Agile Mumbai 2022 Maintenance 4.0: Leveraging AI for Optimization of Maintenance Function Shwetank Sharad Managing Partner, Vitti AI

1. Maintenance 4.0:
Leveraging AI for
Optimization of
Maintenance Function
Digital Intelligence for Optimised Maintenance
Commercial in Confidence
www.vittiai.com
Imagine AI Innovation
http://www.agilemumbai.com/

2. Contents
Industry 4.0 – Evolution
& Global Adoption
Maintenance 4.0
Industry Case
Studies
Augmenting
Maintenance
Technicians
01
03
05
02
04
4.0
Industry 4.0 –
India Adoption
2

3. 6
Industry 4.0
Evolution & GlobalAdoption
01
3

4. Evolution of Technology in Manufacturing…
Role
Impact
"Mass Production for
Global Markets”
“Humans usingmachines
for mass production”
“Cyber Physical System–
Connected Machines”
“Continuous learning /
customized massproduction”
“Assembly Line– Focusing
on One Task/Person”
“Move towardsjob
specialization”
Efficient Manufacturing –
Reduce ManpowerDependency
Faster Product toMarket
profitable manufacturing
18th
Century
Industry 1.0
Mechanical
production
equipment powered
by steam and water
Industry 2.0
Mass prodcution
assembly lines requiring
labor and electrical
energy
Industry 3.0
Automated
production using
electronics andIT
Industry 4.0
Intelligent production
incorporated with IoT,
cloud technology and
big data
19th
Century 20th
Century Today
Industry 4.0 originatedin2011fromaproject inthehigh-techstrategyoftheGermangovt….Transition from 5% of manufacturing IT
spend to 20% by 2021, a 9.6X rise, driven by smart solutions and business sustainability needs during 2011-21…
4

5. Key Emerging Technologies Enabling Industry 4.0 Deployment…
US, China, India, Brazil, UK are planning $100+ Bn new investments in IoT, AI/ML, IT- OT integration, Robotics &Digital Twin
MATURED TECHNOLOGIES
$50 - $60 Bn @10% CAGR
MATURING – NEXT 5 YEARS
$4 - $5 Bn @30% CAGR
EXPANDING TECHNOLOGIES NASCENT TECHNOLOGIES
$30 - $40 Bn @15% CAGR $3 - $4 Bn @25% CAGR
• Cloud Computing
• Industrial Robots
• Internet ofThings (IOT)
• AI inManufacturing
• 3D Printing
• 4D Printing
• Quantum Computing
• Cyber-Physical Systems
• Advanced Human-Machine Interface
• Exoskeleton/Man-Machine
• Cybersecurity Technology
• AR/VR inManufacturing
• Big Data& Analytics
• Wearables &Sensors
• Digital Twin
• 5G inManufacturing
• Edge Computing
• Blockchain inManufacturing
Sources: IIoT World,International Federationof Robotics,The Manufacturing Institute 5

6. Visible Supply Chains Location Agnostic Command andControl Intuitive products and flexible service models
• Traceability of suppliers /material
• Predictability of potential disruptions
• Multisite integration with central
control towers
• CPS-equipped connected products that
enhance usability experience
• Smart Contracts – Digital contracts/SLAs
• Smart Procurement – AI-based supplier
risk management
• Smart Machines – Legacy retrofitting; self-
organizing and correcting machines; digital
twins for remote monitoring
• Smart Process Line – Process automation
to self-optimzing process lines; intelligent
robotics: cobots and HMI; data integration
across MES, SCM, CRM and procurement
• Smart Services – AR/VR- based remote
servicing; predicitve condition
monitoring andmaintenance
• Smart Resourcing – Self-adjusting HMI
and robotic integration; AR/VR based
operator assist
• CPS Equipped – Products equipped with
embedded IoT sensors, self-learning and
self- optimizing capabilities using AI at
the Edge, connectivity tech for M2M
communication
• Smart Logistics – Movement tracing
and ML- based real-time route and
mode optimization
• Smart Warehousing – Autonomous
warehouses with robotics and HMI;
AI-based inventory, returns, reverse
logistics management
• New Data-Driven Business Models – M2M
data led predictive analytics aimed at
innovative employee & customer
experience
Smart Industry: Industry 4.0 is Transforming Operations, Supply Chain,
Customer Solutions…
6
Smart Sourcing
Smart Sourcing
Smart Supply Chain
Smart Services
Smart Operations
Smart Factory
Smart Solutions
Smart Products

