VideoPoints

• ••• • •• • ••••• • •
GEOL7323 Borehole Geophysics - Day7 03-10-2023
Help Tour
© Distribution of this video is restricted by its owner
Transcript ×
Auto highlight
Font-size
00:01 This is, this is one aspect a little bit complicated, but it's
00:09 of good brain exercise. So I'm gonna dwell on it too much.
00:13 it's one of the steps in BS processing and, and also in a
00:17 of other uh ocean bottom seismometers and . And there are a lot of
00:23 areas that, that use this And the idea is that again,
00:32 we're running the V S P, particular, we've got all these receivers
00:36 the well, and typically they're on wire line and the wire line tool
00:42 rotates a little bit as it's coming the well. So you can imagine
00:48 uh there might be calipers but there not be. And so we don't
00:52 know the orientation of the tool. know, we know the well
00:57 And so we know that more or the top part is vertical. So
01:00 know what the vertical channel is in GEO phone, but the horizontal
01:04 I don't really know which way they're . So the way that we can
01:11 do that is to imagine that I've say two positions, the horizontal channels
01:18 sensing at one orientation here at a orientation here. And I, I
01:23 really know the way the instruments So I have to figure that
01:29 So um if we think of the coming down at some angle and it
01:38 the tool and it just moves the and that tool movement gives us the
01:46 . It's like if I had a a wave coming down, I hit
01:49 GEO phone, it oscillates it up down. That outputs a little
01:54 Likewise, if I, if the comes in sideways and hits the
01:58 it oscillates it back and forth. the two sensors measure what is on
02:05 sensor orientation, right? So I've a three component tool and the wave
02:12 coming in at some angle. And these sensors do is say,
02:16 when, when you move on my , I output if you move perpendicular
02:20 axis axis, I don't do anything I'm on a spring and that's,
02:24 the only emotion my spring can Likewise, the horizontal guys say if
02:29 move me up and down, I feel anything that's not moving my spring
02:33 all. So these three sensors basically output the amount of motion that's in
02:42 orientation. So and this, this actually way easier than it seems,
02:54 it's kind of like dancing salsa or when you know how to dance
02:59 which you might, it's really easy you don't know how to dance
03:05 it's really hard. So we're just at the X and Y channels
03:15 So the horizontal one and horizontal So we've, I've got the vertical
03:19 . I'm not looking at that right , I'm just looking at the horizontal
03:22 . And if a wave comes down hits that, then the Y channel
03:28 measures how much motion there is on channel, the X channel or just
03:35 how much motion is on its So if I plot those together and
03:41 give the output at every point which really just a vector that has an
03:49 and A Y coordinate and that X Y coordinate is a measure of the
03:56 . And if I plot this in , just how does the vector move
04:02 you can see here, then it of moves back and forth like
04:08 So that's just what the motion is in the horizontal plane. And in
04:15 , this motion is actually coming about degrees to the GEO phone because it's
04:21 more or less equal amounts on the and the Y channel. So once
04:28 , if this is the full if I just output what's on the
04:32 channel, then the X channel is gonna oscillate back and forth like
04:35 If I just output what's on the channel, it'll oscillate back and forth
04:39 this. That's exactly what we see here. So normally we think this
04:46 and here is what is the But I can also take the record
04:51 plot what the motion is. So we take the two components and plot
05:02 together as a function of time, might remember way back from math,
05:08 are just such parametric realizations or it's a ho gram. So this is
05:17 a hologram. When I plot the of the vector has broken down into
05:23 and Y, I just plot the of that vector and, and where
05:27 moving. And then that in time called a ho gram. So also
05:50 if you were um in the physics in the electronics or electrical engineering or
06:01 , You might have two voltages coming a circuit and you might want to
06:09 how do those voltages vary with each . So if I plotted a voltage
06:18 one area against a voltage from another , and I plotted how they vary
06:22 in physics, they call that a figure. And it always reminds me
06:31 when I was an undergrad at we were doing electronics lab and we
06:36 um big cellos and I, I that familiar with the telescopes and we
06:44 talking about Lisa figures. So I to put the probe of the asco
06:48 the wall outlet just to see what of wave form was coming out of
06:52 electrical outlet on the wall Well, did Turns out that the amount of
06:59 that's coming out of your wallet that really high. The silos copes are
07:03 . So it takes the, the right to the ground. So I
07:07 circuited and blew up about a $3,000 . So good, good work.
07:14 , man. Yeah, that that was kind of embarrassing, but
07:22 never done that again. So I and then, uh, actually we
07:27 the telescope. It did it fried a little bit, but it was
07:30 . We replaced some of the uh some little circuit breakers and stuff,
07:35 it was uh it was pretty dramatic the time. Yeah. OK.
07:41 that's Lisa you figures. So what have to do in the BS P
07:48 because the tool is rotating up well, we want to get it
07:52 make it look like I had one G phone pointing at the so at
07:58 source. And so I can effectively project the data onto this maximum
08:08 So again, the recordings are just X and Y. If I plot
08:13 photograph, it looks like this. I can also say let's take the
08:19 of that hologram and project the X and the Y motion onto this new
08:28 . So effectively, if I had the data with an X primed axis
08:33 and a Y primed axis here, would have almost all the data on
08:37 X axis. So this new axis just projecting the data from X and
08:44 onto this AX primed axis. That's a rotation. And so we get
08:50 and when I rotate to X prime Y prime, then the data puts
08:55 all the energy on the X prime and I can throw the Y prime
09:01 . So I've reduced the data from channels to two channels. And now
09:07 makes it look like it was all in one plane, the X Z
09:13 , the X prime Z plane. so now all the data just looks
09:18 . So here's the uh in this , here's the X channel and the
09:22 channel, here's the first breaking energy down and it looks all kind of
09:29 because the instrument is rotating and it's with the amount of data on the
09:34 and Y s. So I just my ho gram find the direction of
09:39 arrival and then rotate these two channels put most of the arrival on this
09:51 channel or the X prime channel. you can see now there's almost nothing
09:55 the X wave of the P wave the orthogonal channel or the Y prime
10:03 . So now I can just now has other stuff on it, she
10:07 on it. But uh as far the P wave goes, now,
10:10 got a much more continuous X primed channel. And now we can
10:21 continue on and fun with figures. got the vertical channel because this was
10:28 a slightly offset V S P. I've got most energy on the vertical
10:36 and I've got a little bit of on the oriented horizontal ti now incidentally
10:46 that, if my waves are coming like this, most of the P
10:51 energy is on the vertical channel, which energy is on the horizontal
11:06 The energy on the horizontal, would , the P wave is on
11:10 So this would be the S Yeah. So now I've got a
11:13 S wave. So I've on on the vertical channel, I've got
11:19 of the P wave. Mhm But got the horizontal channel there too.
11:27 of course, in a vertically propagating , the S wave motion is largely
11:33 . And so sure enough, I've a little P wave energy on the
11:38 . But now I've got this big event, the downgoing shear wave on
11:43 horizontal channel. So I could pick P wave velocity and the shear wave
11:54 . Now, I don't know whether could see it, but this says
11:56 p waves arriving at .2 seconds at ft. So what's the P wave
12:09 Down to 1655 ft? Um It be so feet per second, it
12:18 be hm 1655 divided by 0.2. 82, ft /s. Yeah.
12:34 that's good. And then you can that later the shear wave is coming
12:40 nice and coherently and what's its So I've got the shear wave coming
12:50 here. Mhm. So that's gonna 6255 x 0.4. So about half
13:02 that. So 41, Yeah. is that kind of standard? So
13:11 a V P over V S of or shear wave velocity is half the
13:16 wave velocity. Mhm Yeah, that's the game. So we, we
13:23 that the shear wave velocity is somewhere half the P wave velocity because it's
13:27 secondary arrival. So, so that makes that all makes sense good.
