PMTx 2.0 multilayer tight gas simulator
Frequently Asked Questions (FAQ)
Frequently Asked Questions (FAQ)
Q: Can PMTx handle a Marcellus Shale well that produces 14 bbl of liquid condensate per mmcf of gas?
A: Yes, you can use PMTx for the Marcellus. Click here for a detailed discussion. See Key to Marcellus
Q: Does your software account for desorption? Is it applicable for shales (e.g., Haynesville)?
A: Yes it does..also pressure dependent permeability and porosity. PMTx is designed specifically for fractured gas shales.
Q: There is a need to design and optimize surface facilities based on reservoir considerations. Does your software include a facilities model?
A: PMTx is the reservoir engine in "WEM Shale" by Phoenix Reservoir Software and P.E. Moseley & Associates. WEM Shale is an integrated reservoir-wellbore-facilities model that enables facilities design and near real-time what-ifs for compression and other optimization decisions.
Q: I cannot seem to get anisotropy permeability case activated. Its always on isotropic permeability. Changed well to horizontal, CBM, Shale, Dual Porosity........etc etc. The option is locked and greyed with isotropic case all the time........am I doing something wrong?
A: The outer boundary has to be either infinite-acting or rectangular for the anisotropy option to be activated.
..For infinite-acting reservoirs, the lateral or fracture can be at any angle with respect to the principal axes of permeability; this allows the effects of non-ideal lateral orientation to be studied.
..For rectangular reservoirs, the sides of the reservoir and the frac or lateral must both be aligned with the principal axes of permeability.
Q: What is the RT Rate, and what is it used for?
A: The "RT Rate" curve is for multilayer reservoirs only. "RT" stands for Running Total. The RT Rate for a layer is the sum of the layer rates from the bottom layer up to and including the layer of interest. I found that comparing the RT Rate (and to a lesser extent, the RT Cum as well) to the corresponding values from data from multiple spinner surveys to be the best tool for simultaneous history matching of total well rate and spinner survey data. This allows me to start with the bottom layer and match two or three layers at a time by hand to get close before using automatic matching to refine the match.
Q: I have a follow-up question regarding RT Rate. Although the uniqueness issue is always one concern of history matching, how do you guess the RT Rate in a reasonable and small range? It looks like a wise guess is important for initial input during history match. How can I qualitatively control a good guess during history match? Is there a tool or function to help start with a reasonable guess?
A: For single-layer models with a single vertical hydraulic fracture in a circular reservoir or rectangular reservoir with fixed aspect ratio, usually the default values for area, frac length, and permeability are sufficient. For more complex single-layer models, and especially for multilayer models, it is essential to have good initial estimates before attempting an automatic history match. This is the reason I included the RT Rate and Cum curves as options. I haven't found a better way to get the initial estimates for automatic matching than to use manual history matching, working from the bottom of the well upward two or three layers at a time. The workflow in the tutorial in the Help file for Ver. 1.1 is pretty close to my current workflow, but I've found some improvements since I wrote the Help. I'm planning to include a revised example using the improved workflow in the tutorial for Ver. 2.0.
Q: The production information and the pressure information do not hold the same amount of points, 1500 and 600 data rows respectively. The well that we are looking at contains 659 data points for both production and pressure. Am I overlooking something?
A: For the vast majority of applications for PMTx it is better to smooth the pressure data to at most 20 or so data points. This is similar to what is normally done in pressure transient analysis with smoothing the rate history.
..An analytical simulator such as PMTx has two or three major advantages over a numerical simulator: 1) ease of use, primarily because of no need to grid; 2) speed of execution, and 3) ease of automatic history matching (primarily a result of #1 and #2).
..There is no performance penalty in having a large number of rate or pressure steps in a numerical simulation model. However, in an analytical simulation model, each change in pressure requires an additional term in the superposition for calculating the solution. The end result is that attempting to use the raw pressure history without smoothing largely negates one of the two major advantages of using an analytical simulator.
Q: Do you have any suggestion on a fast and effective way of smoothing the pressure data down to 20 points?