7. • Digital representation of product or machine, helps in design, testing, simulations
Digital Twin
• Connecting factory objects like machines, vehicles, products for control & optimization
Connected Factory
• 3D Printing can support massive customizations and can increase flexibility
Flexible production
• Automation, Visualization using AR/VR improves man- machine coordination
• Remote monitoring with Sensors and Big Data helps in optimizing maintenance
Visualization & Process
Automation
Predictive
Maintenance
• Helps detect patterns in production or quality data, providing insights for optimization
Big Data
• Factory operating independently on self-learning algorithms, reduces operations cost
Autonomous Digital
Factory
• Sensors to track Products, Raw materials provides full transparency on production process
Track and Trace
Key Technology Features of Digital / Smart Factory…
7

8. Big Data
Decision-
Making
(Bosch
Automotive
China)
• Before: Operational data from the shop floor, such as machine cycle times or part failure
modes required a significant amount of manual collection and pre-processing. Continuous
shop-floor improvement activities were impacted.
• After: The Wuxi site, set up an industrial IoT framework, connecting newly installed
machine condition sensors and individual cutting tool information. They were able to
visualize the data, develop customizable reports with powerful analyses, including
diagnostic, predictive and prescriptive functions, leading to 10% output increase.
Democratized
Technology
At Shop Floor
• A large manufacturer had deployed Autonomous Mobile Robots (AMRs) for a point-to-
point material transfer workflow moving parts from kitting stations to an assembly cell.
• The AMR system employed Cloud Robotics Technology, so it provided a simple
interface that enabled the floor manager to set up & schedule additional workflows
between the kitting area & the new cell with a few clicks, without any support from the
IT staff.
• As a result the workers and local staff were able to increase their productivity.
Examples Of Global Companies Using Digital / Smart Factory Use Cases…
Minimal
Increment al
Cost to Add
Use Cases
(Microsoft
Manufacturi
ng, China)
• To ensure competitiveness of IT products & services (PC & other devices), Microsoft
transformed the manufacturing process at its factory in 3 waves: connection of equipment,
prediction using big data, application of machine learning to create cognitive manufacturing
lines.
• Using connected equipment & the capability to add new use-cases in a short time period,
company added machine learning algorithms for predictive yield improvement based on
production process data of individual components, yield improvement of 30% with the
completion of one use-case.
8

9. 13
Industry 4.0
India Adoption
02
9

10. Discrete Manufacturing – $4.8 Bn Process Manufacturing – $1.6 Bn
Discrete75%
Process
Manufacturing,
25%
Share of
Industry 4.0
Spending,
2021
65% 40%
25% 30%
8% 25%
2% 5%
• Indian Automakers stepped up investments in
Cloud and digital systems, shedding legacy IT
infrastructure
• Electronic component manufacturers in India
have invested heavily in Connected Technologies
like 5G & IIoT
• From retrofitting legacy machines on process
lines with IoT devices, to entirely autonomous
process lines monitored remotely via digital
thread – the discrete segment is capitalizing on
M2M data to manage end-to-end operations
• Indian pharmaceutical companies are
prioritizing Cloud-based modernization
with preference for“pay-per-use”
models
• 50% of the sector spends greater than
6% of its annual revenue on technology
spend and is in early or intermediate
stages of Industry 4.0 adoption
• Other process industries, like Chemicals,
are at early stages of Industry 4.0
deployment
Data andAnalytics Connectivity Tech Intelligent Automation Advanced DigiTech
Most India Industry 4.0 investments are currently in Cloud, IoT,
Big Data Analytics, Connectivity Tech & RPA…
Source: NASSCOMReportFeb2022 10

