VideoPoints

• ••• • •• • ••••• • •
GEOL 6390 3D Seismic Exploration - Day3 11-03-23
Help Tour
© Distribution of this video is restricted by its owner
Transcript ×
Auto highlight
Font-size
00:50 Off. We go. So we right. Um Your time.
00:55 So they may have shown up in different order than what I have them
01:00 because sometimes canvas uh depending on how tie, set it up, it
01:04 have not only a different order that questions are presented, but even the
01:11 can be swapped in that way you say, oh, the number to
01:14 seven is C because, well, , number seven might be a different
01:19 and C might not be, OK, so don't get stressed
01:25 So here's the first one I had on the left, what I've got
01:32 coherence variance. Uh That's a, a, a channel here.
01:41 So I got a channel and I'm little, I'm looking at the area
01:47 for me. OK. OK. Yes, yes. OK. Spon
01:57 seismic camp too. And what hopefully see here is blue in this picture
02:04 positive and red and yellow are Then I have a kind of like
02:14 red, blue, red reflector and it becomes blue, red,
02:18 it becomes negative. So what's It's changed uh the phase in turn
02:26 inside out and that's indicative of a accumulation. Now, the part with
02:33 channel, where does the channel come or channels if the channels are filled
02:37 sand, that's a good thing to the channel with. So there's
02:42 there's a couple of things going on . It's a little brighter in
02:46 but two that much brighter, definitely a different phase change and then it's
02:53 to that channel. OK. So indicator of uh a bright spot,
03:01 gas accumulation, any push back on . Now, multiples you can't see
03:10 water bottom. Multiples are always gonna like the water bottom. Then in
03:16 difficult situations, you might have a generator that is, let's say the
03:22 of a carbonate, the top of volcanic. Well, then it's gonna
03:28 like that carbonate and that volcanic it'll repeat. OK? So a multiple
03:33 gonna have a very specific pattern that across everything. It's not gonna look
03:39 geological at a particular point only. ? Now this one um you're gonna
03:47 better at this when you start picking probably today and tomorrow. So I'm
03:54 which way do you pick? And you, if you're not looking at
03:59 data, you're gonna say, I'm if you're only looking at this bright
04:03 , you're gonna say it goes OK? If you just look at
04:07 pattern, if you're an engineer, that I wanna downgrade engineers,
04:14 But here's what I see on this set. I see, first of
04:21 , a normal fault and this block slid down. So this is the
04:27 wall. This is the foot So you can see here it's slid
04:31 and you can see here that it's down and you can see maybe here
04:37 it's slid down. Definitely here it slid down to the left. So
04:43 I got a normal fault. That this guy has slid down from someplace
04:50 here and how much? Well about much or about that much?
04:55 So you're using a couple of one you're using, you're gonna
05:01 unless now when you have strikes with , it's different. Ok? As
05:06 normal thought, if you have a slip fault or a reverse fault,
05:13 the relative thicknesses do not change across ft unless there's something called syn tectonic
05:25 . So Zack tell us everything, know about sin tectonic deposition, you
05:35 , you should say it in a voice, right? Ok,
05:40 So where is my geologist Carlos? me everything you know about syn tectonic
05:52 because you are right. So the is slipping and sin just means
06:03 in this case with at the same . So as I have more accommodation
06:08 it slips, have more accommodation space I'm I'm filling it. Ok.
06:12 it's, it's not uncommon for the side. I'm sorry, not uncommon
06:19 the hanging wall side, this part sliding down to have thicker horizons if
06:25 in tectonic deposition, but now you're gonna have them all of a
06:30 the same higher up. If, you change in one spot, you're
06:33 change in other spots as well. this one here, we're just
06:38 um, we're gonna move down in . This one is the one that
06:50 of like the obvious wrong answer. right, because the pattern looks the
06:55 if you're just looking at the, amplitude and the peak drop and that
07:00 that looks very, very reasonable. then this one is not moving down
07:05 with respect to this pattern. Look at this one right above
07:09 Look at this guy that correlates to . So this strong event is probably
07:15 one right here. OK. And the one to pick uh this
07:21 It's going in a total wrong OK. Another strong event but not
07:27 with no snowball. Yes, OK. OK. I think I
07:34 this one right. Because OK, because like on the actual block of
07:38 belt, it is wide but then to the right. And yeah,
07:46 changed. It's the amplitude has changed bit. Yes. OK.
07:51 and uh and you don't like No, that's OK. I understand
07:56 don't like it, but here's what have to do. We have to
07:59 two things. We have to honor seismic data and mhm lots of geologic
08:14 and tectonics. So you have to both of them at the same
08:18 So like everything is gonna be a . OK. So in and this
08:28 , what is taking the opposite? reason this one is wrong is moving
08:34 wrong everything else that fall lock to right of the normal box. Thank
08:42 . This, this one is higher this one is higher and this one
08:50 fire and well down here, it's to see. It gets kind of
08:55 . This one is higher up So I gotta be higher on the
08:59 hand. But just, just remember . Now you're gonna be picking faults
09:03 you're gonna after by doing it, gonna get comfortable with it.
09:09 If it were reverse faulk, it be the opposite pa and they push
09:15 . It's OK to get push back . How about Jessica? And um
09:22 year. Yes, for there. , I'm here. He says,
09:43 . OK. Now this one is like the practice one. It's um
09:47 a little tedious. OK? So first thing you need to do is
09:53 what the impedance of the top guy the green one are. So I
09:57 gave you, I gave you that I gave you the little wavelength.
10:02 first of all, 2.2 is greater 23.2 is greater than three. So
10:08 going from high impedance to low I've got a zero phase wait,
10:20 reflection coefficient. That means my wave is trough peak trough. That's what
10:26 is. The first reflection that I you tells me the polarization of the
10:33 . OK. Now, when we across here, 2.0 times three is
10:43 2.0 times three is six. So change in impedance from green six to
10:52 six is zero. So my reflection from green yellow should be zero.
11:03 um oops sorry here that's zero, zero. That 10, that
11:12 This 10, this 10, this , this one's wrong. OK?
11:18 another one that's like that if you to multiply 2.4 by 2.5 you're going
11:26 get sick. So the reflection coefficient Greek to the baby blue is
11:35 So trace four and two. Here roll zero, not zero,
11:43 zero. OK? So it's gonna one of the top two now and
11:48 we have to worry. So we uh oh 2.0 to 3 different
11:57 We still gonna get zero here and we have three and four. That
12:02 be a positive reflection coefficient. This one should be 88 times 2.8
12:12 less than two times. So they be negative. This one should be
12:17 57 and this one positive and So it's gonna be if you want
12:35 knee up, go back you. . With that. Any unhappiness?
12:47 . Remember I taught classes with 100 20 engineers and they would come
12:51 they say, well, you in, in my native language in
12:55 , the number four is actually looks a seven. I mean, they
12:58 told me this. Ok. And starting to go, they're great engineers
13:03 great negotiators. They're gonna be good people. Yeah. But the,
13:18 question those. Right. Right. that tells you what the polarization of
13:24 wavel did. So here, here's happens in the real world. Maybe
13:32 buys your company, never happened, ? OK. Somebody buys your company
13:38 then you got all this data and you're not there anymore. And then
13:46 got to figure out what's the what's the polarity of the data?
13:50 not, no, not all you is the data. So in that
13:54 book, which is kind of like coffee table book on 3D interpretation by
14:00 Brown. It's a very nice the kind of book you borrow from
14:04 don't return. OK. It's too to buy maybe. But anyhow,
14:10 says the way you can figure out polarization is if you got a shallow
14:16 , that's gonna give you a positive coefficient. So figure it out
14:20 from that if you have very, shallow gas, that should have a
14:27 reflection coefficient and figure it out from . If you have an envious intrusion
14:32 will almost always be a positive reflection . And then in many cases,
14:39 definitely not all basement should give you positive reflection co but if you got
14:46 hydrate on top of basement, it be the other way around. But
14:49 , but sands and shales on basement give you that. So you have
14:52 figure it out from the, from data. And uh and the way
14:59 want to do it in an oil , if you can is have a
15:03 log with the sonic and a density , then you really know what the
15:07 is because you generate a synthetic. . So that's the one. Remember
15:12 said, you gotta pick what you see. So this was in that
15:17 , I think I showed, I you, hey, take a look
15:19 that paper, all the papers, at this one. And uh I
15:24 show this picture and what we have a reflection going up dip and we
15:32 see these are going up dip here below this is going up dip and
15:37 guy down here is going up So there's ge going up dip but
15:43 this one reflections disappear. So what ? We're deeper in the section,
15:53 know, maybe 20,015 to 20,000 would clear. We've been used gas in
16:00 system and the gas has made the reflection um match the scale. So
16:10 do what we say is what we . You know, that's kind of
16:18 hard thing that you contribute to, ? It, hard, hard,
16:23 that we're not, we're not gonna able to teach you to do
16:27 I lost my thanks. OK? comments on that? OK. But
16:35 , this one should be pretty It's more about color than anything else
16:40 how organized you were in kindergarten with Crayola brand crayons and whether you put
16:45 in the order or whether you just them all broken in a file.
