Artificial Intelligence and Machine Learning

Unit 1: Introduction to Industry 4.0

The industrial revolutions
The first industrial revolution

The first Industrial Revolution was initiated in Britain during 1760 when the practical steam
engine was invented by James Watt. Thomas Savery (1698) and Thomas Newcomen (1712) also
developed steam engines, which transformed the whole manufacturing scenario globally in what
was aptly called the Industrial Revolution. With the advent of further landmarks in the
advancement of industry, each major stage has been assigned a logical series of designations,
with the first Industrial Revolution being redesignated as Industry 1.0.

The second industrial revolution

This refers to the early 20th century when electric power was extensively used in place of steam
power. Starting with introduction of electric lighting in 1882, and the identification of electric
motors as reliable and economical power transmitter, it quickly became a basic requirement for
normal operation of all industries of developed economies. By the middle of the 20th century,
electric power was seen as a natural monopoly in power generation and use. This greatly enabled
advancement in mass production, and this era is redesignated as Industry 2.0.

The third industrial revolution

This refers to the latter half of the 20th century involving the use of information technology and
electronics in production, as is used currently. This is the stage when computer-integrated
manufacturing (CIM) involving application of computer software to control the entire production
process was developed and adapted globally and aptly redesignated as Industry 3.0. Here, the
enterprise resource planning (ERP) system and machine control are technically two different
nonintegrated processes, which distinguishes it from Industry 4.0.

The fourth Industrial revolution

The fourth Industrial Revolution is designated as Industry 4.0. In addition to the CIM, specially
developed computer software, which integrates the functions of all the departments or activities,
is applied to control the entire production process. Sensors are embedded on machinery and other
salient locations, which allow individual processes to exchange information with other processes
and equipment in the system and initiate actions. Besides enabling manufacturing to be faster
and less errorprone, the integration of computers has the main advantage of the ability to create
automated manufacturing processes. Typically, CIM relies on closed-loop control processes,
based on real-time input from sensors. Hence, it is also known as flexible design and
manufacturing. In fact, the latest concept of the Internet of Things (IoT) is the backbone of

Prof. Sandeep M. Kadane (Dept. of Mechanical Engineering, K.I.T.’S College of Engineering) Page 1
 Artificial Intelligence and Machine Learning

Industry 4.0. With IoT, devices and embedded computing sensors will be communicating,
delivering real-time responses.

● Industry 1.0: Steam power

● Industry 2.0: Electric power

● Industry 3.0: Computing power

● Industry 4.0: IoT power/digitalization

Prof. Sandeep M. Kadane (Dept. of Mechanical Engineering, K.I.T.’S College of Engineering) Page 2
 Artificial Intelligence and Machine Learning

Demand dimensions of Industry 4.0

Some future demand dimensions of Industry 4.0 as follows.

Variety: Companies may introduce multiple models for each product. There may be one or more
standard products consisting of standard modules. These standard modules consist of hardware
and/or software that provide standard functions for general customers. Some customer
participation in product design can be important. There could be other standard platforms where
customers can suggest or request specific personal designed modules and/or components to
realise possible individual customisation. A platform is an unfinished product that consists of
underlying core components or modules to form a common structure from which derivative
products can be efficiently developed and produced.

Time: In Industry 4.0, product life cycles may become more uncertain. The life cycle of a
platform may be short or long. In contrast, life cycles of individual modules that are personally
designed to provide specific functions may be short because of possible frequent upgrades.
Requested delivery time may be short.

Volume: In Industry 4.0, volumes of standard products and platforms may be high or middle. In
contrast, volumes of personal designed modules may be very low. Volumes of standard modules
may fluctuate drastically with a wide range from low to high

Design prerequisites for Industry 4.0

1. Machines, devices, sensors, and operatives must be programmed to communicate with one
another.

2. The information systems must create a virtual copy of the physical conditions through the
sensors so that the information can be contextualized for each situation.

3. The system should be able to make decisions and perform autonomously.

4. The system should also be programmed to support the operatives in decision making.

5. The system should perform tasks that are hazardous for humans.

6. The system should be able to recognize the interferences, conflicting goals, etc., to report them
to a higher level or give indications to humans.

