Autonomous Haulage Trucks - CERM3

Report
Autonomous Haulage Trucks
- the new way to mine
John A. Meech
University of British Columbia
The Norman B. Keevil Institute for Mining Engineering
The Centre for Environmental Research in Minerals, Metals, and Materials
Vancouver, British Columbia, V6T 1Z4
"…only 10-15% of mine sites currently 'leverage
technology well'...what we’re moving...toward
...with...autonomy is a factory-type environment,
...and that’s going to require...a more clinical and
more managed environment.
If you look at a [fully-automated] factory,...they are
very sterile, very structured environments. If you
have robots operating ...the floor can’t be dirty,
it can’t be scattered with empty boxes... "
- Carl Hendricks, CAT Mining Solutions Region Manager for Australia
"A haul road in a modern mine running autonomy
…[has] ...the same issues. You can’t have poorly
constructed… roads ...Some [automation] ...we
use is sensitive to dust ...[which will]...cause the
vehicle to sense...an obstacle...that really isn’t there.
[T]hat’s going to hinder operation of the machine.
There... [must]...be a new level of discipline in how
we maintain that environment – just like a factory.
These are...things...we should be doing,...the sorts
of the things...modern processing plants do. They
maintain their environment and operate with rigour."
- Carl Hendricks, CAT Mining Solutions Region Manager for Australia
Outline
•
•
•
•
•
•
•
Automation and Sustainability
The "New" Mining Engineer
Autonomous Open Pit Haulage Systems (AHS)
Who is in the Game?
Goals (safety, fuel use, tire wear, productivity)
ETF Trucks
Modeling AHS
Motivation behind Automation







Safety (removal of people from danger)
Lack of skilled personnel (training costs)
Loss of equipment and availability
Decreased energy use (fuel savings)
Decreased wear, maintenance, and replacement
Increased productivity
More consistent operations
Mining Truck
Accidents
Somewhere in the World every
year at least two to three truck
drivers are killed because of an
accident from human error.
Automation of Mining Operations
 Batch versus Continuous (start-up/shut-down more relevant)
 Disturbances (environment, nature, human)
 Maintenance Issues
 Key Performance Indicators
 Supervisory Control vs. Autonomous Control
 Data and Process Integration ("Big" Data)
Batch Processes in Mining
Drilling
Loading
Explosives
Blasting
Digging
Loading
Ore/Waste
Maintenance
Hauling
Ore/Waste
Returning
Empty
Dumping
Ore/Waste
Disturbances







Weather (rain, snow, mud, dust, wind, heat)
Equipment failures (breakdowns, accidents)
Maintenance issues (scheduled vs. failure)
Ground conditions (diggability, sticky ore, rocks)
Geology (hardness, rock size, mineralogy)
Human breaks (shift changes, coffee/lunch)
Driving behaviour (passive, normal, aggressive)
The "New" Mining Engineer
Fuzzy Logic Control of a 1/10 scale Autonomous Vehicle
http://www.jmeech.mining.ubc.ca/Mine432/eight_track.wmv
The "New" Mining Engineer
• Understands the role of Automation
– Enhance workplace safety
– Reduce fuel use, GHG emissions, and tire wear
– Increase life cycle time of mining equipment
– Stabilize and Optimize (improve)
– Improve production and productivity
– Reduce haulage costs appreciable
DARPA Grand and Urban Challenges
• 2004 and 2005 –
Mojave Desert
• 2007 - Victorville
• Vehicles drove
autonomously
http://www.jmeech.mining.ubc.ca/Mine432/DPG_highlights1.wmv
Komatsu's AHS at the Gaby Mine, Chile
www.youtube.com/watch?v=fsuRTvK3Nik
Basic Requirements
• Localization
– where am I?
• Navigation
– where do I want to go?
• Obstacle Avoidance
– what is in my path?
• Condition Monitoring – how is my health?