16:50 ? And so I've got different spectral . 1924 35 Hertz. We'll talk
16:56 spectral decomposition later tomorrow. Um But gonna tell her, oh, so
17:06 yellow. So like this guy here kind of yellow here, I'm yellow
17:12 , I'm kind of yellow. This yellow. Well, then the tuning
17:17 gonna be where the um the wavelength at a, at a quarter
17:24 it'll be stronger at that tuning So uh if it's yellow,
17:29 it's halfway between red and green. 22 Hertz. OK. Any problem
17:38 that, OK. And that's the of uh red, green, blue
17:46 of spectral magnitude components as everybody kind knows the order of the colors and
17:55 rain. OK. He learned that five years old, OK. This
18:00 of goes back to the same dim idea where we have lith application of
18:09 as they're older and buried longer through is a little bit of mechanical
18:14 but mainly diogenes. So, we've to figure out, ok. I've
18:19 impedes here of sand and of shale I've got a sand shale reflector.
18:26 I'm going from sand to shale at level. Low impedes to higher
18:32 I should have a positive reflection. . Here's my way. What I
18:37 do it. So this one's Thanks. He is definitely a Bronx
18:47 . Then we can, we can down each one shale on sand.
18:51 , that's gonna be high and beams lower. So that should be upside
18:57 . So it could be uh this or this one. And then let's
19:04 here where they cross. Now, got sand on top of shale or
19:10 , shale on top of sand and are uh the same evening too at
19:20 level. So I have a zero coefficient. So at this level,
19:25 should have a zero. OK. this one's OK. This one's
19:28 This one's not, this one's And then I have sand on sand
19:33 the same level. Well, they same pathology, they're gonna be zero
19:38 coefficient. And so this one's And this one's OK. So,
19:44 far I just, by going that , we can say this is the
19:49 . OK. Makes sense. You don't like it, but it
19:54 sense, right? All right. . All right. This one is
20:01 geological. And so in some uh trying to capture your knowledge of
20:11 because you know geology, but you have a hard time with this,
20:14 you all know some geology and then you're a little bit comfortable time.
20:20 um I got a fault here that picked and then the fault stops
20:26 OK. So, um the simple is I eroded. I had fallen
20:39 up to the surface, whatever the was at at and then I just
20:46 it all the way. All on I have is part of a,
20:50 turban here. OK. And let's some of the wrong answers, the
20:59 of very strong mythology. But that explain why the folk doesn't increase.
21:06 then you're gonna say the folk souls the surface. Now I'm using one
21:11 those deep voice words. I hey, we uh oh we my
21:18 and Raba Newton, they call they call me. So they play
21:24 you, they call me demo and means luing, right? So Cole
21:41 the glue or to weld. And de Cole is to, to
21:47 So when we have ARIC fault and sediments, we have a steep part
21:54 the soft uh the fall and then sold out sole like the sole of
22:02 shoe. OK. And then, those are you? Uh So Jessica
22:07 Javier, I'm waving my arm in ground like a sled runner.
22:12 Like a sled runner and then the part is fine. And then as
22:18 go horizontal, that's my day coma where things just slide. Ok?
22:26 just slide horizontally. So that sounds . But that's gonna be at the
22:30 of a fault. It's never gonna at the top of a fault.
22:34 ? And it would happen maybe for fault or an over thrust fault,
22:38 no, they come on. So a fancy fancy word. And if
22:43 fault were generated after this package, , then default should go right through
22:48 at all. Right. Here's one uh tuning again. And I've got
22:58 reference reflector all by itself on And then I have a positive and
23:04 reflection coefficient uh at the bottom uh a kind of a wedge like I've
23:10 showing you, I get interference of side lobes with the main lobe right
23:18 and I trace eight and 10. my tuning thickness someplace in here.
23:26 then when they get close enough I get weaker, weaker,
23:30 And if I went to zero, would get here. OK. So
23:36 the by the composite wavel, which the two together uh decreases in when
23:43 below his thickness. Hey, this , I went over in clad and
23:51 has to do with picking zero crossing and cross. And I had an
23:58 in the class where it showed where horizon is ladder uh then a small
24:06 of noise can move me, I shouldn't say horizons, but with
24:09 wave and flatter on the signal, little bit of noise is gonna move
24:13 up and down and give me an . Ok. So that's the correct
24:19 . Um, here a well, to a zero crossing. That sounds
24:25 . But when you do, ties, you don't look at a
24:29 , you're looking at the whole right? So it's gonna use the
24:34 crossing and the peaks in the And it's also true for the last
24:39 , use a co a cross correlation . Um It doesn't matter whether you're
24:47 , you wanna pick a peak or trough, you're using maybe 40 milliseconds
24:51 the date. You're gonna tell it we cross, correlate with auto pickers
24:56 tomorrow that, yeah, tomorrow that peaks and troughs you're gonna put in
25:02 number like 40 milliseconds or plus or five sample. So it's gonna include
25:08 , troughs and zero cross. And it's not that and then this one
25:14 de convolution. I'm assuming that two three of, you know what predictive
25:17 convolution is and the other just they this a, a distractor. And
25:24 you do testing and stuff means you something out there that way. I
25:29 no idea what that is. That's be it like dick home. All
25:34 . So it's a distractor. predictive de convolution. Does it add
25:39 to peaks and troughs. It adds . If it, if, if
25:42 adds noise, it adds noise to . But you don't need to know
25:46 they didn't know what the first is he pushed back. Ok. Now
25:56 one, hopefully, hopefully you looked the, um, the tour of
26:02 Cora 3d powerpoint, I sent I, I got a, I
26:06 a couple of pictures in there. . I showed what faults look
26:11 I showed you what uh igneous sills like some shallow uh hydrocarbons that would
26:22 drilling hazards, what they look And uh then I showed some migration
26:28 . Now, this is a very , of course, because everybody here
26:34 terribly enjoyed Professor Howie Joe's class uh 23 months ago. So I'm sure
26:41 drew all kinds of ellipses and showed they constructively and destructively interfere. And
26:47 probably showed you good migrations and bad . He talked about smiles and frowns
26:54 stuff like that. OK. So guy here, these are ellipses and
27:01 just just not migrated, right? not imaged correctly. So those are
27:06 swings, right? Uh Here everything continuous. So it's not a normal
27:15 reverse call. Uh And and they is di would be if you had
27:25 , your site, it would be ugly looking and this doesn't look
27:33 Our Ignas dikes are down and here , we have a hard time imaging
27:43 because you did migration. I know A O talked about imaging salt domes
27:50 the flanks of salt domes and even underside of a salt dome that where
27:57 dye has kind of shifted to to the right or the left.
28:02 to do that, what we do we go through the sedimentary sequence,
28:06 velocity is increasing, increasing, And by fair Mat's principle and snows
28:14 , if that ray is turning, , then it turns all the way
28:18 . It can actually image the bottom the overturned part of the salt dome
28:25 the canopy is the correct word and come back, turn around, come
28:30 the surface. I mean it can that uh to look at the vertical
28:34 can do that. But we have sediments in the igneous volcanoes. We
28:40 have nice sediments around. We have kinds of volcanic stuff. And you
28:44 of your volcano from your third grade experiment, probably everybody here picked volcanoes
28:52 their third grade science experiment. And has this funny like tree kind of
28:57 with little dikes coming off to the . We can't image that yet.
29:04 too hard to image they're vertical, not a plane, there are like
29:09 just doesn't work well. So these migration artifacts. OK. Here's color
29:18 um one of them is gray The other is like a sepia
29:23 So those are the two that are easy see even with this orange thing
29:27 top of it, I can see channels coming through here. I see
29:32 nice channels coming through here. I have this stretch my mind to see
29:41 in the, in the polychromatic color . Oh, is that one?
29:50 . Um OK. This one I to the I, I went through
29:56 lecture talk about um data loading. then I went in, we're here
30:05 Saturday and I added these to No, then I send it to
30:13 and say, hey, let's look the lab one B and I highlighted
30:15 parts that I updated and this was of those parts. OK? Highlighted
30:19 in yellow. So if you didn't at that highlighted part in yellow,
30:24 , now hit yourself in the OK? Um Then what we uh
30:34 done, what I've done is I left the patrol default and this example
30:39 the defaults are a minimum and it's , the data you're playing with the
30:44 100 and 34,000 and the maximum of . And then when I define my
30:51 bar, I'm gonna make it nice pretty minus 25,000 plus 25,000. So
30:58 see how, how good um Yai grading this one because this is,
31:04 kind of questions are really painful to up in canvas. So you're gonna
31:08 at it and scratch your head. ? UT make sure they get partial
31:14 because most folks might get well. work, not all of them.