13:43 . I just want to go through that's called ho gram analysis. That's
13:46 early step of most of the DS processing is just to make the data
13:51 consistent. And we go really from , a three channel without orientation on
13:58 horizontals to a two channel where the channel is rotated to be consistent and
14:09 at the source. OK. Uh covered most of this stuff last
14:26 Um Good. OK. Um let's just begin to look at it
14:45 a little exercise and I'll give you few minutes for this exercise. I
14:54 I think I think I posted this I better check. Mhm If I
15:15 , I will. Yeah, let uh let me just post this I
15:37 that's the best way to do Um. Ok. Um, this
16:24 just one to roll through fairly But I'll, um, these are
16:32 older data but this is just to practice in one of the,
16:38 using the V S P for a interpretation. So there are a few
16:47 , um, products here and you'll able to rotate these stuff, but
16:59 , here's looking sideways and to see I can rotate this. Oh,
17:14 on my home computer. So um my home license with Adobe is
17:21 but here is, here's some circus and it's on its side right now
17:31 it's split at the well location. there's a well drilled in here and
17:40 we've got our surface seismic. There go. Ok. And so we
17:55 see that um the surface seismic is . It's going from 500 milliseconds 5000.5
18:02 down to 1.5 seconds. And as look down, we can see some
18:12 highlights here. This is a little , a reflector. We've got quite
18:17 strong reflector just above B and then got a very strong reflector at
18:27 And so our job and not knowing about this area is to interpret these
18:33 . Now, what we do have uh we picked the first brakes of
18:39 V S P. So I've got the first brakes of the, the
18:44 S P picked and I've got a wave velocity and a shear wave velocity
18:48 of that. Then our trustee geologist given us uh an interpretation. So
18:56 this Pasco limestone that's just under 900 . And then there's the lower bath
19:01 is a sandstone and then there's the , which is a, a carbonate
19:06 limestone, Beaver Hill Lake, And so that's um just a little
19:14 of a really a Strat graphic column a function of death. Then we
19:24 some logs and we looked at these little bit before We've got our gamma
19:28 logs going from 600 m down to m. Were you able to pick
19:34 up there? Ok. Good. you can see we talked about the
19:39 ray before that it's fairly high gam the near surface. And then we
19:43 to a clean area and then we a dirty area and then it's cleanish
19:47 it gets kind of funky at the . P wave slowness is in microseconds
19:53 meter here. So it's fairly slow the near surface. Then we get
19:59 and then slower, fast, it , fast. So those are,
20:04 only got two logs here, but in depth. And then our BS
20:10 we've got P wave energy going down is the depth scale. And we
20:20 see that there's a change in the it goes from slow to fast,
20:25 gives us some reflections. So that's P wave RBS P, we can
20:35 on just the vertical channel here, wave down energy coming back. And
20:44 we got the V S P processed two way time. So here is
20:56 600 m down 1800 m. So see uh our reflections coming back.
21:10 now this V S P is in way time which we could stack and
21:18 we have stacked that into the, got a synthetic size of ground and
21:23 um a corridor stack or a P vertically extracted trace V S P,
21:28 trace. And then we had an V S P. So this is
21:32 little bit of a section. And again, we can see nice events
21:36 the offset V S P, the offset V S P and the synthetic
21:41 grams. We see fairly nice ties and don't worry about this. This
21:45 getting into the converted wave stop. we don't have to worry about
21:51 And the job is to create an plot. So by the, by
21:58 L plot, you, you're, probably gonna have to snip and cut
22:03 paste a little bit here if you do that. So the job is
22:08 get the these well logs 600 to m, just put that on this
22:26 . So the well logs right down . Now we've got a mapping from
22:32 well log to the V S P but these guys in beside the V
22:43 P in time beside this guy and snip out some of the surface seismic
22:54 is at the same times and correlate surface seismic to the V S
23:02 So you really are just going to putting in four of these one,
23:12 This guy 600 to 1800 m On top of this guy 618 cm and
23:22 that beside this guy and then put beside the surface size. So that's
23:32 the L plot. And then the is on the surface seismic. There
23:41 three horizons identified A B and C the job is to correlate those across
23:51 V S P. The synthetics. you're gonna see that the surface size
24:00 it is gonna correlate to some of events. And then we take that
24:04 and we find out what depth it from and we find out what depth
24:08 came from. We look at the and find out what geology it came
24:14 . So, um why don't you quickly assembling that? And let's,
24:23 take a break and assemble it. you, do you see how to
24:27 that, Stephanie? I think Right. Let's take just about,
24:33 give you just about 20 or 25 to try to put that together.
24:37 gonna jump in, we've got a meeting right now too. So I'm
24:40 jump into our faculty meeting and then be back in just about 20
24:44 See how, see how you OK. OK. Beautiful. Um
24:49 gonna stop right now. See if can take a little break, see
24:54 you can assemble this and then we'll we'll chat about it. OK.
24:59 you in about 20 minutes. So this is the, can you
25:08 that now, Stephanie? So this the uh the assembling and you can
25:15 it's an L plot just because it's the shape of an L and with
25:19 with your uh with that little the exercise and I'll leave, I'll
25:26 this with you too. This is answer. But I want you to
25:29 it and You can see that the this axis, the the depths the
25:37 goes from 600 m or so to m. So that's in depth.
25:45 course, the logs are, are in depth going the other way.
25:49 it's good to be able to um these in any orientation. Then these
26:00 and depth are the same depth as V S P. We could imagine
26:03 the, the borehole is here and This log now, in principle,
26:09 , the energy that has been shot zero asset is coming down into the
26:18 , but it's been muted here. we've just muted off that downgoing
26:26 And then as we talked about it we've shifted the upcoming energy to two-way
26:32 . So we, we understand that the fact that we had energy
26:36 down into the earth, it's hitting these interfaces and sending signals back to
26:41 surface. Now, from the sonic density logs, we can create a
26:47 seism gram in time. And then just stacked all this V S P
26:53 across and created a A V S extracted trace or just a stack
27:00 Then we've got an offset image that's little piece of offset seismic. And
27:06 I've taken a little chunk of surface I could put it in here.
27:16 with, with your exercise, you've uh you've got these pieces, I'm
27:21 gonna worry about the converted wave the P DS but the P wave
27:25 , the surface se make the, synthetic this guy and then getting that
27:32 depth. So let's uh let's take example. You can, you can
27:46 how some of this is done. then you've got another piece of paper
27:51 is, it's not of the same , but it's also given the uh
27:54 name of the units as a function depth, that little Strat graphic
28:02 Mhm I was kind of when you on, I had just finished um
28:07 connecting the dots to that. So here's the answer here. But
28:13 could squeeze that other um the Strat column with the velocities. You could
28:19 that and put that here and that just give you the, the names
28:24 these horizons and um rock tapes. for example, somewhere this is,
28:32 can see this change right here. . Mhm. So maybe just describe
28:40 me what's happening as we go from to here. Well, we go
28:48 a, a high gamma to we're decreasing in gamma. And then
28:55 we are kidding faster to er, getting slow to fast on our P
29:08 . Yeah. So our interpretation here that we've got a, a pretty
29:14 shallower section And then uh this is m. So that's 600, that's
29:22 1200 m deep there. So that's around 1000 m right there. so
29:31 a 1000 m depth, we go this kind of clay rich dirty stuff
29:41 has a high slowness or it's not fast and then we hit an
29:54 And, and everything, everything So at about 1000 m, we
29:58 from this cruddy stuff to this very , more regular stuff. And when
30:08 see something that's clean and fast like , like really fast. And this
30:12 um well, we have to divide up but that's was that 500.
30:17 there's 300. So this is like microseconds per meter. So there's a
30:27 exercise for you. What's the you said 100 microseconds per meter.
30:36 It's about 100 and 80 microseconds per . Ok. So that would
30:57 Mhm. Mhm. So about 5555 per second. Yeah, so that's
31:23 . So once again is that assault yeah could be four because four the
31:32 one was at 4000. So we salt's pretty uniform no matter what the
31:43 , it's always very uh low density it's always very moderate velocity, it's
31:52 around 4500 m per second plus or a little bit. Ok. So
31:57 did we say the velocity was 50 555. So that's too fast
32:03 Saul. How about sandstone? that's like around 20. No,
32:14 was in the two. Yeah, 3000 m per second. Too fast
32:19 sandstone. So, could it be ? Yeah. Yeah. So this
32:27 is largely a carbonate section. If go back to your Strat graphic
32:31 you'll see that this is largely So as you just calculated, which
32:35 good, that's, that's pretty It's 5500 m per second, which
32:39 pretty fast. And so mostly what that fast is some kind of
32:56 you know, how, how far is it to your place and from
33:01 to the woodlands to your place, mi. Yeah, give or
33:11 So that's, that's around 65 1.6. So that's somewhere around 65
33:22 away and This rock is going 5.5 . So it takes Something like
33:35 12 seconds, 12 seconds. If were carbonate from me to you,
33:41 would take seismic only 12 seconds to there. So you can really see
33:49 fast that rock is. It's really competent. Very fast.
