Applied Simulation

Applied Simulation: Modeling and Analysis using FlexSim

consumer products company that if they would simply change the transfer logic between
their batch operations and filling line, they could gain a needed 15% increase in production
without any capital outlay. Similarly, a shipbuilder gained 10% to 20% in throughput in a
bottleneck shop by re-sequencing work.

Section 1-6 When and where to use simulation

Before learning how to build simulation models and analyze their results, it is important to
understand what models are and how they can be used. This section illustrates many types
of operational problems that simulation can, and has addressed. The “Baker’s Dozen” of
examples are certainly not exhaustive, but are based on actual experience and demonstrate
the versatility and power of simulation. Each issue includes a screenshot of a simulation
model used to address the issue.

Issue 1: Understanding the bottlenecks that limit an operations system’s performance.

Every system has a bottleneck that restricts
how well the system operates. Some
companies base their performance calcu-
lations on the operation downstream of
the bottleneck. Bottlenecks in processing
operations are often identified and known;
however, they are often transient and
change as the system dynamics change.
There are also many ways to alleviate a
bottleneck. In manufacturing, equipment Figure 1.20: Bottlenecks can appear anywhere.
from another production line can be used.
In a supermarket, a new checkout line can be opened. At times there is no single bottleneck
to focus on. Whenever a step in a process stops, it becomes the bottleneck. For a restaurant,
the bottlenecks may change according to the seasons. Dynamically changing bottlenecks
makes a system very difficult, if not impossible, to analyze with simple spreadsheet methods.
The dynamic focus of simulation makes it an excellent tool for bottleneck analysis.

Issue 2: Addressing the accumulation of material in a manufacturing process.

Modern manufacturing theory focuses on
the elimination of any unnecessary steps
or accumulation of work materials in a
process. Surge areas, which are used to store
materials ahead of equipment or worksta-
tions, are targets of Lean thinking; however,
Lean operations do not necessarily mean
the eliminations of all surge areas. Point-
of-use storage is still an integral part of
modern, efficient operations.
Figure 1.21: Accumulation surge before a filler.

36
 Chapter 1

The sizing of surge areas and buffers are not intuitively obvious as they are based on the
dynamic behavior of the operation. Simulation is used to identify where the surges are
needed and what size they should be so that a minimal amount of inventory is accumulated
in a production area.

Issue 3: Analyzing new concepts.

Simulation is often used to analyze the
impact of modified or new operating
practices or policies. Testing new concepts
with simulation is much faster and more
economical than trying to do it on the actual
system. The im-pact of reliability assump-
tions on equipment can be tested, as well
as the new methods. If there are absolutely
no historical data to use in the simulation,
then the simulation can be used to conduct
a sensitivity analysis in which variables are
modified in order to study that impact on Figure 1.22: Evaluating alternatives before
system performance. The results are used implementation.
to specify the operational characteristics of
the equipment or used to estimate changes in operation.

Issue 4: Studying combined processing and packaging systems.

Systems that both process and package
goods are prime candidates for simulation
analysis. A major issue is that the process
and packaging areas are closely coupled,
making them highly dependent on each
other. The complexity is increased further
if the processing is fluid or batch in nature
and the packaging line is discrete. A simple
continuous blending system for paint
that goes to a packaging operation is an
example of a closely coupled system. Once
Figure 1.23: Paint blending and filling opera- the operation commences, disruptions at
tions are closely coupled.
any point can cause significant production
disruptions.

37
Applied Simulation: Modeling and Analysis using FlexSim

Issue 5: Balancing cost and service level.

At an airport security check point, travelers who have
their tickets and identification verified may be sent to one
of many inspection lines depending on several factors:
ticket status, line length, and need for special screening.
Simulation is usually used to design and analyze these
systems. The simulations help determine the number
of lines, amount of equipment, and staffing levels that
should be used throughout the hours of operation.
Figure 1.24: Improving airport This determination must balance customer service
security system operations. with cost of operations. Simulation is used to evaluate
the efficiency of such proposed concepts as frequent flyer
lines and pre-screening background checks for certain types of travelers.

Issue 6: Validating complex material transfer procedures.