31:27 . OK. That's good. And you choose B then it takes a
31:31 . OK. That's good. that's the way I would like to
31:34 it. Thank you. Um So color inside are different for, for
31:42 each image. Why? Because um the one case, hang on,
31:52 just both sizes, both images are . OK? We gotta fix this
32:05 you guy. So the good thing canvas you can picture for everybody right
32:19 . It's great. It's great. . So the color bin sizes is
32:24 colors and we're going from minus 25,000 plus 25,000. The same number,
32:30 sizes are the same. You're gonna 25,000 minus minus 75,000 divided by
32:38 That's the color in size. Now the the end of five,
32:50 I had it, I I just it. That's what happened.
32:54 So really the bin size, the of inside city that did you take
33:10 ? Yeah, if you make 55 25 51,000. Is that what it
33:17 divided by 255? You should have both cases. Thank you. Then
33:26 data bin size is going to be the eight bit image is going to
33:35 um the difference between these two So about 350,000 divided by 255.
33:44 it's gonna be 1360. It's gonna a big it's gonna be a big
33:48 . Ok, so I've got I'm my arms for those far away.
33:56 maximum is 1 34,000. I'm my maximum is 212,000 minus 134 which
34:09 Z I thought Zack was gonna do in his head. He's using his
34:13 . That's ok. Yeah. Should 380 or something. One's got a
34:28 sign. So you're gonna OK. six or 363 46,000 divided by
34:41 Ok. That's gonna be about So my data are beings when I
34:48 this realiza I push this realization OK? And that means all,
34:55 the data, if it's between zero 1360 it's one call 11 integer and
35:02 from 1362 2720 it's the next color the next B OK? And that's
35:13 what you got. Ok. So you then use this 1360 if you
35:21 I didn't touch it if you Ok? And you're still good if
35:34 use that 1360 divided into 25,500 minus minus minus 25,500. You're gonna find
35:47 I'm only using 38 of those 256 . And the other 220 square aren't
35:55 used at all. And that's why image on the right looks very inferior
36:03 the one on the left. Now, this is the default.
36:07 thought this was important. That's why put it in the lab for.
36:11 if you have it for posterity, know, pass it to your
36:15 whatever. But um so that, know, don't use the default for
36:22 Seismic Amplitude data. And then by , probably not a good thing to
36:28 it for root mean squared amplitude really and also bad for envelope. We
36:35 can't have the same problem when we the spectral magnitude, it'll be bad
36:39 magnitude. And if we do something energy, which is a square
36:43 which is gonna be even worse for . Now, for other things,
36:47 we do the variance coherence, it from 0 to 1, 255 colors
36:55 fine to span that. And if gonna do frequency like what's the peak
37:00 or the instantaneous frequency, it might 0 to 100 and 20 Hertz 255
37:07 of and am I interested in seeing , an eighth of a Hertz?
37:10 , I can't. It doesn't mean much to me. Every one
37:14 every half Hertz is gonna be OK then. Um Yeah, so
37:19 it. So somehow my red box a bit and B is correct.
37:27 A is wrong like, oh, that's the last one. That is
37:35 last one. OK. So Uh Javier, no one here is
37:47 . Is Javier smiling. Can you he's not showing his picture?
37:52 Jessica, she's smiling now. Nobody smiling. OK. So that's the
37:58 of test we'll have. Uh I'll look, I can, what's great
38:01 canvas? I can see which one got wrong. Guess what? That
38:06 be a bonus question next Friday. right. So, all right.
38:13 we'll continue to work on the lab four and anybody stuck anywhere.
38:24 Yes, ma'am. Oh uh that's OK. 150 is gonna be
38:47 middle. You can make it anything want. So go to the players
38:53 uh put intersection player. OK? then um set this to be one
39:02 . Yeah. And that's good. now hit this little VCR button that
39:07 say 1215 1. OK? So 1243. If you got that didn't
39:19 out, that will take you all way to the other side. So
39:22 , we're just looking at different OK? That's all. Now the
39:26 here has nothing to do with the , what, what the has to
39:32 with on your PC. If you this on a, you put
39:42 um, a USB drive and you it on here and then next week
39:49 go and you use machine back everything's on the USB drive. But
39:54 on how that's set up, your drive may be E here and F
39:59 there, you gotta go fix You gotta go tell each one where
40:06 go. And, uh, and that's not a big deal. But
40:11 it comes up with a skull and kind of thing means you can't find
40:15 data because it's looking on A and don't have an app, it's on
40:22 or you have an E but it's of your summer vacation and their data
40:28 . OK. I mean this is real common problem. Trust me.
40:33 this is how when I grade engineer never worried about cheating on the labs
40:39 to figure this out and fix it harder than doing the lab. You
41:00 . Mhm. Thought. Yeah. . Ok. OK. I want
41:13 go for 15 minutes. Walk It was like five. Ok.
41:29 think I, and maybe like No. So you know that.
42:03 right. Oh, wait, he got something else in the,
42:12 you hear the and told them I write this I mhm. When
42:39 is active then we talk about this that I, so you're actually
42:50 so we're going to move first. . So here's the question when you
43:02 the one with you and thank What do you have? Ok.
43:44 . I Yeah. OK. That the moment. OK. That,
44:18 I was going to happen with the . Good bye. Yeah, you
44:31 about no more like testicles all Ok. Um Oh, thank
44:57 Yeah, it was, it was now I'm ready to go seven and
45:24 and then go to the bathroom so farm I so. Mhm.
47:45 that's all. Ok. Here we . You know like that's ok.
48:30 highlight. The highlight. Great. my uncle. Oh no.
49:10 Yeah. So when you make it regular you're gonna see. He wanted
49:18 see me. It's not right. see that, that challenge and the
50:37 . Oh, maybe you got it big size but 33. Well,
50:52 is it going to happen? Um me go ahead and make uh mm
51:18 you. I there was something Ok, but um hey, how
52:12 you? Um so II I visited girl like the next time I got
52:25 pretty it. Ok, good. kind of push back we have with
52:39 do but hey, can you maybe pass a little me in here
52:46 Oh yeah, so far, so . So good. I hear from
52:59 of one or two. Ok. right. One or two. The
53:05 guy, right? All right. . Good. And here the answer
53:15 the question was how do you? , those are pretty f but now
53:31 got them, there's like five, ? But they're both the same
53:39 correct? Yeah, let me if have to do. 01 is kind
53:43 green. Yeah, I think. , that's cool. Ok. So
53:48 trick was when eventually the girl will that with the, with the 3d
54:01 . So going to the 3d, 3d view there, a one
54:08 Ok. So you wanna have different you make. Ok, if they
54:15 a different name, they can have different, they could have a
54:18 that community. So now we can it. So now you're gonna go
54:24 those 20 line and picked one of phones. Ok. OK. We'll
54:34 another one on your of the That right one or not one
55:01 I try to, you know the there are, you know, which
55:06 trying to make without wanting to think whole thing up. So you're gonna
55:14 another one and then what it will you have it activate in your screening
55:30 but they will buy the red one the red, one, green
55:39 put them together. No, I think you could do that but
55:43 think that's exactly the white wine. make that one? All right.
56:37 I not? OK. So Iii I feel good. Thank
57:21 OK. OK, great. If OK, we're gonna do the
57:48 OK. Mhm. Now, are the same all pain? Well,
58:01 need the MRI and all up Yeah. Creating what you want.
58:07 then I would want me to pick line just to read it the quick
58:16 what it's going to do if you to win the fall and you can
59:09 do that. Let me check Mhm. Ok. Ok.
60:22 Ok. Ok. Oh, I know. Ok, to really like
60:47 is not real experience and what five bye. Ok. And walk.
62:02 . And uh you. Ok. , thank you. Good morning.
62:46 . Ok. Ok. Yeah. . So. Mhm. Ok.
64:25 . Are you going to work? Yeah. OK. Mhm.
65:14 Mhm. OK. Yeah. Oh we OK. OK. So
66:41 yeah I need that one doctor to how to put that. Yeah.
66:54 . OK. So for Oh All right. OK. OK.
67:47 like what, what else? Well, no what so OK.
68:48 . Yeah. Yeah. So. . Yeah it it is. Mhm
69:37 . Right. Mm Yeah. Um . So when you get sorry thank
70:26 . Oh four. Exactly. So add this other but still or uh
71:06 . Ok. OK. But they standard probably a lot of stuff.
71:24 . All OK. That's the 10 . Oh and we actually so they
71:45 finish it. OK. Yeah. . You OK. I will.