33:54 That's, wow. Yeah. um, this is, this is
34:00 of your, um, uh, guide to the answer. But now
34:04 we look down here, so we about this that we've got this big
34:08 happening from mushy near surface stuff to carbonate. There's no carbon here.
34:16 it turns out is the top of Mississippian. So we're going from a
34:25 section into a carbonate section. So is a big impedance contrast. So
34:33 do I expect in the seismic A reflection? Yeah, big
34:39 So I'm coming down here in my V S P is in
34:42 I come around to see what happens the energy hits that depth. And
34:46 happens reflection mega reflection. So I'm at the V S P, the
34:53 S P says you've got energy going to the surface big time. And
34:57 synthetic seism gram says there's something Now, do you, would you
35:03 the synthetic seo gram to tell us same thing? Yes. Yeah,
35:09 here's the sonic log. And when look at the change in the sonic
35:14 , there's a big time change So that's gonna give me a big
35:17 reflection in the synthetic seism gram. the synthetic sees it, our seismic
35:24 it and then we go to the seismic stuff and we see it.
35:35 , this is going from uh low to a high velocity. And you
35:42 see that this signature corresponds to a trough peak. Hm It this was
35:57 is not S E G normal So going deeper, we've got an
36:05 increase or decrease. So going across interface going deeper does impedance increase or
36:18 . It was that's a negative, would decrease 50 50 chance,
36:29 Um No, because is that it it would be increasing. Why is
36:35 increasing? Because it's getting faster? , we're going from a low velocity
36:42 high velocity generally the density is going follow. So the delta V the
36:48 in velocity is positive here going So there's gonna be a positive reflection
37:00 . Now getting into a few more the details, they have plotted all
37:06 this vass Anchorage, but that's the that slummer used to plot all this
37:14 . So a positive impedance contrast here plotted as a negative spike. Getting
37:27 some of the details, this gets to acquisition. There was a reason
37:30 could say, well, why, did you bone heads do that?
37:33 doesn't make good interpretive sense. But the seismic world originally, it
37:39 And the original was because when we a GEO phone on the surface to
37:45 out how it was wired. People a tap test, you tap the
37:52 phone on the top and it gave positive output. So that was just
38:00 make sure that was a check in field that checked to make sure that
38:04 was wired properly and then nobody plugged their data on or anything was
38:09 So the tap test was the gold tap in the top of the GEO
38:13 . It should give you a positive on the instruments. So that being
38:18 right? Mhm OK. So if have a an explosion explosion, first
38:27 down and I'm going down and the coefficient is to say in this
38:36 um negative, then first motion would down. In this case though the
38:45 changes to the positive, we just that the velocity is increasing. So
38:49 means that the first motion is the uh the reflection coefficient is positive
38:54 V is positive. So the reflection is positive. So it sends the
38:58 back up as a compression. So sending the energy back up. When
39:04 hits the GEO phone, it pushes GEO phone up. So what should
39:08 GEO phone read? It should be negative. That's right. So the
39:17 reflection coefficient from a positive impedance contrast the field would read? Mhm So
39:29 is all plotted in field polarity. . If that makes sense, this
39:37 now getting into a little bit more the details, um we, we've
39:40 to worry about uh the polarity and shape and the, and the beauty
39:45 the V S P is that now know unambiguously the polarity of this
39:54 So again, with the surface I don't really know whether this is
39:59 impedance increase, decrease, whether there three layers there. I don't know
40:03 going on. But with the V P I can see there's a clean
40:08 right there it comes down, there's signature of that interface and it's a
40:13 trough peak because I have a band instrument, I don't have a spike
40:18 into the earth. I have this shape that I'm measuring. So this
40:26 agrees. So I completely understand that of these data are plotted in field
40:34 good. So if we go down farther, so this was an impedance
40:46 . If I go down here into layer, which is the salt,
40:49 a little layer of salt there. I go from the carbonate into the
40:54 , is that an impedance increase or that be a decrease, it's a
41:00 . So now the decrease corresponds to kind of impedance. So my recorded
41:07 in the surface is a decrease. that correspond to in geology? A
41:14 density, a higher deft well, this case, just a higher
41:20 So OK. Well, I'm I misstated that we're going from a
41:25 impedance, it's getting slower to a impedance is an impedance decrease.
41:34 And there's nothing um sophisticated about An impedance decrease manifests as a peak
41:42 the field. OK. So we that an impedance increase was a
41:49 So a impedance's decrease has to be out as a peak. And you
41:56 see, well, is that a fantasy lie? Well, no,
41:59 it is, we've got an impedance going into the salt from the
42:03 And I see a nice big positive coming off that impedance decrease the top
42:10 the salt because it's relatively a lower and it's a much lower density.
42:17 that much lower density times the lower gives me a much lower impedance coming
42:23 of a high velocity, higher density . So now we understand a little
42:33 more, not just about the timing the events and the type of events
42:37 can actually say about whether the it's impedance increase or decrease corresponding to the
42:44 of the seismic data. So let's another one. We've got this layer
43:05 . You can see we got carbonate . Then something funny is going on
43:10 here. It's dirty. Mhm It's . So could that be another salt
43:27 ? No, no, you're exactly . Why not? Hm I
43:34 according to the mythology law, we've a shale but the um velocity is
43:41 the, it just doesn't make Well, the, yeah, the
43:46 is not so much the velocity, the, it's the gamma rate.
43:51 , gamma, the the velocity of is relatively low. So this,
43:56 seen the velocity down here that So velocity alone this could be
44:05 But the gamma ray says no, is, this is quite dirty.
44:10 so I don't really know what it . All I know is that it's
44:15 and it's dirtier. So it's probably straight from these two lines.
44:32 So, but let's look at this , that's, so that's our interpretation
44:35 I'm going from carbonate here. Mhm what I'm interpreting to be a slower
44:44 Shelley garbagey and it might, might probably not pure shale, but
44:51 it's dirty. So this shay sandy which is characterized by a slower
45:02 So once again, I'm going from velocity to slow velocity. So that's
45:06 impedance decrease. And if I've got decrease, what kind of reflection is
45:14 la it's gonna manifest as a So if I come down here,
45:19 come right down, down, down, down, boom. There
45:22 is. Now, do you believe we're seeing that on the surface seismic
45:30 I go over here, there's this event. Is that event real?
45:35 it a multiple? Is it What's going on? Um Would that
45:50 um a multiple or is it just offset? Because I can kind of
45:56 like that, that the lighter one under like that big dark one.
46:02 . Um, so maybe, what, what did we say
46:07 What's a dark, uh, big correspond to? Oh, that's gonna
46:12 the, um, that's that other that's the trough, the load
46:16 load of high velocity and pence Yeah, that's that one for coming
46:25 here in depth. It's there. . It's bang on the top of
46:33 low velocity layer. Mhm. So is very definitive. I believe
46:41 And what is it? I go and I look at my logs.
46:45 It's the top of that sand which a a low impedance. Now,
46:54 bottom of the sand is going from to high impedes. So how should
46:57 manifest on our seismic? A little ? That should be the the peak
47:05 were just talking about. We went high to low as a peak.
47:10 low, I'm sorry, low to is the. So I come over
47:15 and I look at my seismic, got a peak trough and so how
47:18 I going to interpret that peak trough as a in and secrets? Is
47:31 what you're asking? Well, I that as this layer. OK.
47:37 . Yeah. No. OK. wasn't really sure what you are.
47:40 . And so the the layer has peak and in my field polarity here
47:45 peak corresponds to a negative reflection coefficient is an impedance decrease. So I
47:51 the impedance decrease here. And then going into a trough which in seismic
48:00 to an impedance increase in the log . And so I come down there
48:05 this line and I see when I out of this sand and I,
48:08 hit the bottom, I hit the of the carbonate. That's the impedance
48:13 which in field polarity is a So I see the top and the
48:18 of that sand lay. OK. what this is this is doing arm
48:34 and weightlifting exercises and not. So go out and play sports. So
48:43 we're looking, we're going from the Time Domain here, that signature to
48:52 borehole log domain and that signature. what we're doing is we're just going
48:57 and forth here, but it's exercising brain because I'm going from seismic wiggles
49:02 time to rock properties in depth. that's what this plan is intended to
49:11 because we've got these seismic wiggles in . I don't know what that means
49:14 all. But when I correlated it the, the V S P extracted
49:20 , the synthetic and I look at it evolves and then I go back
49:24 that depth. I can see exactly it means. That peak means that
49:39 decrease. So that's when you assemble , that's what you want to look
49:46 and look at the character as we into depth, we can see the
49:50 character. So we've done a we've looked at a lot of
49:53 So we're beginning to understand what this means. That's the log, the
49:58 jet folding in depth. But as geophysicist, generally, we're looking at
50:02 sections. And so we've got to transforming in our mind from the seismic
50:09 to the log and then from the to the seismic section. And that
50:13 the, that's really the job of interpreter is to be going back and
50:17 like this all the time. You , the geologist is gonna spend his
50:22 her time doing this. Most seismic are gonna do this. The
50:27 the geophysical interpreter is running back and between these two. Mhm Good.