In an effort to make operations very flexible, material
is often routed from one operation to another based
on various rules. In a cheese plant, material is passed
from fermentation units to processing areas based on
the history of the receiving unit. These rules have to
take into account the nature and age of the material
in each location and the cleaning record of the
receiving vessel. Simulation of such operating rules
is used for training as well as ensuring that the rules
do not become unintentional bottlenecks.
Figure 1.25: Multiple transfer paths.
Issue 7: Checking batch operating sequences.
Batch operations contain a sequence of events that often change with the product. In fluid
operations, the batch is usually described in a way prescribed by industry standards. The
description starts with a recipe table listing the amount of each material to be added and a
separate table lists the steps in the sequence. The use of a recipe and step table has become
the standard for specifying batch operations. For multiple products, separate recipe and step
tables have to be maintained. Logic is employed to handle any additional cleaning required
and to direct the output of the batch to its destination. Simulation is needed to analyze the
number of batches that can be made especially if some of the steps can be subject to random
events, such as the requirement of an operator or delays from test results.

38
 Chapter 1

Figure 1.26: Batch specification tables.

Issue 8: Analyzing the impact of production schedules on operations.

Experience has shown that the greatest impact on resources and performance is the sched-
uling of products or people through the system. In manufacturing, a production facility is
often at the mercy of a marketing or sales organization that dictates the order of material
flow. Simulations that only focus on producing a single item or a set sequence of items are
not robust enough to analyze performance over an extended period of time.
Additionally, when simulating a schedule,
operations usually have to be adjusted for
each item going through the system, and
monitoring of the line is needed in order
to know when the scheduled task has been
completed.
Operations where more than one area
that can be scheduled and the separate
production areas can interact can be
complex to analyze. Similarly, complexity
occurs if the operations must be carried out Figure 1.27: Two independent production lines
in a particular sequence without exception sharing storage and packing equipment.
and shared equipment is involved. In both
these cases simulation provides a robust method that takes complexities into account.
At times the steps required to clean or set up an operation can be more complex than the
operation itself. Food processing is an example of this as it has extensive requirements for
sanitation and allergy cleaning that depend on the particular product as well as the elapsed
time of operation. In addition to food processing, hospital requirements for sanitation and
cleaning are also extensive. In all these cases, the time required to perform the activities
varies. Simulation takes these dynamics into account and allows performance metrics to be
accurately calculated.

Issue 9: Improving raw material distribution within a facility.

The beginning or end of a system containing many operations is often a shared resource.
A common example is the use of a shared raw material facility or shared warehouse area.

39
Applied Simulation: Modeling and Analysis using FlexSim

While bottlenecks are known for the

individual operations, the impact of the
shared facilities is often neglected until it
actually causes a loss in performance.
The speed of getting material from an
incoming dock to its destination is often
dependent on the routing and availability
of transportation resources. The same is true
of moving finished goods to their correct
warehouse or loading dock destination. In Figure 1.28: Distributing raw material to
medical simulations, shared resources of multiple plant sections.
X-Ray or lab areas cause similar problems.
Because of the interrelated nature of these
operations, simulation is required.

Issue 10: Effectively coordinating a production line.

Coordination with other areas in a system
is similar to the problems caused by shared
facilities as discussed in Issue 9. In that case,
however, each processing operation is still
independent of the shared equipment. With
coordination, two or more independent
operations must interact at the exact
same time. For example, in an automobile
production line, components manufactured
on independent lines must arrive at an
Figure 1.29: Automotive assembly lines require assembly point at the same time in order to
coordinated operations. fill a particular order.
In the food industry, various components
must join to form “variety packs” of a finished product. The traditional way of dealing with
such coordination is to use large surge systems. In a lean environment, this process is not
considered efficient, and simulation is used to develop line configurations, operating proce-
dures, and production rate strategies, as an alternative to surge systems.
Coordination of production areas is further complicated by the stopping and starting of
parts of the production lines. Equipment downtime is generally considered to be the result
of some failure; however, predicting downtimes is complex as there are often competing
failures in different sections of a machine. Downtimes can also be caused by outside influ-
ences, such as operator break times, scheduled meeting times, and even preventive mainte-
nance. Taking competing downtimes into account and correctly entering them into a
simulation is important for representing system behavior and conducting a comprehensive
analysis. Simulation is a valuable tool for studying the impact of downtimes and how to
overcome them.

40
 Chapter 1

Issue 11: Working with operations that involve people.

