A major revolution in the field of instrumentation and control technology is well underway. Research, development and deployment activities are focused on making quantum leaps in industrial automation performance. Called Industry 4.0, this includes a new generation of low-cost wireless sensors, improved real-time data analytics and control systems, and advancements in high-fidelity process modeling. These innovations will include systems that improve industrial manufacturing efficiencies, and integrate and network subsystems across manufacturing processes.
2. Overview
1) What is Instrumentation
• Basic Instrumentation
• Modern/Advanced Instrumentation
2) What is Industry 4.0?
• Introduction into 4th Industrial Revolution
• What it actually is
• Case studies/Examples and it’s speculated future
3) Industry 4.0 + Instrumentation
4) Why Manufacturers need this transformation?
5) The Role of Humans
3. What is Instrumentation?
Instrumentation is basically the branch of engineering that
deals with sensing, measurement and control.
With a bit of task-specific signal processing, instrumentation becomes
that potent tool in any system (even with dead time and unpredictable
non-linearity), which convert sensor and other data into usable format,
and thus acute control of the process(es).
The FCE could be any device that changes the state of the process,
according to controller output.
4. Modern Instrumentation
Modern Instrumentation consists advanced sensors, centralized adaptive
control systems and panels (DCS, PLC) and the newest addition,
Machine Learning (and Deep Learning, in places) and Data Analytics.
Advanced Sensors are enabled with signal processing techniques (some
equip DSP algorithms) for a better, cleaner and more desirable output.
It’s able to adjust to varying conditions through self-monitoring and self-
calibration, and use various Industrial Network Protocols for inter-
communication.
Adaptive Control are various control techniques that uses sensor derived
data to continuously monitor the state of the system through one of
many non-linear and adaptive control techniques. (Iterative learning,
Gain Scheduling, Model Reference, etc)
It changes the controller dynamics and output of the system, according
5. Machine Learning, simply put is a software technology often referred to
as AI, that helps the computers learn about a certain task (through
programming logic) based on extensive existing data on that task
(example determination of counterfeit bills). What the model has then
learned, will be then applied to the real world for similar tasks (checking
whether the payment is counterfeit or not).
Analytics (science of analysis) is largely useful in process monitoring,
anomaly detection and adaptive control algorithms that can respond in
real time. It stitches together the Digital Threads created by Internet of
Things to generate actionable intelligence, achieving optimal
performance.
Hence, with all these, Instrumentation transforms into a modern
marvel.
6. What is Industry 4.0?
Industry 4.0, or Fourth Industrial Revolution, follows the Steam Engine
(Industry 1.0), Mass Production (Industry 2.0) and Automation and
Computer (Industry 3.0).
From a core industry
perspective, it’s the
interconnection of all
devices on the floor and the
entire process line (from
raw material to consumer
feedback) in general,
merging the cyber and physical worlds. Artificial Intelligence drives this
change
For simplicity, consider a medicine manufacturing industry. If the raw
plants being procured, steps being taken for extracting the juice and
making the medicine, wrapping and being sent to shops, and doctors’
7. • This may seem like the process already in action. But Industry 4.0 is just
a cut above it; it monitors and controls the entire process on it’s own by
it’s own intelligence.
• Production managers having a bird’s eye on the entire chain (or loop, in
this case).
• WHY THE CHANGE?:
Raw materials procured must depend on market demand and R&D.
Customers should be able to order specific designs, which machines
themselves process and make. Factory floor should come alive and
collaborate with each other.
Machine intelligence should also help in reducing offsets and predict
failures and maintenance requirements.
Further analytics would improve the production design.
Finally consumer feedback would add more data to the existing model, both
improving it and predicting future possibilities and designs.
The most exciting part (possibly even the most difficult to handle) of this
8. • Imagine a “vehicle” as a factory with many number of working blocks,
such as engine, body, tires, light, stereo, etc.
Say, vehicle tyre wear, is mainly caused by low tyre pressure.
So, as a part of an Integrated Vehicle Management System, a simple
pressure sensor can continuously monitor tyre pressures, and keep alerting
the driver whenever required.
Similar continuous monitoring predicts exactly when parts in use might fail
– so that you could replace it just in time.
Most replacements would involve relatively minor maintenance and could
be performed without bringing down the ‘factory’ as a whole.
