Seems like it happens all the time, doesn’t it? You’ve been craving your favorite dish at your go-to restaurant all week. You finally get there on Friday night only to find out they don’t serve it anymore. What do you do? You’ve already been waiting for a table for almost an hour, you’ve already pierced the wall of people around the bar and gotten the first round of refreshments, and your group even managed to find free parking. Then, somebody says the words, “We might as well just stay here and try to make the most of it.”
You order a different plate, but is it as good? Of course not. And sprinkling extra hot sauce or parmesan cheese on top of it doesn’t make it much better. Still, it is instinctual to try to make the best out of any given situation, even after finding out the results will never measure up to something more ideal.Oil firms are no exception.
While working with Senior Geologists & Senior Engineers at a small but successful onshore operator a few weeks back, one of the more surprising statistics that an Analytics workflow yielded was the fact that total proppant usage was actually inversely related to 12 month CUM production. Now, it should be said that the focus area involved fewer than 100 wells and that the relationship was minimal. Even still, pumping more proppant yielded smaller predicted production CUMs.
Much to the frustration of the whole group involved, we reconvened with a logical conclusion. They revealed that when they drill the pilot wells and analyze the log data, sometimes they learn that they’re in bad rock. The interval is thin, the porosity/permeability is low, the reservoir is heavily faulted and structurally complex, etc. Geology matters! So, after they learn this, the two engineering parameters they make changes to are lateral length and total proppant. They drill farther out, and they push more sand. Ultimately, what you have are very expensive, very long, heavily-propped wells that have low CUMs. Did production get better after their enhanced completions strategy? Probably. But did it yield a healthy ROI? Likely not.
And this is hoping that the Geosteering process goes well. For 250ft thick Bone Spring sand intervals, this may be no problem. But thinner plays like the Marcellus are a different story. And unless you’ve got the most up-to-date logging technology, you may be logging 90+ feet behind the drill bit. And correcting an 18-wheeler worth of pipe is a huge challenge, an incredible time expense, and dollars lost. If they were able to look at acreage and rock quality prior to drilling in certain areas, a lot of these frustrations might have been alleviated.
Let’s take a look at the data. At Drillinginfo, we have normalized Engineering parameters and leveraged both existing production & geology to develop our Graded Acreage maps, assigning entire play extents a letter grade. ‘A’ Grade is the best acreage, and it goes all the way to ‘J’ Grade which is the worst. One of the areas we have Graded Acreage calculations for is the Eagle Ford play (FIG. 1).
The Eagle Ford is a highly prolific play; the sheer number of wells that we have data for are well north of 10,000. This includes wells that have production and completions data. And depending on what factor you’d like to focus on (spud date to look at the industry at that time, targeted formation, operator-by-operator, county of focus, minimum production/production type), we can bias our analysis accordingly. First, we’ll look at some plots focusing on oil wells drilled post-2009 in LaSalle & McMullen counties in Texas (FIG. 2).
The upper left graph shows Total Proppant vs. Cum Oil 6 month. The upper right shows Lateral Length vs. Cum Oil 6 month. Both plots have almost 1400 wells worth of data. And while both relationships have positive correlation coefficients, they are low. The R-squared figures for these two plots are not good at all. And yet, for unconventionals, these two attributes are normally huge factors in improved well performance.
We can also look at Graded Acreage comparisons. The bottom left graph shows Total Proppant vs. Graded Acreage at the wellhead on the y-axis, and the bottom right shows Lateral Length vs. Graded Acreage. In this case, A Grade = 1, B Grade = 2, etc. What’s interesting about these two plots is that the longest, most propped wells appear in the D, E and F grade acreage. In fact, there are no +10,000ft wells in A or B grade acreage in these two counties!
Let’s take a closer look at two separate wells (FIG. 3).
Well #1, spudded in 2012 has a lateral length under 5,000ft, its total proppant was under 3.5 million lbs, and its 6 month CUM oil was over 150,000bbl. Let’s then take a look at Well #2, also spudded in 2012. Its lateral length was just under 10,000ft, the total proppant was over a staggering 13 million lbs, but its 12 month CUM oil was only a hair over 57,000bbl. And yet the wellheads lie within 10 miles of each other. The biggest glaring difference? Well #1 lies within C Grade acreage, Well #2 lies within F Grade acreage.
And these aren’t isolated incidents. The data show that while mean CUM oil production for long, heavily-propped wells may be higher, firms that are optimizing well performance by first looking at the WHERE component are spending less time and, more importantly, money, than their competitors who are not taking into consideration the geological & spatial component.
Latest posts by Ryan Martin (see all)
- Increasing Proppant: Not Always The Best Solution - December 3, 2015
- Using Analytics to De-risk Drilling Practices: an Offshore Example - April 16, 2015