The behavior of people taking part in
a simulated system is the most difficult
aspect to characterize or model. Individual
performance rates and logical choices
vary widely. Simulation can take many of
these characteristics into account when
operators, fork truck drivers, bank tellers,
healthcare personnel, ticket takers, airport
security, or other people are involved. The
improvement of healthcare facilities is of
Figure 1.30: Healthcare simulations track people particular importance today. Simulation
and their medical needs. provides a method to analyze how patient
needs can be both effectively and efficiently
handled in facilities such as hospitals, emergency rooms and clinics.

Issue 12: Developing logic for material transfer.

Making significant improvements to opera-
tions can involve specialized logic normally
applied through control systems. Such logic
includes dynamically changing rates that
depend on operating conditions, manip-
ulating flows to maintain coordinated
operations, or using photo eyes to control
movement of material. Such logic can be
complex, and accurately predicting its impact Figure 1.31: Conveyor photoeyes and logic
can be difficult without simulating the system. control material traffic.

Issue 13: Designing for just-in-time manufacturing.

Providing goods and services for a just-in-
time production system means producing
materials at a rate that matches the market
demand. Such an operation is also called a
“pull”system and requires precise control of the
manufacturing process. Whenever demand
changes, the processing systems must adapt
to the demand in both quantity and quality.
Significant logic may be employed to assure Figure 1.32: Pies are made to order.
that the desired product is correctly managed.
Additional logic may require that one operation be suspended to produce another product
and then return to the prior operation. A pull manufacturing environment is best designed
and analyzed with simulation.

41
Applied Simulation: Modeling and Analysis using FlexSim

Exercise 3-1 Coasting around

Background

As the manager of the “Super Rocket”

roller coaster ride at the Grand Bay
amusement park, you are responsible for
the day-to-day operation and financial
performance of the coaster (see Figure
3.4). Your parent company owns the
coaster and leases the space from the
park. The coaster at this park is only
three years old and is the lowest in
financial performance of all the parent
Figure 3.4: Super Rocket roller coaster.
company’s assets.
This year you, the manager, must show improved financials for the ride or find
another job. There are many trade-offs to consider. You are in competition with all the
other rides at the park. You get paid by the park based on the number of riders served.
Customers may love the coaster, but if they must wait in line too long, they will walk
away or not bother to ride again that day. Potential customers who see long lines will
opt for other rides with shorter lines. Most choices to increase riders involve increased
costs of equipment or staffing.
Your brother-in-law happens to be familiar
with simulation methods and offered to help. You
explain your problem and provide base data about
the operation. You are appreciative of the results
he gave you and decide to run the simulation
yourself.

Problem statement

How can the profitability of the coaster be

improved during the operating season?

Operating data

Riders enter the coaster line at different rates

during the 12-hour operating day. Operating
data show the rates per minute to have a Poisson
distribution; their means are provided in the table
in Figure 3.5.
A minimum of three workers are needed to
operate the coaster. One person takes tickets, one Figure 3.5: Rider average per hour.

98
 Chapter 3

supervises loading the cars, and one supervises unloading the cars. Extra people can
be added.
The first additional person helps to unload the cars, the next helps to load the cars.
In each case the unloading or loading times are reduced.
The coaster is currently operating with a base of four cars, each with a capacity of
four riders. New cars can be added to the base with the depreciation costs added to
the operating expense.
Rider capacity of the cars can also be
increased with a retrofit paid for through a
depreciation charge. Other operating costs
include depreciation costs for the present
cars, hourly wages, and general maintenance
costs. A summary of the operating costs is
provided in Figure 3.6.
The following options are available to
increase financial performance:
• Increase number of cars
Figure 3.6: Coaster operating costs.
• Increase car capacity (upgrade)
• Increase staffing level for an hour

Expected results

Provide an operating plan that includes staffing levels. Determine if there is sufficient
return to request additional cars.

Modeling and analysis issues

• How long should the simulation run? The simulation stop time
is set for 12 hours; selecting the run button will continue the
simulation from the stopped point.
• What variables should have the most impact?
• Is there a plan for trying different changes?
• What is most important about the financial report?
• While financial results are the primary metrics, do you feel long
wait times will reduce future riders?

99
Applied Simulation: Modeling and Analysis using FlexSim

Exercise 3-2 Martian Transfer Station

Background

The space exploration task force on

Mars has seven laboratory stations (see
Figure 3.7). Supplies are brought in
on cargo rockets and then transferred
to their destination by surface vehicles.
Pallets are recycled and brought to the
transfer station where two robotic cranes
separate the incoming supplies. When a
pallet is complete for a station, a guided
vehicle picks it up and carries it to the Figure 3.7: Martian transfer station.
station. The transfer station is operated
by two locals.