11. Industry 4.0 Use Cases by Value Chain Stages, Key Technologies Involved…
Real-TimeSupplier
Management
Real Time Order
Management –
IIoT and
MES/SCADA
integration
Predicitve Supplier
Performance– BDA,
AI/ML
Supplier Scenario
Planning,
Vulnerability
Assessment – AI/
ML, BDA
Sourcing Mix
Modeling/
Dynamic orFlexi-
Sourcing Strategy –
AI/ML, AR/VR,
Blockchain
Supplier Financing -
Blockchain
Predictive Planning
Predictive Demand
Planning – Edge
Devices, IoT, Big
Data,AI/ML
Real Time
Replanning and
Scheduling – ML,
BDA
Outcome-Based
Decision Modeling
– Blockchain, BDA,
AI
Traceability – IIoT
Platform (Cloud,
Edge Devices,
Sensors), Robotics,
AR/VR, Digital
Thread
Planning Production Operations
Upstream –Supplier Warehouse/Logistics Downstream –
Customers/Partners
Smart or Dark Factories: Smart Machines
Predictive Maintenance – Big
Data, Cloud, AI/ML, Edge
Devices
Remote Controlled Supervisory
or Maintenance Operations –
Connectivity Tech, Robotics,
Automation, Digital Twins
Smart Lines
Self-Optimizing Assembly Lines
–IIoT Platform, Automation,
AI/ML, Edge Devices, and
integrated OT Platforms
Flexi-Assembly Lines –
Digital Twins,Additive
Manufacturing
Smart Operators/ Services
Remote Floor Shop Monitoring
– Robotics, Automation, Digital
Twins, AR/VR,Drones
Integrated Logistics: Smart
warehouse/Logistics
Predictive Warehouse
Management – IIoT,
Robotics, Automation,
Connectivity Tech,Edge
Devices
Real-Time/ Predictive
Inventory
Management– IoT,
Edge Devices, AI/ML,
Drones, AR/VR,
Robotics
Freight-Sourcing
Decision and
IntegratedMulti-Modal
Logistics – IoT, AI/ML,
Connectivity Tech,BDA
Digital Customer Experience:
Smart Partners
Predicitve Distribution Planning
– Integrated CRM and SCM with
MES, BDA, AI/ML based
optimization
Customers/Partner Decision
Analytics – BDA, IIoT, Edge
Devices, Connectivity Tech
Hyperlocal or Last Mile Services
Micro-Fulfillment – Big Data,
Edge Devices,AI/ML, IoT
Real Time Location Data – IoT,
Connectivity Tech
Direct-to-Customer (D2C)
Services
Omnichannel strategy – Cloud,
BDA, AI/ML
Traceability – IoT, AR/VR, Digital
Thread
11

12. Industry 4.0 Case Study 1: Bajaj Auto…
Shopfloor Efficiency
Improvement
– Lowest running costs,
Can operate without a
cage in space constrained
areas.
Reduction in Ergonomic
Risks- Usage of Co-Bots,
thus reducing manual
stress, providing
Compact movement,
extremely flexible (all
axes + or – 360-degree
rotation) and
lightweight.
Safety - Eases work for
women workforce, with
30 patented force limiting
features built in
compliance with ISO TS
15066, Ceiling mount,
Wall mount or Floor
mount Co-Bots.
Zero annual
maintenance costs -
Reduced power
consumption and
retention of IP within the
company, organically
driving growth of the
organization.
SOLUTION
PROBLEM STATEMENT IMPACT – Smart Lines and Smart Operators/Services
Two-wheeler assembly
lines were highly labour
intensive, spatially
challenged. Around 50%
of the workforce were
women, who found it
difficult to operate
intensive assembly lines.
Bajaj auto wanted to:
• Reduce ergonomic risks
to the employees.
• Find a standardized
automation solution.
Tech Solution Deployed – Partnered
with Universal Robots after 3 months
of extensive testing of Universal
Robots’ cobots:
• Ceiling Mounted Cobots –
Diminished the challenge of space
constraint .
• Reduction in Redundancy-Led
Fatigue and Errors – Completing the
repetitive movements that required
precision.
• Standardization & New Decal
Applications – Catered to multi-
modelling adaptability and tasks
that required flexibility,productivity
and reliability.
12