50:38 you've got a bit more of a of, of that. OK.
50:56 , here's another example. So we see that again, we've got logs
51:03 depth here. This is shallow and this case, uh we've got an
51:08 P log and a gamma ray log we can see that our gamma ray
51:14 goes from 0 to 1 50 and S P log goes from zero to
51:20 like minus 50. So if we look at the gamma ray, there's
51:31 of character on here. It's generally up as we go deeper, but
51:36 all pretty chilly, classy garbagey. . The S P shows that we've
51:43 a few areas of big permeability, we can go down into the we've
51:51 here, the product, the impedance . And so the impedance log goes
51:56 low impedance, low density, low to high impedance, high density
52:03 Then we've got our processed BS P then circus seismic and then synthetic and
52:11 an offset. And so we just the same thing. We have to
52:15 to interpret this kind of poor surface , but we can correlate that event
52:24 figure out exactly what geology it corresponds . So that, that's the job
52:29 this plot. It correlates geology and with seismic in time. OK.
52:47 then there's more of the same OK. Well, that's uh that's
53:04 of the V S P stuff um just as one little other example as
53:10 case we can look at. Um was again to hammer home a few
53:20 of these ideas. We imagined that got to say our V S P
53:28 the well in the middle. And we've got a lot of different shots
53:32 we can make a V S P of all those shots. And this
53:36 , this was one of the cases we were involved with. So we
53:38 a well in the middle, then had two offset shots uh in various
53:45 around this guy was shot because the was drilled and it missed the target
53:51 they couldn't get in to shoot a 3d size because it was too expensive
53:54 they had a rig on site. they quickly put a GEO phone in
53:59 well did a bunch of shot points then created the offset V SPS.
54:04 here is the idea. Uh we a well log are zero offset V
54:10 P and then they walk away the V SPS. And now we're gonna
54:15 at the events and interpret them. these horizons And then correlate this is
54:25 of a bad picture, but this an old one and then pick that
54:29 and create a structure on the top that um map. Now, in
54:38 case, we picked an event that interpreted as the base of the
54:42 And we're looking for a high in below it. So in other
54:51 we pick the base of the salt we're looking for a reef sticking up
54:55 the salt. So it was a , the salt got deposited on top
55:00 the reef. But I'm looking for structure on the reef. And sure
55:06 , if you see just in this here there is a high on the
55:10 of the salt and that's the reef looking for. And so with
55:15 they took the well and then deviated well to hit the top of the
55:19 and get the hydrocarbons in the So that's uh the kind of thing
55:26 can do with multiple offset V Hm. And that can be done
55:32 fast. So that's, that's Ok. Well, great. Uh
55:43 , let's continue on for just a of minutes then take a little
55:50 Um We're gonna, we're gonna move uh another geometry. That's another borehole
56:20 . So we've talked about, logs down the, well, we've
56:25 put a seismic shot. We've got seismic log with velocity and reflectivity and
56:31 we walked away the seismic source that us an opportunity to create a
56:35 So we've created a picture. Then done a little bit of interpretation of
56:39 we use that new picture. And we're gonna put shots and receivers in
56:44 well. So now I, now I have two wells and I'm
56:48 gonna have just receivers in the I'm gonna have shots in a
56:52 So that's called the crosswell geometry. before we get to that, here's
56:59 little uh and before you see the , you might see the answer.
57:06 this is just uh I think it to be called Chinese squares or something
57:12 that. I'm sure that's illegal But the um yeah, but the
57:19 idea was that they Are four unknown in this pattern. So there's one
57:29 in each of those squares. So here, number here, number
57:33 number here. And then the sum these two numbers is five. The
57:37 of these two numbers is nine da da. So if you haven't done
57:42 already can you quickly guess. What's number here, here, here,
57:47 . What are the numbers that go here to give these sums? It
57:54 should, when I pull up my , the numbers are right here.
57:57 , no, it took the fun . Yeah, that's ok.
58:02 that, that's because it's so much to guess, you know, I
58:05 we every day. I don't know you do that at all.
58:08 I play word. OK. I, I do we every day
58:11 but it's, it's, it's really . These little puzzles are really
58:16 But um, so there you But we could create some different
58:21 Uh But as you can see, you've got, you've got all
58:25 Yeah, because the PDF smashed them together. Mhm To tell you the
58:30 when I was first working on way, way, way back.
58:35 It had just come out and I trying to figure out everything about tomography
58:41 I was actually giving a talk with lot of people and explaining a little
58:47 about tomography. And I had an like this and I did the case
58:52 I had everybody else do the case everybody gave different answers and I had
58:59 just calculated one of the answers. so when everybody gave me the
59:02 I thought, uh-oh, they're uh-oh there are all these different answers that
59:10 you can see work. So when get into inverse theory, this is
59:15 of a classic problem. And the problem is non uniqueness. So there
59:21 an answer but it's not a unique . So in other words, this
59:26 a, in a sense, this or these measurements are not enough,
59:31 not adequate to give a unique So that's an interesting detail. So
59:39 I'm gonna process this data, say gave this data to a computer to
59:45 , it's probably gonna blow up because going to have multiple answers, then
59:52 not gonna be able to figure that . So you can see that all
59:55 these, all these guys work. if I give it one more observation
60:00 I give it one more equation, can see now which of these does
60:07 have to be A different one Yeah. So all of these were
60:17 with just four equations and four but they weren't the right observations of
60:23 right equations to make it unique. if I add another equation in
60:27 then that selects a unique answer. it exists and it's unique. Um
60:37 where we're going with this is because is effectively the case for the crosswell
60:46 . So we're gonna use this a . But before we get there,
60:51 one other place that we did was work in Belize and we were trying
60:59 do imaging of pyramids. And so took seismic down there and I'm just
61:08 show you where it was in So here's the kind of structure we
61:19 seismic receivers and sources around the structure then we're trying to nondestructive, we
61:25 if there's anything inside this structure. that's, that's the idea. In
61:31 , here's the pyramid that we were on. Here's one of the archaeologists
61:35 were setting up and so we put ring or a belt of geophones around
61:41 pyramid and then we tap and we trying to make a picture of what
61:46 inside of it. So let's go now, that's where we're going.
61:51 here's the math. So you can that if we're taking a shot and
62:02 receiver, this is a slice through pyramid or it could be a slice
62:07 anything that the time to go from shot to the receiver is the sum
62:16 all the transit times through each picture picture element or little square. So
62:23 what we get here. So the transit time from a shot to receiver
62:27 just the sum of all the interval times for each one of those little
62:31 or pixels. Then you can see the travel time through each pixel is
62:40 the distance over the velocity of that . That's the sum of the transit
62:46 , which is just the sum of distance times the slowest. Does that
62:55 sense? All good? So that's we get. And now we could
63:05 that this is just for one shot receiver, but I've got many shots
63:09 receivers. So say for all the shots and receivers, I've got J
63:13 are all around. So each one those is now an equation of a
63:18 travel time from each receiver. So I have a matrix and a couple
63:28 actors. So the vector here T strictly the total travel time from one
63:37 to receiver multiplied by all of its . The second line is just the
63:46 receiver. And so this T is vector of times P is a vector
63:53 slowness and A is the geometry. to make this simple, we're just
64:11 say T is the vector of all different travel times P is the slowness
64:16 all the different um pixels arranged into vector. And A is just the
64:23 the distance through each one of those . So we're expressing it this
64:27 which is easy. And then we how to solve this equation. We
64:31 by a transpose and then I take inverse of a transpose A which is
64:39 that gives me this equation. And is how I solve that matrix
64:53 This is tomography. So what we do is just go in do our
65:16 which takes days trying to avoid scorpions malaria and lances and all kinds of
65:25 . Looking forward to getting out to beach. And uh we do that
65:29 we come back, we pick all first breaks. We know the
65:35 Mhm I just assume straight rays for pixel. So I know this,
65:40 arrange it like that, do the . And that gives me the velocities
65:46 slowness. So that's exactly what we . And here's the resulting picture.
