Disclaimer: I have no formal (or, for that matter, informal) training or experience in the oil sector, so I'm probably getting ahead of my headlights with regard to some of the comments I will be making later.
First, this. Earlier this morning I mentioned that there seemed to be a "quantum" jump in the quality of the Bakken wells (based on 24-hour, 30-day, 60-day, and 90-day IPs) when going through the 2017 Bakken permits. It appeared that somewhere between 2016 and 2017, things in the Bakken changed significantly with regard to better wells.
One almost wonders if the "near-death" experience visited upon the Bakken operators by Saudi Arabia (2014 - 2016) forced "them" to improve their strategies and technologies.
I remember early on, during this time, maybe even as late as 2016, really, really smart oilmen simply said, "more sand, more oil." It was all about simply adding more and more sand.
But look at the completion strategies in 2018 and going forward: maybe they're using more and more sand in the Permian, but in the Bakken, it appears we're leveling out at around 8 million lbs and fewer, rather than more, stages. Granted, "we're" drilling in the "best" spots in the Bakken but that would be true in the Permian, too, I assume, and yet some of the best Bakken wells I've seen used less than 8 million lbs sand and only 35 stages.
Much could be written, but let me post what a reader sent me. It's obvious he's been thinking about this even more than I (have been thinking about it). Here was the reader's note [the comments in red/bold are mine, not the reader's]:
Somewhat brief, hopefully accurate, theory concerning the increase in IPs in the 2016 - 2017 timeframe ...Isn't that interesting?
Many operators had, by that time, started to declare that diverters were being used in completions.
EOG most probably started this earlier as they depicted controlled, HIGHLY increased fracture intensity in their presentations prior to 2016.
Fast forward to today and the ramifications of Extreme Limited Entry (XLE) ...
I have studied the September, 2018, article from aogr.com from Liberty Resources regarding Extreme Limited Entry in attempts to fully "get" what is happening and project future implications. [To repeat: the September, 2018, article from aogr.com from Liberty Resources regarding Extreme Limited Entry.]
Exciting stuff.
In a ultra condensed nutshell, what these operators are doing is maintaining a VERY high pressure "bubble" underground so as to induce fractures (technically, opening up pre-existing fractures. This is a crucial - yet often overlooked distinction).
This pressure "bubble" functions best if it can be maintained at 1,500-2,000+ psi to induce connectivity (create complex fractures).
By identifying homogeneous rock that will tend to fracture with similar pressure levels, thief fractures will be minimized. (This directly ties in with the number, length, and placement of stages as each stage is a somewhat 'stand alone' entity. [This is so incredibly cool.]
Using extremely sturdy metal, the perf entry points - precisely located - will not enlarge as a consequence of proppant friction. A ton of sand per linear foot is a LOT of scouring material.
Perf clusters now number 13 to 15 per stage, up from 3 in earlier years.
Diverters, both near wellbore and far field, are used to temporarily block undesirable frac propagation so as to bring about an optimal case of Stimulated Reservoir Volume (SRV).
Speculative bonus regarding halo effect?
I am starting to suspect that later frac jobs are deliberately "touching" earlier wells' drainage area by fracturing and propping unstimulated areas of the parent wells' rock and enabling the newly frac'd/propped reservoir to flow into the older wells' wellbores.
Now this, for free. I am providing a link at no extra cost -- free to all subscribers -- JPT Extreme Limited Entry Perforating. I had completely missed this paper. Again, thanks to a reader I've learned more about the Bakken. I find this absolutely fascinating.
Yes, I know. I am inappropriately exuberant about the Bakken.
By the way, the article the reader was referencing: an "editor's choice" article from September, 2018.
I've been focused on the halo effect, what is now called the "parent well uplift" phenomenon. It's very possible the oil industry was more concerned with damaging parent wells when drilling new wells, which ultimately led to extreme limited entry perforating. I agree with the reader: my hunch is that the operators are choosing very, very carefully where best to put daughter wells and how to best complete (frack) them.
The operators are probably more concerned about preventing damage to existing wells. I'm fascinated by the parent-well uplift phenomenon. It appears these may simply be two sides of the same coin.
Again, as mentioned earlier, I'm probably getting well ahead of my headlights.
Good stuff. Bakken wells using 10,000 psi to fracture, maybe for at least a year? Northern white sand also.
ReplyDeleteGood stuff. I agree. Very, very interesting. There are just so many interesting things about the Bakken. One of the other interesting things is that despite the seams (the targets) being very, very thin, they seem to be very, very "regular." This makes it easier for the geologists to find the target and keep the wellbore in the target. No one talks about that but when reading the geologists' reports, over and over and over, it is amazing how "predictable" the targets seem to be. But again, I'm well ahead of my headlights.