72:26 that the, the very as we the exercise, the structural work in
72:40 and then he said OK, I this image and I got here's an
72:44 of your hand with our structure when feel like and then with as well
72:54 and you know, so you see here at this point you need to
73:01 , you gotta put some hours. did you say? 954 millisecond,
73:08 like that. Ok. There I do a couple and they
73:17 gotcha. Give me five minutes. play from the more violent it's to
73:35 . We have the right that, , then what do you wanna look
73:41 ? Are the, oh, um, I didn't mean to step
74:07 you Jessica and so you're gonna get . Uh you're gonna look at the
74:14 elements in that turbo and around there also saw a nice little kind of
74:18 bright spot, maybe uh 50 bins 20 bins. And that was clearer
74:27 structural warning is filtering. So the you judge whether a a process,
74:32 filter process is good or bad does it help me see more or
74:41 it smear everything out? So there's of filters and the filters aren't always
74:46 filters. You got to try them , you're gonna judge through looking for
74:52 , right? That's a, that's basic, that's a real fundamental question
74:58 you need to get your head OK. It's helping me.
75:10 ma'am. I think I got, don't know. Yeah. OK.
75:25 what you use. Three by 35 five. OK. Sometimes. So
75:36 is a three by three filter and is a five by five 00
75:46 OK. In lab three, this good. Now I'm gonna, I'm
75:52 gonna go over uh the lecture. have um data conditioning because I felt
76:04 more fun to just do it. . I do have it. I
76:08 it with a voiceover as well. . So listen to my melodious voice
76:14 the Cougar Den while you're drinking a this evening, then they might find
76:21 annoying. But uh so put earphones . Um, so I asked you
76:29 generate the structure in a filtering with traces, then take that output and
76:43 the filtered version. Now in seismic , we'll call those cascaded filters.
76:50 . When you have mapping, you say, oh I'm gonna do a
76:54 filter and then I'll follow that with mean filter. So those would be
76:57 filters you apply to a map. if I have a three by
77:03 the second time I apply by three three, I'm using the data and
77:13 do 23 by threes. I'm using data, a five by five.
77:19 that obvious? Or do I need show a picture? Oh, I
77:24 to show a picture. Mhm. . Sure. Um So if I
77:32 a dream ticket, think of it two dimensions, I run through and
77:37 gonna do it at just the I'm gonna do one third, one
77:41 , one third, then I'm gonna the average of the averages, one
77:45 , one third, one third of averages. And then it will turn
77:51 if you look at the original It will be 36, it will
77:56 1626362616 on the original data. What I doing? I'm convolving first,
78:06 convolving a one third, 11 3rd the data. Then I'm convolving a
78:11 third, one third, one third the filtered data. And then Zach
78:15 loved convolution, right? Zack. says, well because of the Associate
78:22 Law or the, yeah, because the associative law, I can convolve
78:30 one third, one third, one with a one third, one
78:33 one third, come up with a filter which is 162616362616. So
78:41 that was the day I did the thing. Now, if I do
78:46 five point filter right off, what you gonna have? 1/5 1
78:50 1/5 1 5th, 1/5. So one, the cascaded builder is tapered
78:59 the simple filter is not tapered. know what I mean by taper means
79:09 amplitude at the edges of the Let's see if I have a picture
79:15 guys. I'm I'm saying Jessica, seeing glassy eyes. She might be
79:23 by now. Is she there I'm still here? OK, I'm
79:28 . OK. Let's see if I find, I see if I have
79:38 in my lecture. I don't. , nothing. Yes, sir.
79:50 . I might but let me see I have it here and then we'll
79:52 look for it. Ok. No over. Let me see if I
80:12 it cascaded. Like, yeah, come, let me find some
80:33 Put it in. Oh, so am gonna blame Stephanie, I'll blame
81:05 . I've forgotten now because it made mind muddle. So the question
81:11 no, it wasn't. Stephanie might been Zach. It was Zach and
81:17 to do with like, OK, this Casca of operating? What's going
81:22 here? Why do I, why you want me to apply a three
81:27 three operator twice versus a five by or five by five operator once?
81:41 . Now there's going to be two , it's structure oriented. So if
81:47 take my plane and I have my and it's gonna go a long
81:56 well, a small plane three by , there's gonna be tangent to that
82:03 and a bigger one is still but it's sticking off in the
82:08 So if I do a three by and then I apply it iteratively,
82:13 really adapting to a curve structure. that make sense for the first
82:19 It's tangent? The second time it's , but it's, it's using the
82:23 that's in between. The second thing tapered and the third is cheaper.
82:32 . So let's do the tapered Thanks, Zack. This is really
82:38 paying to do in powerpoint. So on the left you don't have
82:43 , but I'll, I'll update the, uh UT I've got a
82:47 filter on the left with 111 multiplied a third, a three point
82:53 OK. Five point average, 11111 by multiplied by 1/5 seven point
83:00 1111, no, no, no, no, divided by a
83:03 by 17. Easy, right? would be a single filter.
83:09 single filter. Well, if I'm cascading once, it's the same
83:13 the second one, I don't have right? That they don't think you
83:20 have, you should have seen this I gotta move this guy over a
83:27 . Hang on. Did I say is painful to do in powerpoint?
83:32 think I did. Uh OK. . And this why? OK.
83:55 now I've got my original filter then gonna cascade it. So I'm going
84:06 convolve the one third, one, , one, third filter with the
84:11 filter. So with this location one 11, right? That one time
84:19 one has one, that date. got it. OK. Then I'm
84:24 slide it. The last one is time is one when they overlap one
84:32 , one plus one times one plus times 13, when they overlap
84:39 one time is one time, one is one. OK. So here
84:42 guys don't like this. So I'm do this. Hang on just for
84:46 . Here notice I like to explain time ago and I'm gonna make
84:57 There's a fast in the slow, I can make it. No.
85:05 . Ok. And then that's what can make the longer too.
85:17 So as I move that guy ok. Now I get two.
85:21 I get three. Now I get . Now I get one again,
85:25 my filter. 12321, a triangle in terms of weights. OK?
85:32 let's take one third, one one third and cascaded with this guy
85:40 three times. I do the same . The first one is gonna be
85:47 then. Oh, I'm gonna get . I'm gonna get six. I'm
85:51 get seven. I'm gonna get gonna get three, gonna get
85:57 I get that as my joke in limit. It turns out to be
86:02 Gian. I don't want to do limit. It took me probably an
86:06 to make this simple slide in two . I've got this. I did
86:12 for the book. I'm working on filter. See everything's all one.
86:19 , I involve this guy just like did before except now I convolve it
86:23 two dimensions, not one dimension. ? So I did 9 81st in
86:29 center and 1 81st on the edge then three dimensions. I get 49
86:37 in the middle and 1 to 729 the edge and in the limit I
86:42 a two G Gaussian. So, I'm saying is this is a paper
86:49 instead of a constant boxcar filter with edges. Howie Joe probably told you
86:55 are, they can have nasty gibbs , right? Do I have to
87:02 it again? You don't even know I'm talking about. Do you,
87:11 you never watch the bee? 0000. Stay in and out,
87:18 in, hang on. You gotta your, your hair show out on
87:23 shirt. More through the beach. guys did? You weren't around in
87:27 seventies? Were you disco? You've heard this one? No,
87:32 has heard more than a woman. more than a woman. To me
87:40 the Gibbs phenomenon. BG. I'll it in here so you can see
87:46 . Ok. Um Now a three three filter. Oh, I got
87:57 multiplies. Nets. Ok. I it twice. What do I
88:06 Nine plus nine, 18. A five by five filter. 25
88:14 in that. Which number is 25 or 18? Good. And
88:21 I can do the same as I to cascade. It becomes cheaper and
88:26 to cascade. Plus I adapt the plus I taper. Best way to
88:36 it. Thanks for asking me that . You just burned up an
88:44 Ok. I'm gonna take a three break and turn this off and then
88:51 start on with the next lesion. this is good. This is the
93:42 . So uh I wanna go to six tomorrow morning. We'll start talking
93:48 attributes. But here, what's the of acquisition and processing on seismic
93:57 Surprisingly, your interpretation typically as good the data you have to work
94:02 OK. So it has a pretty effect. So I want you to
94:08 able to weigh the value of den acquisition versus repeated sweeps at the same
94:17 and suppressing source correlated seismic noise. , one of the uh platitudes of
94:26 seismic acquisition is we need to get energy in the ground. So what
94:33 do is we'll use bigger sources at same location or if it's a vibrator
94:41 , we'll stay at the same sweep sweep four times to get more energy
94:45 the ground. Ok. What we'll out is, ah, that's not
94:50 best way of doing it. If have surface acquisition with access, then
94:56 gonna use the concept of migration aperture justify acquisition of seismic data beyond the
95:02 of your acreage. So you um let's say Liberty, you have
95:08 over uh Liberty Township, uh just east here of Houston. I wanna
95:15 everything underneath the Liberty Township. do I need to go to the
95:20 towns to image it? And it out you do you need to go
95:25 maybe five or six miles from your ? Of interest. In order to
95:30 a good image inside the area of , we want to justify acquisition of
95:36 offset and wide. A as seismic identify the improvement in data quality and
95:43 processing when using higher fold data. it turns out when you have more
95:51 , the computer works harder, the work is easier. So there's a
95:59 off between human effort and need to clever and just burning up more electricity
96:12 , and then evaluate some of the . There's one called DRA or imaging
96:17 . That's quite interesting. Uh If doubtful of, let's call it an
96:23 and use it in the same subjective . What we do as an attribute
96:27 , oh this can enhance certain features interest like FF and each out in
96:35 , whether you like the theory of imaging or not. OK. So
96:41 comes from uh my buddy Chopra who for a big processing company, Alberta
96:47 here's his boot stack and in seismic . Oh That means I go in
96:54 um I uh I kind of guess the velocity using understanding of nearby surveys
97:03 acquired before. So I might use or two velocities for the entire survey
97:08 to get something quick. OK. then here is the residual static.
