Induced Polarization (IP) - storage.googleapis.com · Induced Polarization n Current injected into ground and the voltage continues to increase. n Recognized in 1950’s: it was termed

Slide 1

Induced Polarization (IP)

n Basic principles

n Data Acquistion

n Pseudosection

n Inversion

n Case Histories

Induced Polarization

n Current injected into ground and the voltage continues to increase.

n Recognized in 1950’s: it was termed Over-voltage. n Understand the effect in terms of charge accumulation.n The phenomenon is called induced polarization.

I source V potentialNot chargeable Chargeable

Source(Amps)

Potential(Volts)

Chargeability is a microscopic phenomenon

Thoroughly understanding what is happening at the microscopic levelis scientifically challenging. In practice we work with the concept of “chargeability”

Chargeability

pyritechalcocite

copper

graphite

chalcopyrite

bornite

galena

magnetitemalachite

hematite

13.4 ms

13.3 ms

12.3 ms

11.2 ms

9.4 ms

6.3 ms

3.7 ms2.2 ms

0.2 ms0.0 ms

Minerals at 1% Concentration in Samples

Chargeability: rocks and minerals

Earth materials are “chargeable”

Initial situationNeutrality

Apply an electric fieldBuild up of charges

Net effectCharge PolarizationElectric dipole

Induced Polarization: Over-voltage

Not chargeable Chargeable

Source(Amps)

Potential(Volts)

Chargeability Data: Time domain IP

Intrinsic chargeability0<n<1 (dimensionless)

Integrate over the decay

Sample a channel

(msec) mV/V

IP data: frequency domainn Percent frequency effect:

n Phase:

low freq. f2high freq. f1

Sourcecurrent

Measuredpotential

V1 V2

I I

⎟⎟⎠

⎞⎜⎜⎝

⎛ −=

1

12100a

aaPFEρρρ

Sourcecurrent

Measuredpotential

Phase (mrad)

phase (mrad)

Data acquisition

n Data are acquired along with DC resistivity data (just sample a different part of the waveform)

n Data are plotted as pseudosections (exactly the same as DC resistivity)

n For IP the data plotted in the pseudosections will have units (mV/V, msec, mrad, PFE).

Earth

Energy Source In Measured signalsOut = “Data”

Plotting plane

~ v

Plotting plane

~ v

Plotting plane

~ v

Plotting plane

~ v

IGV

aΔ

=π

ρ2

Each data point is an apparent resistivity:

~ v~v

(Click for animation)

DC resistivity and IP data

Example IP pseudosection

2) A chargeable block.

2) A chargeable block and geologic noise.

Example IP pseudosection

3) The “UBC-GIF model”

Example IP pseudosection

Pseudosections … conclusions

n Except for very simple structures, geologic interpretations can not be clearly made directly from pseudosections.

n Interpretation is even more difficult in 3D

Given:- Field observations- Error estimates- Ability to forward model- Prior knowledge

Choose a suitable misfit criterion

Design modelobjective function

Discretize the Earth

Perform inversion

Evaluate results Iterate

Interpret preferred model(s)

Summary: what is needed to invert a data set?

( )∫ −= dxmmsm2

0αφ

∫ ⎟⎠

⎞⎜⎝

⎛ −+ dxmmdxd

x

2

0 )(α

∫ ⎟⎠

⎞⎜⎝

⎛ −+ dzmmdzd

z

2

0 )(α

Summary of IP data types:

n Time domain:q Theoretical chargeability (dimensionless).q Integrated decay time (msec).

n Frequency domain:q PFE (dimensionless)q Phase (mrad)

n For all data types, Jη = d .

n where J is a sensitivity matrix that requires that the electrical conductivity σ is known. We find σ by inverting the DC resisitivity data.

DC / IP datagathered together

Use σ model forforward mapping of

chargeability

IPData

Invert potentialsfor conductivity

model

Potential (i.e. voltage) dataConductivity model

Invert for chargeability models

Chargeability model

IP Inversion

Inversion of IP data

Step 1: Invert Vm to obtain σ.

Step 2: Generate sensitivities

Step 3: Invert the IP data (any form) by solving:

j

i

ijJ σφ

lnln

∂

∂−=

obsdJ =η subject to η > 0.

Example 1: buried prism.

Chargeability model

Data with 5% Gaussian noise

• Pole-dipole; n=1,8; a=10m; N=316; (αs, αx, αz)=(.001, 1.0, 1.0)

Recovered chargeability

Predicted data

Example 2: prism with geologic noise.

Chargeability model

Data with 5% Gaussian noise

• Pole-dipole; n=1,8; a=10m; N=316; (αs, αx, αz)=(.001, 1.0, 1.0)

Recovered chargeability

Predicted data

Example 3: UBC-GIF model.

Chargeability model Recovered chargeability

Data with 5% Gaussian noise Predicted data

• Pole-dipole; n=1,8; a=10m; N=316; (αs, αx, αz)=(.001, 1.0, 1.0)

Field Case History

n Cluny deposit, Australia

n 10 lines of DCIP data acquired

n Inversion carried out in 3D

Data set #1:

Apparent resistivity,dipole - pole.

Cluny: 3D resistivityn Eight survey linesn Two survey configurations.

Easting (m) Easting (m)

mS/m

10500 11500 12500

13000

14000

15000

16000

400

450

500

Easting (m)

North

ing

(m)

Surface topography:ElevationMeters

10 lines surveyed

Easting (m) Easting (m)

mS/m

Data set #2:

Apparent resistivity,pole - dipole.

Data set #1:

Apparent resistivity,dipole - pole.

Cluny: 3D resistivityn Eight survey linesn Two survey configurations.

Easting (m) Easting (m)

mS/m

10500 11500 12500

13000

14000

15000

16000

400

450

500

Easting (m)

North

ing

(m)

Surface topography:ElevationMeters

10 lines surveyed

Easting (m) Easting (m)

mS/m

Data set #2:

Apparent resistivity,pole - dipole.

Conductivity model from 3D inversion of DC

Conductivity model from 3D inversion of DC

Apparent chargeability,dipole - pole.

10500 11500 12500

13000

14000

15000

16000

400

450

500

Easting (m)

North

ing

(m)

Surface topography:ElevationMeters

10 lines surveyed

3D Induced polarization (IP)

Click image to see the AVI movie

Chargeability model from 3D inversion of IP

Click image to see the AVI movie

Chargeability model from 3D inversion of IP

Volume rendered resistivity model Volume rendered chargeability model

3D conductivity and chargeability models

Coming Up

n TBL DC resistivity and IP

n Quiz