Komatsu – VHMS >>> KOMTRAX
Komatsu/Modular Mining Approach
IEEE 802.11 n , ac , ad WLAN
Computer hardware on-board
Central data processing system
Supervisory Software
Modular
Mining
Systems
Front Runner
Modular Mining’s DISPATCH®
fleet management system and
MASTERLINK® communication system
Caterpillar's Approach
IEEE 802.11 n , ac , ad WLAN
Computer hardware on-board
Central data processing system
Supervisory Software
COMMAND
for hauling
CAT’s MineStar System
Sensors – Localization and Navigation
IEEE 802.11 Communication network
Sensors for
• Navigation
>>> GPS and Radar
• Object- Avoidance
GPS accurate to
10 cm (D-GPS)
Sensors – Object Recognition and Avoidance
IEEE 802.11 Communication network
Sensors for
• Navigation
• Object- Avoidance
>>> Radar and LIDAR
Radar range to 80 m
• Front
LIDAR range to 20 m
• Sides and Rear
mm-wave Radar Obstacle Detection System
Sensors – Object Recognition and Avoidance
IEEE 802.11 Communication network
Sensors for
• Navigation
• Object- Avoidance
>>> Radar and LIDAR
Radar range to 80 m
• Front
LIDAR range to 20 m
• Sides and Rear
IBEO and SICK scanning laser instruments
Sensors – Object Recognition and Avoidance
IEEE 802.11 Communication network
Sensors for
• Navigation
• Object- Avoidance
>>> Radar and LIDAR
Radar range to 80 m
• Front
LIDAR range to 20 m
• Sides and Rear
CAT’s Radar and LIDAR-based
Obstacle Detection System
Obstacle Detection - System Reliability
Reliability
Actually
Present
Not
Present
Detected
100
1-5
Nothing
Detected
0
95-99
^
^
Goal
<<
Measure
of Success
Elements of an Autonomous Haulage System
Additional Sensors
•
•
•
•
•
•
Wheel Speed
Steering Angle
Road Edge Guidance Lasers
Payload Monitoring
Tire Temperatures (embedded in tread)
Status Lights
Requirements for Success
• A Project Champion is essential at the highest levels
• Long-term organization commitment based on benefits
• Significant workplace culture change is needed
• Revolutionary vs. Evolutionary
• Small steps better than one large step
• Develop new core competencies first
• Engage with workforce personnel
• Replace labour by attrition and promotion
• Build-in system redundancies
Implementation of a Successful Project
Manual
Moonshot
KPI
(core)
0
% Autonomous
Robotic
KPI
(core)
100
Implementation of a Successful Project
Manual
‘Baby’ steps
KPI
(core)
0
% Autonomous
Robotic
KPI
(core)
100
Implementation of a Successful Project
Manual
Robotic
KPI
(core)
KPI
(core)
?
0
% Autonomous
100
Implementation of a Successful Project
• KPIs may decrease initially until full adaptation
• Which plan is best?
1. Replace MHS with AHS in one step – no interaction
2. Isolate AHS from MHS : Separate routes, staged introduction
3. Integrate AHS with MHS: Significant safety concerns
• Safety concerns require careful design and planning
• Is a back-up or fall-back system necessary or desired?
Develop Core Competencies
•
•
•
•
•
•
•
•
Process Control fundamentals
Understanding control stability
Supervisory control hierarchies
Software algorithms
Artificial Intelligence methods
Managing large databases
Sensor knowledge and maintenance
Remote operation of equipment
Change Management Requirements
• Mine Personnel Issues
– Truck Drivers >>> Hardware/Software Maintenance
– Introduce AHS with all affected personnel involved
– Humans in-the-loop must be accounted for
• Machine Issues
– Monitoring health of sensors on regular basis
– Soft-sensors to confirm operational effectiveness
– Data Collection to integrate into planning/scheduling
Change Management Requirements
• Mine Management Issues
– Must be on-side with all decisions about the changes
– New safety/traffic rules required (some are positive)
– More maintenance / less operational activities
– Drilling and Blasting practices must change
• Headquarter Issues
– Move to Central Control must be done with care
– Initial focus on integrating massive data collections
– Decisions must support local mine site personnel
Who is in the Game?