13. Industry 4.0 Case Study 2: TVS Motors…
Traceability – IoT-based
product traceability
through the flow cycle to
assess quality of the
material in real- time, for
upstream and
downstream information
and associated
decisions.
Skill Matrix - Maintain a
digital trace of operator
performance. Enable the
identification of a skill
matrix and identify any
exceptions that could
impact product quality.
OEE Improvement –
Real-time insight into
parameters that impact
line productivity, such
as line rates, loss, and
quality analysis across
multiple levels of
operations.
Predictive
Maintenance –
Statistical analysis of
product quality
parameters, coupled
with real-time machine
condition data
enabled predictive
maintenance and
minimized costly stalls.
TVS Motor’s assembly
line machines were not
connected, and data
from machines was not
flowing into the data
lake,
impacting traceability,
visibility and
predictability at the
shopfloor
TVS wanted to Build an
integrated
manufacturing data lake,
Integrate machine data
on shop floor, Move data
from other IT systems
on the shopfloor into
the data lake
Tech Solution Deployed – Partnered
with Altizon and deployed the
provider’s proprietary IoT platform
and Digital Factory hybrid solution
with an Edge solution inside the TVS
network. The solution stack included:
• Edge Computing: Distributed
computing platform that allows IIoT
data to be processed closer to the
edge of thenetwork.
• Connected Work: Integrated data
lake for storing and processing all
machine and manufacturing data for
further analytics.
• Digital Factory: Unified digital
manufacturing platform powered
by IoT and out-of-the box apps for
monitoring, measuring, analyzing
and predicting outcomes using AI.
SOLUTION IMPACT – Digital Customer Experience
13
PROBLEM STATEMENT

14. Industry 4.0 Case Study 3: Kia Motors…
Real-Time Transaction
Visibility Via Digitalized
Showroom – Live Stream
Showroom capability
demonstrated continued
commitment to tailor the
car-buying journey to the
demands of the
customers with virtual
viewings.
Transparency – Customers
could digitally make
buying decisions along
with their family members
logged in from multiple
geographies at the same
time, recreating a physical
showroom experience.
Customer Connectedness –
Digital consultation
services by established
dealers gave customers a
sense of reliability and
security while making
purchase decisions during
a pandemic.
SOLUTION IMPACT – Digital Customer Experience
During the pandemic,
sales and services
practically ceased
overnight, affecting
customer connect and
demand forecasting.
Challenge was to keep the
potential customers
engaged so that once the
industry picks up, they
turn buyers.
Biggest challenge that KIA
faced as a new player is
that they were not able
to demonstrate their
product due to the
restrictions set during the
coronavirus lockdowns
3D Configurators – Kia Motors deployed an
AR/VR based 3D configurator solution to create a
digital catalogue of the showcased vehicle and a
digital specifications board for every vehicle
category in their product portfolio at the Mumbai
showroom.
3D Configurator Customer Zone – Enabled
customers to customize and design their favorite
Kia cars and witness their intricate details. The
content displayed in the showroom was remotely
controlled centrally.
‘Kia Digi-Connect’: Anindustry-first
video-based live sales consultation solution
website integrated with the company’s CRM
system, provided customers options of 360-
degree virtual experience through video calls and
screen sharing, along with sharing of digital
brochures and dynamic pricing.
6000+ pre-bookings made on Day 1 of opening from pandemic lockdown 14
PROBLEM STATEMENT

