- [Serge] Little bit about my background. At Shell, I working for organization called P and T, Production Technology, so what we're doing is building a static and dynamic models in order to forecast production, which will be used in operational plans, and we usually build it on brownfields, so where you have a lot of constraints, like you have to match for defects, you have to match PLT logs, cased holes, so not only operation, production as it usually happened in a more or less greenfields. That's why cutoffs definition is kind of important in my work. Today, I'll talk about petrophysical cutoff definitions. What is usually used as a cutoff? It's porosity, permeability, saturation, and the volume of shale, and there is no any consistent physically meaning workflow, to define these cutoffs. Usually, it's analysis of histogram distribution. You can see couple humps, so you kind of cut one of them, another decide that that's your cutoff, like in a porosity, saturation, et cetera. Analysis of a course, plotting's another way, like a histoplot, when you're analyzing a relationship between shale content and resistivity, and you can define a cutoff between your shales and reservoirs. Based on local knowledge about reservoir performance, I've heard of different answers, like, "It has always been done this way," or "It's the industry standard." Internal shale publication, I found this one this one very smart person put it, you can read it on the screen, so it's regularly claimed that all rock types are included in dynamic model, so you give it your static model, with the all kind of rock types reservoir engineer. He put it in dynamic model. In a dynamic model, using a prevailing production/recovery mechanism will decide which rock types contribute to ultimate recovery, and which one does not contribute, but it is based on the assumption that the dynamic scale model, in a scale at which cutoffs are defined, is the same, but it's a very rarely the case. Usually dynamic model is something quick, and five, 10, 20 meters cells, well obviously, the physical definitions of layers is much smaller. What's happened is that upscaling will blur, other words, or average out, the difference between good and bad rocks, and as a consequence the dynamic model cannot discriminate between rock types and cannot decide what rock types do or do not contribute to ultimate recovery. What we're proposing is a workflow based on application of saturation-height calculation, combined with a fractional flow computation to estimate initial water cut. A saturation-height calculation use property-based Techlog capillary pressure models, and link between relative permeability, with a fractional flow calculation. This approach provide a consistent way to introduce a dynamic element to cutoff definitions, for instance for producing only dry oil, which is really important for deep water wells. A few examples of why we cannot use, not always can, let's say permeability cutoff. Here, it's a actual core datum, and you can clearly see it's two different relationship between porosity and permeability, two distinctive trends. If we will put it on two wells, adjacent wells, you can see that here it's poor rocks, kind of on the upper upper side of reservoir, while here it's at the bottom side, so two different trends, and obviously you cannot put a constant permeability cutoff in order to distinguish between the good rocks and the bad rocks. Another example is looking at Vhale cutoffs. Here it's a Gulf of Mexico whole well. It's not very easy to see from 16,000 to 20,000 feet, in Gulf of Mexico. If we will look at the Vshale, we can see about three best possible intervals, opportunities for perforate these wells, based on the Vshale cutoff only. Actual perforation intervals, I will switch back and forth. Up here, so if we will zoom in, you can see it's Vshale, zoom in. We'll use a Vshale cutoff 0.3, so all these intervals will be gone. That's also some very smart person decided to perforate this interval, and what we can see is why resistivity is so low, because we have a very high porosity, high saturation, high permeability, but net to gross is low in this part of interval, and this is because of thin bed environments, so our resistivity is suppressed, but here you can see it's a production log, and you can it's producing as much from a thin beds as from bottom interval, where you'd have 100% net to gross in blocky sand. If we'll even use a 50% Vshale cutoff, still most of this interval, within the interval will be not in the net, will be left out, and it's very important to differentiate between dispersed shale and laminated shale. Another approach is to use relative permeability. This is a standard presentation of a relative permeability output model. It's a linear scale and a logarithmic scale, so if we will assumes that our accuracy is 1%, which is obviously very difficult to attain, at relative relative permeability experiment, but even when we assume the best case scenario, the range of saturation will be from 60 to 69%, so we already can see that we cannot use a single value for saturation cutoff, and even we will assume that it's 69% of water saturation is our cutoff, and we can define porosity and permeability associated with it, so this is example how I usually do that. That's a whole hydrocarbon column, and this particular reservoir, where I create a constant value of porosity, sets of porosities from five porosity units to 25 porosity units, and calculate constant value of permeability using porosity/permeability relationship, and I using saturation-height model to calculate what kind of saturation we have at this