Prof. Sandeep M. Kadane (Dept. of Mechanical Engineering, K.I.T.’S College of Engineering) Page 3
 Artificial Intelligence and Machine Learning

Benefits and Limitations of Industry 4.0

Benefits of Industry 4.0

To the management

1. Less occurrence of crisis management situations.

2. Better business monitoring from the central office.
3. Ease of decision-making process.
4. More time available for critical thinking process

To the manufacturing personnel

1. Planning and execution can be done more efficiently.

2. Efficiency increased by manmachine collaboration.
3. Shift/working hours coordination online is facilitated.
4. Remote control of production equipment aids the control personnel.
5. Any interference or impedance to the smooth production flow will be detected instantly
and will alert personnel. For the plant maintenance personnel
6. Remote-controlled maintenance.
7. Computer-based support for decision making.
8. Extensive use of sensors to monitor the state of production equipment.
9. Improved maintenance due to machine learning will lead to lesser machine variation and
consequently better quality. For the production control function
10. Paperless logistics.
11. Traceability of products and parts.
12. Optimization of production with Big Data.

For the quality control function

1. Measuring instruments will be interconnected and the measurement data will be stored in
the cloud automatically.
2. More time is available for analysis rather than data collection.
3. Analysis will also become automated to a large extent.
4. We will have more mind space to make clear decisions, and the decision making will
become faster and accurate.
5. Controlling quality will be easier. Quality personnel policing will considerably decrease.
6. Implementation of poka yoke will be facilitated.
7. Defect generation can be prevented at the source
8. Adherence to the quality system can be better ensured.
9. Machines will remember past quality data and alert all stakeholders.

Prof. Sandeep M. Kadane (Dept. of Mechanical Engineering, K.I.T.’S College of Engineering) Page 4
 Artificial Intelligence and Machine Learning

Limitations of Industry 4.0

However, many firms face difficulties in their transition to the fourth Industrial Revolution, with
challenges including the following:

1. Data security and privacy

2. Integration of different data sources
3. Financing the implementation
4. Management of arising complexity
5. Breaking down established structures and processes
6. Implementing projects without cutting back existing operations
7. Need to maintain integrity in the production process
8. General resistance to change by stakeholders

Smart Factory

A smart factory is a digitised manufacturing facility that uses connected devices, machinery and
production systems to continuously collect and share data. This data is then used to inform
decisions to improve processes as well as address any issues that may arise.

The smart manufacturing practices used by a smart factory are enabled by a variety of
technologies including artificial intelligence (AI), big data analytics, cloud computing, and the
industrial Internet of Things (IoT).

Smart factories connect the digital and physical worlds in order to monitor an entire production
process, from supply chain management to manufacturing tools and even the work of individual
operators on the shop floor.

Fully integrated, collaborative manufacturing systems provide a range of benefits for operators,
including allowing operations to be adaptable and readily optimised.

Benefits

1. Smart factories use connected equipment and devices to allow for evidence-based decision-
making to optimise efficiency and productivity throughout the manufacturing process.
2. Delivering an agile, iterative production process can extend the capabilities of both devices
and employees, leading to lower costs, reduced downtimes and less waste in the
manufacturing industry.
3. Identifying and then reducing or eliminating underused or misplaced production capabilities
increases efficiency and output with little investment in new resources.
4. The benefits of digitalising a factory include those related to planning, quality control,
product development and logistics as each is assessed and optimised based on real feedback.
5. There are also long term benefits to be gained through the introduction of machine learning
to the process. By collecting and analysing data, it is possible to schedule preventive and

Prof. Sandeep M. Kadane (Dept. of Mechanical Engineering, K.I.T.’S College of Engineering) Page 5
 Artificial Intelligence and Machine Learning

predictive maintenance - based on accurate real-life information - to avoid production line

shutdowns.

The Four Levels of Smart Factories

There are four levels that can be used to assess your journey through the improvement process to
becoming a smart manufacturer:

1. Level One: Basic Data Availability

At this level, a factory or facility is not really ‘smart’ at all. There is data available but it is not
easily accessed or analysed. Data analysis, where it is done, is time consuming and can add
inefficiencies to your production process.