15. Industry 4.0 Case Study 4: Nokia…
Real-Time Visibility for Central
Control - Screens display real-
time information from the
various sensors that monitor
every process across the
factory floor. The data from
these sensors runs through
Microsoft’s Azure platform, and
the system allows managers to
track parts by serial number as
they move through the factory,
physically or via a digital twin
platform
Automation of Quality
Testing Processes –
Maintains a digital trace of
operator performance. The
system allows the company
to pinpoint exactly where
something went wrong and
fix the problem quickly.
Low Latency and Real-
Time Data Capture
- Deploying a private
wireless network
helped in greater
agility on the shop
floor to
accommodate the
rising need for line
configuration
changes.
Fully Remote-
Controlled Operations -
Digital twin of the
factory enabled
automation of the
production flow and
remote operation and
maintenance.
SOLUTION IMPACT – Across Value Chain
31% labor time reduction through robotic automation. 31,000-man hours saved through RPA. 16% OEE improvement
Nokia’s factory in Chennai,
yielding 16 billion chip
mounts per year, faced
severe external supply
chain shocks due to
competition from China.
Needed to cut costs and
drive efficiency in the
supply chain.
Pressure to be agile and
responsive in a volatile
market was high.
Nokia battled a monolithic
IT system as a result of
merging legacies of
Siemens, Alcatel- Lucent,
Nortel, Motorola and
Panasonic.
Tech Solution Deployed – Nokia has built a
private wireless network based on 4G LTE.
• Autonomous Guided Vehicles/Autonomous
Intelligent Vehicles: Material flows
warehouses driven by intelligent, autonomous
vehicles. To enable the seamless movement of
the AGVs, AIVs and also to track the assets
moving around the shop floor, High Accuracy
Indoor Positioning (HAIP) system using
sensors, IoT gateways and private LTE
platform.
• ‘’Pick to Light System” for Inventory Control–
All parts stored in racks across the store, and
when the part if requested at a production
station or testing area, an operator enters
the data into the asset management system
and a light goes on at the specific rack in the
warehouse to make it easy to locate the part in
the specific storage rack, and further transport
it to the required place on the shop floor.
15
PROBLEM STATEMENT

16. Major Technology Investments by Global and Large Manufacturers…
Ola Electric with Siemens - $300 Mn for building India’s
most advanced electric vehicle manufacturing facility
Bosch Home Appliances - €100 Mn spend by 2025 on IoT-
based solutions + smart refrigerator factory in India
Henkel Adhesives - €50 Mn into a smart factory in Pune,
equipped with end-to-end quality and track-and-trace
capabilities using digitalized workflows
M&M and Bosch – Partnership to develop Mahindra’s
connected vehicle platform “AdrenoX Connect”with
integrated platforms enabling flexible swichovers
4
Vedanta and GE – Partnership to digitalize India’s first
Aluminium smelting plant deploying Digital Twin technology
built on GE’s Predix Platform
1
2
3
5
2
5
16

17. Indian Industry 4.0 Provider Landscape: Illustrative, Not-Exhaustive..
• Connected Building Blocks
• Hosting Industrial IoT Platforms Analytics
• Microchips Sensors Connected Hardware
• System Integrators Cybersecurity
Aiding I.40 Technologies
Collaborative Robots/ Robotics
Universal Robots
AR/VR Drones/ UAV’s
Additive Manufacturing/Connected Machine
Vision
AR/VR Drones/ UAV’s
Microchips Sensors Connected H/W
System Integrators Cybersecurity
Additive Manufacturing/Connected
Machine Vision
Connected Building Blocks Aiding I 4.0 Technologies
Hosting Industrial IoT Platforms Analytics Collaborative Robots
17

18. 39
Maintenance 4.0
03
18

19. Maintenance 4.0: SMART MAINTENANCE …
Preventive and
Proactive Maintenance
Condition Monitoring
Leaner Maintenance
Automation of Clerical
Maintenance Tasks
Maintenance 4.0 describes a specific stream of innovation within Industry 4.0 focusing on Maintenance.
Cornerstone of Maintenance 4.0:
19

20. Maintenance Strategies: A Continuum…
*
Original
equipment
effectiveness
Poor maintenance strategies can reduce a plant’s overall productive capacity between 5 and 20 percent.
Unplanned downtime costs industrial manufacturers an estimated $50 billion each year.
Predictive Maintenance (PdM) is the most efficient maintenance strategy available – a Gold Standard. PdM can
increase equipment uptime by 10–20 percent and reduce maintenance costs by 5–10 percent…
20