A: The easiest way is to graph both raw and smoothed in Excel, and adjust the smoothed rates until the smoothed pressure history reproduces the raw history "close enough". Most low-perm wells are produced with declining WHP for the first few weeks or months until the WHP reaches line pressure, then produce at close to constant WHP for the remainder of their life. Pressure changes during brief shut-in periods or system upsets can generally be ignored. Longer shut-ins (weeks to months) are difficult to handle with a specified pressure model like PMTx.
..If you are using Office 2003 or earlier, the smoothed points can be adjusted using graphical goal seek. Let me know, and I'll send a template spreadsheet for you to use.
..If you are using the most recent version of Office, you have to adjust the data points manually.
Q: I am simulating two cases, a well with no hydraulic fracture and another case with the fracture. The fractured one produces less than the non fractured.This doesn't make sense; can you explain?
A: Review of your file shows that you are running the forecast from 1/1/2000 to 1/1/2025. Since the well begins producing on 10/05/1980, the first 20 years of production are skipped in the forecast. The unfractured base case correctly shows the higher production rate from 2000 through 2025 because of slower reservoir depletion during the first 20 years. The cumulative production graph shows the fractured well case has higher cum production at every point in the life of the well.
Q: There is a file with the gas composition and doesn't work. I also put the well schematic to compute pressure drops and doesn't work either. Am I doing something wrong?
A: Review of your file shows that you have a casing size that is too small for the tubing size. You also have only a single pressure point in the pressure table. It is a bit confusing, but production data are entered with the date at the beginning of the interval, while pressure data are entered with the date at the end of the interval. This corresponds closely with the way production and pressure data are conventionally reported, but it makes the data entry potentially confusing. When surface pressures are input, PMTx assumes that the calculated bottomhole pressure is constant throughout the time interval. For the low permeability cases for which PMTx was designed, this is not much of a handicap, but can be problematic when modeling higher productivity wells. For best results in these cases, I recommend 1) using the linear pressure changes option, and 2) entering as many data points as necessary for the surface pressure to force accurate flowing bottomhole pressure history.
Request a Free Trial
Return to PMTx Home
A: Yes, you can use PMTx for the Marcellus. Click here for a detailed discussion. See Key to Marcellus
Q: Does your software account for desorption? Is it applicable for shales (e.g., Haynesville)?
A: Yes it does..also pressure dependent permeability and porosity. PMTx is designed specifically for fractured gas shales.
Q: There is a need to design and optimize surface facilities based on reservoir considerations. Does your software include a facilities model?
A: PMTx is the reservoir engine in "WEM Shale" by Phoenix Reservoir Software and P.E. Moseley & Associates. WEM Shale is an integrated reservoir-wellbore-facilities model that enables facilities design and near real-time what-ifs for compression and other optimization decisions.
Q: I cannot seem to get anisotropy permeability case activated. Its always on isotropic permeability. Changed well to horizontal, CBM, Shale, Dual Porosity........etc etc. The option is locked and greyed with isotropic case all the time........am I doing something wrong?
A: The outer boundary has to be either infinite-acting or rectangular for the anisotropy option to be activated.
..For infinite-acting reservoirs, the lateral or fracture can be at any angle with respect to the principal axes of permeability; this allows the effects of non-ideal lateral orientation to be studied.
..For rectangular reservoirs, the sides of the reservoir and the frac or lateral must both be aligned with the principal axes of permeability.
Q: What is the RT Rate, and what is it used for?
A: The "RT Rate" curve is for multilayer reservoirs only. "RT" stands for Running Total. The RT Rate for a layer is the sum of the layer rates from the bottom layer up to and including the layer of interest. I found that comparing the RT Rate (and to a lesser extent, the RT Cum as well) to the corresponding values from data from multiple spinner surveys to be the best tool for simultaneous history matching of total well rate and spinner survey data. This allows me to start with the bottom layer and match two or three layers at a time by hand to get close before using automatic matching to refine the match.
