How to build a smart factory
Harigopal Suyambu Amnbihai
in Mechanical Engineering
Date of submission
Production has taken
various transformations, starting from the first industrial revolution where
steam engines and water turbines reduced human effort and led path for the
mechanical production process. During the second industrial revolution conveyer
belt and split of work led to mass production. It is that point of time when
scientific management was introduced. Then the automation played its role in
the third industrial revolution by the invention of the programable logic
circuit. CNC machines were also introduced during the third industrial revolution.
When it was believed that the production process is optimized to its maximum
level, the cyber physical systems and the Internet of Things gave way for the
fourth industrial revolution, “The Smart Factory”. This paper will deal with
what a smart factory is? Its features and characteristics and the key physical
components to construct a smart factory.
What is a Smart factory?
the industrial revolution, automation in production has achieved a tremendous
growth. The next giant step in production will be construction of smart factory
which uses the concept of industry 4.0 and artificial intelligence. Smart factories
work on the principle of interaction between the physical object and the
control software which are interlinked by Internet of Things (IoT). In this the
information technology is interlinked with the operational technology, right
from the start of inventory control till the delivery of the fully finished
object. In some necessary places the robot interacts with the human. All the
data are digitalized and stored in the database for future reference. These
factories are so flexible that the production processes could be varied automatically
according to the product being produced which is controlled by the centralized cloud
network. By this highly standardized production process, highly precise and high-quality
products is delivered with high fill rate and high variance.
Smart factory features
construct a smart factory a central hub of digital supply network is necessary.
Unlike the traditional production supply chain, the centralized digital supply
network can interact between various supply chain systems and the production
process can be customized. This helps in achieving a highly flexible production
and supply chain system. All the information’s are collected, processed and
transmitted by this digital core, which decides the sequence of operations to
be carried out and hence the line production system can be eliminated.
as key players
manual and semiautomatic production process the split of human resource and
utilization of the machineries determines the efficiency of the process. In a
smart factory each employee’s skill set is fed in the central digital core
system and once the employee enters the work place the system identifies the
worker (with a barcode or magnetic strips) and allots work according to his
skills. It also monitors the work done by the worker and provides him with
appropriate tools. Ergonomics of the worker is also monitored and provides
recommendations to ease of work.
central control system is then decentralized, and the information are send to
the sub artificial intelligence, which splits the work and get it done. The
main central system sends the product specification to the subsystem. The
subsystem then makes autonomous decision and starts the process by controlling
the drives, sequence and operations. The artificial intelligence continuously
monitors the changing requirements and working conditions and adopts itself
according to the atmosphere. There should be sufficient capacity to produce
high variance in products.
smart factory must always have fixed and open standard, which is one of the 5S
principle (standardization). These standards should be the same all over the
process and system. Once the standards are decided, there should be no
deviation from the standards to avoid malfunctioning or production of low
quality products. The standards of the Ethernet and connections should also be standardized.
Two machines with two different standards will interact between them which are
not understandable by each other and produces high scrap or rework.
complete elements and process of the smart factory are in virtual reality time
representation. All the virtual parts interact with the physical components to
perform the operation. For instance, the central hub collects all the
information from the surrounding through software and sensors. These data are
then send to the hub system through Ethernet and processed. The feedback is
send to the control system and the above procedure repeats without any
intervention of humans.
the computer system decides the process to be carried out, it analyses with
various data in nature and finds the easiest and the efficient way to solve it.
Hence the development or planning time, setting time, process time and cost can
also be reduced significantly. The customer can also easily track his product
when it is getting manufactured. Variance of the products been produced can
have good scope with minimal non-value-added time.
Smart factory characteristics
main characteristic of a smart factory is its connectivity of the material with
the process, relationship between the departments of the factory, interlink
between the customer, supplier and production process along with the capability
of Artificial Intelligence to take decisions in real time. Sensors should be
setup in all possible sub system to extract data and keep the control system
updated. The system also retrieves the historic data of the past and relates
them with the current data. It should also retrieve information from the
customer and the supplier end to create a highly efficient supply chain and
factory should be designed in such a way that all the process should be
optimized to have its least process time. Automation should be given higher
importance and human interaction should be reduced as much as possible. In the upcoming
years energy consumption is also more important, so renewable energy sources
should be preferred. The other two things to be noticed is quality and wastage.
Highly precise and high-quality products should be produced with minimum
wastage as possible. All these optimizations will automatically reduce the cost
of the product. Optimization should also be done in setup time by implementing
SMED, having optimal lead time and having minimal inventory and stock.
the real-time data should be transparent for decision making. When the process
and data are transparent, the faults can be found easily and rectified. It is
possible for the system to provide with notifications and alerts. It also helps
the customers to track their products and monitor the process.
should be made before an issue or failure occur. Due to the transparency of the
process, decisions can be made before the fault. It also includes monitoring
the restocking, inventory, quality issues, maintenance and safety. The historic
data’s can also be used as benchmark for maintenance and prevent breakdown. This
helps in reducing the down time of machines.
to flexibility of the smart factory, scheduling of the operations and material
requirements are planned automatically and executed. Then the data are cross
checked with data of real time, and the system updates automatically. Advanced
smart factories should be capable to plan the requirements according to the
change of product. The best possible way should be selected by the system, so
that the setup time, scheduling time and changeover time can be optimized.
within a smart factory
helps in producing rapid prototype and low volume products.
and planning reduces
waste and cycle time using real time inventory and production data.
Autonomous robots can be used in
recurring process which is repeated and again with high accuracy.
Digital twin can be used to
integrate the predictive analysis by digitalization of operation.
Augmented reality assists humans in
picking and positioning the articles.
