►
From YouTube: DISC Lab Delft Afternoon
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
C
B
B
D
D
D
Them
not
needless
formats
for
maintain,
but
anyway,
my
colleagues
at
that
time
and
they
identified
clean
and
selling
water
fresh
salon,
water
interface,
just
for
your
to
know
from
where
they
come
where
it
come
from.
That's
in
in
about
after
the
90s
infect
these
vertical
electrical
soundings
cups,
they.
D
D
A
D
But
nowadays
my
colleagues
say
now:
we
have
these
kind
of
problems,
so
we
have
a
clay
layer,
another
one.
Maybe
there
is
some
hole
in
between
and
you
can't
really
say:
if
you
make
it
borrow
here
and
the
borehole
there,
you
would
say
that's
one
and
the
same
clay
layer
but
I
want
to
know
whether
there
is
a
hole
inside
or
not
so.
I
was
I
have
questions.
D
A
D
C
A
A
D
D
A
C
D
D
See
ya
in
resistivity,
but
then
I
thought:
let's
go,
let's
forget
about
the
SP.
If
I
do,
if
I
have
horizontal
electrical
current,
my
magnetic
field
that
the
service
cost
by
all
the
currents
will
be
rather
silly
now
I
have
this
deviation
of
the
electrical
current
and
that's
what
I
always
look
at
the
current
patterns.
Just
like
you,
I
can't
see
any
people,
people
doing
it,
they
just
model
and
look
what
the
result
is,
but
I
love
the
currents
so
the
current
density
field.
D
A
A
D
D
D
C
D
D
A
C
A
D
A
E
D
G
G
D
D
D
E
D
A
D
A
They
might
make
you
shake,
because
even
though
you're
you're
still
going
to
model
the
effects
with
that
cable,
but
you
know,
if
you're
out
of
a
meter
or
something
it
doesn't
make
any
difference.
But
if
you're
right
along
the
lines
that
you're
measuring
then
first
of
all
you
magnetic
fields
are
gonna
high
because
so
close
to
the
wire
and
that's
gonna,
make
it
tough
and
then.
Secondly,
you
try
to
subtract
that
out.
You
don't
know
exactly
what
wire
is
some
way:
you're
pregnant,
that
wire
from.
C
A
This,
where
we
use
an
inductive
sort
of
use,
both
grounded
and
inductive
sources
and
you're
looking
for
self-advocate,
not
target,
and
you
know
you
the
electric
fields
over
the
air
exp.
You
mentioned
magnetic
fields,
yes,
and
you
know
in
work
the
minute
in
you
can
convert
just
electrics
or
you
can
hear
from
headaches
and
you
get
out
answer.
You
can
put
them
all
together
and
you
know
in
cases
like
that,
we've
got
out
more
or
less
the
same
master
that
that
was
because
we
had
a
lot
of
measure
measuring
over
creating
with
receivers.
A
F
A
D
E
E
D
D
A
E
E
A
C
D
A
D
D
D
Yeah
and
I
was
just
thinking,
so
what
what
would
I
do
try
to?
It
seems
so
simple
if
you
do
it,
for
instance,
20
40,
60,
80,
Hertz,
or
something
or
maybe
you
only
need
one
frequency
and
get
three
coils
doesn't
seem
to
be
too
expensive
and
I
for
recording.
You
can
do
that
in
the
frequency
range
with
high
quality
audio
recorders.
These
days,
which
are
very
cheap
and
soon
doesn't
need
to
be
very
expensive.
D
And
I
only
need
a
source
that
can
be
applied
for
24
these
frequencies
or
something
quality
speakers.
It
looks
so
simple
to
do
and
to
test,
but
with
it
too,
you
know
it's
still
complicated
to
get
things
don't
know,
but
you
want
to
just
find
the
rhythm.
That's
the
point
at
the
moment.
Yeah
awesome,
yeah
I
built
my
first
coil
Oliver,
but
I
didn't
do
any
mushrooms.
That
was
in
my
time
you
anything
so
simple.
C
D
What's
going
on
the
thing
here,
it's
also
that
if
you
use
an
inductive
source
to
tackle
this
in
the
Netherlands,
you
have
fences
everywhere.
You
have
cables
everywhere
and
with
an
inductive
source
you
you
light
them
all
up,
but
with
a
grounded
source,
you
don't
do
it.
So
you
you
have
much
more
possibilities
with
these
kind
of
sources
than
with
inductive
sciences.
D
D
D
C
D
On
all
sides
and
I
connect
these,
and
these
are
connected
a
sort
of
injected
sort
of
homogeneous
field.
Current
field
in
here,
but
here
at
all
the
borders
I
get
deviations
in
the
magnetic
field,
so
the
magnetic
field
would
just
never
out
these
boards.
That
is
what
would
be
my
dream
cells,
not
only
with.