Mines using AHS
Codelco
Radomiro Tomic, Chile
Komatsu Cu
2005
Codelco
Gabriela Mistral, Chile
Komatsu Cu
2008
Rio Tinto
West Angelas, Australia
Komatsu Fe
2010
BHP-Billiton Navajo Coal, NM, USA
CAT
coal 2012
BHP-Billiton Jimblebar mine, Australia CAT
Fe
2013
Fortescue
Solomon mine, Australia
CAT
Fe
2013
Stanwell
Meandu mine, Australia
Hitachi
coal 2014
Komatsu – Codelco
Radomiro Tomic mine - 2004
Komatsu – Codelco
Radomiro Tomic mine - 2005
Komatsu – Codelco
•
•
•
•
•
•
•
In 2006: 5 AHS 930E trucks; 32,000 tpd; 256 days
Mechanical Availability:
> 90%
Cost per tonne reduced from $1.36 to $0.50
Est. maintenance reduction:
7%
Est. depreciation reduction:
3%
Gaby mine AHS trucks:
2008 – 11 2012 – 18
Safety issues (accidents): 2006 – 0 2007 – 2
Komatsu – Codelco
•
•
•
•
AHS trucks operate in an "electronic bubble"
Each truck is aware of all other machines on site
Unknown machine in AHS area causes shutdown
Navigation is a hybrid of
– High-precision GPS, and
– Dead-reckoning IMU (accelerometers/gyroscopes)
Komatsu – Codelco
• Change how mine operations are planned & implemented
• Must consider all vehicles, not only AHS trucks
• Complexity increases exponentially with number of trucks
"There are hardware restrictions...Information
exchanged between trucks and central control is
enormous. At Gaby, 11 trucks and 30 pieces of
equipment...limit...information transfer."
– Jeffery Dawes, Komatsu Chile
Komatsu – Rio Tinto
• Rio’s “Mine of the Future” concept
• Began in 2008 at West Angelas Mine, Australia
• First 24 months
– 42,000,000 tonnes
– 145,000 cycles (290 t)
– Short haul distance ~1.5 km
• 5 trucks – 25 min. cycles
• Ave. Velocities (initial trial):
– Loaded = 7-10 kph
– Empty = 14-18 kph
Rio Tinto's Mine of the Future –
the Future is NOW!
Caterpillar - BHP-Billiton
• Joint venture at 2 mines since 2007
– Mt. Keith Nickel Mine in Australia
– Navajo Coal Mine in New Mexico
•
•
•
•
Initial 2 truck trial in Arizona and at Mt. Keith - 2010
Planned a staged implementation from 5 > 55 > 150
Plan was adjusted after 2008 Financial Crisis
Planning an Integrated Remote Operations Centre
(IROC) in Perth to schedule/plan/control Pilbara mines
http://milltongroup.blogspot.ca/2011/09/bhp-plans-autonomous-mining-operation.html
Caterpillar - Fortescue
• MOU with Fortesque and WesTrac in 2011
– Solomon Iron Ore Mine in Australia
• CAT MineStarTM system & Command for hauling
• Initial fleet - 12 AHS 793F trucks – 2012
• At full capacity, 45 AHS trucks by 2015
http://www.miningmagazine.com/reports/cat-signs-haul-truck-deal-with-fmg
European Truck Factory – the next step?
• Decoupling Maintenance from Operation
• 95% Mechanical Availability
http://www.etftrucks.eu
ETF MT-240 Truck - Haul Trains
ETF MT-240 Truck on Empty Haul
ETF MT-240 Truck Turning Circle
ETF MT-240 – Oscillating Axle Advantage
ETF MT-240 – Stability on Rough Roads
ETF MT-240 – Simultaneous Tipping
ETF MT-240 – Engine Change-out (15 min.)
ETF MT-240 – Axle swap-out (45 min.)
ETF MT-240 – tire change (15 min.)