66:02 these are low velocities in blue and high velocities in the red. And
66:08 is just one slice through that ancient . So we thought that maybe we
66:14 start to see some of the internal . And ultimately, we did a
66:18 of these slices and we reconstructed the . And what we're looking for is
66:29 that's an indicator of cavities or tombs burials or riches. Mhm. That's
66:37 we're looking for. Again, these were all built uh over the years
66:43 then they were built upon and then over and then drilled and then built
66:48 da da da da. So there's lot of structure inside here, but
66:52 were trying to get the uh internal in a nondestructive noninvasive way because with
67:01 these structures, you don't want to them and you don't want to cut
67:08 them. Now, looters did because through Central American coast, there was
67:11 of drugs going through and still are then, and then lots of arms
67:17 then lots of looting of these So people would go into the
67:20 they would search all over talk to locals and then find a pyramid and
67:25 uh cut it and try to loot . And so that, that happened
67:28 lot. You can see that big like this, this is cleaned
67:32 But where we were, people would these structures and then cut into them
67:37 see if there's any goodies inside and often was. But anyway, that's
67:43 , we were just doing the archaeology the uh the imaging. So that's
67:46 basic idea. Um Now we could out that was taking a horizontal
67:52 but we can do the same idea vertical slices. And if I've got
68:08 well here and then another while here receivers, I can just get the
68:16 times to each one of these receivers the shot, divide up the earth
68:23 take these sums that travel time is sum of all the slowness is through
68:28 and then do the inverse problem with guy and create an image between the
68:38 . So, and that's, that's we're gonna do. But let's take
68:41 quick break, Stephanie, then we'll back, we'll talk some more about
68:46 . OK. All right. Toxin. OK. Great. Welcome
69:09 . OK. We're just gonna kind spot check in here, Stephanie to
69:14 to get some of the basic concepts in and then wrap up. So
69:22 , the basic idea with Crosswell is got to have some kind of source
69:27 the, well, now, so got to create this kind of disturbance
69:31 the, well, and then we've to have a receiver in the other
69:35 , to, to record that So once again, some kind of
69:41 deform or something in, in well, and then some kind of
69:45 that's vibration or pressure sensitive in the . Well, and so uh here's
69:51 of the sources that we've done just the, in the shallow case down
69:55 the mark. So one is uh sparkler, what's called a spark
70:04 and this uh for this guy, is just a big stack of batteries
70:12 then it has a spark plug on end. So the way it works
70:27 that we're gonna have uh batteries on surface and then we're gonna charge up
70:32 set of capacitors and batteries down So we're gonna charge this all
70:36 put a big voltage across here and release it across just a little metal
70:45 here. So you've seen spark plugs a car or from a lawnmower or
70:50 like that. And so it's just release across the little um gap between
70:58 two pieces of metal. So what does is that this is in the
71:02 in the water and it basically uh , it's very, very hot,
71:10 a huge electric current going across, heats up the water, basically
71:14 it just produces steam and a And so quick discharge, electrical discharge
71:22 the water heats it up. And so it just creates a,
71:27 pressure pulse in the borehole. And our shaker. Uh, another,
71:34 technique that's mechanical. This is another that we have is this is uh
71:40 little cylinder. The cylinder has a inside it and then a plunger and
71:43 weight on the end. And this into the borehole, we inflate the
71:50 here so that it presses against the of the borehole and then you just
71:55 it up and drop it. it just creates boom, a sheer
72:01 the borehole wall. And that is disturbance of vibration that propagates across between
72:10 wells. So this is just an of one of the ones that um
72:13 we shot at Lamar. So we our two wells. I think you've
72:21 those guys uh Lamar and then we've the uh the sources on one
72:30 And then the receivers now as a , we can put download geophones again
72:35 sometimes we use a hydrophone cable, hydrophone cable just has uh pressure sensors
72:44 it. We just lower it into hole. It has all these pressure
72:47 and then we're gonna detect the energy goes from one more hole to the
72:58 . Did we do Croswell at Lamar , with you at all?
73:06 because I was the semester that COVID . So we couldn't do anything.
73:11 But I did do a well exercise Doctor Wiley. Uh, we did
73:18 water or like hydrology, geophysics or with him. We did something but
73:26 was my first song out there to . Yeah. Um, so
73:33 you've been there once? Just the ones with, uh, Bob
73:37 with ones with Doctor Wiley? And I did another one for,
73:40 when I took, uh, applications GPS and we did a couple different
73:46 out there with Bob Wang. Doctor Wang. I was like,
73:50 was his name? I can't It was so long ago. Um
73:53 yeah, we went out there with too, but that was it.
73:56 , ok. Oh, good. , you know the area and then
73:59 we've also drilled two wells there or wells and that's where we do a
74:02 of little tests and instruction. So was the idea and just to look
74:08 some of the data, here's the of data that we receive. So
74:13 we're again looking in depth and this the Scorpion. So this was the
74:24 , the spark or source in Well, and then We had the
74:31 array in the fluid in the Well, and the hydrophone was going
74:37 22 m to 118 m. So , and you can see that uh
74:45 know, the shot, the shot something like 30 m down in one
74:51 , and that um propagates across So , the uh the, the shot
75:00 in one well, at about 30 and then we've got the hydrophone string
75:03 the other. Well, that's down and 18 m. And so we
75:07 and then that data is recorded on hydrophone string. And then we were
75:16 part of this exercise, we were testing different sources. The other,
75:21 , the weight dropping source, the one with the bladder is called a
75:25 source. And so we were testing spark versus the Ballard. You can
75:29 the sedate is pretty clean and you actually start to see some reflections coming
75:33 up to the surface now. So Ballard gave some very nice, some
75:40 nice data. So once again, our geometry, we've got uh a
75:47 , say one shot here, the is going down here and you can
75:52 see the energy propagating across and arriving all different times. So in this
76:07 , the uh thesaurus was at 38 . So the thesaurus was just right
76:14 about there. So in around here we can see the energy all coming
76:24 . Now, there's other, there's energy forms because what happens is the
76:28 goes across, it hits the fluid , this hydrophone string is in a
76:32 filled borehole. And so we get low velocity wave going down the,
76:37 fluid which isn't too instructive to We're looking for, we're looking for
76:42 going on between the wells, not going on in the fluid here.
76:47 we're gonna suppress that in processing and pick all this stuff now as a
76:54 project with a student right now, , we're looking at uh DA
77:01 which is a fiber optic data. the fiber optic sensor in a horizontal
77:07 with shots on the surface. So got the wells drilled horizontally, there's
77:13 all in the horizontal well inside the , then we've got surface shots.
77:18 I think you can imagine that's again crosswell geometry just now on its
77:23 So shots receivers and so we can that whole thing is also crosswell and
77:32 what we're starting to do right now uh another one of the grad
77:38 That's his part, his phd. . So um that's the basic idea
77:52 um with the crosswell geometry. Uh you can see how it, how
78:05 works is we've got all of our and for every shot we can pick
78:14 times of the receivers. And so you can see that much of this
78:20 has been interrogated or traversed by different . So we're gonna do our standard
78:29 here that we imagine we're gonna divide area up into squares or pixels and
78:36 sum all the interval travel times to the total travel time. For every
78:40 of these guys put it in a matrix equation and do the inversion and
78:46 out with the velocities that are And so that's the idea we could
78:53 different pixel sizes. We run lots different tests and then ultimately come out
78:58 these velocities. So this is the of the slowness that fit all of
79:06 travel times. Now, you can that sometimes there are no rays.
79:11 the the area of confidence is right the middle here. So I'm confident
79:16 these velocities, but this gives us background. So you can see that
79:21 again, the sediments at Lamarque are around 2200 m per second, that
79:28 number that we know and love. that's the P wave velocity of all
79:32 sediments around here. So that's just the first arrivals in this particular geometry
79:47 the tomography kind of concept good. . So that's uh that's the cross
80:04 which we can put, we can it flat to look at pyramids,
80:08 can put it on its side to at dams in the near surface.
80:12 can turn it the other way and at surface to a horizontal well.