D
G
C
D
D
D
C
D
It's
a
long
time
ago,
but
somehow
I
think
it
still
contains
valuable
information
that
I
also
don't
see
back
in
many
recent
paper
surfaces
and,
of
course,
for
now,
I
will
focus
on
the
time
domain
en
and
vertical
electrical
standing
and,
of
course,
we
think
they
said
so
to
speak.
The
same
riff
resistivity
model
subsurface
has
only
one
resistivity.
Isn't
it
is
it
or
is
it.
E
D
D
The
nice
thing
is
also
that,
after
we
did
all
the
measurements,
there
was
a
bore
hole,
drilled
and
evaluated
on
permeability
even
and
we
tried
to
forecast
permeability
based
on
the
measurements
on
the
geophysical
measurements
that
must
be
of
interest-
oh
yeah,
but
yeah.
We
have
other
skills
yeah,
then,
indeed,
anyway,.
D
As
I
told
you
already,
we
were
used
doing
vertical
electrical
sound
waves
for
decades,
so
we
started
off.
We
have.
This
is
an
example
of
a
measurement.
We
have
their
vertical
electrical
sounding
and
we
started
to
interpret
depth
but
wrong
phenomena,
and
it's
colleagues
and
say
this
is
the
best
we
can
do
it
with
this
vertical
electrical
sounding
with.
C
C
D
D
D
A
D
C
D
D
D
D
D
C
D
D
E
E
D
D
And
I
can
also
show
you
the
seismic
just
to
give
you
an
idea
of
what
can
be
achieved
in
the
Netherlands
we
have
about
350
meters.
We
have
a
lot
of
detail
in
there.
You
see
dipping
layers,
you
see
a
continuous
layer
is
here
complex
layer
which
is
the
anisotropic
layer
or
layers
part
of
it.
The
shadow
is
part
of
that
complex
there
and
then
the
deeper
part.
You
have
many
arts
on
to
its
source,
to
use
here
only
in
state
I,
a
detonator
four
meters
depth
or
something.
D
D
D
Before
him,
but
we
couldn't
do
it
in
another
way,
we,
it
was
amazing,
to
see,
and
then
here
you
guys
play
layers
again.
Players
come
out
also
of
the
board,
so
I
think
it
says
it's
a
very
interesting
to
see
how
this
combination
can
confine
your
model
much
more
much
better,
at
least
in
distance,
and
has
happened
to
be
rather.
D
D
A
D
A
E
B
B
C
E
G
D
B
Took
a
juicy
weaving,
the
famous
athletes
for
for
logging
tools,
you
package
online
and
you
have
a
single
measurements,
and
then
you
write
an
antibiotic
and
needed
to
vote.
We
can
see
all
those
light
bits
we're
actually
getting
some
averaging.
You
would
make
me
say
that
that
layer
doesn't
represent
the
reservoir
with
in
factories.
It
will
go
back
and
Constantine
like
it,
and
this
whole
enjoy
live
now,
that's
right
hand,
side,
the
whole
law
died
on
acoustics,
but
with
over
the
crew
speaks
and
subject
put.
A
D
A
A
D
D
A
D
E
F
F
G
A
F
D
F
B
D
D
F
D
A
B
C
F
F
C
B
B
A
D
D
D
D
F
D
B
D
D
D
F
D
D
D
D
If
the
water
is
deeper
and
water
is
extracted
for
drinking
water
purposes,
and
then
you
have
coning
of
this
selling
water,
so
it's
everywhere
and
there
the
cost.
You
have,
of
course,
a
line
water
very
shallow,
also
because
of
the
polar
areas
where
you
have
low,
where
the
deep
water
yeah
everywhere
silly,
but
there's
lots
of
fresh
water
on
this
boat,
but
still
it's
an
issue
everywhere
in
Zealand
here
in
these,
it
is
around
a
few
meters
only,
but
still
the
farmers.
D
D
D
C
D
A
B
D
D
D
D
E
F
F
Actually,
what
I
do
this
father
I
really
enjoyed
your
talk
about
the
fact
where
you
said
that
you
updated
seismic
velocities
using
yeah?
We
actually
for
a
lot
of
things
that
we
do
with
science
fiction
or
migration,
and
we
need
a
good-sized
loss
tomorrow,
but
for
the
method
that
I'm,
using
which
ultimate
champion
method,
most
models,
XP
very,
very
important,
because
what
we
do
with
a
check
our
Mathis.
We
can
basically
change
the
location
of
the
receivers
of
our
data
from
any
location
in
the
medium
from
heaven,
service
I
mean
to
any
location
inside
medium.
F
You
have
another.
Take.
We've
got
a
response
from
that,
so
that
we
need
to
reflect
the
response,
but
also
very
poor.
We
need
to
have
an
estimation
of
the
first
arrival,
the
first
wave,
the
direction
of
location,
and
for
that
you
need
something
to
estimate
and
that's
what
we
need
the
velocity
for.
This
is
very
easy
to
do.
