- [Narrator] So it's my pleasure now to show you a number of log examples where the principles that Jeff was just introducing to you are applied for the sake of combining these two measurements, NMR and resistivity that we're discussing here today. Most of these examples are actually published in the literature, so you find references on the slides as well as in the paper, which is uploaded in the handouts area. Many of the examples actually come from logging by drilling applications, which is the main area I'm working in, but the basic principles of course apply for wireline data just as well as for LMBD data. And I guess for those of you who are working in the petrophysical interpretation area, dealing with NMR and resistivity data, I would sort of expect that you have been coming across cases like the ones we're showing here, and maybe you can actually relate to some of the effects that we have been discussing and that you will be seeing here in the upcoming slides, as you have seen them on your own, or it may trigger some of the thoughts on your side for going back to some of the data sets and revisit them. So we're just starting off with very simple straightforward conditions assuming that we know the T2 cutoff for the NMR data, we know the R2 parameters, and as Jeff already mentioned, if you are in a reservoir irreducible water saturation, you'd actually expect NMR bound water to be the same as the resistivity-based fluid volume phi t times s w. That of course then means that if you produce the reservoir, all the movable fluids will be the hydrocarbons and you have nice water free production. If you move downwards in the reservoir into the transition zone or some water leg, obviously the resistivity-based water volume will increase, but the bound water from the NMR will pretty much stay the same, and you have a nice indicator by comparing the two volumetrics for expecting some water production there. The first example I'm going to show is an oil-bearing sandstone in the Gulf of Mexico, data were acquired with wireline. Now with this example, as well as with all the following examples, I will not have the time to go through all the tracks that you're seeing, I will just point you to the main data and the key findings on each of these examples. So in the middle track here, we see the hydrocarbon saturation as derived from resistivity, and you can already see that there are some large variations in the hydrocarbon saturation across this reservoir section here. Translating that into volumetrics, in the track to the right, we see the oil in red, the water is shown in blue, and there's already the clay-bound water saturated out in the brown color. And one of the key questions now here is, if you perforate this top section here, do you have to expect a lot of water being produced from some of these adjacent intervals here with a high water saturation? And this is when you bring in the NMR information, where you then see the NMR here, the bound water being represented by the clay-bound plastic reducible water. You actually see that all the water in this top zone that you've been seeing before, it's all bound water, so, whatever you perforate in this top section, there will be no water production coming out of this section. That will be different of course in the lower section, where there's a lot of movable water, and you see that nicely also here the transition of water contact from the hydrocarbon-bearing reservoir section into the lower section, where you were then in the water leg. Another interesting finding from the NMR then is all these sands up here with a high water saturation, they need to be some very fine grained sands in order to have the sufficient capillary forces to keep all the water in place, for causing this to be bound water. Second example is a gas-bearing, sand-shale turbidite formation drilled in Italy, so the data comes from logging while drilling. And in the third track to the right, you see the volumetrics again. In red, there is the gas-bearing volume coming from resistivity, so you see that the gas occupies 30, 40, maybe up to 50 percent of the pore space, but not more. And again, you are back to the question, with that relatively low gas saturation, how is there a risk of producing water when you're perforating these stones? And again, it's the same story as in the previous example, the NMR actually shows you that all the water in these sand layers is actually bound water, either clay-bound brown or capillary-bound light blue, so there is no water production to be expected despite the high water saturation of 50 percent and above. Let's go to the next scenario. In some cases, the NMR can actually also separate filtrate and native fluid types by either a simple T2 cutoff approach as we've seen before, or some more sophisticated 2D-NMR applications, which are basically only available for wireline applications. And again, the starting point would be the identical fluid volumes bound water versus the water-filled volume from resistivity, but if we now introduce some water-base mud filtrate, we'll affect primarily the NMR as a shallow reading tool, but the resitivity as a deep reading