80:17 it's a very, very powerful concept how to process data and make a
80:25 good. So there's a bunch of stuff here. But I wanted to
80:30 um just a little bit more um taking this into the micro seismic world
80:41 this is uh also a big application a down hall source and uh surface
80:47 receivers in the well. And so a, here's a kind of example
80:56 and to bounce around not just oil gas because who knows where you're gonna
80:59 up working. But this was in , the geotechnical world. And you
81:06 see this nice mountain here and a valley. It turned out that uh
81:12 can see a little bit of rock here and there was a peak here
81:18 that 100 years ago failed and caused huge rock slide and buried an area
81:27 the middle of a valley. So was uh actually the most fatal landslide
81:36 North American history. And it buried a little village, a town in
81:45 crow's nest pass as this is And uh this is called the Frank
81:52 . So there was some worry that of these other peaks might um also
81:58 at some point. And so we asked to investigate this and put some
82:04 on and monitor the peak. So we actually ended up, it was
82:13 um uh a quick, quick So we bid on the contract and
82:18 we turned out we had to put this equipment on the mountain in
82:22 in the winter. And so uh quickly did that. I remember the
82:29 helicopter ride in the winter that we in to get all this equipment.
82:34 We're bouncing around and they let us on the top of the mountain
82:40 And I really started praying to every or goddess that I knew. I
82:47 this, this is really a bad . But here we are. So
82:53 uh we installed uh the equipment on mountain and here's an example. So
83:01 this case, we had six stations we started monitoring cracks inside the
83:12 So now we're listening to the These are micro seismic events inside the
83:19 . And so once again, our is pick the P wave arrival
83:23 the shear wave arrival times and locate source and find out the velocities.
83:32 there are only a few stations in case. And that's um again,
83:36 big inverse problem like we've been talking from the travel times, estimate the
83:43 and the velocity. So some of signal processing is we don't want to
83:48 around and pick this all ourselves. we have to figure out how can
83:51 automatically pick events. And incidentally, just saw an advertisement for micro seismic
84:02 . Did you see that? Oh, sorry, I was on
84:11 . I didn't even know. Um , I haven't. They're looking,
84:15 looking to hire people and your profile probably be very hirable with them.
84:22 . Micro seismic ink and they do stuff that I'm just about to show
84:26 . That's exactly what they do. they've hired a for a few of
84:31 former students there already. So Uh think Jose Basil is there right now
84:48 uh I think Elena might be there . I think they've hired several of
84:54 students. Hm. Ok. So kind of work that they do is
85:01 this, they monitor hydraulic fracturing and enhanced geothermal systems. They've just started
85:10 whole new uh subsidiary company called, think it's called Micro Thermal Inc,
85:17 is doing geothermal work with uh down fracturing and hydraulic fracturing, et
85:22 So what they do is they do from geophones that are on the surface
85:27 in the borehole and they're monitoring fracturing or, or seismicity underground. So
85:35 what you do. They've got all sensors and they, they first of
85:41 have to detect events. So say got some seismic, we're continuously recording
85:54 a surface station or in a down station and then there's a, a
86:00 fracture or a micro seismic event or mini earthquake or something. And that
86:05 a vibration and that vibration transits to sensor, our geo phone and boom
86:13 it is. So I would like be able to automatically detect that.
86:26 it looks pretty easy when uh I see it, you could go in
86:29 you could say, OK, there's onset I can, I can determine
86:34 . But now you've got to stay all night 24 hours and you're only
86:37 to staying up 20 hours, OK. You're accustomed to staying up
86:43 hours, you get two hours but you want to go to sleep
86:46 those two hours. Um So we to do this all automatically, but
86:54 would tell you that this used to be picked by hand. So when
87:00 was, when I was a there were whole companies that just took
87:05 like this, they just had dreams people that just picked these events all
87:08 long. And as grad students, was our job. And even up
87:12 about 15 years ago, people still to do all this stuff by
87:16 And grad studentss probably still do for earthquakes. But uh there's so many
87:22 these events they gotta to be And so you can see how we
87:26 begin to do this. There's a of character here, we can
87:31 first of all, the simplest thing can do is say, you know
87:34 the amplitude changes a lot. So defining the event by an amplitude
87:38 And this is the the simplest way do it. There's something called a
87:43 term average that's just gonna take the as it runs along here of a
87:50 number of samples. And then there's long term average, it's gonna take
87:56 average of a whole broader window of . So if I run the short
88:05 average across this signal, Then I the little line in red here,
88:11 can see that this all averages out almost zero because it's kind of like
88:16 noise. So we're getting zero there then all of a sudden the short
88:21 average when it hits this, it up like crazy. Now, the
88:29 term average is just another averaging. so it's coming along and as it
88:34 to get in here, it's starting feel that the aptitude is increasing,
88:38 it's got such a long window that doesn't feel it too much and then
88:43 it gets bigger and bigger and then down. So one simple way to
88:52 a detection would be to just take ratio of these two guys and once
88:59 ratio exceeds a certain value, I'm declare that's a trigger, that's an
89:08 . Mhm Now, in fact, could have just set a noise level
89:15 and then gone in right in this said whenever it gets above a certain
89:19 level, just pick it and people that. So that would be one
89:22 , but it's often it runs into . So you can see what happens
89:26 . We take our short term our long term average, take the
89:31 and when that exceeds some trigger say two, we're going to say
89:37 that is the event and then we're gonna calibrate a little bit because I
89:47 there's a bit of a delay but the event is somewhere around
89:51 And so that's how we can first of all, just calibrate
89:56 give it its threshold and then from on, just let it go in
90:00 do all of its picking and just me the time. So that is
90:06 event picking. And then once I've all those picks, I can start
90:10 process them. So here's what micro ink is gonna do or some company
90:19 it, there's a hydraulic fracture that breakages, that break causes a vibration
90:25 we monitor that and get the arrivals all these different levels. So the
90:37 , one of the problem with this that actually we don't really know the
90:43 or the location of this micro seismic . But I've got a monitoring
90:50 So I could have measured the So I might have I measure the
90:54 here. I've got that and now the velocities, I've got to determine
90:59 location. So this is the hyper location problem. And so let's just
91:10 one. Here's a simple case. was actually real data. So we
91:23 a, we had receivers in the and then there was an event someplace
91:35 I don't know where it was, from just one receiver in the,
91:42 , I'm gonna try to determine where event was. So that the,
91:48 technique is that we find an we've got an automatic pick. So
91:54 get an arrival of the P wave some time P this event went off
92:02 some unknown times, say, and it traveled some distance with a known
92:11 . So the P wave time, you can see here is T P
92:14 equal to T plus D over B P. And you can see the
92:30 S if there's a shear wave it happened at the same time,
92:35 it just took longer to get there it traveled more or less the same
92:38 , but it was traveling to the wave velocity. So if I can
92:42 the P wave arrival and the shear arrival, then I can determine how
92:50 away it was. And the way we're gonna do that is to subtract
93:07 two guys. What does subtracting them ? It gets rid of the T
93:10 term, the origin time term because don't know what that was. And
93:15 I'd like to get rid of And the way I can get rid
93:17 it is to pick the P wave the shear wave time and subtract
93:22 And that difference gets rid of the time. So you can see the
93:42 equation here, I pick a P time. I pick the shear wave
93:49 . I know the velocities because I've say a V S P measurement or
93:54 characterized the subsurface. So I know P and I know V S and
93:59 all I need to do is calculate distance. So what that's going to
94:14 is give me um a distance from receiver to the source. Now,
94:28 I had a whole pile of I could do various things. But
94:31 , for example, I've just got receiver, but it's a three component
94:37 . Then we talked about photographs at angles and so I can use how
94:42 energy is on each channel to figure the direction too. So I get
94:52 distance from the difference in arrival times I'm gonna get the direction from how
94:57 is on each channel. So you imagine here if I've got a vertical
95:02 and a horizontal GEO, this energy gonna arrive. And how is it
95:09 to express on the vertical and the geo how it's gonna express like what
95:23 you mean like positive or negative Yeah, so we'll get, we'll
95:27 some energy on the vertical and some on the horizontal and it's going to
95:33 up and down like this. So should, for this case here,
95:38 should see energy on the vertical and should see energy on the horizontal.
95:43 it's gonna look, if I plotted together, it would look like a
95:48 . But separately, I'm just gonna energy on the vertical and I'm gonna
95:51 energy on the horizontal. If that was more or less equal, then
95:56 would tell me the arrival angle is I saw equal energy on the vertical
96:05 equal the same amount of energy on horizontal. What arrival angle would that
96:11 me Uh is that 45°? Yeah, be about 45°. Yeah, but if
96:24 , whatever the ratio is, I just take the inverse tan of that
96:31 ratio. And that's gonna give me angle just like you pointed out.