ETF MT-240 – decreased tire scrubbing
Conventional 240 t truck
ETF 240 t truck
ETF MT-240 –tire size comparison
ETF Haulage Truck Capacities
Modeling Open Pit Haulage Operations
Key Performance Indicators – KPIs
–
–
–
–
–
–
–
–
–
Production
per day
per truck per month
Fuel Use
per hour
per km
per tonne
Tire Wear
per hr
per km
per tonne
Cycle Time
increased/decreased
Truck Speeds
Increased/decreased
Cycles per day
increased
%Mechanical availability
increased
%Utilization
increased
O&M Costs
decreased
Batch or Discrete Process - Loading
(a State)
(an Event)
(a State)
Shovel
Waiting
Shovel begins
to Load
Shovel
Waiting
Truck
Driving to Spot
(a State)
Truck stops
(an Event)
Truck being
Loaded
(a State)
Truck starts
(an Event)
Truck
(a State)
Driving from Spot
after: John Sowa, 2001. Processes and Causality, www.jfsowa.com/ontology/causal.htm
Components of the Model
•
•
•
•
•
•
•
•
Driver Behaviour module (fuzzy)
Road Conditions module (fuzzy)
Fuel Consumption module (deterministic)
Tire Wear module (thermodynamic/fuzzy)
Truck Movement module (deterministic)
Lateral Displacement module (probabilistic)
Loading, Dumping, Queuing module (stochastic)
Maintenance and delays module (stochastic)
Road Segment Characteristics
Route
Waste Shovel
to Dump
(5.203 km)
Dump
to Parking
(4.330km)
Ore Shovel
to Crusher
(6.041 km)
Crusher to Parking
(1.427 km)
Segment
Length (m)
Grade (%)
Speed Limit
(km/h)
Maximum
Acceleration
(m/s²)
Stop at end?
1
2
3
4
5
6
7
8
9
9
8
7
10
11
12
13
14
15
2
3
16
17
12
13
18
19
19
18
14
98
537
680
761
44
1,502
500
944
137
137
944
500
1,201
300
353
554
359
78
537
680
1,904
867
353
554
968
100
100
968
359
0
5
7
10
5
10
2
5
0
0
-5
-2
0
0
0
0
0
0
5
7
10
10
5
0
0
0
0
0
0
25
40
40
40
40
40
40
60
25
25
60
40
40
40
40
40
40
25
40
40
40
40
40
40
60
25
25
60
40
0.42
0.42
0.21
0.21
0.42
0.21
0.42
0.42
0.42
0.42
0.42
0.42
0.42
0.42
0.42
0.42
0.42
0.42
0.42
0.21
0.21
0.21
0.42
0.42
0.42
0.42
0.42
0.42
0.42
no
no
yes
no
no
yes
yes
no
yes
no
yes
yes
yes
yes
yes
yes
yes
no
no
yes
yes
yes
yes
yes
no
yes
no
yes
yes
Equipment
• 9 CAT 793D haulage trucks (7-10)
• Initially 3 assigned to waste
• Initially 6 assigned to ore
• Command system reschedules to address
– Stripping ratio requirements (set to 0.5)
– Queuing delays due to maintenance
• 2 shovels (one digging waste / one digging ore)
• Auxiliary equipment (grader, water truck, dozers)
Truck Movement
ExtendSIM Software
• Dynamic modeling of real-world processes
• Uses building blocks to explore processing steps
• Benefits
• Easy to use
• Inexpensive
• MS-Windows environment
• Handles both Discrete and Deterministic Models
Discrete and Deterministic
• Discrete Events
• Probabilistic method
• Maintenance, Loading, Dumping
• Deterministic (100 msec)
• First Principles
• Truck movement
– Fuel consumption
– Tire temperature
• Fuzzy Models (A.I.)