Believing
it's
pathetic
day
that
occurs.
A
you
can
accept
must
be
model
problem
is
when
we
want
to
do
this
for
real
data,
so
you
of
course,
have
things.
F
What's
with
this
stated
in
Saudi
Arabia
believe
yes,
then
you
can
get
your
first
rifles
are
coming
more
accurate
and
then
also
yourself
in
which
echo,
through
the
proof
and
using
that
you
can
actually
they
use
no
change.
I
lost
the
move
on.
We
also
use
data
to
improve
your
Yap
sort
of
like
getting
a
bit
of
a
jolt
in
first
I,
said,
and
that's
something
actually,
because
that
is
one
of
the
larger
problems
that
we
have
with
methods
which
is
very
interesting
but
I
think
that
could
be
a
very
big
benefit.
F
I'm
working
right
now,
data
that
I
have
very
mobile
votes,
but
I
am
hoping
actually
also
to
work
with
polar
field
data
which
may
have
a
bit
more
complicated
structure.
A
very
curious
if
this,
for
example,
could
improve
the
velocity
model
in
case
you're,
dealing
with
a
fault
of
course.
After
reaching
that
there's
quite
a
drastic
change
in
velocity,
because
the
layers,
certainly
in
lateral
sense
being
different.
So
I
was
thinking.
If
you
have
sensitive
you
can
be
to
that
with
yeah,
and
this
is
what
a
display
really
like,
for
example,
you're
stopping.
F
That
that
was
doing
on
bachelors
and
did
my
masters
sort
of
didn't
have
time
to
be
involved
with
it
anymore,
which
was
a
real
shame.
It
thought
was
really
interesting.
This
others
all
selection
inspired
to
join
us.
We
don't
really
have
much
information
about
the
field
I
do
now
there
is
in
active,
or
was
in
the
oil
field
supposed
to
say
where
it
is,
and
everything
but
I
think
it
would
be
nice
to
look,
maybe
I
assume
that
they
have
wells.
There
were
where
they
probably
also
this
son.
F
So
the
parameter
estimation-
maybe
you
can
see
if
there
is
any
relation
between
community
and
the
layer
parameters,
because
if
you
can
see,
if
can
get
more
estimations
like
that
in
different
parts
of
the
field-
and
you
can
maybe
see
in
the
false
model
model
that
we
have
right
now,
who
can
link
that
connects
lots
of
cognitive
email
see
if
I
D
especially
began?
Certainly
I'm
young
I,
actually
gotta
definitely
be,
of
course
schedule,
and
that
was
problem,
but
when
the
first
animal
for
us
I
can't
be
looking
we're
gonna
jump
evenly.
F
B
A
A
F
And
if
you
got
a
bit
of
information
like
this,
like
there's
this
big
plate
in
here
Billa
instead
of
saying
life
is
just
a
gradual
increase.
Here,
just
happens
bacon
over
here
to
sort
of
set
this
model
Israel.
You
can
find
a
different
solution
with
me
that
they
wanted
and
again
that
would
be
interesting
to
test.
So
so
maybe
I
will
look
at
some
data
from
you
from
your
modeling
clocks,
which
Universal
captain
providing
on
the
right
side,
which
I
think
it's
very
nice
and
compare
them
with
the
seismic
data
of
the
same
thing.
F
E
Imaginary
Cheryl
model
systems
and
shillings
receptively
the
size
of
these
different
political
operatives
generated
model.
The
NSF
and
material
is
also
fine
nation,
yes
spacesuits.
Actually
that's
generally
better
for
my
checkup
and
then
skip
the
day
job.
You
writing
articles
about
not
that
stupid
thing.
I
was
always
very
much,
usually
the
Stegosaurus
before
so
much
sense
to
exclude,
rather
than
maybe
yeah
using
that
database
and
let's
see
if
there's
any
defeasance.
E
F
F
F
C
A
A
A
F
A
F
D
F
E
A
D
C
A
A
C
F
C
D
C
C
D
A
D
B
D
C
A
D
A
A
C
C
C
G
The
surface:
wonderful,
yes,.
G
E
G
A
A
G
B
E
E
B
The
noises
description
for
me
to
it
is
imagine
a
stretched
service,
and
then
we
throw
our
hitting
balls
some
rather
certainly
at
this
dip
yeah
that
bending
of
space
finally
arrived
to
the
ball
that
I
have
to
dip
and
then
what
it
you
get
an
object
and
you
spit
it.
Do
some
angular
momentum,
it'll
rotate
around
in
all
of
that
body.
Obviously
at
some
point
in
the
grabbing
Saturday,
because
the
light
of
all
civilians
are
colliding
them
the
other
day
slide
and
the
other
thing
that
you
almost
be
like.
Yeah.
