COMMUNITIES OF PRACTICE
Lean: Simulation Technology
Simulation technology has come a long way and is a key element in lean
implementation. Lean improvement teams now have the ability to analyze
manufacturing and supply chain operations, and simulation capabilities
provide the next step in the traditional lean improvement toolkit.
Simulation adds the dynamic
representation that shows a process over time and the impact of changes on
that process. It is visually powerful to actually see material flowing
through a process and metrics calculated that mirror the real operation.
Product mix is typically a key driver of performance in an operation, and
simulation is unique in its ability to model mix impacts.
and financial flows may all be modeled as an interdependent system. For
instance, how much will the consignment inventory for a customer really
affect the cash-to-cash cycle if the payment terms don’t change? What is
the impact of information delay when importing raw materials?
Simulation has been used
successfully in factory floor improvement, inventory management, capacity
analyses, and process design. Benefits such as the following are not
- Postponing final packaging
results in 30 percent inventory reduction with substantial service
improvement to more than 97 percent.
- Shift/work center change
allows reduction of staff from 7 to 6 days per week.
- Adding staff to a bottleneck
work center reduces overall number of workers by 10 percent
- Synchronized operations
reduce cycle time from 18 to 13 weeks.
A traditional tool such
as value stream mapping is a good first step in a lean process analysis.
However, it is a static representation and typically not completed for the
lower volume products that cause many problems. Because simulation models
are dynamic, managers can see the impacts of variability in demand
patterns, production processes, and downtime. It is also possible to test
schedules, kanbans, and placement of inventory buffers before employing
them in the actual operation.
Variability and product
mix have significant impacts on throughput and capacity investments.
Statistical distributions derived from historical records may be used to
represent operations, which is important to develop sound confidence
intervals for the results. Many real-world parameters have skewed
distributions resulting in maximum backlogs, process times, wait times,
outages, and so forth, which are much longer than the average. Things do
not even out over time when it comes to the impact of variability in a
Lean analysts have used simulation to test process improvements including
- Kanbans and constant
- Every-part-every intervals
for continuous flow
- Batch and campaign sizes vs.
- Line balancing
- Setup reduction
- Routing changes
- Shared resources
- Yield and scrap
- Material lead times
- Work cell design and
- Cross-trained workers to
perform multiple tasks.
For any change being
tested, analysts may view results graphically or in a spreadsheet.
Traditional and lean metrics such as takt times/rates, overall equipment
effectiveness, and end-to-end cycle times may all be incorporated.
The technology has broad
applicability in process and discrete industries, multilocation supply
chains, job or flow shops, make-to-order and make-to-stock environments,
and engineering design activities.
A credible model is
essential; a baseline model is normally tested against current operational
metrics using actual historical data. After that, the model is validated
for demand and supply processes against those metrics and only then is it
ready to be used to test changes in processes, capacities, lead times, and
With the software
advantages of Excel spreadsheets and graphic user interface/database
technologies, models may be fed with actual data from enterprise resources
planning systems, advanced planning and scheduling systems, and
manufacturing execution systems. It also makes it possible to keep them
up-to-date for use by planners, engineers, research and development staff,
and process excellence teams.
Achieving a fully
demand-driven pull process remains a challenge for lean programs.
Simulation is a powerful iterative tool that can have an important impact
today in implementing the pull process, but it is not the silver bullet.
The next breakthrough will be integrating optimization algorithms to test
solutions by simulating them for a period of time with all of the
real-world variability applied.
—Jim Curry, CEO, OpStat
Group, Inc., an operations improvement company, can be reached at (203)
431-3905 or via e-mail at JimCurry@OpStat.com.
Douglas R. Kelly
Mary H. King, senior manager, APICS
Industry Content Division