21. Predictive Maintenance: The Physical-Digital-Physical Loop…
Real-time access to data and intelligence is driven by continuous, cyclical flow of information
between physical and digital world through iterative series of three steps, physical-to-digital-to-physical loop
Source: Deloitte analysis. 21

22. The Predictive Maintenance Process…
Jim a factory floor supervisor in a manufacturing plant in charge of maintaining numerous machines.
Source: Deloitte analysis.
22

23. Understanding Technologies That Enable PdM Process Deployment…
Data Integration + Augmented Intelligence + Edge Computing + Augmented Behavior Using Wearables and Mixed Reality
23

24. (PredictionWithPrecision)
Vibration3D
Acoustic Emission
MagneticFlux
Humidity
TrueRPM
Temperature
World's First 6-in-1 Sensor
.
MachineDoctor is easy to configure, It feeds directly into Analysis Software RotationLF
Insights Diagnosis Action
Anomaly
Detection
Fault
Diagnosis
Time toFault
Prediction Action
MachineDoctor™
RotationLF™
Analysis Software
=
AUTOMATED
End 2 End
SOLUTION FOR
REMAINING USEFUL
LIFE PREDICTION
WITH 99%
ACCURACY
24

25. THE SOLUTION&RESULT
BACKGROUND
L&T Nabha Power Plant is 700 MW thermal power plant in Punjab. Unplanned shutdown maintenance impacts profitability.
THE CHALLENGE
The Condensate Cooling Water (CCW) pump is a horizontal vane pump operating at up to 1650 m3/hr. Each day this pump is offline, it costs
the plant $250,000 in lost revenue. L&T needed a predictive maintenance solution to detect faults at an early stage and provide a reliable
prediction of Remaining Useful Life (RUL).
Nanoprecise proposed rotation LF system, installed 24 wireless sensors as a part of a pilot project on air compressors, CCW Pumps, Fans. Sensors
sending data to SaaS-based platform using Edge and Cloud computing. RotationLF platform analysed data using algorithms. Six weeks later
AI alerted that a vane fault had been detected on the pump, causing cavitation. The fault frequency depicted in the below plot indicates an
early stage failure. As cavitation damage to vanes and housing progressed, the amplitude increased and the RUL decreased. Anomaly in the
pattern alerted plant staff about this unusual trend automatically through mobile text & email alerts. The maintenance team used a hand-
held vibration meter to verify the fault detected by RotationLF and then partially disassembled the pump to visually confirm that the vanes
were damaged. Atemporary repair was made to the damaged vanes before putting the pump back in service. The RUL prediction of
25 days to failure provided sufficient time to schedule the part replacement and prevented shut down.
PdMCaseStudy1: 25 Days Of Remaining Useful Life (RUL) Prediction…
25

26. Trenitalia was able to maximize
the brake pads’ useful life while
reducing the number of needed
spares.
Decrease downtime by 5–8
percent, reduce annual
maintenance spend of $1.3 billion
by an estimated 8–10 percent,
saving $100 million per year.
More trains have run on time, so
more passengers are happier.
SOLUTION
PROBLEM STATEMENT IMPACT
Italian train operator,
Trenitalia, had to
remove each one of its
1,600 trains from
service not just for
scheduled maintenance
and when a train failed
unexpectedly.This
created delays,
performance penalties,
annoyed passengers.
Trenitalia added hundreds
of onboard sensors on
1,500 locomotives as part
of a three-year
maintenance improvement
initiative. Data were
transmitted to private
cloud storage in near-real
time, where diagnostic
analytics provided advance
warning of the failure of
parts such as brake pads.
PdM Case Study2: TransportationIndustry…
The Impact Of PdM: Not Just Operational Efficiency Improvement, But Also Business Growth Through Better
Product Quality Resulting In Differentiation & Higher Customer Satisfaction…
26

27. 21
04
Augmenting Maintenance
Technician
27

28. The Challenge For Utility Operators…
To optimise the availability / uptime of generating
and distribution assets
Equipment downtime results in:
Revenue loss
Brand
damage
7
Additional costs
Diminishes customer
confidence
28