Q: I have a follow-up question regarding RT Rate. Although the uniqueness issue is always one concern of history matching, how do you guess the RT Rate in a reasonable and small range? It looks like a wise guess is important for initial input during history match. How can I qualitatively control a good guess during history match? Is there a tool or function to help start with a reasonable guess?
A: For single-layer models with a single vertical hydraulic fracture in a circular reservoir or rectangular reservoir with fixed aspect ratio, usually the default values for area, frac length, and permeability are sufficient. For more complex single-layer models, and especially for multilayer models, it is essential to have good initial estimates before attempting an automatic history match. This is the reason I included the RT Rate and Cum curves as options. I haven't found a better way to get the initial estimates for automatic matching than to use manual history matching, working from the bottom of the well upward two or three layers at a time. The workflow in the tutorial in the Help file for Ver. 1.1 is pretty close to my current workflow, but I've found some improvements since I wrote the Help. I'm planning to include a revised example using the improved workflow in the tutorial for Ver. 2.0.
Q: The production information and the pressure information do not hold the same amount of points, 1500 and 600 data rows respectively. The well that we are looking at contains 659 data points for both production and pressure. Am I overlooking something?
A: For the vast majority of applications for PMTx it is better to smooth the pressure data to at most 20 or so data points. This is similar to what is normally done in pressure transient analysis with smoothing the rate history.
..An analytical simulator such as PMTx has two or three major advantages over a numerical simulator: 1) ease of use, primarily because of no need to grid; 2) speed of execution, and 3) ease of automatic history matching (primarily a result of #1 and #2).
..There is no performance penalty in having a large number of rate or pressure steps in a numerical simulation model. However, in an analytical simulation model, each change in pressure requires an additional term in the superposition for calculating the solution. The end result is that attempting to use the raw pressure history without smoothing largely negates one of the two major advantages of using an analytical simulator.
Q: Do you have any suggestion on a fast and effective way of smoothing the pressure data down to 20 points?
A: The easiest way is to graph both raw and smoothed in Excel, and adjust the smoothed rates until the smoothed pressure history reproduces the raw history "close enough". Most low-perm wells are produced with declining WHP for the first few weeks or months until the WHP reaches line pressure, then produce at close to constant WHP for the remainder of their life. Pressure changes during brief shut-in periods or system upsets can generally be ignored. Longer shut-ins (weeks to months) are difficult to handle with a specified pressure model like PMTx.
..If you are using Office 2003 or earlier, the smoothed points can be adjusted using graphical goal seek. Let me know, and I'll send a template spreadsheet for you to use.
..If you are using the most recent version of Office, you have to adjust the data points manually.
Q: I am simulating two cases, a well with no hydraulic fracture and another case with the fracture. The fractured one produces less than the non fractured.This doesn't make sense; can you explain?
A: Review of your file shows that you are running the forecast from 1/1/2000 to 1/1/2025. Since the well begins producing on 10/05/1980, the first 20 years of production are skipped in the forecast. The unfractured base case correctly shows the higher production rate from 2000 through 2025 because of slower reservoir depletion during the first 20 years. The cumulative production graph shows the fractured well case has higher cum production at every point in the life of the well.
Q: There is a file with the gas composition and doesn't work. I also put the well schematic to compute pressure drops and doesn't work either. Am I doing something wrong?
A: Review of your file shows that you have a casing size that is too small for the tubing size. You also have only a single pressure point in the pressure table. It is a bit confusing, but production data are entered with the date at the beginning of the interval, while pressure data are entered with the date at the end of the interval. This corresponds closely with the way production and pressure data are conventionally reported, but it makes the data entry potentially confusing. When surface pressures are input, PMTx assumes that the calculated bottomhole pressure is constant throughout the time interval. For the low permeability cases for which PMTx was designed, this is not much of a handicap, but can be problematic when modeling higher productivity wells. For best results in these cases, I recommend 1) using the linear pressure changes option, and 2) entering as many data points as necessary for the surface pressure to force accurate flowing bottomhole pressure history.
Request a Free Trial
Return to PMTx Home