The various sensor data from the vehicle would also send feedback to the
designer, who will look forward to improving the design (which includes raw
materials, r&d, manufacturing tech, market strategy, etc) of frequently
disturbing elements.
Just like the vehicle, if every block of the entire chain of every industry
communicates within and with each other about their issues, and work
towards them, it will undoubtedly increase efficiency, without increase in
cost.
9. Instrumentation+Industry 4.0
1. Instrumentation, as we learned, deals with sensors, measurement,
signal processing and finally, control.
Intelligent machines or control logics use sensor-derived data to
determine and work on various aspects of manufacturing.
Remote sensing with data transmission and aggregation, real-time
communications throughout the factory, distributed control systems
and machine learning algorithms – all of these function together to
monitor electromechanical systems such as motors, transport systems
and robots, along with field transmitters and actuators, and help
detecting anomalies in complex and uncertain manufacturing
environments (like the newly developing renewable energy sector), to
optimize present and future operations.
Cloud-based algorithms may evaluate millions of parts currently in
use worldwide and through their full life cycle in the field.
The scope of Instrumentation pertains not only to a core industry, but to
any self-sufficient system (such as biomedical inst. devices)
10. 1. .
2. Briefly, Industry 4.0 derives and processes and analyses data from
every block of the chain, to enhance the capabilities of the entire
process line. It not only consists of technical expertise, but logistics,
sales, market research and prediction and business intelligence, as a
whole.
However, the base of the 4th Industrial Revolution is purely data,
machine learning, and advanced control algorithms, based on which all
auxiliary teams make their decisions.
The idea of device/machine-factory floor-control room integration is
impossible without advanced sensors and telemetry. Complex and
custom manufacturing would be possible with adaptive control. A
vehicle suggesting repairs will not be able to do so without ML and
data analytics, and the team receiving faulty parts’ data would also
require data analytics.
Furthermore, a paper suggests a ML algorithm collect faulty designs,
analyze their design patents, and suggest revamped designs combining
all of them.
11. Why manufacturers need this
transformation?• Helps become more flexible and react to changes in the market easier
(adaptive manufacturing, just like adaptive control with external
changes)
• Increases the speed of R&D and being very consumer centered, it
leads to faster design processes. It allows every customer to have
specific customizations.
• Having connected devices mean lesser headache and more execution
of maintenance. Usage of efficient robots increase and workers
evolve.
• Industry 4.0 market scope, practicality and maturity will drive
clients to gain comprehensive knowledge of the entire structure, from
birth to recycling.
• Industry 4.0 along promises more revenue at lesser running costs,
lesser downtime, along with other promises. Eg., by using sensors’
12. The Role of Humans
With increasing computer dependency and decisions being taken by
artificial brains, it is natural of humans to pay sudden heed to Darwin’s
‘Survival of the Fittest’.
But to think about it, all the advanced robots (which do fail and need
human maintenance) can never perform tasks it has not been
trained for.
An object picking (the simplest of tasks) algorithm cannot control AC
temperature and vice versa. Artificial General Intelligence is a
distant future, with self-driving cars unable to distinguish man from
lamp posts.
Humans are the ones training these algorithms. We need these robots in
places we cannot be, but very much need to be in. And they need us, to be
controlled, monitored and adjusted according to our needs. ‘A
collaboration, using machinery to support human intuition, is the
actual aim’-CISCO.
AI is replacing repetitive jobs, and making room for more innovation from
13. Bottlenecks
Large data sets are needed for this, hence data storage capacities
must be huge.
Data gathering, cleaning and processing becomes a tedious task, only
to be followed by algorithm training.
Initial cost of setup and integration would be massive.
Since it’s a cyber-physical system, it needs to be vigilant of hackers.
Environmental and ethical dangers are present.
However, although there can be a few more bottlenecks to the
realisation of Industry 4.0, the net product of closed loop system
and feedback, as was illustrated in theory and test cases,
increases efficiency many-fold, from various aspects like market
value, investments, process line, logistics, and more.
Notes de l'éditeur
ML-Hence, such an algorithm develops the intelligence of a system set for specific task(s). With the sole purpose of enhancing human intuition, such ML algorithms can help detecting anomalies, present and future, in complex and uncertain process lines.
ANALYTICS-Provides us an opportunity to address issues that come up from the advanced sensors and ML algorithms
Intelligence of a system is it’s innate ability to analyse, decide and be capable of achieving sustained desired behaviour