E
E
E
E
B
Exactly
through
petroleum
engineering,
it's
a
between
around
University
and
here
again,
because
your
team
was
your
unit
open
up
with
restrictions
and
since
it's
not
need
to
you,
don't
come
on
professor
deeply
established
a
connection
with
around
because
they
have
the
problem.
This
is
a
very
blue
mysterious,
not
for
the
reason
I
can
never
acquire
experimental
teeth
to
do
experimental
things,
so
they
did
everything,
sort
of
the
theoretical
on
step
and
now
they're
very
interested
by
brackets
last
yeah
Keith,
because
I
couldn't
get
keep
so.
E
B
E
D
E
E
A
E
G
G
So
we
have
a
number
of
links
here.
The
GP
g
is
geophysics
for
practicing
geoscientists.
This
is
a
resource.
That's
pretty
similar
to
EMGs
I
like
a
textbook
style
resource-
that's
more
so
been
developed
for
an
underground.
Of
course,
we
use
it
every
year
for
students
they're
mostly
geologists
and
engineers,
so
non
geophysicist.
So
it's
an
entry
level
resource,
but
there's
some
good
things
there
and
there's
also
links
to
some
similar
styles
of
maths
for
sort
of
more
introductory
geophysics
and
that
expands
all
methods.
G
Emg
osai
is
the
resource
that
we've
developed
for
this
course
and
so
I'll
jump
into
that
in
a
moment,
and
then
we'll
come
back
and
also
look
at
simpe,
which
is
the
software
that
we've
been
developing.
It's
all
completely
open
source
I've
written
in
Python
and
very
welcome
to
join
in
and
contribute.
So
I'll
show
you
how
to
do
that.
That's!
But
let's
start
from
IAM
chisaem.
G
So
there's
a
few
different
things
that
you
might
be
interested
in
checking
out.
We've
got
topics
on
sort
of
all
of
the
fundamentals
discussing
Maxwell's
equations
and
some
of
these
pages
are
under
construction.
So
you
will
come
across
pages
that
don't
exist
yet
if
you're
interested
in
or
if
you
find
typos
or
things
like
that,
you're
we'd
love
contributions
in
terms
of
fixes,
and
so
you
can
do
that.
Just
by
clicking
this
edit
on
github
button
and
it'll
bring
you
to
a
page,
you
can
edit
it.
G
One
thing
you
might
want
to
have
a
look
at-
and
this
is
a
great
place
for
contributions-
is
the
case
histories,
so
most
of
the
ones
that
we
presented
have
a
write-up
associated
with
them.
So
most
of
these
actually
have
a
publication
somewhere,
it
might
be
leading
edge.
Your
first
break
or
geophysics
article
and.
C
G
G
Current
links
not
available,
but
if
you
email
us
we'd,
be
happy.
A
A
G
G
The
other
thing
I
want
to
show.
You
is
the
apps
page.
So
here
there
are
there's
some
instructions
on
how
to
actually
run
the
apps
and
there's
also
links
to
get
up
and
running
with
them
on
the
web.
So
there's
two
different
styles
that
we've
developed
the
e/m
ups
are
what
Doug
showed
yesterday
and
I'll
jump
to
these
in
a
minute.
These
are
very
much
developed
to
be
basically
like
interactive,
GUI
style.
Apps
I,
see
you
open
it
up.
You're,
not
really
gonna
see
much
code.
G
It's
basically
just
get
straight
to
a
GUI
with
slide
bars.
That's
gonna
run
some
equations,
the
other
style
that
we
have
are
the
simulation
notebooks,
and
so
these
are
much
more.
If
you
are
interested
in
seeing
the
code
we've
gone
through
them
and
done
a
better
job
explaining
step-by-step
what
the
code
is
doing.
So,
if
you're
interested
in
running
a
frequency-domain
e/m
simulation,
for
example,
this.
G
To
go,
or
if
you
have
you
know
a
custom
model
that
you
want
to
input
where
you
see
an
app,
and
you
know
the
what
the
parameters
that
we've
provided
aren't
sufficient
for
whatever
example
you
want
to
run.
There
should
be
some
examples
here,
so
both
of
these
you
can
download
them
if
you're
familiar
with
github
and
python
and
installing
stuff
there
are
links
on
how
to
get
the
notebook,
so
you
can
run
them
locally.
G
G
G
B
G
G
So
sorry,
the
link
that
I
just
went
to
was
this
here
so
Microsoft
Azure.
So
if
we
follow
that
you'll
come
to
a
page
that
looks
like
this,
so
it
has
a
listing
of
all
of
the
notebooks
right
now,
but
you
can't
actually
run
them.
This
is
grayed
out.
So
in
order
to
run
them,
you
actually
have
to
sign
up
for
a
Microsoft
account.
It
is
free.
You
just
need
to
email.
It
password
sign
in.
G
There
will
be
another
way
to
host
them
in
the
near
future
as
well
through
it's
called
my
binder.
This
is
a
project
out
of
Berkeley
and
they
run
it
without.
