clemrissa / combined_nmr_conventional_logs-shaley-sand-analysis-using-petrogg

Mihai's PetroGG modified to be used with our shaly-sand Gulf Coast NMR data.

Basic-Well-Log-Interpretation-using-PetroGG-Adapted-to-Gulf-Coast-Shaley-Sand-Example-using-NMR-and-Conventional-Logs

This repository started with PetroGG written by Mihai in GitHub that has been modified to be used with our Gulf of Mexico NMR shaly-sand data. These data are in the NMR.txt text file provided. PetroGG provides for an excellent foundation for our work. We appreciate all the fine work of Mihai on PetroGG and the suitability of his code to be expanded upon for this application of PetroGG.

We have made a few changes to PetroGG. First of all, we have modified the code to use Vshale and not Vclay as was previously employed. Almost all shales have less than 100% clay volume. We have personally sampled some of the greasiest, gumbo shales to find that the maximum clay content was only about 65% according to the XRD or FTIR mineralogy results. Therefore, estimating the endpoint parameters for clay points is a challenge since these 100% clay points are imaginary points and do not exist in the data. Instead we use shale point parameters which can be determined from most shaly-sands data sets. However, with sufficient mineralogy data from core samples or elemental log data like Schlumberger's Elemental Capture Spectroscopy (ECS) we can model the volume of clay for future applications.

We have included two additional Saturation models suitable for shaly-sand analysis. Mihai has another excellent repository with various Petrophysical routines including saturation models. We started with his Waxman-Smits and Dual-Water saturation models from this repository and then made a few modifications for our saturation code found in this project.

For Dual-Water we are using the George Coates (1) MRIAN method that was primarily used at Numar to integrate conventional log and NMR log data for the interpretation. NMR data has been found to be quite useful in assessing reservoir quality and defining the volume of non-movable Capillary Bound Water (MBVI). MBVI is quite different to the volume of Clay Bound Water (CBW) or CBW saturations (Swb) associated with the waters chemically bound to the shales. CBW is determined from the difference between Total (PHIT) and Effective (PHIE) porosity.

In this example we are using NMR data from the older Numar tool that was only capable of measuring a NMR effective porosity (MPHI) and NMR Bulk Volume Irreducible (MBVI) in the effective porosity system. This tool did not have the capability at this point to employ the PR06 activation for the determination of the volume of Clay Bound Water (CBW).

MBVI is calculated from the NMR T2 distribution by partitioning this T2 distribution into MBVI and Free Fluid (FFI). FFI pore volume has larger pores with movable fluid where MBVI has smaller pores with non-movable fluids. This partition point is called the T2 Cutoff and for this clastic well the T2 Cutoff was 33ms. We typically compare our log analysis based Bulk Volume Water in the effective porosity system (BVWe) to MBVI, in the effective porosity system, to determine what intervals have hydrocarbon saturations at irreducible water saturation with no movable water. When BVWe is higher than MBVI, then we expect some movable water depending on the relative permeability of these fluids.

This movable water can be seen in the depth plot below. From a depth of 4610' and below the Waxman-Smits Bulk Volume Water (WSCBVWE) is greather than the NMR MBVI (shown as the dark blue fill) and the shading then becomes cyan in color. This cyan fill is where the bulk volume movable water calculated from our log analysis is greater than MBVI and we would expect movable water. Above 4610' the WSCBVWE is at MBVI with no apparent movable water. Above 4610' we would expect water-free hydrocarbon production. Above 4610' all water is non-movable, capillary bound water even though there are some relatively high water saturations. This demonstrates one of the the key advantages of using NMR data in our petrophysical interpretations.

For our Waxman-Smits saturation model we use the Hill, Shirley and Klein technique(2) to solve for Qv from Swb as shown below:

Qv = Swb/(0.6425/((Fluid_Density*Salinity(kppm))**0.5) + 0.22) 

We then use the Waxman-Smits saturation equation provided by Crain in lieu of an iterative approach.

We have found that Waxman-Smits method is a bit more flexible to use in shaly-sand interpretations in that we can model the Waxman-Smits cementation exponent m* to vary with Swb according to the trends observed in the wet intervals from our Swb vs. apparent m* cross plot as shown below. Apparent m* increases as Swb increases in the wet intervals which leads to less apparent hydrocarbon saturations being calculated with higher Swb. This is not true if one uses a constant m* for the analysis. The m* at Swb=0 y-intercept is the Archie m, but in this case m* increases with increased Swb. We have also found that a wrong Rw can lead to an unrealistic y-intercept to give us feedback on our choice of Rw too.

In the plot above we calculated the apparent m* data in another program to make this plot using the following equation:

m* apparent = log10(Rw /(Rt*(1 + Rw*B*Qv))) / log10(PHIT)  

This will be included in a future update.

Depth Plot:

1. Coates, G.R., Gardner, J.S., and Miller, D.L., 1994, "Applying pulse-echo NMR to shaly sand formation evaluation", paper B, 35th Annual SPWLA Logging Symposium Transactions, 22 p.

2. Hill, H.J., Shirley, O.J., Klein, G.E.: “Bound Water in Shaley Sands - Its Relation to Qv and Other Formation Properties”, Log Analyst, May-June 1979.

3. Dacy, J., Martin, P.: "Practical Advances in Core-Based Water Saturation Analysis of Shaly Tight Gas Sands", SCA, SCA2009-29, 2009.