2. Level Two: Proactive Data Analysis

At this level, the data can be accessed in a more structured and understandable form. The data
will be centrally available and organised with visualisation and displays assisting with its
processing. This all allows for proactive data analysis, although there will still be a level of effort
involved.

3. Level Three: Active Data

At this level, the data can be analysed with the assistance of machine learning and artificial
intelligence, creating insight without as much human supervision. The system is more automated
than at level two and can predict key issues or anomalies to proactively predict potential failures.

4. Level Four: Action-Oriented Data

The fourth level builds on the active nature of level three to create solutions to issues and, in
some instances, undertake action to alleviate a problem or improve a process with no human
intervention. At this level, data is collected and analysed for issues before solutions are generated
and, where possible, actioned with very little human input.

Technologies used in a Smart Factory

Smart factories use a variety of different technologies related to the fourth industrial revolution
(Industry 4.0) to optimise smart manufacturing processes.

These technologies include:

Prof. Sandeep M. Kadane (Dept. of Mechanical Engineering, K.I.T.’S College of Engineering) Page 6
 Artificial Intelligence and Machine Learning

Sensors

Sensors on devices and machines are used at specific stages of the manufacturing process to
collect data that can be used to monitor processes. For example, sensors can monitor temperature
or other variables and either self-correct any problems or alert staff. These sensors can be linked
to a network to provide joined-up monitoring across several machines.

Cloud Computing

Storing and processing data collected from the sensors is achieved through cloud computing.
This if more flexible and cheaper than traditional on-site storage, allowing large amounts of data
to be uploaded, stored, and assessed to provide feedback for decision-making in real time.

Big Data Analytics

As more data is collected, it is possible to use it to provide insights into how a production
process is performing. Big data allows for error patterns to be spotted and predictive quality
assurance undertaken with a greater degree of accuracy. This data can be shared between
different factories or even organisations to solve common problems and further optimise
processes.

Virtual and Augmented Reality

Augmented reality is a digital technology that involves digital information being overlayed
across reality and viewed via a smartphone, while virtual reality is a more immersive virtual
world that requires special glasses. Both of these technologies can help smart factory operators to
organise products, production tasks and the maintenance and repair of equipment.

Digital Twins

A digital twin can be used to represent a process or physical object and simulate performance in
the real world. This can lead to efficiency improvements while also aiding control and operations
planning.

Smart Manufacturing

Smart Manufacturing is a subset of Industry 4.0 that is focused on the use of smart technologies
to improve the manufacturing process.

Prof. Sandeep M. Kadane (Dept. of Mechanical Engineering, K.I.T.’S College of Engineering) Page 7
 Artificial Intelligence and Machine Learning

The main difference between Industry 4.0 and Smart Manufacturing is that Industry 4.0 is a
broader concept encompassing the entire value chain from the initial product design to the
manufacturing process to after-sales service. Smart Manufacturing is more limited in
scope, only focused on the manufacturing process itself.

Smart Manufacturing involves the use of cutting-edge cyber-physical technologies, big data
analytics, and automation to create a manufacturing process that is not only efficient but adaptive
and intelligent.

Core Principles of Smart Manufacturing

Smart Manufacturing is guided by a set of principles:

 Interconnectedness: facilitating seamless communications between machines, devices,

systems, and people, forming a network that shares real-time data and insights.

 Data-driven decision-making: every action in smart manufacturing should be informed

by predictive and prescriptive data-driven decision-making.

 Adaptability: the ability to quickly adjust production processes and strategies in

response to changing conditions. Processes should be optimized continuously based on
data-driven insights and real-time user feedback.

 Visibility: technology should enable real-time visibility into operations to facilitate more
transparency and more informed decision-making.

Benefits of Smart Manufacturing

1. Efficiency and productivity: With automation, real-time data analytics, and integrated
manufacturing solutions, your teams can work faster, smarter, and safer. And predictive
maintenance and automated workflows can help your IoT connected machines and assets
operate at peak level with more streamlined outputs.
2. Agility and responsiveness: At each stage in the process, smart manufacturing solutions
and data analytics allow you to respond quickly to market changes – pivoting your
manufacturing workflows and personalising your products with speed and acuity.
3. Sustainability: The data gathered by smart manufacturing technologies can help enact
strategic and cost-effective plans to streamline operations and lower energy usage. From
sustainable product design to greener logistics, smart solutions can support your
sustainable initiatives.
4. Improved quality control: From the frustration of bad reviews to the catastrophe of a
product recall, smart manufacturing solutions can integrate across you supply chain and
manufacturing operations, to ensure that quality standards are visible and verifiable and
every stage.