You
know
having
a
login
wall,
but
right
now
it's
still
sort
of
under
development.
So
this
is
the
most
stable
way
we
can
share
them.
So
we've
got
in
total.
There
are
25,
I,
think
25
or
26
notebooks
that
are
available.
G
So
the
way
did
they're
organized
is
we
have
all
the
DC
ones
ABC
up
front
and
then,
if
you
go
to
the
next
page
over
you
see,
we've
got
frequency
domain
iam
marine
CS.
So
all
the
methods
are
sort
of
first
and
then
there's
a
bit
of
a
description
afterwards,
and
so,
if
I
want
actually
run
in
this
library,
I'll
clone
it
first.
E
G
E
G
So
now
we've
got
our
own
coffee
and
we
can
go
in
and,
for
example,
look
at.
Let's
look
at
this
one
here,
so
they
start
up
a
server
for
you,
and
so
the
first
thing
I
think
there's
two
ways:
you
can
run
this,
so
what
I'm
gonna
do
here
is
go
cell
by
cell
and
you
shift
enter
and
that
executes
it
one
at
a
time.
The
other
thing
that
you
can
do
is
just
do:
cell
runnel
and
I'll
just
run
them
all.
G
So
when
I
go
one
at
a
time,
the
first
thing
this
is
doing
is
just
installing
the
dependencies.
So
eeehm
examples
is
a
library,
Python
library
that
we've
written
that
basically
just
has
all
the
code
to
hook
up
the
plots
to
the
computation.
So
it's
just
going
to
install
that
and
it's
dependencies
and
so
e/m
examples.
The
way
that
we've
built
these
up
is
they
really
are
basically
plotting
functions
on
top
of
simpe,
which
is
the
underlying
physics
simulation
engine.
G
E
G
Or
if
you
want
to
do
it
from
the
menu
you
can
do
cell
at
all,
okay,
so
this
guy
key
yeah,
and
so
this
example
is
we've
got
a
thin
layer.
And
so
what
are
the
one
of
the
questions
that
came
up
yesterday
was
talking
about
charges
and
where
the
charges
are
with
the
layer.
So
that's
something
that
we
can
actually
go
ahead
and
look
at
here.
So
do
we
want
a.
G
G
So
what
this
one
actually
does
is
we
have
a
uniform
background
and
there's
a
cylinder
in
there.
So
this
is
we've
got
some
sort
of
background,
which
is
the
same
over
top
and
bottom
of
that
in
between
okay,
so
we'll
leave
that
layer
there
and
then
we
can
go
ahead
and
look
at
the
charges,
and
so
it
first
plots
the
total
charge,
as
we
expect
there.
Electrode
locations
fluid
total
current
as
expected
for
lighting
up
our
layer
and
then
let's
go
ahead
and
look
at
the
secondary
charges.
G
So
that's
something
that
you
can
play
around
with
and
if
you
run
into
any
issues
or
questions
with
things
like
that
feel
free
to
fire
off
emails,
we're
always
happy
to
respond
to
those
and
I
would
like
to
show
you
now
the
other
style
of
apps
that
we
have.
So
if
you
are
interested
in
jumping
into
the
code,
these
will
be
these
simulation
notebooks
and
so
the
way
I'm
going
to
do.
This
is
let's
first
actually
have
a
look
at
simpe,
and
so
the
way
you
can
get
to
simpe
is
again
from
geosite.
G
G
So
right
now
we
mostly
have
electro
magnetics.
We
have
DC
IP,
mag
and
gravity,
there's
also
some
fluid
flow
with
Richards
equation,
and
so
in.
In
this
case,
with
all
of
these
methods,
we
have
both
the
forward
and
the
inverse
problem,
and
if
you're
interested
in
getting
up
and
running
the
first
place,
I
would
look
is
looking
at
the
examples.
G
So
there's
a
few
that
have
been
published.
So
you
can
see
those
and
reproduce
results
from
the
papers
and
then
there's
my
method.
You
can
have
a
look
at
some
of
the
examples
that
there's
only
there
is
only
a
handful
of
examples,
and
so,
if
you
do
run
into
something
that
you
want
to
try
and
do
and
have
a
question
for
us,
the
best
way
to
get
in
touch
about
simple
is
either
through
github,
but
I
would
actually
recommend.
G
G
G
So
here,
if
we
have
a
look,
there's
a
handful
of
examples,
the
so
we've
got
both
we've
got
a
DC
example.
We
have
an
MT
tutorial,
so
this
actually
goes
through
dis,
critizing,
a
set
of
equations
and
forming
all
of
the
operators
doing
the
forward
simulation
and
the
inversion.
So
we've
got
a
little
leading-edge
tutorial
out
on
that
and
then
these
notebooks
go
with
it.
So
that's
something
you
can
potentially
look
at
if
you're
interested
in
the
example
that
I
think
I'm
going
to
show
you
now
right.
