: one in which the fluid interphase level changes (meaning the liquid level falls or rises depending upon conditions). In the second, the liquid level stays static (more or less), but the dissolved gas within the liquid comes out and establishes its own equilibrium at the top of the well column. Both affect the transient (build up/fall off...), and makes our type curve matching (and our lives!) more difficult. And life gets a lot worse when it is a water injector.
So, when you are doing regression analysis/type curve matching in PTA, remember:
1. Selects rough value of CD (dimensionless storage constant) from your first unit slope of derivative curve
2. Input that into your model and see how the fit is.
3. Then start inputting very small numbers for CaD (Dimensionless apparent wellbore storage coeff) and CpD (Dimensionless storage pressure parameter), and see how the curve moves.
4. If your PTA package is high end, do an APE (automated parameter estimation), but remember to keep a low range on the CD number otherwise it'll keep matching till new year's day !
Try this out and let me know
Second question: DCA and RTA are both "indirect" means since they treat one phase at a time. You could be able to identify banking by productivity shifts on your flowing material balance and Blasingame plots.
Thank you oilsave for opening this threadoilslave wrote:Awright !
I'll answer Q1 first:
As you know, there are two principal types of storage: one in which the fluid interphase level changes (meaning the liquid level falls or rises depending upon conditions). In the second, the liquid level stays static (more or less), but the dissolved gas within the liquid comes out and establishes its own equilibrium at the top of the well column. Both affect the transient (build up/fall off...), and makes our type curve matching (and our lives!) more difficult. And life gets a lot worse when it is a water injector.
So, when you are doing regression analysis/type curve matching in PTA, remember:
1. Selects rough value of CD (dimensionless storage constant) from your first unit slope of derivative curve
2. Input that into your model and see how the fit is.
3. Then start inputting very small numbers for CaD (Dimensionless apparent wellbore storage coeff) and CpD (Dimensionless storage pressure parameter), and see how the curve moves.
4. If your PTA package is high end, do an APE (automated parameter estimation), but remember to keep a low range on the CD number otherwise it'll keep matching till new year's day !
Try this out and let me know
Second question: DCA and RTA are both "indirect" means since they treat one phase at a time. You could be able to identify banking by productivity shifts on your flowing material balance and Blasingame plots.
I have two comments:
1- The well bore storage doesn't affect the whole model (i.e. the well model, the reservoir model and boundary model)
2- The automated parameters estimation (i.e. Regression analysis done by Saphire or Pansys) will never be accurated. i preferre to use manual trial and error
Fanarco, I agree with you.
It is just that in most water injection wells, the pressure fall-off tests demonstrate storage periods for a considerable portion of test time, and the model has to accommodate this aspect more fully/precisely.
Secondly, I always advice my juniors to start APE only after they have a fair idea of the working parameter values, and a good image of the reservoir in their minds. After that, an iterative approach can give you good results if the transient data is sufficiently long. Otherwise.......oh gosh!!!
It is just that in most water injection wells, the pressure fall-off tests demonstrate storage periods for a considerable portion of test time, and the model has to accommodate this aspect more fully/precisely.
Secondly, I always advice my juniors to start APE only after they have a fair idea of the working parameter values, and a good image of the reservoir in their minds. After that, an iterative approach can give you good results if the transient data is sufficiently long. Otherwise.......oh gosh!!!
which software you are using for PTA ??oilslave wrote:Fanarco, I agree with you.
It is just that in most water injection wells, the pressure fall-off tests demonstrate storage periods for a considerable portion of test time, and the model has to accommodate this aspect more fully/precisely.
Secondly, I always advice my juniors to start APE only after they have a fair idea of the working parameter values, and a good image of the reservoir in their minds. After that, an iterative approach can give you good results if the transient data is sufficiently long. Otherwise.......oh gosh!!!
Hi Fanarco. Was away on tour for quite some time.
I've had the privilege of using most of the popular PTA packages. But really, after ten solid years of computer-based interpretation, the best is Fekete's FAST welltest. They have an excellent support system, and for licenced users like us, the upgradation and platform change is pretty much seamless.
Still, anything is better than the old Lotus-1-2-3 and the older log-paper hell, eh!
I've had the privilege of using most of the popular PTA packages. But really, after ten solid years of computer-based interpretation, the best is Fekete's FAST welltest. They have an excellent support system, and for licenced users like us, the upgradation and platform change is pretty much seamless.
Still, anything is better than the old Lotus-1-2-3 and the older log-paper hell, eh!
Two lessons:
1. You're never too old to learn
2. In the oil industry, not every question has a solution
Yesterday, I was stumped after quite some time: I was looking at a build up in a depleting naturally fractured reservoir with known matrix porosity. But on the diagnostics, the dual porosity trough was totally absent and the transient response was linear all the way thru to almost end of test. Worse, the initial test some years back had shown the trough in almost classical detail. Took me two coffee's and two cigarettes to figure it out.
Can you?
1. You're never too old to learn
2. In the oil industry, not every question has a solution
Yesterday, I was stumped after quite some time: I was looking at a build up in a depleting naturally fractured reservoir with known matrix porosity. But on the diagnostics, the dual porosity trough was totally absent and the transient response was linear all the way thru to almost end of test. Worse, the initial test some years back had shown the trough in almost classical detail. Took me two coffee's and two cigarettes to figure it out.
Can you?
mmmm,
It is good exercise to share. i would like to ask you if these two PBUs in the same well??
You said it is in the same reservoir,
What is production data during these two tests because as you know may the shape affected by the phase either oil, gas or water
The P* in the old test is around 2844 Psi and the new test was 3077 Psi, how come ??
Also as i know in the conventional PBU derivative plot you should plot dP Vs dT and derivative Vs dT which is not clearly represented. i see the y axis is not dP
Again, thank you for sharing this interesting problem and we are waiting for more investigation
It is good exercise to share. i would like to ask you if these two PBUs in the same well??
You said it is in the same reservoir,
What is production data during these two tests because as you know may the shape affected by the phase either oil, gas or water
The P* in the old test is around 2844 Psi and the new test was 3077 Psi, how come ??
Also as i know in the conventional PBU derivative plot you should plot dP Vs dT and derivative Vs dT which is not clearly represented. i see the y axis is not dP
Again, thank you for sharing this interesting problem and we are waiting for more investigation