97:13 now I'm gonna apply statics and really at it. And then here's my
97:18 migration with the best migration velocity. . So here is the amplitude data
97:24 it's been processed and stacked and then the coherence on these amplitude vs,
97:31 see as you go along up, starting to image these channels here and
97:39 bulks. This is clearly better than one in the middle and this is
97:45 than the one on the left. , all right. So big.
97:51 now your colleagues who are seismic they're not geologists, not, some
97:56 them may have had a class or in geology, but just as likely
98:01 were music majors or other people who the right side of the brain that
98:05 do mathematical things with. OK. then you can see if my velocity
98:11 me a nice sharp vault interface on vertical section. They can see if
98:20 wavel is hier less smeared, less and pick a decon algorithm that comes
98:29 with a better image. OK? if I ask them, hey,
98:33 , I want you to image this build up. Um One, what's
98:38 carbonate two? What's a build up to look like? Or that uh
98:44 first question on the test today where had the little bitty channel.
98:49 what's a channel look like on vertical ? On vertical sections? They're really
98:53 to see that you got to have eyes. So here with attributes,
98:58 case happens to be coherent. You stuff that looks kind of like
99:03 You know, mom and dad can which one better? OK. Then
99:11 coherence on the migrated volume with original . Here's one with improved velocity.
99:16 , ah, clearly the one on right looks better we're going in the
99:22 direction. So, again, most you I think are geology is background
99:27 three quarters. And you think seismic has all these equations in it.
99:34 . And you think, oh, a right way to do it.
99:38 mean, they're good workflows and so . But when it comes down to
99:42 , to pick parameters, it's like way I pick eyeglasses in the,
99:46 the optometrist's office. And she'll ask , is this one better,
99:50 better, worse, better, you know, I don't know what
99:54 doing, but I can tell which the better, which worse, which
99:57 I like. Ok, that's how pick decon operators. That's how we
100:05 statics. That's how we improve We try it. Ok? And
100:11 when we go to the customer, may show two or three of the
100:14 we think are best and let the make the final decision. Ok.
100:20 it's wide as a you so white music has become very common on show
100:28 the US over the past the 15 I would say. And this one
100:36 to be a cartoon from a survey the Fort Worth Basin. So,
100:42 oh, near, near that big you guys like to climb 150 miles
100:47 of here. Watch the big Yeah. Enchanted Enchanted Rock.
100:55 So that's where the basin comes up finishes. So, here I've got
101:01 receiver lines laid out. So we junk hustlers. You know, people
101:05 geophones and they're in the big they might weigh £50 and they're,
101:10 hustling, they're putting geophones in the . Somebody's already surveyed them. This
101:15 time. Surveyor takes time by putting in the ground. You don't wanna
101:21 , you know, you want, got a north arrow on it.
101:23 wanna make sure they're oriented in the direction, et cetera. So they
101:27 be put in accurately and that costs then perpendicular to that. So I
101:32 four lines with maybe 0, 1000 laid out. Ok. And then
101:39 have a uh a source, a source. OK. And the
101:47 oh, and then when I wanna to the next source location, I
101:52 to lay out the next line of . So there's always two to the
101:57 and two to the south. So maybe I have five lines
102:02 Well, while I'm recording one, , I'm busy weighing the next one
102:08 then I get this kind of ray . Ok. So this is narrow
102:14 me, it's most of the rays going up and down, but more
102:18 and west than north and south. . Now here's wide a acquisition instead
102:27 four lines, I got 16 laid 10,000 ft long. Um And
102:35 uh my shot point interval is 880 . I'm sorry, my shot point
102:41 100 and 10 ft. So as go between the lines, so eight
102:45 them, I'll have eight shots going the lines. Then I have to
102:49 the line. But you see, take one from the bottom, move
102:53 to the top and then I did attribute. Ok. So which one
103:11 of picking on Jack? Because he's the front. Ha. Right.
103:21 , Hayden. Yeah. Point to . Tell me that guy's name
103:26 Anthony, the guy who was gone week. Right. That's right.
103:31 , Anthony, which one's more expensive look to lay out? Forget about
103:38 equipment. I own the equipment. one takes longer to lay in
103:43 in the field narrow as you through line up White House. Why,
103:52 , how much more? Yeah, got four times as many lines.
104:01 . What do you think of Hayden? Four times? Ok.
104:05 four times, um, where it from. So you're gonna, what
104:21 shot? Ok. I'm gonna have same number of shots. Now,
104:26 have all these surveys and I'm gonna going, you know, 1020 miles
104:30 and south. How many new receiver do I need to survey?
104:47 I'm, I am gonna, every is going to be occupied by either
104:52 four narrow avenues roll along acquisitions or 16 roll along. Why? That
105:00 , I have to survey all I have to plant geophones the
105:05 So the acquisition and I'm gonna shoot same source location. The acquisition is
105:10 the exact same time in the Now, with CGG, they're gonna
105:15 you for having more equipment. they, they gotta make a
105:19 Ok. But it's the same number people in the field. So the
105:23 wall clock time, therefore, for same expense. Bye. I'm getting
105:30 , full and more important than I'm getting signals from different. No
105:37 important. I'm getting noise from different . OK? My signal is kind
105:43 going down and back up. But if I'm doing narrow as a
105:49 this way, I'm bouncing off of wall, I have no leverage against
105:55 wall. If I'm collecting data in direction, if I have data over
105:59 , oh, I'm closer to the . I have a different move out
106:03 the noise. So I'm going round, roll with dispersive velocity.
106:14 his part of his phd dissertation. coming from different directions. OK.
106:20 when I have more fold, what's important is the random noise?
106:30 we're really good at getting rid of noise. We got dozens and dozens
106:36 filters to get rid of random What we have problem with is coherent
106:42 noise that is correlated with the So here is the source and I
106:47 that source off and I got a coming from the side. I can't
106:54 if that hyperbola coming from the side different than a hyperbola coming from
106:59 But if I have more, move by coming closer and further away,
107:07 , note to those far away, is moving left and right across the
107:11 of the classroom. Did you notice you think? Ok. So that's
107:18 we do. OK. So here's older survey from uh Devon Energy.
107:26 one acquired in 1995 using short offset Asia 1997 using long narrow as we
107:37 1999 using white as music. Pretty low fold like 60 fold.
107:48 here, I've got a dip magnitude from the P horizon garbage better,
107:56 better. Oh And the data were by the same person using the same
108:00 at the same time or the same at the same time. OK.
108:04 it's not like uh this was He, he did this one and
108:09 did the other one. No, the same person. OK.
108:14 here's an example from I if I correctly, Egypt, North Africa for
108:23 . So here is wide versus narrow , here's a narrow ASU one,
108:29 was white Asia pretty darn big OK? So you have better
108:37 coherent noise and multiple attenuation with wide algorithms and you have better velocities and
108:48 . So Lily tell my geology buddy , geology buddy. Right. Tell
108:55 geology buddy, what do, what we mean by statics in seismic
109:00 statics? Ok. And what, the geological reason for statics?
109:14 Lower velocity of what? Ok. . On top. So, basically
109:20 to the weathering zone. Ok. we're gonna have a weathering zone,
109:24 zone might be filled with uh clay uh up in Canada with uh
109:33 you know, different uh low velocity and it can be quite irregular.