Prof. Sandeep M. Kadane (Dept. of Mechanical Engineering, K.I.T.’S College of Engineering) Page 8
 Artificial Intelligence and Machine Learning

5. End-to-end savings: Digital integration from one end of your supply chain to the other,
allows for better forecasting, inventory management, and logistic solutions. This means
fewer risks, less spending and – best of all – happier customers.

Lean manufacturing
Lean manufacturing is a methodology that focuses on minimizing waste within manufacturing
systems while simultaneously maximizing productivity. Waste is seen as anything that customers
do not believe adds value and are not willing to pay for. Some of the benefits of lean
manufacturing can include reduced lead times, reduced operating costs and improved product
quality.

The Toyota Production System laid out seven wastes, or processes and resources, that don't add
value for the customer. These seven wastes are:
1. Unnecessary transportation
2. Excess inventory
3. Unnecessary motion of people, equipment or machinery
4. Waiting, whether it is people waiting or idle equipment
5. Over-production of a product
6. Over-processing or putting more time into a product than a customer needs, such as
designs that require high-tech machinery for unnecessary features
7. Defects, which require effort and cost for corrections.

LEAN MANAGEMENT TOOLS

5S

5S is a method that uses a list of five Japanese words: seiri, seiton, seiso, seiketsu, and shitsuke.
It describes how to organize a work space for efficiency and effectiveness by identifying and
storing the items used, maintaining the area and items, and sustaining the new order. The
following are the translations of the 5 Japanese words

1. Sort (Seiri)

1. Make work easier by eliminating obstacles.

2. Reduce chances of being disturbed with unnecessary items.
3. Prevent accumulation of unnecessary items.
4. Evaluate necessary items with regard to cost or other factors.
5. Remove all parts or tools that are not in use.
6. Segregate unwanted material from the workplace.
7. Define Red-Tag area to place unnecessary items that cannot immediately be disposed of.
8. Dispose of these items when possible.
9. Need fully skilled supervisor for checking on a regular basis.

Prof. Sandeep M. Kadane (Dept. of Mechanical Engineering, K.I.T.’S College of Engineering) Page 9
 Artificial Intelligence and Machine Learning

10. Waste removal.

11. Make clear all working floor except using material.

2. Set In Order (Seiton)

1. Arrange all necessary items so that they can be easily selected for use.
2. Prevent loss and waste of time by arranging work station in such a way that all tooling /
equipment is in close proximity.
3. Make it easy to find and pick up necessary items.
4. Ensure first-in-first-out FIFO basis.
5. Make workflow smooth and easy.
6. All of the above work should be done on a regular basis.
7. Maintain safety.
8. Place components according to their uses, with the frequently used components being
nearest to the work place.

3. Shine (Seiso)

1. Clean the workplace on daily basis completely or set cleaning frequency

2. Use cleaning as inspection.
3. Prevent machinery and equipment deterioration.
4. Keep workplace safe and easy to work.
5. Keep workplace clean and pleasing to work in
6. When in place, anyone not familiar to the environment must be able to detect any
problems within 50 feet in 5 secs.

4.Standardize (Seiketsu)

1. Standardize the best practices in the work area.

2. Maintain high standards in workplace organization at all times.
3. Maintain orderliness. Maintain everything in order and according to its standard.
4. Everything in its right place.
5. Every process has a standard.

5.Sustain (Shitsuke)

1. Not harmful to anyone.

2. Also translates as "do without being told".
3. Perform regular audits.
4. Training and discipline.
5. Training is goal-oriented process. Its resulting feedback is necessary monthly.
6. Self discipline
7. To maintain proper order

Prof. Sandeep M. Kadane (Dept. of Mechanical Engineering, K.I.T.’S College of Engineering) Page 10