DeleteGreat post. General comments:
ReplyDelete1. Yes, productivity growth has been impressive. What's even more impressive is getting such improvements NOW, ten+ years into the ND Bakken play. You would think that such improvements would tail off. Late innings and all that. But instead, perhaps a lot of the earlier wells were not done optimally so there was room to improve.
2. How long improvements continue, who knows. In addition to just running out of knobs to twist, there is the issue of exhausting the Tier 1 land. For instance, EOG drilled through their extremely awesome land by 2010 or so (Sanish or Parshall, can't remember). But then after that average EOG productivity went down, not up. So sweet spot exhaustion can happen.
3. In general, you have to look at number of completions as well as average production. The reason is "high grading" (and the opposite) when prices move down/up. IOW, if you drill much fewer wells, you tend to drill the best ones. Thus, average productivity goes up, just from "culling the herd" to keep the best cows only, not from making them all better at making milk. However, there are other analyses, you can do, that tend to show there is MORE than just high grading going on. The wells in similar locations are just better now than earlier.
4. I would be very careful of assuming that IP translates all to higher well EUR. I am a Bakken booster and so is Rystad Energy. But we both agree that decline is faster for bigger wells. Sometimes companies try to glide past this point (not lying, but not explaining). On the opposite side, you have peak oilers, who say it is all front loading, no EUR improvement.
4.1. My seat of the pants estimate is 50% front loading, 50% EUR improvement. So if the IP doubles, figure the EUR goes up 50%. Note that even just the acceleration helps the well make a better investment because of time value of money. But if you are a landowner and see the first month double, don't expect the total well to be double forever. Will also decline faster.
4.2. I don't have any deep reason for the 50%/50% guess on acceleration/EUR. It's just half-way between the extremes and feels right by gut. Intuitively, when you do a bigger frack, you are probably doing BOTH cracking rock that would never crack (new oil) as well as making it easier for oil that would take a long time to get to the wellbore, to get there a little earlier.
5. By ~2010, perhaps earlier, the whole basin had mostly moved to two-mile laterals. So the improvements are not from getting longer, but better.
Specific comments:
ReplyDeleteProduction average IP Numbers
Per shaleprofile.com, plus analysis:
3-month IP* (#wells)
2005: 6,750 (30)
2006: 9,502 (74)
2007: 17,529 (163)
2008: 25,184 (428)
2009: 24,036 (469)
2010: 28,304 (776)
2011: 27,389 (1,235)
2012: 28,408 (1,798)
2013: 29,926 (1,984)
2014: 32,063 (2,160)
2015: 32,865 (1,423)
2016: 40,445 (723)
2017: 46,733 (971)
2018: 49,197 (1,229)
Comments:
a. 2006: significant improvement in productivity, more wells drilled also. Still very small amount so presumably this was all real improvement.
b. 2007: more wells and higher production both. This is getting better and not from down grading.
c. 2008: Another large improvement and in spit of doing even more wells.
d. 2009: Similar productivity and number of wells to 2008 (a stall in improvement).
e. 2010: 15% improvement in productivity and almost double the well count. Great year.
f. 2011: Flat productivity. But much higher well count. So keeping things on track despite down grading.
g. 2012: Slightly more productivity, but many more wells.
h. 2013: Similar to 2012. Slightly more productivity and wells.
i. 2014: About 10% productivity increase and with slightly more wells. Good year.
j. 2015: Same productivity, but a drop in well count. Not a great year.
k. 2016: Large increase in productivity, but big drop in well count. This was partially high grading. But also some improvement of technique (can see in other analyses of top wells, coming). But hard to disentangle how much was high grading and how much completion improvement.
l. 2017: 15% improvement in productivity and slightly more wells. That is getting better despite downgrading. Great year!
2018: 5% more productivity improvement and in the face of doing more wells. Another good improvement year, but not as big a jump as 2017.
[Net, net: yeah, 2017 was the big change.]
*This is the 3 calendar month IP. So, the first "month" will average being a half month long. Also, the well is offline some amount of time from operations, etc. These make the 3 month IP smaller than in a perfect test. But if you look year to year, it's apples to apples, since the average impact is probably similar. In addition, this is the 3 month oil cum. No gas, no BOE. But gas is a pain anyways and gets almost no money. ND only. Bakken and TF only (no Red River, no Madison). And horizontal wells only.
I really appreciate these long notes. I'm on the road traveling and won't get back to these until tomorrow or Wednesday, but I will certainly spend some time with them. But I wanted to get them posted.
Delete