109:38 don't even know it's irregular because we see it. Ok. So
109:43 just like we, we said, try to align adjacent to shots in
109:47 , in a shot, uh adjacent in a shot gather after we correct
109:51 for move out, then we do the same thing in common receiver
109:56 OK. So we try to align and we're gonna do cross correlations and
110:01 to figure that out. So a thing is something called surface consistent
110:07 So we say, OK, I shift, there's a weathering zone above
110:11 source and that source is going to the same correction for all 15 geophones
110:21 this room. Thanks. Now, is the receiver. Each one of
110:28 are different sh shots. The weathering correction over Lily is going to be
110:36 same for all 15 shots. So a least squares problem we iterate solved
110:43 . So this is called surface consistent . We put the static correction to
110:48 physical location on the earth. If have 15 traces, uh it's pretty
110:55 . A human being, you need get involved if I have 1000 traces
111:00 a computer problem. Pretty easy. . So more redundancy in the
111:06 even though there's more data, there's human intervention. So it becomes
111:13 comes cheaper to process the velocities. here picked velocities. You haven't,
111:22 haven't lived. OK. So you go, you know, do move
111:26 and then you know, look for ups and um how about residual velocity
111:31 ? You've done that, right? you do prestech and version?
111:38 John Castano needs to have you do conversion in a lab. Then you
111:43 to correct these things. It's OK? If you have more
111:49 it's easier. OK? The thing have to worry about is aliasing.
111:57 . So are you comfortable with aliasing the name White Carlos from Colombia?
112:09 the jackal. What was his real ? The Jackal? I he was
112:19 he was like AAA terrorist back in 19 sixties and seventies, right?
112:28 . His name was Carlos de Anyhow. Anyhow. So his alias
112:32 Carlos de Jacko. My alias on street is my street name is
112:41 What? You're gonna laugh at that ? Oh, you, you don't
112:46 at somebody whose name is Spider. wanna see all my tattoos?
112:52 Anyhow, that's, that's, that's alias. So alias is Latin for
112:58 . So you're gonna think one thing something else. OK. So this
113:05 from a textbook by Larry Lines and Newick. And um they've got a
113:10 Hertz and uh 20 Hertz sign a and here the dip is uh eight
113:18 per trace. So here's the dip to pick. OK. Here's the
113:26 dip. But I bet most of see A OS T, OK.
113:33 that pink one is the alias di it'll turn out that when you're going
113:42 take and think of Howie Joe's class take my diffraction hyperbole and I'm gonna
113:49 the data and push it perpendicular to hyperbole. I'm gonna push it backwards
113:55 into the earth. Well, I wanted to pick that yellow one and
114:03 I'm gonna put, I'm gonna put in the pink direction, gonna have
114:08 serious problems. OK? So here's example of alias you don't have
114:17 but when you play it, it's helicopter and take a picture, take
114:22 video of a helicopter with your cell . OK? And there's something wrong
114:29 that picture, right? How was flying? The rotor is not
114:37 right. So what happens the repeat the, on the the the the
114:47 rate of the cell phone is equal some fraction of 1/5 of the rotation
114:53 of the helicopter. OK. So have the, we're interpreting it as
114:59 else. So here's an ex and got a bunch of others in there
115:05 are kind of fun, but here's example of source generated noise, got
115:11 coming up. That's what I'm interested . I got multiples and then I've
115:22 shallow things coming back bouncing off, cetera. Ok. So I've got
115:31 diffraction. They're in particularly bad because high amplitude because they haven't gone very
115:37 . They haven't decayed by one over and their velocity is slow. Uh
115:43 their high frequency. So uh what Bill prem when he was a
115:48 he did this model and he generated a synthetic. So he's got generated
115:55 this model. So he's got kind a slope here. A corner
115:59 corner here, corner, corner, , corner, corner corner, each
116:03 these corners is gonna give me a hyperbole like how we talked about.
116:08 my diffraction hyperbole. Why do I gaps uh use rate tracing ray tracing
116:15 some over it. That's all. these are steeply dipping. So they're
116:20 be aliased right? To the problem the fractions. They have a linear
116:27 out traveling at very low velocities. tend to have very broad bandwidth because
116:33 have less absorption. They didn't go far and they're frequently hire in the
116:39 from depth. OK. And they're fly alien with respect to the
116:46 So here is the image he he gets a pretty good image here
116:51 everything that's flat. But the diffraction , the signal was imaged kind of
116:59 . But she always garbage. That's to alias. Thank you. So
117:08 way to get around that is to your data on a denture grid.
117:15 means you're gonna spend more money, ? And management doesn't like to hear
117:21 , but that's what you have. . Here's Oklahoma. Oklahoma is just
117:25 of you guys in Texas. I about here. This is the Wichita
117:29 front, Wichita Mountains are here and um bunch of surveys that PGS ran
117:37 . So he, he's gonna look this Carter Knox survey uh and they
117:42 data and newer data, but really show this, they're gonna like the
117:46 data and then decimate it like the data works. So here's conventional acquisition
117:54 the time. Uh So this is 2010, a 222 ft 330 ft
118:03 shot. So this gives you a size of 110 by 165. So
118:08 of you working with data in the States, most of you are probably
118:12 with data. That's 100 and 10 100 and 10 as a bin
118:15 So that's real common. The dental is gonna have 55 by 55.
118:21 many receivers per line will more, many active lines or more uh fine
118:28 uh space in, in not too . The big difference is the number
118:35 great first W mile. So like times the traces really increase it.
118:46 . So here is the data uh conventional acquisition and then here is with
118:55 A was it really is the same ? Just the one on the left
118:59 uh and the same velocity is One on the left has been decimated
119:04 migration. And while you look at area, it looks OK. But
119:10 , oh now I see a nice coming through here real easy uh to
119:17 . OK. And then you look front here at the the peak of
119:24 uh the fold. I don't have any data. Now the signal is
119:29 there. It's just be noise in alias noise from other events that are
119:36 . So it's overprinted with stock. here you've got a nice image and
119:43 the front, you've got a good here and a poor image there.
119:49 a called a strike line but it's complicated geology that no simple striker
119:56 Yeah, the alias noise from the diffraction. He hardly sees anything
120:05 I've got a nice clean both block then up here war image, same
120:15 , same signals being recorded just that don't have the alias noise on top
120:21 it. That's the difference. And in the front in the foreground uh
120:27 resolution then OK. Now wide and as acquisition to image a car striker
120:36 and this is some work, work done by uh some folks in
120:40 card basin in, in northwest. know. So I don't know
120:46 where's home? You're, you're from China, right? From Xinjiang,
120:50 said? Ok. So it's kind home. Not here, hopefully have
120:54 trees than here. And how about ? You're from Taiwan? OK.
120:59 where he is. So um half oh biggest company in the world,
121:11 of money but take a guess. Apple P Amazon, they're pretty big
121:30 big Chinese oil company, they own kinds of stuff, they own all
121:36 of stuff. So they're the biggest in the world. Half of their
121:40 comes from here and almost all of comes from cars collapse from or division
121:47 . So you know, you oh well, who cares about cars
121:51 ? Well, they care a whole , right? So cars think
121:55 all right, caves and the collapse . So here's the thorough as in
122:01 survey. So this was like a project they did to justify uh 3d
122:09 wide avenue. So here is the of the ray coverage like I tried
122:15 show earlier and then here's a wider mute that's not full, but it's
122:21 . And uh so the ratio here 0.3 to one here is 0.58 to
122:28 . And now these little guys that putting my mouse on, that's the
122:34 . That's what they're trying to OK, then uh they drew an
122:40 line, you'll be drawing arbitrary you know, but they're trying
122:43 you know, connect some of these features, these cars features, here's
122:47 wide AU and here's the call it ASU and I'll define what we mean
122:52 that. Now, you see these are nicely resolved. So here at
122:59 , maybe I see them here, very, very distinct. Thank
123:05 Here is coherence variance if you will from the narrow ASM of data.
123:12 , I love cars. So you , I I can see cars in
123:17 , I can also see Elvis in data and many other things I can
123:20 lots of stuff. That's my my hypothesis. Thanks. Then here's white
123:30 wider 0.58. Well, now it to look more like cars collapsing channels
123:37 then here's the full ASU. Now way they computed full ASM they had
123:40 much acquisition equipment, they acquired a first North South and they acquired it
123:45 West. So they acquired it twice then put them back together. But
123:51 , I think everybody here sees elliptical collapse features and some kind of channel
123:57 , right? That's their time darrow wide and that's how they acquired
124:06 They just acquired the surveys twice uh be only merge surveys. Did they
124:14 better lateral resolution less mixing? Here's narrow avenue survey, looking at a
124:20 area and here's their White Asia. one do you want to interpret?
124:29 ask Zach because he said, why are we doing this? Which
124:32 are you gonna like agonize over? which one is easy or unambiguous?
124:44 . Yeah. So, you you're, you're, you're always better
124:48 clearer images, you know, and you're gonna focus on geology and
124:54 what's the environment of deposition and where hydrocarbons go instead of like 00 where
125:00 I put that horizon? Where do put that phone? A long
125:08 Gonna help more with multiples. Hopefully I talk about multiples. And we've
125:13 multiples above that Cora 3D survey. I know I sat with Stephanie here
125:18 we looked at some of them and energy goes down, bounces back up
125:24 the surface where you measure it, it goes down again. Now,
125:27 I have some strong reflectors in the or in the subsurface and for Cora
125:33 , the strong reflection is shallow So that bright spot, I see
125:38 not once, not twice, not times, I think I see it
125:40 or seven times. OK. So complicates your image the farther offset.