G
So
this
is
the
a
frequency
domain
sounding
over
a
sphere.
So
I
think
this
showed
up
in
fundamentals
right
here:
yeah,
so
we've
got
a
sphere
and
we're
gonna
put
our
transmitter
right
over
top.
So
in
this
case
it's
a
cylindrical
asymmetric
example.
So
we
can
use
the
cylindrical
mesh,
that's
very
fast
and
will
offset
our
receiver,
and
so
here
the
the
dependencies
are
simpler.
So
if
we're
just
bringing
in
simpang
and
then
we're
going
to
go
ahead
and
work
with
that,
so
just
installing.
G
It
and
then
just
import.
These
are
fairly
standard
scientific
Python
packages,
so
numpy
does
like
matrices
it's
effectively
like
MATLAB
matplotlib,
to
sell
the
flooding
and
then
Syfy
we're
just
bringing
right
now
and
constant,
but
Syfy
also
has
all
the
sparse
linear,
algebra
they've
got
solvers
optimization.
Well,
that's
where
it's
done
and
then
from
soompi
we're
bringing
in
a
few
different
modules
so
we'll
bring
in
the
mesh
some
utilities
maps
I'll
show
you
what
that
is
in
a
second
and
then
from
electromagnetics,
we'll
bring
in
the
frequency
domain.
Yeah
I'm
code,
I.
G
E
G
And
so
here
we'll
do
we're
gonna
set
our
transmitter,
20
meters
about
the
surface
and
have
our
receiver
8
meters
offset
and
we'll
do
25
frequencies
logarithmic
weights
based
and
then
we
set
our
source
location
so
we'll
do
it
at
the
origin
horizontally
and
then
up
20
meters.
And
then
our
receivers
are
just
gonna
be
offset
so
because
we
need
the
3
3
coordinates
there
and
then
one
of
the
important
things
when
you're
actually
setting
up
a
mesh
is
to
make
sure
that
it's
going
to
be
appropriate
for
the
computation.
G
And
so
specifically,
when
we're
looking
at
eeehm
simulations.
You
want
to
be
looking
at
the
skin
depth,
so
our
minimum
cell
size.
We
need
to
capture
the
quickest
things
going
on,
and
so
in
that
case
we
want
to
look
at
the
highest
conductivity
and
the
highest
frequency,
and
so
in
this
case
that
gives
us
a
minimum
skin
depth
of
5
meters.
So
we
want
to
make
sure
that
our
finest
selves
are
probably
4
times
smaller
than
that
and
then
our
maximum
skin
depth.
G
G
And
then
one
of
the
things
that's
important
when
you're
thinking
about
computation
time
is
just
how
many
cells
do
you
have
so
in
this
case
we're
just
printing
out
how
many
cells
we
have
so
python
is
object-oriented.
So
that
means,
when
you
form
a
mesh,
there's
a
few
different
things
you
can
do
with
it,
and
so
in
this
case,
once
we
formed
our
mesh,
we
have
all
of
our
operators
with
it.
So
we've
our
edge
curl,
for
example,
we
know
what
it's
dimension
is
or
we
can
get
a
cell
grab
all
those
sorts
of
things.
E
G
Us
a
cylindrical
mesh
and
then
here
when
I
pulled
up
this
menu,
all
I
did
was
hit
tab.
So
this
is
a
nice
thing
in
the
jupiter
environment.
I
think
so.
Jupiter
is
what
this
is
running.
Is
this
notebook?
It's
the
computational
environment,
and
so
when
you
do
that,
you
can
see
all
of
the
methods
and
attributes
that
your
object
has.
So
if
we
wanted
to
see
the
number
of
edges,
for
example,
we
can
see
our
number
of
edges.
G
G
So
the
next
thing
we
want
to
do
now
that
we've
got
a
mesh
and
we
know
our
model
parameters.
We
can
put
our
model
on
the
mesh.
So
here
all
we're
gonna
do
is
first
assign
the
conductivity
at
the
air
everywhere
and
then
we'll
insert
our
half
space
below
and
then
I
can
stick
our
sphere
in
and
then
we'll
just
plot
that
to
make
sure
that
it
looks
reasonable.
G
G
Secondary
fields
so
we're
just
setting
up
to
frequency
domain
receivers,
and
then
we
put
those
in
a
list.
Those
are
the
all
the
things
that
we
want
to,
listen
to
you
basically
and
then
from
there.
We
can
then
construct
our
source.
So
in
this
case
we
want
to
use
a
magnetic
dipole
source
at
the
frequencies
that
we
listed
earlier.
So
that
was
just
this
parameter.
Freeze
will
give
it
a
set
orientation,
because
that
keeps
everything
is
literally
symmetric
in
this
case
on
its
location
and
everything,
and
then
here
once
we've
formed
our
source
list.
G
G
G
The
other
thing
that
we
often
pass
the
problem
is
a
mapping.