Autonomous robots do all warehouse
operations with the help of artificial intelligence.
Sensors can be used to
know the real-time position of the raw material, their storage in warehouse,
their stock, current work status and tracking of finished products.
Analytics to reduce the
inventory, to monitor them and automatically order them when there is demand.
Optical based analysis
systems to be used to find defects.
to find the defects in the machine which produces the product.
Special sensors which identifies
the flaws and does the quality check.
Augmented reality, the robotic
system which helps the maintenance people in repairing and maintaining the
Sensors to detect the
malfunction of the machine and conduct the predictive maintenance.
Environmental, safety and health
Sensors to monitor the health
of the person and send notifications in times of emergency or saturation
Sensors to predict the
environmental changes and take precautions before the system gets affected.
Sensors to monitor the
machines to prevent accidents due to machine failures.
Key physical components of a smart factory
are detection devices which monitors the machine around the clock and provides
digital data for the system. Countless sensors are available in market, but the
sensors characterized according to the application should be used to monitor
Thermal sensor (thermometer, thermocouple, pyrometer, net
Pressure sensor (barometer, pressure gauge, piezometer,
ionization gauge, etc.)
Force sensor (piezoelectric sensor, strain gauge, force
gauge, torque sensor, etc.)
optical sensors (photo detector, photo diode, optical
position sensor, etc.)
position sensors (gravimeter, odometer, tilt sensor,
tachometer, flex sensor, etc.)
navigation sensors (air speed indicator, altimeter,
variometer, turn coordinator, etc.)
sensor (cloud chamber, Geiger counter, neutron detection, etc.)
Flow sensor (air flow meter, gas meter, mass flow sensor,
Environmental sensors (frequency domain sensor, rain sensor,
soil moisture sensor, etc.)
Electric sensor (current sensor, voltage sensor,
magnetometer, electron multiplier, etc.)
Chemical sensor (pH sensor, hydrogen sensor, optode, ozone
Sound sensor (geophone, seismometer, hydrophone, microphone, etc.)
and actuators are the basic electro mechanical components which controls the speed
of the process. These basic components of automation are controlled by
software, in which the process requirements are fed.
Example: ball valve, butterfly
valve, clapper valve, check valve, choke valve, pinch valve, piston valve, plug
valve, poppet valve, safety valve, etc.
programable logic circuit is the base of automation. The information collected
by the sensors are read by the PLC and analyzed. The sequence of events can
also be programmed in the PLC, according to which the process is carried out.
The actuators and valves are also controlled by the signals from the PLC
circuit. The most hazardous and complicated works should be replaced by
automated devices controlled by PLC’s. PLC’s are used in
Highly reliable activities
High precision activities
control and data acquisition can be used to control the activities or progress
of the activities. This could also be used to control the efficiency of the
systems. Once installed to the main system, it controls the PLC’s and retrieves
data from them and commands the PLC’s for the further actions. SCADA is the
backbone for the automation and control systems in all the smart factories. The
main applications of SCADA s in the fields of oil, energy, road and transport,
power and grid, water and resources, manufacturing, recycling, air and sea,
The object linking and embedding for process
control is the bridge between the physical component and the control system. The OPC server translates the data collected
from the hardware to through PLC into the OPC control. The OPC historic data
access is used to access the data from the past and analyze it to know if it
could be used as a bench mark for the new process.
Internet of Things is the gateway between the cloud of data and the accessories
in the real world. The data transfer is bidirectional, and it can send
thousands of information per second. The gateway also has the capacity to
process the data before it is send to the cloud or to the component. Due to
bidirectional data transfer, the data security is high.
Augmented reality and the virtual reality could be used to troubleshoot faults
and failures in minimum time. The other features are
Increase efficiency by increasing the workers safety.
The virtual reality collects information during the virtual
reality experience and computes it with the historic data saved in the
Development cycle reduces with the augmented reality.
Planning time can be
reduced using augmented reality and high-quality planning can be achieved.
Production system need not be modeled full.
Steps to build a smart factory
Create a concept of
the smart factory in a manageable environment. Planning should be done from the
base, starting from the product to be produced, followed by the machine and
machine tools required to produce the product, and then the dimension of the
factory required to hold all these machineries along with the storage and
Scale the basic concept
into a prototype by increasing the performance of the tools and machineries.
This prototype should hold all the basic information required to construct the
factory. This prototype can be made physical with trial methods or in virtual
by simulation software.
Once the prototype succeeds, the additional
assets should be added. This include addition of all supporting process. The
connections between the sequential operations should be validated.
production line by improving the performance of the dependent processes. The
implementation should be done such that the setup time, waiting time, and the process
time is reduced.
Improve the factory
by the better utilization of the resources and assets. The major thing that
should be considered is lowering the inventory and stock in the mean time
utilizing the machine to its optimum level.
Networking should be
done by linking all the process to the central hub. The supply chain should
also be connected to the central system. The connections may be through wires
or through Internet. In this case every machine will have a transmitter and a
receiver that would help to code and decode the commands and messages.
Due to the variability in
customer requirements and high degree of automation, it is obvious for all the
modern industries to move towards smart factory. Even though most of the
companies make business by naming themselves smart factory, they aren’t smart. A
real smart factory should have the following properties.
Cyber physical systems and the Internet of Things should be
It should be capable enough to produce variability in
There should be a clear track for the inventory, supply
chain and delivery of the product.
The data should be stored in cloud and should be retrieved
for future reference.
All these four
requirements should be fulfilled without the threat of cyber attack or privacy
Ø Factory Physics for
Managers by Edward S.Pound/ Jeffrey H.Bell/ Mark L.Speraman, PhD