125:48 the multiples are typically in the shallower of the data, the velocities in
125:54 shower part of the data are the velocities in the deeper part of
126:00 data because of diogenes and so forth to be faster. So the what
126:06 call the move out. Uh If the event from sha going shallow at
126:17 slow velocity, even though if it four times as long because it's 1/4
126:21 multiple, it's still going at that velocity in the water call.
126:28 So it's gonna have more move The deeper reflection comes in at four
126:34 the multiple level or at four times water bottom level, that's gonna have
126:37 higher velocity. So what you'll see is if I have the multiples,
126:44 I have near offset, I don't as much leverage against the multiples if
126:50 have fire offset. Now, this like I gather it's not, it's
126:54 a carbonate build up in, I it's um Abu Dhabi. And what
127:01 wanna look at are is this, knew I'd do that. They wanna
127:06 at that structural that structure. So here is um conventional acquisition with
127:21 and then here is long offset So where I have the arrows,
127:26 you see the stuff cutting across? ? Like like this thing here,
127:33 a multiple. And now it's OK. This one, this blue
127:41 , it cuts right across. That's multiple. OK. And the other
127:47 that are dipping that structure. So up top uh here's a high density
127:57 in the Midwin Basin and um 0.16 traces per square mile, 16.7 per
128:09 mile. Oh When you don't have traced as close enough shot. You
128:19 really image the shallow structure, especially the middle days. Yeah, Midland
128:25 , you have a lot of salt anhydride up shallow in the section.
128:30 . You can actually start to image we're looking at this image down
128:37 Uh This to me that's a mass complex. If it's sand, that's
128:43 of the really sweet uh producers in Midland Basin. So a lot more
128:54 . Then here is the Viber same guy Viber size sweep. So
129:00 or 2009 survey, he swept for seconds. And then here's the 2019
129:08 recent survey swept for 24 seconds. what do they wanna do? They
129:13 get more low frequency. So if go down to uh five hers.
129:23 if I have five Hertz, it's cycles per second. So uh if
129:31 have attenuation, I wanna go if I wanna have 10 cycles at
129:35 Hertz, it's gotta vibrate 10 times long as 10 cycles at 50
129:40 OK. So it takes a little sweep. So here he's going,
129:44 said what 24 seconds we instead of uh eight second sweep in the
129:51 um The low frequencies are really, important for impedance inversion. OK.
129:58 you in order to estimate things like , OK. So here is a
130:06 image legacy one and then a high or high trace density. And now
130:15 can image a nice fault in Actually, it looks like a reverse
130:22 . Then he ran some attributes on . This is coherence. OK,
130:30 you like this color bar or that's up to you. But here
130:34 seen parts of a mass transport deposit and here you can see all the
130:42 pieces of a mass transport deposit. we'll talk about mass transport deposits and
130:47 next week, but just think of landslide right where I got slumping going
130:56 the slope. And then here coherence on the top. And then
131:03 talk about curvature tomorrow and how it . But look at the difference in
131:09 you can map. Look at the right now. We're gonna drill horizontal
131:14 for this. I think I want know about that detail and then he's
131:18 dip a dip angle. You're probably with that right now. OK.
131:24 do we increase the bandwidth? we can do that through broadband acquisition
131:36 nobody here likes disco. In I never liked Disco Banana, but
131:40 don't know about the peaches, but of you like music, right?
131:47 Coolio, you're gonna look, you cool. You look at cool out
131:53 , staying alive. Good with the . So it's a mixture of rap
132:01 disco. It is excellent. In music, we talk about
132:09 That's the way we want to think terms of resolution. If I go
132:16 40 Hertz to 80 Hertz at 40 80 Hertz. That is one
132:25 That's the same as going adding from to 5 Hertz. That's still an
132:30 . OK? So that's what's And this work done in Oklahoma of
132:36 places and um what they found they actually get data down to two and
132:44 Hertz and then up come up at Hertz or so. So they got
132:51 , maybe 5.5 octaves of data. most of the data you guys work
132:57 uh might be 18 to 70 something like that. That's pretty
133:02 But here to do this, they a yeah, they had to make
133:05 vibrators vibrate longer. But what they had to do and this is something
133:11 did 40 years ago, we would our geophones and put the geophones on
133:20 table and shake the table. So shake it at two Hertz, three
133:25 , four Hertz, we call them table and we measure the output.
133:30 the input, here's the output and gonna be a little phase change.
133:34 ? And both, especially below the of the geophones. So geophones,
133:38 kind of resonate at 10 Hertz. it's a two Hertz, they're too
133:42 to carry. So they kind of at 10 Hertz and below 10
133:46 the phase rotates a lot. And theory, the convolution should take care
133:53 all that in practice, the convolution do that well. So what these
134:00 did, they actually measured their geophones they said, all right, let's
134:06 a correction to the phase of the , 10 Hertz first. And then
134:12 gets us pretty close and then de should work. And that, that
134:16 their big breakthrough. So he's got good data at 2 to 4 Hertz
134:23 . OK? And this is the data we get. This is what
134:26 like to have all the time. . So that's state of the art
134:33 eight years old. Now, another with white as if I skipped,
134:45 , I probably have a, I a bunch of skip slides in these
134:49 , not skip hidden songs that if in, what's your least favorite part
135:01 Texas? I don't know. West Texas like the mountain part or
135:08 flat part, flat part. So we're down there at the Texas
135:16 . It's Yeah. Have you been ? Yeah. OK. So you're
135:21 there in Texas that basically darn I weigh my geophones out uh in
135:29 880 ft apart. OK. So kind of cool. Then I wanna
135:35 as much energy in the ground as . And traditional, what we used
135:40 do is my vibrator would sweep four at the same location to get more
135:47 . It's kind of a macho you know, more energy, more
135:52 then now what we do, best is sweet once, then drive to
136:01 new source location and have more force . Ok. And just sweep
136:10 So the amount of energy we have the ground is one half as
136:15 Ok? Because we only swept but we swept in two different
136:20 Now, if I have incoherent like the wind blowing through the trees
136:24 the birds chirping and the traffic on 45 well, repeating the data four
136:30 helps against random noise. However, problem is the energy bouncing off of
136:39 wall to the side of me and repeat four times at the same
136:44 the signal to noise ratio is exactly same because I get the exact same
136:48 coming back, right? I got ground roll coming back and forth each
136:54 . So I don't have any leverage I move it to a different
136:59 Now, I have a little leverage the travel times are distant difference.
137:04 horizontally and through processing the travel times a deeper reflector are about the
137:11 So that, so the way we it now is we try to have
137:16 sweeps at a location. I have bunch of slides in there that,
137:19 show that as well. Oh One time instead of two, it
137:27 time to drive your Viber size to other location. You want to do
137:33 the best data you can acquire but best data for the same money,
137:39 best data you can get for the your manager budgets for you. So
137:46 , that's the, that's the correct answer. OK. All right.
137:52 factors of affecting lateral resolution side to , there are things we can control
137:58 little bit, the source bandwidth size location of migration aperture. I'll define
138:04 velocity depth model. How accurate are alias noise? How dense are
138:10 Uh the migration algorithm are we gonna uh Utah's migration algorithm? We're gonna
138:15 some garbage that we got. We're gonna use a good one.
138:20 things we can't control subsurface illumination due the overburden increasing velocity with depth means
138:27 gonna lose resolution with depth. I'm have loss of frequency due to geometric
138:34 intrinsic attenuation. So geometric attenuation scattering of a Ruba surface like the ceiling
138:41 , intrinsic continuation like the squirt I talked about last week with three
138:47 four points. All right. Uh just gave a talk down the hall
138:54 at before when we came into here she's a Chevron now. So here
139:03 the um here is the oh here's down going way from the source.