96:38 let's just do this little calculation. actually was from a uh this is
96:48 data from a site. And we that the P wave velocity on average
96:57 this site is about 2000 m per . The shear wave velocity is around
97:02 m per second. So I've got velocities and what I know,
97:25 that I had my three component geophones and then I recorded this data.
97:49 I want to figure out I've got recording here. How far away was
97:56 rupture or that micro seismic event? we want to use our little equation
98:02 was T S minus T P is distance times the reciprocal of these
98:21 So here's the vertical element, Horizontal and Horizontal two geophones. And we're
98:34 to interpret that there's the first arrival is the P wave arrival and we
98:40 see how it it expresses on the channel and the X and Y
98:50 So first of all, what time , what time is this arrival
99:00 Hm. A little so P wave a little less than three.
99:05 So like maybe like two point 7 . Yeah. So let's take it
99:17 .2 seven. Ok. And then s wave is more than three,
99:31 .26. Somewhere around there. So the S wave is Maybe like
99:40 . Yeah. So somewhere around Ok. So just plug that into
99:51 little equation. One over 701 over . Mhm. I got negative.
100:24 , I did not get pulled Wow. Miles. So two About
100:49 , 97.8 uh meters. OK. oh 835 Oh I guess 20.26.
101:23 , .26. So that should be . Yeah, which is OK.
101:37 we're just, we're just doing this . So it's somewhere around 90
101:46 And so that that tells us how away it is. It doesn't say
101:55 it is, it just says how away it is, but we can
101:59 a little bit better than that by idea of we had three components here
102:05 I have an arrival angle that's got , for example, big time on
102:15 of the components. So first of , if say I've got vertical and
102:23 horizontal is north X and then Y east. So if I've got Z
102:39 vertical and say X is east and is north? Why is north?
102:57 a yeast more or less? What do you think this is coming from
103:07 you can write it down. I've Z why is north? So just
103:39 make sure, let's see. Um I'm seeing where it's ok. Um
104:20 not sure how I would figure it . I mean, I know it's
104:30 I know that um hm would it ? Mhm. No, I'm not
105:11 . Well, um you can see there's very little arrival this on this
105:20 horizontal channel. Ok? So if no arrival on the North channel,
105:26 no energy in that channel really? the energy be coming from that
105:36 Yeah. No, there's nothing on North channel. So it's not coming
105:41 the north. If it were, have, I'd have energy there.
105:46 there's there actually is a little, bit but not very much.
105:52 So we can see on the East now this guy there's a fair amount
105:58 energy. So it's coming from the then on the vertical channel, there's
106:07 fair amount of energy. So it's upward. Mhm So when it,
106:22 know that there's a half of the is on the vertical, half is
106:28 the east. So It's coming up of as 45° on the vertical and
106:38 channels. I don't even see that X Y at first because our chat
106:46 was like kind of covering it because was like, how am I supposed
106:49 know? But it makes sense Oh, good. Oh, that's
106:55 bad about that. Yeah. that, that's key, that's key
107:00 . Yeah, I was like, am I supposed to know?
107:05 I've noticed that sometimes the, various things are being covered up by
107:11 or something. So, but it is so that now we
107:18 we can even check this out a bit because that was the P wave
107:25 was oscillating on these channels. But had the shear wave coming up
107:30 which we assume is coming up sort in the same direction it was from
107:34 same source. So if I have shear wave coming up like this,
107:54 remember its particles are oscillating Ortho So I should still see primarily.
108:02 it's a lot bigger. Mhm So a little bit on the north channel
108:13 little bit. So we know that source is not directly east, it's
108:20 little bit east northeast because there's a bit of energy there. And then
108:29 I look at this, I've picked Z, here's X. So the
108:37 wave is coming up, we can some of it on Z, we
108:39 see some of it on X, shear waves coming up. I could
108:45 some of it on Z and some it on X. Now, if
108:48 look just in plan view, as said, most of the energy is
108:52 X, but there's a little bit Y or North groups. So we
109:12 check. And that ultimately is what seeing so that here's the energy we
109:21 its distance, we got its direction we can do all that just from
109:24 single G phone. Now, in practice, you're gonna use a
109:32 array of geophones to nail this a better or you're gonna put some on
109:36 surface like micro seismic does. But is an idea of sort of an
109:40 in situ shot and an in situ in the crosswell geometry. And if
109:47 trying to detect events, then this how we can do it. So
109:53 is called a hypo center location. means beneath like a hypodermic needle beneath
109:58 needle. So this is a hypo , the epicenter is on the
110:05 the hypo center is at depth. so this is how we can do
110:09 micro seismic analysis. And so what going to do is you've got a
110:25 fractured area, you've got sensors in , well, it breaks it
110:33 we pick all of those arrivals and we estimate where the break was and
110:41 big it was. And this is people do with micro seismic analysis.
110:47 so they're trying to find out where the rock break and then how big
110:51 it, how big of a fracture it cause? How big of a
110:55 was it? And this is the of micro seismic analysis that's used to
110:59 hydraulic fracturing for oil or as micro thermal is now doing or micro thermal
111:09 geothermal production. So that's um uh application. Good. Oops.
112:52 Sorry, I was dealing with screens and I think I lost you.
112:59 still here. Ok. Oh, , good. OK. So that's
113:07 that's a run through of crosswell. just a quick, there's lots of
113:11 stuff there, but that's just a overview of some of the cross view
113:17 concepts. Uh Let's just take one case and put this all together.
113:35 is all all posted. Stephanie, uh there's a bunch here, but
113:38 just go through one case. I you probably remember this time. Uh
113:45 years ago, I was taking pictures my uh my view uh here and
113:51 could really see, you could really the water Here was, here's Highway
113:59 88 here, completely flooded. Then buildings outside my building were under about
114:08 ft of water. Um But let's on will just pick one of these
114:17 cases. Uh I want to quickly, quickly talk about Geats and
114:51 OK. Let's uh let's take this that ultimately, a lot of the
115:04 um what our job is is to how much uh fluid or, or
115:14 in place or gas in place or resource or something, how much is
115:21 place. And again, our job to calculate basically an oil column and
115:28 that oil column times an area and gives us the volume of resource in
115:33 . So this is in, in reservoir world. So what we've been
115:48 with our log analysis is we've been the thickness of the area of interest
115:55 the iso pa. So we've been log ins interest. We've determined sand
116:02 say from the gamma ray log. been looking at the porosity logs and
116:07 with resistivity, we've been calculating the of hydrocarbon fill. So if we
116:14 all that together, that gives us oil column or just the thickness of
116:26 resource, and then if we multiply times the area, that's just the
116:31 of fluid there. And so that's original oil in place or original gas
116:37 place or we could convert that probably original heat in place or our original
116:43 water in place. So this is whole idea. And so we imagine
116:49 seismic is gonna have to be interpreted give us the area. All of
116:54 log analysis is gonna give us the column multiply it together. And then
117:00 turn that over to the production engineers the accountants and they're gonna tell you
117:05 you get any more of this or a bonus this year. So that's
117:09 idea. So we can put that more of a sophisticated equation that the
117:13 oil in place is conversion factors, area times thickness times porosity times hydrocarbon
117:24 . Um Then when we take the out of the ground, typically the
117:30 is gonna come out of it. the oil is going to shrink or
117:37 , It will expand because it's under pressure, but it's gonna shrink because
117:41 gas will come out of it. so that's a little factor. Might
117:44 .9 and it's not too big, it might be some little factor.
117:50 . So what I wanna do is wanna put together our log analysis and
117:55 seismic and try to estimate the original in place and we'll do that with
118:02 study. So in this case, going to interpret our seismic and then
118:07 well logs try to put it all and get a volume. So we've
118:14 a 3d seismic in the area and we're taking that volume of echoes reflectivity
118:21 I'm Cutting across it, I'm looking the volume at one time and in
118:31 , there's sand channels in this area we're looking at the top of the
118:36 channels at that echo time. So you looking deeper in the volume cutting
118:43 and I've got the amplitude of the reflections at that time. And I'm
118:49 these bright amplitudes as my sand Now as it turned out, these
119:03 drilled and oil was discovered. But very top and what, what the
119:11 is here is just the amplitude of reflection. So this is a big
119:14 reflection. So as we've talked, we go down the section, what
119:21 cause the biggest amplitudes that we could what kind of rock change or can
119:30 the biggest reflections like from you like from like a low density to
119:37 high density or something? Yeah. . Yeah. Because like we smell
119:43 the salt and then when I went that shale to limestone, like those
119:49 pretty big differences. Yeah. So could be a big mythology change.