Problem statement

The stations are receiving increasing amounts of supplies—some of them with time-
critical contents. Can the work-in-process inventory be reduced by changing opera-
tional parameters?

Operating data

The transfer station has two working cranes. Empty pallets are delivered by a trans-
porter vehicle and transferred to the loading station with a shuttle. A crane will only
start loading the next box of supplies if there is a pallet in the appropriate transfer
station and the correct load is available.
Each pallet contains four crates. When a pallet is filled, it is moved to the distri-
bution side by the shuttle, and a call is made for pickup. When the transporter that
was called stops, the full pallet is loaded and taken to the experiment station. Trans-
porters already en-route will not stop for the pallet.
Only a certain number of transporters are available on the planet. They are sent out
at specific time intervals. After going through the transfer station, they continue to
one of the seven lab stations. After depositing any load, they continue to a transpor-
tation center, where they are brought back as a group to the sending station to once
again be sent to the transfer station.
The supply rocket is normally packed randomly on earth as supplies are brought to
the spaceport. Requests have been made to stage the rocket load first, and then group
them on the rocket by sequence, with several boxes grouped by their destination.
One time unit is one Martian01 second. As on Earth, there are 60 Martian seconds
per Martian minute and 60 Martian minutes per Martian hour.

100
 Chapter 3

The operating rules can be expressed as follows:

• Transporters leave the sending station at a fixed interval.
• If empty pallets are needed or a pallet is in place and ready for
pickup, a request is sent to the sending station and the next trans-
porter is given the assignment. A transporter can have only one
assignment at a time.
• Only one transporter can be at the transfer station at a time—
others will bypass the station.
• Crates from the supply ship arrive either in a random sequence or
grouped for each destination.

Expected results

• Run the simulation for a period of time and observe the system
dynamics. How do the cranes work to position crates on
pallets?
• Devise a plan to maximize the distribution of materials to
the laboratory stations. What are the tradeoffs with cost per
crate?

Modeling and analysis issues

• What is the impact of the arrival sequence?

• Can you tell by observation what logic is employed in moving
material from the space ship to the pallet layout?

101
Applied Simulation: Modeling and Analysis using FlexSim

Exercise 3-3 Slime Inc.

Background

As the new engineer working for Slime, Inc., you’ve been asked to familiarize yourself
with the company’s operations (see Figure 3.8). This plant only bottles the green variety
of slime and ranks in the middle of all plants in performance. Overall plant perfor-
mance across package sizes is a key metric, but performance of packing the popular
12-pack of slime is critical from a market perspective. You’ve found a simulation of
the plant that was built last year and decided to try it as a way of understanding the
operation.

Figure 3.8: Slime Factory.

Marketing is pushing the plant to agree to higher speeds and smaller tray pack units
because people can only handle so much slime at a time.
As an incentive, bonus points are awarded based on meeting production targets and
result in additional pay at the end of the fiscal year.

Problem statement

What are the best operating conditions to meet marketing and plant needs?

Operating data

Starting from raw plastic, the bottles are created with an injection molding machine.
The plastic forms are puffed into the bottle shape by a blow molder. The empty bottles
are put into baskets that transfer them to an un-scrambler, which in turn lines them

102
 Chapter 3

up to go into the filler lines. There are two filler lines that clean, fill, cap, and label the
bottles. The two lines join at the single tray packer. Trays are conveyed to a case closer
and brought to an operator who loads a hand truck to put them into waiting trucks.
One operator is responsible for the molding and unscrambling operations. This
operator unloads the baskets of clean bottles and places the basket back on the return
conveyor. The operator also must attend to any downtime on the equipment in the
area. The second operator carries cases to the trucks and fixes any problems with the
filling and case packing operations. Operators are concerned about their workloads
under the higher production rates.
Once a truck is filled, the facility’s doors will close and re-open only when another
truck has arrived.
The plant can operate at molding speeds between 45 and 85 bottles per minute. At
higher speeds the machines become less reliable and jam more often. This requires the
operators to stop their production-related work and fix the equipment.
The tray packer can handle 12, 16, and 24 count trays. Production standards for the
plant were set based on the 24 count trays when the plant was built. At a given bottle
rate, more 12 count trays are produced, thus requiring the second operator to make
more trips into the trucks carrying 5 cartons per load.
Plant efficiency is rated at better than 97% based on their standard 24 count tray
and 55 bottles per minute. Lower efficiencies will result in lower bonus points.