particular interval, above free water water level. If, let's say, I'm interested in this interval, where I have a 0.69 cutoff, and I can calculate, obviously, porosity and permeability. Let's say porosity 15% and permeability 16 millidarcy, I can apply for this interval constant cutoff values for porosity, permeability, but it's not the case. If we will zoom in for this particular interval, we can see that at about 15,000 feet, with a constant porosity and permeability, 15% and 16 millidarcy, we will have a saturation about 0.435 at about 15,000 feet, a 0.517 at about 17,000 feet, so if we will calculate a fractional flow, we can see that at the bottom of the reservoir, we will have a fractional flow of water about 50%, while at that upper part, it will be about 16%, so if you're in a deep water environment, 50% of water is not what you're looking for. You may not be able to treat such amount of water. Why it's happened? Because of, obviously, relative permeability. If we will look at the relationship between relative permeability experiment, and in a previous picture, you can see that at 0.517, you have a water saturation, you have relative permeability of oil 0.1, while at 0.435, it's the same permeability group, but you have relative permeability of oil 0.5, so significantly higher. It tells us that the only way to do a proper cutoff, based on the dynamic parameters of our reservoir is fractional flow. This is a fractional flow equation, and what goes into equation? I develop a script in Techlog, so from Techlog side, it's porosity, permeability, temperature, water saturation from saturation-height function, obviously lithology we have to take into account, and from saturation-height function, irreducable water saturation. From PVT experiment, obviously you need the viscosity of oil, this would be salinity, et cetera, whatever you can get from a PVT experiment, and from Relate, you need to know wettability of reservoir, Corey exponent, and tuning factors, for oil from experiment, so everything goes into fractional flow equation, and as a result, we have a relative permeability of oil, relative permeability of water, pseudo-flow rate of oil and water, and the fractional flow, oil, water, and a calculated viscosity of water as well, so this is a graphical presentation, what I can get out of script. It's porosity 27 porosity units, permeability associated with this porosity group, water saturation, fractional flow of water, relative permeability of oil. It's a parameter of Corey exponent, maximum relative permeability of oil, relative permeability of water, water viscosity, and a pseudo-flow rate of oil and water. It is important to define a pay cutoff, but sometimes if you're building a static and dynamic model, you have to define size of aquifer and parameter of aquifer, because it gives you energy for producing hydrocarbons. You have to define the aquifer properly, so you have to differentiate between shales, thin beds and sands, so a fractional cutoff also helps you to define, not only pay cutoff, but to properly define aquifer parameters. Another application of this methodology is determination of optimal perforation intervals, which is very important, again, for deep water environment, so here, for example, we have a reservoir with a different porosity and permeability saturation, so we applied the fractional flow script, this methodology. At the bottom, you can see two tables. One is a normalized flow rate of water, so it's that ratio between flow rate of water and certain cutoffs, and normalized but a flow rate of water without any cutoffs, so if we will perforate whole interval. This is a matrix of SW and permeabilities. Same thing for oil. Different SW, different permeability matrix, what kind of normalized oil flow we can expect in a different case, with a different porosity and permeability cutoffs, within this interval. If we will put this data on a graph, you can see it's a normalized flow, where solid lines is normalized flow for oil, and dashed lines normalized flow for water, at a different permeability cutoff, so you can see, if we use a cutoff for Sw, 0.6, at different permeabilities, we can see that if we will perforate anything, any interval with a Sw more than 6, your flow of oil will not increase. It'll be constant, while we can clearly see increase of water production, so it clearly tells to you that 0.6 is the cutoff for perforation, where you can get a maximized production of oil and a minimized production of your water, so it's a optimal cutoff. Summary of the presentation is that parameters, at this moment, for selection cut-off values in the industry are not rigorously defined. Constant cut-off parameters may not be applicable in a reservoir with significant hydrocarbon saturation column, and that's pretty much where my experience is, in the Gulf of Mexico, where have a significant hydrocarbon saturation column. If your reservoir is flat, then it's a different story, so proposed methodology includes saturation-height calculation, utilize property-based Techlog capillary pressure models, and the fractional flow calculation, linked through Python Script to the relative permeability models, used by reservoir engineers. It is developed as workflow for cutoff definitions. This provides a consistent way to introduce a dynamic element to cutoff definitions, for instance by requiring pay to produce only dry oil. I'd like to thank Shell for permission to present this paper, and the Peter Doe who helped me to develop it, and also for SPWLA, our leaders Sharon, Zoya, And Shipeng, who organized this Distinguished Speaker Program, and helped me to present it. Thank you.