tool may not see the invasion effect at all. So if we can actually separate out these different fluids by the NMR, we have a nice shallow water-base mud filtrate invasion indicator, and we'll see an example on the next slide. The NMR response that we've seen in this case is actually very similar to what we would expect also in a water-flooded or in a transition zone, but the difference here is then coming from the resistivity data where the resistivity data then also will pick up all the water in the nature formation, and thereby you have a nice differentiation of these two cases. The third case outlined here is the case of oil-base mud filtrate invasion. Again, sometimes the NMR can actually separate the filtrate from oil-base mud versus the native hydrocarbon and then we are back again to a nice indicator, using this methodology for shallow oil-base mud filtrate invasion. The example I am showing here is a light oil-bearing chalk formation in the Danish North Sea, data acquired while drilling, and this one was drilled with water-based mud. In the middle track, the T2 distribution, you see a nice bi-modal T2 distribution. The middle peak water, the late T2 peak is actually oil. If you calculate the volumetrics from the NMR, and put that on a saturation scale, you actually get the green-shaded area, which is the NMR-based oil saturation. If you then add the Archie saturation, that is the blue shaded area, and then you see in the lower section, there is actually a very nice match of the NMR hydrocarbon saturation to the Archie water saturation. But in the top section here, you see there is a large discrepancy between the two, and that is actually a fact of the shallow water-base mud filtrate invasion that the NMR experiences. And thereby you now really have, given in real time, a water-base mud filtrate invasion indicator. This information, you can then in real time use for adjustment of mud pressure or mud weight. You can also use it for identification of reservoir compartments, with different pressures or permeabilities. And in this case, the mud pressure and mud weight were not changed during the acquisition of the data. And also, if you look at the permeabilities, they don't vary substantially across this interval, so the outcoming interpretation was that these are actually reservoir compartments with different pressures, which is consistent with what the local operator knows about this reservoir here. Let's now look at the situation when some of the Archie parameters are maladjusted. It's pretty straightforward. If any of the parameters: a, m, n, or R w, are selected too low that the resulting water volume from the resistivity calculation will be too small. The opposite is actually then taking place if any of these values is taken too large and you will get a too high estimate of the resistivity-based total water volume. If you now have a reliable T2 cutoff and you can estimate the bound water from NMR, you can then use that type of information to assess whether your Archie calculation may require some adjustment in one or multiple of these parameters. An example is shown on this slide here, so this is a light oil-bearing chalk formation, again in the North Sea, but this is the Norwegian North Sea, data acquired while drilling. There were actually two sections of data, one drilled with water-base mud, and the lower section with oil-base mud. Now in the top section, this is where we see pretty much the same effect as before. The Archie saturation, shown in orange here, a high hydrocarbon saturation, and in comparison, the NMR shows the low NMR, a low oil saturation, and the discrepancy actually resulting from the shallow water-based mud filtrate invasion. Now if we go to the second half of the well, in the oil-base mud filtrate drill section, there is obviously no water-base mud filtrate that could invade, but we still see the consistent discrepancy between the Archie saturation and the NMR-based oil saturation. And this one we actually attribute to some inaccuracies in the Archie parameters, and if you do the calculation, you can actually get a good match of the Archie saturation with the NMR saturation by introducing some disconnected pores, disconnected works by means of an increase in the exponent n here, up to 2.6. The interesting observation now is, in the top zone, we had some saturation from core data, and the reduction that we suggest for the Archie saturation calculation in the lower zone is actually pretty consistent with the observed over calling Archie saturation calculation also in this top zone up here. Now, let's take a look at heavy oil and tar, as Jeff already introduced. The amount of bound water or apparent bound water in heavy oil situations can be overestimated by NMR, and even more so if you have these very high porosity tar components, you may even end up seeing an underestimation of the NMR total porosity, so you're really missing a part of your total porosity in addition. Example is shown from an oil-bearing carbonate formation in Saudi Arabia, data acquired while drilling a horizontal well here. In the third to the right track, we actually show the difference between bound