Expected results

Considering changes in bottles per case and the production rate, prepare a recommen-
dation for the next engineering group meeting indicating what you’ve found out about
the operations and what might be done to improve performance and meet marketing
needs.

Modeling and analysis issues

• For what period of time should the simulation run?

• Does the production or utilization report give any indication of a
bottleneck?
• Note what happens to operations as line speed increases.
• How busy do the operators appear to be?
• Is there a tradeoff between the plant’s operating performance
metric and the need for smaller sizes?

103
Applied Simulation: Modeling and Analysis using FlexSim

Exercise 3-4 Danson Electronics

Background

The Danson Electronics plant (see Figure 3.9) in Pembroke has been assembling high
quality DVD players for two years. To improve their operations, they instituted a
continuous improvement program with teams active on the plant floor.

Figure 3.9: Danson Electronics.

Problem statement

Which proposal should be adopted, or should all proposals be considered with a

priority established. What else should be done?

Operating data

The plant is divided into three buildings. In the first, double circuit boards are populated
with components and then sent to a wave soldering machine. After inspection they are
grouped onto a tray and manually transferred to the next building.
In the second building, the double boards are separated using a router, and the
individual boards are then put through a series of tests. After testing, the boards are
coated and put in racks to dry. Dry boards are manually transported to the third
building for final assembly.
In the assembly building the boards go through a series of assembly steps and tests.
The assembled units are placed on a table and moved to an aging box. After sitting in
the aging box, the units are tested again before moving on to shipping.
Four proposals have been put forth to improve plant productivity:

104
 Chapter 3

• The first building team notices that boards are backed up through
the wave soldering machine. They attribute the backup to the
fact that the router operator only has another batch of boards
brought over when the current tray was empty. They want a
surge added before the router so that boards can be moved more
effectively.
• The second building team wants to purchase a new router that
would decrease the routing time by 20%.
• The third building team wants new assembly machines to increase
their throughput by 20%.
• The union wants another person hired for the second building to
help move boards between buildings 2 and 3.

Expected results

Suggest how the improvement team recommendations should be handled.

Modeling and analysis issues

• What should the simulation run time be?

• Where do backups occur?
• What priority should be given to the suggestions?
• What improvement level would be considered good?
• What is most important: total production, WIP, or quality
numbers?

105
Applied Simulation: Modeling and Analysis using FlexSim

Exercise 3-5 Maritime Ltd.

Background

The town of SeaSide is situated on the

coast and enjoys a vibrant tourist trade.
Last year a hurricane badly eroded a
significant part of their shoreline and
exposed their main beaches to further
damage. The town council has hired
Maritime Ltd. to build a new breakwater
using large rocks from an excavation
further up the coast (see Figure 3.10).
Maritime has quoted the job for $11
million with significant penalties if the
job exceeds the amount and a reward
based on coming in under the price.
To plan the project, Maritime has
engaged a consulting firm to build a
simulation to help plan their options. Figure 3.10: Breakwater construction.
Since Maritime doesn’t want to give
the consulting firm access to their cost information, they have decided to have their
engineers input the cost data and run the simulation themselves. The consulting firm
acquired tidal data to use in the simulation.

Problem statement

Use the simulation to determine the best approach to finish the job under the quoted
price.

Operating data

The new breakwater will require 165,000 tons of rock carried by coastal barges
(coasters) for 833 nautical miles from the quarry to the town. The quarry is open for
the loading coasters 12 hours a day on weekdays only. The loaded coasters sail to a
position just outside the coastline where the rocks are transferred to a barge. The barge
then sails to the coastline and discharges the stones directly in front of the location.
Transfers can’t be made if the wave action is too high.
It is possible to use a pontoon to buffer rock material off shore. The coaster can
discharge to the pontoon while the barge is making a transfer so that the barge can
then load from the pontoon if the coaster has left. The pontoons are available in two
sizes.
The choices and costs associated with renting equipment are shown in Figure 3.11.
Site costs are $37,250 per day.

106
 Chapter 3

Figure 3.11: Marine project costs and choices.

The simulation has three views. One for the quarry, one for the coastal travel, and
one for the project site. Charts accumulate dynamically and the costs are calculated at
the end of the simulation when the transfer of rocks is completed.

107