So
when
you
think
about
doing
an
inversion,
you
generally
don't
want
to
actually
invert
for
electrical
conductivity
directly.
You
usually
want
to
invert
for
something
like
log
conductivity
and
then,
if
you
have
an
air
earth
interface,
for
example,
you
don't
want
to
invert
for
the
air.
You
know
where
it
is,
so
you
want
to
fix
those
cells,
so
those
steps
can
be
accounted
for
through
maps
and
then
that
also
keeps
track
of
all
the
appropriate
derivatives.
I.
G
Okay,
so
now
we've
got
basically
all
the
pieces
that
we
need
to
go
ahead
and
solve
the
forward
simulation,
so
we'll
just
solve
for
the
fields,
so
we're
going
to
solve
twenty
five
four
problems
and
that
finished
in
just
over
a
second
and
then
we
can
go
ahead
and
plot
out
the
fields.
So
this
is
just
a
giant
bunch
of
plotting
code.
G
It's
all
from
that
problem,
so
I'm
not
going
to
go
through
and
then
what
we
did
here
is
actually
then
we
wrote
it
as
a
function
and
then
that
lets
you
easily
use
these
widgets.
So
if
you're
curious
about
how
to
use
the
widgets,
all
that
I
had
to
do
to
make
this
plot
is
basically
provide
a
function
that
takes
two
parameters.
So
it
takes
a
frequency
index
because
I
want.
G
Slice
through
frequencies
and
then
a
toggle
for
real
or
imaginary,
and
then
the
I
PI
widgets
library
provides
this
interact,
which
basically
just
wraps
your
function,
and
then
you
provide
your
input
parameters.
You
tell
you
tell
iPad
widgets
what
you
want
them
to
be.
So
in
this
case,
my
frequency
index
is
just
an
integer
slide
bar.
So
that's
all
this
is
and
then
toggling
between
real
imaginary
is
just
toggle
lens
and
that's
all
we
had
to
do
to
get
now
an
app
to
look
at
our
result.
G
G
G
G
D
G
Brain
it
would
be
independent
of
yeah,
and
so,
as
we
go
higher
and
higher
in
frequency,
you
can
also
see
that
the
the
maximum
starts
right.
It's
moving
up
as
always
we're
at
our
highest.
It's
mostly
concentrated
right
near
the
top
of
its
game,
and
then
we
can
also
have
a
look
at
what's
going
on
with
the
imaginary
part,.
E
C
G
E
C
D
G
Think
the
next
plot,
I'm
gonna,
sure
you
might
get
yeah
I-
might
help.
So
what
we're
gonna
plot
here
is
you
see
the
this
pink
dot
yeah?
This
is
where
our
receiver
is
located.
So
we're
gonna
do
is
plot
the
magnetic
fields,
the
sounding
so
across
frequency,
and
so
do
you
have
a
picture?
What
do
you
think
that
should
look
like
as
we
increase
frequency
Ford
will
see
a
real
magic?
Do
you
remember
the
characteristic.
C
C
A
G
G
D
G
And
so
for
your
example,
we
could
do
it
in
ecstatic
where
we
could
actually
solve
the
the
frequency
domain
problem
at
the
frequencies.
You're
planning
on
looking
at
I
would
suggest
starting
on
the
static,
and
so
there
is
a
3d
DC
code.
So
in
this
case
all
you
need
to
do
is
set
up
a
tensor
mesh.
Stick
your
layered
model
in
and
then
go
ahead
and
and
simulate
for
a
single
source,
it's
a
fairly
light
problem
and
then
to
get
the
magnetic
field,
the
integration.
So
you
just
need
to
integrate
and
I.
G
G
D
G
E
G
So
if
you
have
right
now,
if
you're,
specifically
looking
for
just
magnetic
data
in
that
case,
I
would
come
to
the
sympathy.
Examples
and
there's
examples
here
for
magnetics,
and
so
here
is
an
example
of
a
linear
inversion,
and
this
one
looks
at
a
few
different
norms
that
you
can
employ
in
an
inversion.
So
this
here,
then
shows
you
on
the
output,
as
well
as
the
code
that
was
used
to
run
this.
C
F
E
G
G
B
G
So
we
show
you
all
of
the
finite
volume
up
or
this
one's
finding
different,
but
we
show
you
how
to
basically
build
up
all
of
the
operators
that
you
need
from
just
the
the
mesh
box.
So
that
gives
you
a
bit
of
a
handle
on
like
what
you
can
do
just
with
the
simple
mesh.
So
if
there's
a
problem
like
seismic,
for
example,
that
we
don't
have
implemented,
but
all
the
operators
are
there,
you
can
go
ahead
and
follow
this
to
basically
piece
together.
Your
set
of
equations.
G
The
next
one
is
then
going
through
and
looking
at
now
using
that
code
and
exploring
the
forward
modeling.
