Department of Forest and Wood Science

1
Department of Forest and Wood Science
Tree Length Mechanised Harvesting System
Simulation
Glynn Hogg, Reino Pulkki Pierre Ackerman
2
(No Transcript)
3
Reasons for the study

• Mechanisation trend in SA.
• Introduction of multi-stem to SA benchmarking
  required.
• Low cost, high productivity drive for forest
  operations.
• Potential to apply Operations Research techniques
  to these issues.
• Simulation Most popular Operations Research
  tool.
• Forest operation simulation models been around
  internationally since late 1960s.
• Simulation studies done abroad, but none in SA.

4
Why Multi-stem?
5
Objectives of the study

• Assess how accurately a forest operation
  simulation model can represent reality within
  case study conditions.
• Gauge potential system balance, production and/or
  cost improvements achievable through simulated
  system adjustments, as well as identify
  beneficial equipment operating and application
  practices.
• Evaluate the usability of commercial simulation
  software in modelling forest operations.

6
Simulation defined
Experimentation with a model of a real world
system, given certain starting conditions, to
observe behaviour of the model and relate the
behaviour back to the real world system which the
model represents
Models are abstractions of reality
Real World System
Assumed Real World System
Model
7
Potential of Simulation
Simulated Production Frontier
C
Production Rate
B
A
Current Production Rate
Number of Machines
8
Simulation in the Corporate Sector
9
Advantages of Simulation

• Allows modelled study where direct
  experimentation would be costly, disruptive or
  impossible.
• Facilitates comparison between systems.
• Alternative working method comparisons.
• Simulated time compression.
• Experimental conditions can be tightly
  controlled.
• Identification of problems before they happen.
• System modelling shows effects of a single change
  on all parts of the system.
• Some systems are difficult to understand in their
  entirety due to the greatness of their scope
  until abstracted in a simulation model.
• Random number generators artificial
  observations.
• Study time reduced considerably over traditional
  study methodologies.

10
Shortcomings of Simulation

• Not an optimisation tool.
• Each model is specific to a certain system and a
  defined problem.
• Interpretation of simulation results requires a
  sound statistical background.
• Simulation is an experiment, meaning that it is
  not guaranteed to solve the defined problem.
• Model accuracy depends on model quality and input
  data accuracy.
• Data should be up to date and accurate, which is
  not always possible.
• Software can be expensive.
• Analyst needs to have good understanding of the
  system being simulated and the simulation
  software.

11
Methodology

• Conduct a case study of a multi-stem harvesting
  operation using Arena 9 commercial simulation
  software.
• Made up of 4 steps
• Time study (field work).
• Model construction in Arena software.
• Verification and validation using statistical
  analysis and tools.
• Multiple scenario studies within constructed
  model.

Tree-to-mill simulation.
12
Case Study System Matrix
13
Operating Conditions

• Location Kwambonambi, Zululand.
• Period January/February 2007 (summer).
• Terrain classification Flat terrain.
• Logistics Roadside loading.
• Work object Eucalyptus urophylla pulpwood.
• Maximum extraction distance 850m.

14
Step 1 Time Study
Time study observation periods
Total time study period 191.14hrs
(11,468.37mins).
All observations resulted in collected data
which greatly exceeded the required amount to
describe the respective means with a 95.45 level
of confidence and a margin of error within 5 of
the true mean
15
Step 2 Model Construction

• Distribution fitting
• Kolmogorov-Smirnov (K-S) test for continuous
  distributions.
• Chi-Square test was used for distributions
  describing integer data.
• Flowchart construction in Arena.

Travel Speed 0.579 0.181LN(dist) -
17.279(1/dist)
16
Step 3 Verification and Validation

• Verification (debugging).
• Arenas built-in error report function.
• Counters.
• Animation.
• Model was run for 2000 simulated hours no
  runtime errors.
• Validation (determining model accuracy).
• Model output data contrasted with real world
  systems outputs.
• All simulated outputs found to adequately
  represent reality.
• Difference of 0.85 between simulated and real
  system production incurred.
• Manual sensitivity analysis conducted by
  adjusting input times and evaluating outputs to
  ensure model robustness.

17
Step 4 Model Manipulation

• Alternative operating methods.
• Alternative systems.
• Requires
• Reworking of model inputs.
• Re-running of simulations.
• Model verification and validation.

Iterative process
18
Results

• Modelled system production differed from observed
  production by only 0.85 for the model of the
  real-world system.
• Simulated production increased by 31.1 as a
  result of simulated adjustments to operating
  techniques (no change to equipment
  configuration).
• System balance improved reduced machine waiting
  time.
• Cost reduction of 12.3 per unit of timber.

19
Operation adjustments

• System
• Refuelling and greasing of bottleneck equipment
  always during shift change.
• More, shorter operator rest breaks.
• Feller Buncher
• Feller buncher was least utilised unit more
  time to present bunches for skidder
• Skidder
• Log recovery grapple mounted on the skidders
  blade to overcome the problem of stopping while
  travelling loaded to pick up stems which were
  dropped in previous cycles.
• Blade size increased, meaning to less indexing
  cycles required.
• Bunches all collected at 90 to skidder travel
  direction increases payload.
• All processed stems indexed before the slasher
  crosscuts them, resulting in the slasher not
  having to move butts out of its path while
  travelling up and down the landing, as well as
  spending less time on butt alignment before
  crosscutting.
• Slash only collected from the same processor at
  which the skidder dumped the extracted bunch.

20
Operation adjustments

• Processors
• Time delay between felling and processing reduced
  by making the operation slightly hotter and
  adjusting feller buncher and skidder interaction,
  leading to lower bark adhesion and faster
  processing times.
• Slasher
• Increased buffer between processors and slasher.
• Less slasher movement (not moving from one
  processor pile to next).
• More consistent timber flow for the slasher.
• Less risk of trucks not being loaded if a machine
  other than the slasher breaks down.
• Less indexing time for skidder due to only one
  pile of timber.

21
Commercial Simulation in Forestry

• Good identification of bottleneck machines.
• Compressed time means many results in a short
  space of time.
• Expensive harvesting systems not affected during
  study.
• Positive simulated production and cost results
  from case study.
• Forestry works on far bigger areas than
  industrial facilities.
• Forestry machines not fully automated the human
  factor.
• Many variables in forestry operations
  (operational and site considerations) complex
  models.
• Forest operations are mobile (including both
  within-stand moves and between-stand moves).
• Machine movements have to follow specific,
  sometimes unconventional rules.
• Simulation works off a snapshot of operating
  conditions extrapolation of results often not
  possible.
• Time studies a slow, manual process in forest
  operations.

22
Conclusions

• Commercial simulation software used is an
  acceptable tool for modelling forest operations.
  
• Positive simulated results in terms of cost and
  production obtained as a result of system
  manipulation.
• Results are specific to the conditions under
  which the study took place.
• Taking real world into simulation is acceptable.
  
• Taking simulation into real world room for
  future study?

23
Thank You