Digital Twin + Industrial Internet for Smart Manufacturing: A Case Study in the Steel Industry
quality, productivity, delivery time, etc.) and
optimization (e.g. achieving zero-inventory
on-demand production).
In summary, the key to the industrial
internet, including its applications in a
manufacturing setting, is how to implement
data-driven optimization via the data,
analytics
(model)
and
closed-loop
application to solve specific problems in
various industrial scenarios.
A PPLYING THE I NDUSTRIAL I NTERNET
FOR S MART M ANUFACTURING
Applying the core idea of the industrial
internet in a manufacturing environment
requires data, analytics and application in
the following ways:
A RCHITECTURAL AND S YSTEMATIC
C HALLENGES
Manufacturing systems are complex
systems, often involving a large number of
interconnected equipment and many
intertwining processes working in concert.
For example, in a typical setting in the iron
and steel industry, a continuing process
manufacturing sector, a steel plant has a
long and complex end-to-end production
process consisting of many sub-processes
including sintering, ironmaking (blast
furnace), steelmaking (converter),
continuous casting, heat treatment, hot
rolling, cold rolling and strip processing.
Each of these sub-processes operates
dozens of various equipment pieces in a
complex production process. Furthermore,
these processes run at various rhythms and
paces ranging from a continuous process at
an early stage (e.g. iron making) to a
discrete process at a later stage (e.g.
striping). Through the end-to-end
production process, there are intertwined
Data is about connecting to the
various types of equipment and
systems—including PLCs, SCADA,
DCS and PCS—and other
manufacturing software systems,
such as MES, QMS, ERP and PLM, to
collect data about the production
material and parts, the products as
they are being manufactured, the
production equipment and
processes, the workers, the product
design and the business processes.
Analytics (Model) include building
and applying various analytic models
to analyze the data and gain insights
about the operational states of the
equipment and production
processes. The depth of the
analytics increases from descriptive
(e.g. what happens in remote
monitoring), diagnostic (e.g.
understanding why it happens),
IIC Journal of Innovation
predictive (what and when it will
happen) and prescriptive (how to
respond to a predicted event)—and
the analytics have become more
sophisticated.
Application involves implementing
business logic that transforms the
insights from the analytics into
optimal decisions and actions, either
providing recommendations of
action to the operators (humans in
the loop) or directly instruct the
systems to complete the closed
feedback loop of optimization in the
production processes.
- 13 -