Borehole Radar

Download Report

Transcript Borehole Radar

SAGA, September 2009
1
ADVANCES IN DRILL RIG
DEPLOYED RADARS
Mr Tim Sindle, ARCO/CRC Mining Imaging Lab, The University of Sydney
Dr Carina Kemp, Business Development Manager, GEOMOLE
11th SAGA Biennial Conference and Exhibition, 16-18 September 2009
Outline
2


Introduction
Method and Results
 Survey
Gear Minimisation
 Analyzing Drill Deployed Data
 Automatic Algorithm Development

Conclusions
SAGA, September 2009
Introduction – In-mine geophysics
3
The Good


Anticipate problems
ahead of mining
Improve efficiency of
mining operations
The Bad


Bulky gear
Time consuming
surveys cause delays
in production
No matter how good the results, if any technique cannot be
easily and reliably implemented in the mining environment,
it will not be used mainstream.
SAGA, September 2009
Introduction – Borehole Radar (BHR)?
4



Ground penetrating radar
(GPR) in a drillhole
Reflections indicate a
contrast in the electrical
properties of the rock.
BHR provides high detailed
continuous reflections from
lithology contacts and
structures.
GeoMole BHR



10 – 124 MHz Bandwidth
Resolution: less than1m
Range: up to 50m or more
(depending on rock type)

Probe diameter: 32 mm
BHR Profiling at ~10 m/min

Omnidirectional antenna

SAGA, September 2009
BHR then….
5

Survey trials of BHR
showed very promising
results, but the gear let
us down.
 50
kg optical fibre
winch
 20 kg push rods
 10 kg probes
SAGA, September 2009
BHR Now - Minimal Gear
6

Radar Tool
 1.6m
 3kg


each
Drill attachment
PDA
+
IQ
 2kg

+
Non-conductive spacers
 1.5m
PDA
Radar
SAGA, September 2009
Spacers
Drill
Attachment
BHR Now – Drill rig deployed
7
drill bit
Core barrel
spacers
SAGA, September 2009
IQ
Drill Rig
Deployed
Borehole
Radar
The radar tool continuously
records data.
The motion of the rods is
discontinuous as the rods
pulled and removed.
- Pumpdown
Spacers
Radar Tool
Deployment Motion…
OTR Survey
Moving
Depth
Measurement (Stationary)
Measurement (station)
Winch Survey
Moving
Stationary
Depth
Radar Data…
Winch Survey
OTR Survey
Moving
Same 40m section of a horizontal borehole
Stationary
Raw Data
11


Aim:
To understand
the motion in
order to work
out how to
recompress it.
Different motion
for each type
of drill-rig
Boart LM75 Diamond
SAGA, September 2009
Recompressing Radar Data..
Raw Data
Recompressed Data
Movement Log
Amplitude

Logging procedure
tracks accurately the
motion of the drill rig.
User records
‘MOVE’,
R
R
‘STOP’ and
‘ROD-CHANGE’
following the motion of
M S M SM S M SM
the drill.
Trace Number
These events are time
stamped and recorded Accelerometers were installed in
for data processing
the radars to assist with movement

logging
SAGA, September 2009
Accelerometer Data

Radar Data
13
Time Log processed Data
14

Vulnerable to
human error
SAGA, September 2009
Automatic Algorithm Development
15
Using the accelerometer data for automatic
processing:
 Statistical deviation measurement
 Fourier Spectrum Analysis
 Velocity integration calculations
SAGA, September 2009
Statistical Processed Data
16
0.2
0.1
Standard Deviation
Threshold 0
Amplitude
-0.1
-0.2
-0.3
-0.4
-0.5
-0.6
700
705
710
715
720
Traces
725
SAGA, September 2009
730
735
740
745
Accelerometer Processed Data
17

Suffers from
random
accelerometer
events
SAGA, September 2009
Fourier Spectrum Analysis
18



Examine the power in
various regions of
motion
Difference observed
between some moving
and stopped traces by
examining the higher
frequency content.
However, drill
vibrations cause wide
band energy gains.
METHOD ABANDONED
Frequency Spectrum
Single-Sided Amplitude Spectrum of y(t)
0.04
*
0.035
Stopped with drill shock
Start of move
Stopped
Constant velocity move
0.03
0.025
Amplitude

0.02
0.015
0.01
0.005
0
0
5
10
15
SAGA, September 2009
20
25
Frequency (Hz)
30
35
40
Velocity Processed Data
19

Noisy environment causes
spurious accelerations and
accurate velocity is hard to
gather.
 acceleration  velocity
Positive = Moving, Negative = Stopped


A high pass filter
distributes the
velocities around
zero.
Then the mean
representation of
the velocity is
calculated
Amplitude
1
0.5
0
-0.5
-1
700
705
710
715
720
725
Trace
SAGA, September 2009
730
735
740
745
750
Velocity Processed Data
20



Copes well with
the sharp drill
shocks and
vibrations as they
often have equal
positive and
negative
direction.
Captures the start
and stop of the
movement well.
Particularly
violent jerks can
cause a trace to
be lost.
SAGA, September 2009
Comparison…
21
Accelerometer
Raw
Time
Velocity
Data
Log
SAGA, September 2009
Conclusions…
22



Drill deployed radars can
be run with minimal
disruption to normal work
flow.
Using the time log alone
can be vulnerable to human
error
Yet all automated methods
investigated so far are
vulnerable to sharp
spurious drill movements.


A combination of a time log
together with statistical and
velocity methods will result
in smooth “winch quality”
images being produced.
Development in this area
continues
SAGA, September 2009
Conclusions
23


The ultimate aim of a tool knowing its own position
automatically is theoretically possible, but only
within well defined constraints, and there will
always be the unknown events on the drill rig that
can cause inaccuracies.
The above progress makes it possible for quick
data turnaround from survey to seamless integration
of BHR data into mine planning packages, to
enable day to day mining decisions to be made
using such tools.
SAGA, September 2009
Acknowledgements…
24



The authors would like to thank DeBeers Canada in
particular Kevin Smith, for their ongoing feedback
and use of the tool.
The funding contributions of ARCO, CRC Mining,
and GeoMole are gratefully acknowledged.
Many thanks to the tireless work by Sydney
University ARCO Lab members including; Andrew
Bray, Steven Owens, and Phillip Manning.
SAGA, September 2009