• Road conditions (rolling resistance and traction)
• Tire wear
• Driver behaviour (velocity, acceleration, reaction time)
Truck Movement
• Model accounts for all forces that affect motion
• Rolling resistance and traction
• Grade and Gravity
• Aerodynamic forces (wind)
• Motive power from the engine
Truck Movement
• Acceleration and velocity are limited by road conditions
• Significant influence of the driver behaviour
• Wind effects are included to account for head-winds,
following winds, and side-winds (significant above 25 kph)
Truck Movement - Rimpull
Truck Movement - Retarding
Comparison of Manual vs. Autonomous
Comparison of Driver Behaviour Output
Economic Comparison
– 9-truck manual fleet and 7-truck AHS fleet
(AHS set to same production as manual)
– 9-truck manual fleet and 9-truck AHS fleet
(different production for both cases);
Significant Assumptions
1. AHS infrastructure costs = $6,690,000
2. AHS Start-up costs = $500,000
3. AHS incremental capital costs = $1,000,000
Incremental Economic Analysis
∆CF = (1 – t) (∆Opex) + t∆D
where ∆CF
t
∆Opex
∆D
= cash flow difference between two projects
= Tax Rate of 50% (conservative analysis)
= operating cost difference
= depreciation difference (straight-line)
Net Present Value is given by:
NPV(@ i) = [∆CF/(1+i) y] – ∆Capex + ∆S/(1+i) d
where
i
y
∆Capex
∆S
= interest rate of 10%;
= year and
d = Project life (years);
= capital cost difference
= salvage value difference at end of life
Assumptions
Initial tread depth
Ave Time to Scrap a tire (hours)
Maintenance costs ($/h)
Operators per truck
Labour per AHS
Turnover = 35%
Truck Depreciation
Purchase price to site (Manual)
Purchase price to site (AHS)
= 97 mm
= 5,300 hrs
= 130
= 4.2
= 0.45
= 7.0 yrs
= $4,000,000
= $5,000,000
Source: mine (shop visit)
Source: mine (shop visit)
Source:mine report
Including vacation/training
Source: mine report
Source: mine (visit)
Assumptions – common costs
Village Cost (fly-in/fly-out)
Flight Cost
Tire Cost
Fuel Price per L
Training cost (simulator)
Mining
Quarterly wage
Labour costs - HR overheads
Hiring Cost
US$/person/night
US$/person/flight
US$/tire
US$L delivered
US$ Real Qtr
US$/t
US$/person/Qtr
% of wage
US$/new starter
62.73
169.86
33,000
0.90
25,000
2.30
30,000
15%
3,200
Assumptions – AHS infrastructure
Infrastructure Telecom / IT
Basic transmission station
Servers (with redundancy)
Routers (24-ports/PoE)
Switches
Energy System (with Redundancy)
Network Adaptation (Cables CAT 6)
Monitoring System (Camera, SW specific, etc.)
Positioning System with redundancy (DGPS, antennas, etc.)
Quant.
30
8
10
20
1
1
1
1
Unit $
$30,000
$12,500
$40,000
$5,000
$150,000
$200,000
$1,500,000
$200,000
Subtotal
Total
$900,000
$100,000
$400,000
$100,000
$150,000
$200,000
$1,500,000
$200,000
$3,550,000
1
4
2
2
20
1
$700,000
$180,000
$100,000
$10,000
$50,000
$500,000
Subtotal
$700,000
$720,000
$200,000
$20,000
$1,000,000
$500,000
$3,140,000
Total
$6,690,000
Services
Installation and Commissioning
Consulting (12 months)
Project Manager (6 months)
Transmission Link
Training
Transport/logistics
AHS Trucks to match Manual Production
(37,867 tpd)
Element
Manual
(9)
7 AHS
8 AHS
9 AHS
V full - km/h
17.4
18.5
13.4
V empty - km/h
27.1
28.6
Ave Fuel Use - L/t
0.83
Ave Fuel Use - L/cycle
% Change (AHS-Man)
7 AHS
8 AHS
9 AHS
12.7
6.2
-23.1
-26.8
23.4
17.2
5.7
-13.6
-36.4
0.76
0.83
0.78
-8.1
-0.1
-5.1
185.3
168.4
209.3
223.9
-9.1
12.9
20.9
Ave Fuel Use - L/hour
218
235.5
252
239.4
8.1
15.6
9.9
Ave tire wear - mm/cycle
0.015
0.014
0.014
0.013
-5.0
-7.8
-13.8
Ave tire time to scrap - hrs
5,504
4,876
5,834
6,029
-11.4
6.0
9.6
Ave. Number of Cycle/day
18.9
24.5
21.1
19
29.4
11.9
0.4
Ave. Total Cycle Time (min)
51
42.9
49.8
56.1
-15.9
-2.3
10.0
Ave. Queuing (min)/cycle
1.8
0.9
0.9
1.0
-49.4
-50.8
-47.8
Percent Utilization (%)
65
78
73
74
19.4
12.7
14.1
AHS Trucks to match Manual Production
(37,867 tpd)