So
what
we
do
here
is
go
through
and
explore
some
aspects
of
mesh
design.
So
when
I
mentioned
you
know,
we
need
to
make
sure
that
our
smallest
cells
are
sufficiently
small
to
capture
what's
going
on
in
the
physics
and
then
our
domain
needs
to
extend
far
enough
so
that
everything
is
the
key.
So
we
set
up
this
notebook,
so
you
can
explore
the
impacts
of
what
happens
when
you
violate
those.
G
So
there's
that
and
then
there's
also
an
example
of
non
uniqueness
and
empty.
So
one
of
the
classic
examples
is
varying.
If
you
have
a
default,
if
you
have
any
thin,
highly
conductive
layer,
what
we're
really
sensitive
to
you
in
MT
is
the
conductance,
and
so
it's
that
conductivity
thing
it's
product.
So
if
you
double
the
thickness
of
the
layer
and
have
the
conductivity,
you
can
hardly
tell
the
difference
in
them
to
do
anything.
B
A
B
C
G
G
Okay,
so
that's
the
second
notebook
and
then
the
third
one
is
the
one
that
shows
inversion.
So
here
we
actually
set
up
a
five
layer
model
and
go
through
and
run
the
inversion
and
look
at
some
of
the
different
impacts
of
regularization
parameters,
as
well
as
your
choice
of
beta,
and
so
that
is
the
parameter
that
controls
how
well
you
fit
your
data
versus
your
opera
killer,
education
and.
G
To
then
go
through
and
explore
some
of
those
things
so
yeah
and
there
is
a
publication
that
goes
along
with
these.
That
sort
of
gives
you
a
once
over
lightly,
so
next
to
like
appendix
notebooks.
If
you
want
to
really
see
how
we
did
the
derivation
of
the
sensitivity
as
well
as
how
we
tested
the
code,
those
notebooks
are
here
as
well,
so
they're
in
a
lot
more
sort
of
gory
detail,
which
is
why
they're
my
new
ten
years
and
then
there's
also
there's
an
example
of
a
so
that
same
sounding
example
that
we
did.
G
E
G
In
this
case,
yeah
exactly
absolutely
really
so
simple
when
you
actually
so
this
is
the
main
website
for
senpai,
but
if
you
actually
go
and
look
at
the
code,
it's
all
on
github,
and
so
this
is
all
version
controlled,
it's
very
well
tested.
So
anytime
somebody
pushes
a
change
to
the
code.
Yeah
there's
a
two
hours
of
tests
that
are
running
touching
I.
Think,
like
80%
of
the
code
base
right
now,
the
only.
C
G
G
D
G
We've
had
a
number
of
contributors
from
so
mainly
at
UBC.
There's
been
a
number
of
contributors
and
users
from
Calgary
from
Colorado
School
of
Mines,
and
it
is
yeah
we're
starting
to
grow
and
we've
been
in
touch
with
other
open-source
software
package
developers.
So
there's
a
package
that
does
gravity
and
magnetics
and
called
fatty
and
oats.
So
we've
been
in
touch
with
Leo
who's,
the
core
developer
there
and
we're
working
on
trying
to
identify
some
core
pieces
and
work
together
on
those.
B
G
A
B
B
It's
developed
a
lot
and
that's
quite
powerful
and
easy
to
use
and
has
a
graphical
interfaces
without
images
easily.
Actually
it
has
its
own
strange
code
that
lets
you
develop
measures,
it's
sort
of
like
see,
sort
of
like
art
and
the
less
you
write
up
your
there's
geometry
and
you
can
also
import
from
other
3d
programs
into
G.
Mentioning
the
fine
I
would
match
dinner.
Imagine
it.
C
B
E
G
C
E
A
Yeah
it's
over
the
large-scale
airborne
problems,
especially
when
you've
got
tens
of
thousands
of
sources
really
want
to
have
three
e
inversion
can
cover.
Seven
teams
do
not
have
a
patch
things
in
the
approach
there.
That
seems
for
each
forward.
Mom,
you
just
can't
worry
of
a
different
sounding
location,
the
small-scale
forward,
longing
that
kind
of
works
forward
and
then
calculate
your
sensitivities
and
then
maybe
uses
a
multi-core
processors
so
that
you're
actually
solving
the
inverse
problem,
a
very
large
Nash,
but
each
individual.
D
This
thing
is
rounded:
we
have
one
meter,
so
this
thing
has
a
specific
length
or
two
dimension:
let's
think
two
dimensions
own,
and
so
this
acts
like
a
wave
guide
for
some
frequency.
This
has
a
particular
own
frequency
where,
for
which
X
as
a
wave,
that
isn't
that
correct
and
if
I
start
to
produce
this
oil
field,
the
shape
changes.
This
changes
and
I
get
a
different
frequency
for
which
this
is
a
wave
cut,
I,
so
I'm,
looking
yeah
I
I
love
resonance
yeah
with
my
button.
Also.