water and the total water from resistivity, in green, and that is actually an indicator for the heavy oil in this reservoir, which is consistent with also the changes in the T2 distribution as we see them in the rightmost track. The shorter T2s indicating the heavier components, compared to the longer T2s indicating all the movable and light components. In the next track to the left, we also show the porosity deficit on the NMR porosity in black, and that is a tar indicator which is pretty consistent with a heavy oil indicator in the neighboring track. A validation of the tar and heavy oil component actually comes from formation tested mobilities in the next track on the right hand side, which show very low mobilities in these zones where heavy oil and tar are actually identified from the NMR and resistivity combination. The application in this case was to avoid tar mats for water injector placement, but I guess you can easily imagine that this type of information on heavy oil and tar is also of general interest for reservoir characterization and completion decisions. A little bit less common, but nevertheless something you should also consider in your interpretation are changes of wettability towards mixed and oil-wet. They will actually result in an underestimation of the NMR-based bound water, so as a compensation you would need to increase the T2 cutoff to be applied, and the opposite actually happens if magnetic minerals are present. They will actually cause an overestimation in the bound water, which you then can compensate by decreasing the T2 cutoff. Again, if you have a reliable water volume calculated from the resistivity, that comparison can actually lead you to the identification and quantification of these types of effects. The example I am showing here is for an oil-bearing glauconitic sandstone reservoir in Egypt. The data will actually be presented at the upcoming AETC Conference in Nairobi, at the end of this year, but we were fortunate to get a pre-release of the data so we can already show this to you today. Data acquired also by logging while drilling. In the T2 distribution track here, you actually see that the initial processing was done with a constant T2 cutoff of 33 milliseconds, translating into the volumetrics in the rightmost track. So there is again the clay-bound water in brown and the irreducible capillary-bound water fraction in light blue for three different zones. On top of that you see, hopefully you can actually see it, there's a blue line which represents the total water volume calculated from Archie saturation. And you see a good match in the lowermost zone, but in the middle and in the top zone, there is actually an overestimation of the NMR-bound water compared to the total water from Archie saturation. You can actually compensate that effect by introducing different T2 cutoffs. In the top zone, 10 milliseconds, in the middle zone, 20 milliseconds. And now you get a very good match between these two volumetrics. And the reason for introducing this would then be a change in the glauconite content, and that can actually be validated by looking at the crossover between density neutron in these three zones, where you see that the presence of glauconite is highest in the top zone, sort of intermediate in the middle zone, and there is hardly any glauconite left in the bottom zone, as you can also say from gamma ray and some other measurements. The whole exercise here of identifying the reservoirs and doing an accurate fluid volumetrics quantification was actually extremely challenging from using triple combo data, so there's another aspect of introducing the NMR and resisitivity combination for doing very accurate fluid volumetrics quantification in this case. The last slide with examples actually comes from a recent publication at the SPWLA Conference in Iceland by a colleague of us. Now he took the whole approach, let's say one step further, by not only comparing the results from NMR on one hand versus the resistivity on the other hand, but he actually introduced a joint inversion NMR resistivity and he also included density data. So on the left hand side, you see an example from a Middle East formation with some medium/heavy oil, and on the right hand side there is a gas-bearing sandstone formation from Argentina, data coming from wireline. And the result of the joint inversion, you can on one hand see the NMR T2 distribution where the heavy oil in the left example and the gas in the right example can be nicely separated out and quantified. If you then look deeper into the data, you again see the effects similar to what we've observed before, NMR being a shallow reading device, so the NMR-based oil saturation here is not making up the complete movable fluid, but there's some water-base mud filtrate invasion here, as you see from the comparison to hydrocarbon saturation to deep resistivity. So you basically get a similar type of information to some of the applications we have been seeing before, but in addition joining all these data together, NMR density and resistivity, will also end up to provide you more consistent and ultimately then also more reserved estimates.