Manual
9 trucks
7 AHS
8 AHS
9 AHS
7 AHS vs.
Manual
8 AHS vs.
Manual
9 AHS vs.
Manual
CAPEX (M$)*
$36.00
$42.19
$47.19
$52.19
-
-
-
OPEX (M$/year)
$50.17
$44.63
$46.08
$47.11
-
-
-
-
-
-
-
$6.19
-$5.54
$0.88
$3.21
$0.00
$9.45
48.7%
$11.19
-$4.09
$0.76
$2.42
$0.65
$1.26
11.7%
$16.19
-$3.06
$0.66
$1.86
$5.23
-$1.92
-5.2%
Element
∆CC (M$)
∆OC (M$/year)
∆D - Depreciation (M$/year)
∆CF (M$/year)
∆SV (M$) - Salvage Value @ DCFROR
After Tax [email protected]%
After Tax DCFROR
AHS Trucks running at Default Speeds
Element
Manual*
7AHS
8 AHS
9AHS
Ave Fuel - L/tonnes
0.83
0.78
0.78
0.78
Ave Fuel - L/cycle
185.27
172.53
172.91
172.89
Ave tire - mm/cycle
0.015
0.014
0.014
0.014
Fuel burn rate (L/h)
218
236
252
239
Tire life (hours)
5504
4876
5834
7029
% Utilization
65%
78%
73%
74%
Maintenance (%)
4.0%
3.4%
4.3%
4.9%
Annual Material Moved (t)
13,821,605
13,821,605
16,185,560
18,156,560
Years of Mining
7
7
5.98
5.33
AHS Trucks running at Default Speeds
Manual
9 trucks
7 AHS
8 AHS
9 AHS
CAPEX (M$)*
$36.00
$42.19
$47.19
$52.19
-
-
-
OPEX (M$/year)
$50.17
$44.63
$51.51
$57.08
-
-
-
∆CC (M$)
-
-
-
-
$6.19
$11.19
$16.19
∆OC (M$/year)
-
-
-
-
-$5.54
$1.34
$6.91
∆D =Depreciation (M$/yr)
-
-
-
-
$0.88
$2.72
$4.65
∆CF (M$/year)
-
-
-
-
$3.21
$0.69
-$1.13
∆SV (M$) = Salvage Value
-
-
-
-
$0.00
$0.00
$0.00
After Tax [email protected]%
-
-
-
-
$9.45
$3.36
$15.59
After Tax DCFROR
-
-
-
-
48.7%
14.8%
9.5%
Element
7 AHS vs. 8 AHS vs. 9 AHS vs.
Manual
Manual
Manual
Trend due to advancement of revenue from later years into early years
Comparison of KPIs
AHS
AHS
AHS
+ 20%
AHS
AHS
+22%
-5%
-8%
-7%
Manual
Manual
Manual
Manual
Manual
Investment cost
per truck
Truck haulage
speeds
- 7 - 14%
Fuel
consumption
- 5 - 10%
Mechanical
Availability
+15 - 25%
Tire Wear
+20 - 30%
- 5 - 15%
Comparison of KPIs
AHS
AHS
AHS
+ 22%
+10%
AHS
-15%
-8%
-80%
Manual
Manual
Manual
Increased
Productivity
Maintenance
costs
Increased
Truck life
Manual
Labour costs*
* Labour savings depend on current mine circumstances – union and turnover issues
Conclusion
• “New” miners of the 21st Century have new skills
– Environment
– Socio-political
– Automation
•
•
•
•
Mining industry is a new frontier for automation
AHS is the first step in Open Pits
Komatsu & CAT are well "down the road" to robotic mining
Mine site management change is an important element
Conclusion
• Modeling a mine is useful in determining change benefits
• AI methods can play a major role in developing these models
• Benefits of AHS
–
–
–
–
–
Production/productivity
Fuel Consumption
Tire Wear Rate
%Utilization
Maintenance
+ 15 to 20 %
– 10 to 15 %
– 5 to 15 %
+ 10 to 20 %
– 8%

similar documents