139:13 the receiver, here's the midpoint. the migration aperture is, how far
139:20 the side do I take this ellipsoid Howie Joe talked about, how far
139:26 I propagate it? OK. And may say, oh I wanna make
139:31 as big as possible. Well, good. If your velocities are good
139:37 it also the bigger the migration the greater. If you double the
139:44 aperture, you quadruple uh computation And for migration of big 3d
139:51 we're talking one and $2 million. there's a difference between $1 million.04 million
139:57 you need to think about that. notice that the waves turn sideways as
140:03 go deeper. So all of a vertical resolution becomes lateral resolution and here's
140:13 in migration velocity. So if I a, if I have an error
140:17 the migration velocity and how I almost talked about this, I want to
140:21 this with a hyperbola. Well, low frequency part, the peaks and
140:27 will be constructive but the high frequencies be destructive. OK? So I'm
140:33 have a lower resolution and then my are gonna be in the wrong
140:39 OK? And here's high resolution. you. Then here's a merge survey
140:47 this comes from a Fairfield advertisement one company, two and each of
140:55 were required and processed by different So uh Fairfield uh Cel uh even
141:04 , they're data brokers. OK. BN Field, you're working in the
141:14 , right? OK. So, Field, Chevron says we don't want
141:20 stinking bin. OK. So they're sell their acreage and the seismic data
141:27 it and uh maybe they'll sell the data to or they'll license the seismic
141:36 to Cel and Cel will take the survey and the Continental Resources Survey,
141:48 is one of the big competitors in Bakken. Who else is up
141:53 You know which one? OK. there's, let's say three of them
142:02 there. OK. Now we're gonna those three surveys and we're gonna reprocess
142:07 , but before, before migration. . So each one is gonna have
142:11 source, different kinds of sources. are dynamite. Some are vibrators,
142:17 are acquired in different orientations at different and frequencies in the ground. Uh
142:23 of the ugliest thing is every survey a shot 0.1001 1002 like,
142:29 I gotta remember everything and that, the way is a total nightmare.
142:33 one of the most tedious part. never look your type. Don't choose
142:40 merged seismic processing problem as a OK. Don't, don't do
142:48 It's really tedious, really tedious. now they do do that. This
142:55 part of their business and then they it. OK. So you give
143:01 one library book and then they reprocess , mix it with other library books
143:10 then they sell it to maybe 10 the new operators who want to get
143:15 there. So you look at this , let's say this is the Texaco
143:20 , careful with my fingers. let's say the one on the left
143:26 uh a Texaco survey and the the one on the right is an
143:32 survey. OK? And I've got folks here don't see them very
143:39 Let's merge them to really image these . The Texaco guys need the amical
143:45 because those ball plane reflectors and the of the diffraction are actually required in
143:52 survey. So now I need those . OK? So after the
144:01 I get this image, I mean look at the difference here between the
144:04 and the right on the right. see those faults really nice. So
144:09 is the same piece of garbage data into it. It's just I comb
144:16 and increased the migration aperture. That the diffraction that were required in survey
144:23 two, they're being put in the place and survey number one.
144:31 So here's one from Anna Darko Basin Oklahoma and uh George Annis. Uh
144:39 uh ran coherence and here's uh a 96 survey and it looks OK?
144:46 nice channels in here. I got footprint and then theory I'm never
144:56 So the same input data from what call the Watonga Oklahoma survey and then
145:03 or four neighboring surveys. OK? here we're after Strat Democratic features.
145:10 two things are valuable with a mega . One is, I'm sorry,
145:18 see you that the channels are much , much sharper that more geology.
145:31 kind of nice too, right? I can put things in proper
145:35 So on the vertical section, you better. Thank you. And here's
145:40 decomposition. So we're looking at three looks looks fine but then merge.
145:49 , really helped a lot better signal noise, more geology. So mega
145:57 surveys, they show more geology such your play can now be put in
146:02 regional context. The data quality is due to improved processing technology, you
146:08 , maybe they were required 1015 years and the data from neighboring acreage images
146:15 dip and bulks and your acreage. why Asmus land data cost about 5
146:22 10% more than narrow Asmus data required because you have to pay for more
146:28 . Ok? But they're cheap pro made an isotropy as well with Leon
146:36 , we'll talk about when you take class from him. Long offsets provide
146:40 leverage against inter bed multiples or less by ground oil. Uh Surveys should
146:48 extended far enough beyond your acre acreage good images of steeply dipping structures to
146:56 at the uh Gulf of Mexico. The steep sal combs, you might
147:00 to go 10 miles away from that comb to record and state to image
147:05 right. Um Cut out some merge and large spec surveys often have excellent
147:12 aperture. So you're working for a company, Bill Bob Oil Company.
147:19 . And they've got acreage here. got, you know, let's say
147:25 square miles and they go to CGG CGG has a mega merge. That's
147:34 , 0, 10,000 square miles. mean, they get really big.
147:38 call them gig emerge surveys at that . And you just wanna get
147:42 Ok. They're gonna pay, they're charge you per, per square
147:47 Are they gonna reprocess just for No, they're gonna go in their
147:52 room and say, well, cut for Bill Bob Oil Company, cut
147:57 this square or this funny shaped polygonal . So you're gonna have the advantage
148:03 all the neighboring data to image your . So you're much better off having
148:10 from a mega merge than repros data over your area. Then uh if
148:21 , agriculture terrain allow, um then provide superior resolution and less footprint at
148:29 cost to conventional surveys. So that's why I asked Zack about his
148:34 not favorite place in Texas. If went and wanted to do identity up
148:43 Sam Houston Forest, north of it would cost more money. Or
148:49 big thicket east of here would cost money because you'd have to somehow magically
148:57 your, you, you, you get your vibrators in there. You'd
149:01 to helicopter in a dynamite. So would go way way up. But
149:07 you're on the, the Prairie of and it's not wheat season or if
149:12 had a big fire like a lot places recently, it, it doesn't
149:18 any more to get the high Ok. So any um questions?
149:33 ? Sorry, I gotta come All I'm hearing is uh ok.
149:40 you know where the 100 some for came from? The, the name
149:45 why picture? Oh, what which between traces or between the the bin
149:54 interval? Oh, ok. Oh, you want, you want
150:06 to share that with you? so we need to have some cocktail
150:15 . Why are there 5000 280 ft a mile? Anthony? Nobody
150:29 Ok. Well, so now you're be fun at the cocktail party.
150:33 drinks cocktail anymore. Oh, El Patron. You drink El
150:36 right? Um, so mile go Spanish neither. What's that mean?
150:49 , go to Spanish beer. Yeah, 1000 so left,
151:00 I had a home and I left paces. Mile passos, Latin.
151:09 . It's the way the Romans measured , but they would have a mile
151:14 every 1000 paces and their pace was ft. Ok. So that's,
151:22 where the word mile come from. , that's the, that's the cocktail
151:25 conversation. Did you know that? , then, yeah. And then
151:30 start walking away, you know, jokes with equations and doesn't go over
151:34 either, you know. Yeah, didn't know that. You wonder why
151:39 walked away, right? Um, let's break it in half, breaking
151:45 half. Again and break it in again. So that's got his phone
151:54 , uh divide 5280 by, I know what 16, what do we
152:01 ? 330? Ok. That's where comes from, uh, divided by
152:09 . No, 50 to 80. is 220? Ok. That's the
152:19 short and receiver 48. So what , there are intervals of a
152:25 So in the, in the United you go to school here?
152:31 Ok. In school, did they about townships and ranges? Ok.
152:43 . Well, no, that's not and 10, but in the United
152:46 in 1789 when we have the northwest like Ohio, right? You
152:54 you don't like anybody from Ohio. . Ohio. They said we're gonna
152:59 the northwest territories which were, we're gonna let it be claimed by Virginia
153:05 Massachusetts and New York and Carolina and forth. We're gonna let them become
153:09 States. We're gonna break it on grid. So the grid and the
153:17 and western United States is recto and grid. Ok. And they're
153:22 they're made into sections. Each section 36 square miles. So they're six
153:27 by six miles and then one square of each of those 36 uh square
153:40 is for public education and you can it or, you know,
153:46 but you can't sell it and that for the land grant school like Texas
153:50 and M and Oklahoma State University quite that. So, um,
153:58 that's why we have miles. Like here in Houston. Houston is really
154:07 in terms of Elgin becomes Lockwood becomes , right? Like, what kind
154:19 city is that? Where the same has three names? No wonder why
154:25 always lost. Ok. But other like Oklahoma City, Tulsa,
154:31 they're on a rec to win your and I'm going to name the
154:35 first street, second street, third , and then ABC D street or
154:41 of Mississippi, western Mississippi cities coming that, you know, real,
154:45 common. So that all that grid part of it and that's part of
154:49 system. So that's, that's where o comes from and why everything is
154:53 in the quarter mile except for older of Texas because that was Spanish and
155:01 of that was using the Spanish, Spanish system. So when you look
155:05 the counties in West Texas, they're rotated and stuff anyhow. OK.
155:13 any other questions you're afraid? I got you scared, right?
155:18 I might talk for 10 minutes. . Ok. If not, we'll
155:24 at 830 in the morning, we'll across the street again if you can
155:28 in right in the Fleming. at the car tonight. And if
155:35 were ill, send email and we do to you. Ok? You
155:40 , you rather meet here. It's to you guys. Yeah.
155:44 You just have to remind me not go waving at the board.
155:50 Ok. So we'll meet here tomorrow at 830 and you'll make coffee out
155:55 the hall. Good. Ok. , good. Midday. The next
156:01 , Beak Mozart, you know the ,
-
+