119:57 . But what else can cause big competing contrasts? Um Hydrocarbons.
120:12 exactly. Especially gas. If the are gas saturated, that's gonna really
120:17 the density and lower the velocity. so that's another big contrast. And
120:23 it turns out here, it's actually . So we've got a mythology change
120:29 shale to sand and then the top the sand is gas saturated. So
120:36 gives us these boomer reflections. Those reflections were drilled now as it
120:53 out uh you drilled through them and was oil but the reflection is actually
121:00 on the top. Ok. So that was in time and I
121:10 to convert it to depth. we could just stretch it like we
121:18 before. But this is an aerial . So I've got, I've got
121:23 number of wells in here and from well logs, I know the depth
121:30 that reflector say. So I know depth of a reflector from the well
121:40 . And I know the time of reflector from the seismic, I picked
121:45 seismic and then I can do that multiple horizons. So I can
121:50 say one event here, pick another here and take the time difference.
121:55 that's called an isochron, that's a thickness. But then I know from
122:01 well logs, I've got the interpretation the that top and the base.
122:06 I know the actual depth thickness So from multiple well logs, I
122:13 that all the depth thicknesses and then my seismic having picked two surfaces that
122:19 to those, I've got the time or the isochron. So I can
122:28 the Isochron thickness. That's just the between two layers and the actual depth
122:35 two layers. So I want to this Isochron map to in time to
122:44 iso pack or an actual thickness map depth. So what we're gonna do
122:54 use some just statistical ideas to, do that. So here's the one
122:58 idea for Friday afternoon. The idea that I've got all of these wells
123:13 for each, well, I've got thickness. Now, you can imagine
123:23 the wells are really close together, difference in thickness is gonna be what
123:34 if the wells are really close Yeah, then the thickness isn't gonna
123:42 very big. Yeah, for It should be pretty much the
123:47 And in fact, as I put wells exactly together, the difference is
123:51 to go to zero and then as take the whales farther and farther
123:58 well, anything could happen. But what I can do is I
124:08 look at how much does that difference as I go farther and farther
124:15 So, the whole concept here is a function of distance. How much
124:25 do I get? Mhm. So , I can look at The
124:42 the one thickness and then I could at all the other distances away and
124:49 how different those are and construct what's a Vario gram, which is just
124:57 difference in that thickness. So here's plot of a vari gum as I
125:03 farther away from one, well, variation gets bigger. Mhm Stand to
125:12 . But it turns out that when scale that variation, it usually levels
125:18 . So once I've got a certain away that the that the variation is
125:22 a certain number, so that's, kind of a limit. So what
125:30 wanna do is take this Isochron which is a measure of the time
125:43 from the seismic between say this Mississippian a Manville at the tops. And
125:47 got a time thickness and I've also a thickness from the log. So
125:55 I'm gonna take each log and compare to each other log and see how
126:01 changes as a function of distance. how I construct the vari gram and
126:06 I want to convert this whole map a thickness using those vari graphs.
126:19 the question is suppose I just have data and I've got these sects that
126:24 going 100 and 50 m, 100 60 m, 100 and 80
126:27 And I look at the variation as function of distance. Then I can
126:32 you I can give you a number any point here. That's based really
126:39 all these other points. And the in GEO statistics is that from looking
126:49 these pairs and seeing how the variation as distance, that gives me a
126:56 . So if I want to output point here, I'll use all these
127:01 and add each of these values weighted how far away they are. So
127:06 are the two concepts in GEO statistics I want to output. I want
127:12 get a, a point here, I only have all these other
127:18 So how do I use all these points to predict what's here? And
127:23 way you do it is it'll take these known points compute how they vary
127:27 distance. And then for this point , I will wait the distance from
127:32 point to all the other ones, their values and get a value that's
127:36 weighted sum. Oh that's a But that's the GEOS statistical approach and
127:44 works pretty well. So I'll do , not just with the well logs
127:52 are in depth, I'll do that with the time thickness from the
127:57 which is in time. But I've a relationship between time and thickness.
128:04 so I'll convert everything and geo statistically each one of these points using that
128:11 . So this is called core. now I've converted, using all the
128:16 information and all the seismic information. converted this whole map now to depth
128:23 thickness. And I'm looking for thick in the reservoir. Then I can
128:40 leave one of those points out. the whole analysis. Look at my
128:45 compare to the actual point and get error and begin to understand how wrong
128:50 am. This is called a validation . So with this thickness, It's
129:03 somewhere around 170 m thick, but can leave one of these real well
129:08 out do the all the analysis compare real to the predicted output. And
129:13 I can say that in general For map, I'm within plus or -10
129:20 . And so that gives me a . OK. So we go through
129:30 whole analysis and we take all the from seismic, all the a all
129:39 I've got from the log and I predict the thickness with confidence. I
129:46 predict how much sand there is with . And I can also do the
129:50 same thing with porosity and even So we put all that together.
129:56 then ultimately jumping ahead, I can the oil column conjure that up from
130:08 and then predict how much oil is place. So that's, that's a
130:17 . And it would um that's a course in itself. But that's the
130:21 idea of using logs with seismic and it all together. Now, the
130:29 we want to do this is that got a few logs and I've got
130:33 size back and I would like to exactly how much oil is it in
130:39 and I can do that. And people are going to, the engineers
130:43 gonna pump it for a year or and they're gonna history match and do
130:46 their simulations and they're gonna get a , but their number came two years
130:51 our number. And so we came with a number that said, you
130:56 , you've got eight million barrels of in place there after a couple of
131:01 of production, they came out with similar number. So, uh I
131:14 say that jumping there, I gave talk some time ago about 10 years
131:19 we'd done the work, a guy up and he said, you know
131:21 , I just wanted to check and how with 10 more years production,
131:25 this all turned out. So he and he looked at this whole area
131:29 he got the government statistics on how had been produced. So ultimately,
131:37 had predicted that you would be able say that this particular pool there
131:43 I think what we predict, we that there was 90% chance that there
131:51 4.5 million barrels of oil. And after 10 years of production, they
131:55 actually there was 5.5 million barrels of . So that was pretty good.
132:03 Again, uh with a little bit time, AP 90 is what the
132:08 industry uses for reserves. And that's probability that your answer Is 90%
132:20 So we said that it's 90 there's 90% chance that there's more than 4.5
132:26 barrels of oil there. And then the end after 10 years of
132:30 they said, well, we've found there's actually 5.5 million barrels of
132:35 which is good because that would be our p 50. But when you
132:41 it to the bank, they don't to say, I don't want 50
132:43 odds. I want a 90% So that was good. So Stephanie
132:50 wrapping up with our work. Um looked at some core, we looked
132:58 well, logs, we did petro , we did some of the basic
133:02 for saturation. We did V S ties and then we looked at the
133:08 geometry really fast. And then how do we can roll all this
133:13 together to get uh, a reservoir . So, um, that's
133:22 that's a quick run through and I your brain might be partially full,
133:30 saturated. Good. So, I we'll, um, I'll leave it
133:43 . Um, just got the one , one little exercise that we talked
133:48 earlier, so, maybe finish that and then, uh, I'll
133:54 uh, Utah and you the, test probably Tuesday night, I
133:59 And then Tuesday or Wednesday and then got three or four hours to go
134:04 it. It won't be that It'll be something like an hour long
134:11 . Do you want me to have back to you by Tuesday or send
134:15 to you Wednesday night? Um, , how's your schedule looking for
134:21 Do you have a couple hours Wednesday ? Yeah, I mean, I
134:26 home from work around what time I get home about five and then everybody
134:33 home at about six so they can the baby while I kind of just
134:37 myself in office. So that should fine. Well, why don't
134:41 I'll send it to you Wednesday morning something and then just set it in
134:46 Wednesday night when you finish it. . Ok. That'll work. That's
134:52 . All right, good. we'll do that then. Um,
134:55 then, uh, I, I know, does I, are,
134:58 you online when you do it or it just, or just,
135:03 by yourself? Or how, how been doing that? Oh, I've
135:07 been doing it by myself. So I'll get it to you Wednesday
135:13 and then you could take three or hours however long you want it only
135:16 an hour long, but you can some more time and then just fire
135:19 back to us Wednesday night. I can do that. Ok,
135:25 . We'll, we'll talk to uh, in the interim and
135:29 uh, wrap things up for late night. Ok, that'll work.
135:35 , beautiful. Have a good You too. Thank you so
135:37 I appreciate everything. You bet. to you later. See you
-
+