29. Business Impact For The Energy Industry…
At a medium-size
power plant, 1%
change in plant
availability could have
a $3.5 million revenue
impact
Fewer than 24% of
operators describe their
maintenance approach
as being a predictive
one
According to
Aberdeen,
unplanned Utility
downtime can cost
companies as much
$260,000 per hour
Capital expenditures
continue to rise, with an
increase in 2018 of 14
percent, to reach an all-
time high of $133.8
billion for the 50 electric
and gas utilities S&P
Global tracks annually.
72% of Organizations Target Zero Unplanned Downtime…
29

30. Root Cause Analysis for High O & M Cost…
• Inaccurate diagnosis & prognosis
• Skills, knowledge, experience and support limitations
• Insufficient service procedures
• Correct information not available at point of failure
• Premature component/subsystem failures
Causes:
• Prolonged System Downtime
• Large no. of Unplanned Maintenance
Events
30

31. ERP Maintenance
Management
Document
Management
Handwritten
Forms
Electronic
Forms
Manual
Reports
Spares
Management
Tool
Requests
Senior
technicians
retire: Knowledge
Retention
Sensors
SCADA
Inspection
Management Emails
Multimedia
(e.g. Photos)
Phone calls,
Memory recall
The Challenge In Perspective…
Critical knowledge is inaccessible due to
Fragmented Systems
Man-hours are wasted on
Manual and Disconnected Processes
Inefficient use of technicians’ time with
Lack of Automation
Young digital natives require rapid technician
Upskilling and Guidance
Expectations of rich
digital experience
31

32. Transforming Maintenance With AI & Machine Learning…
AI works like a human brain,
but with advanced analytic and
processing power
Artificial Intelligence (including Machine Learning and
Neural Networks) enables a system to learn, self-improve and
interpret as it performs a task, refining over time through
strategic trial and error.
Natural Language Processing fills the gap between human
communication and computer understanding.
Our unique utilisation of these technologies allows us to
perform tasks such as skilled analysis, pattern recognition,
image and speech recognition, analysis of massive amounts of
data, and sophisticated decision making.
Big Data allows systematic extraction of information from large
and complex datasets.
32

33. Requires a system that:
Provides the right technical
information at the right
time and place to reduce
risk of maintenance error
Provides all technical
information at one place,
Instructs and guides the
technician
Advises if the right tools
are available and their
location
Enables real-time
collaboration between
technicians
Achieving Optimized Maintenance…
Ensures optimisation of
the use of spares
through data driven
insights
Continuously learns and
optimises the
maintenance process
33

34. Provides a natural language interface to ERP,
Maintenance and Document Management Systems,
so as to extend and enhance their capabilities
Utilises any device (phones, tablets, etc.) to provide
solutions at the right time and place
Is a digital intelligent system that learns from:
• Asset behaviour
• Technician behaviour
• Organisational Data
Guides the technician at all stages of the maintenance
process, providing advanced troubleshooting
capabilities
The LexX Platform: Solution for Optimized Maintenance…
LexX is an intelligent digital
colleague, empowering
technicians by bringing
knowledge, information
experience to their fingertips
34

35. How This Technology Works…
LexX utilises AI and Machine
Learning to learn over time
from your organisational
data, technician interaction,
and equipment behaviour, so
that troubleshooting is
continually refined.
LexX Ingests all technical
information (digital and
handwritten) and
structures the data in a
unique manner to
facilitate the AI and NLP
capabilities of LexX.
This information is used to
provide solutions and guide the
technician through Natural
Language interaction (in
Conversational style, like
having an instant messaging
discussion with a human)
35

36. Intelligent
Maintenance
Assistant
LexX Core
(Eg. User-friendly
interface, troubleshooting,
repairs, parts lookup )
LexX Platform Modules…
Data Pipeline Automated ingestion
Handwritten (OCR)
Structured Storage
Representational State
Transfer Protocol
Standardized
Database
Learning
Systems
Search
Algorithms
Intelligent
Knowledge Base
Specialised
Business
Functions
Eg., Classification Reporting
Data Analytics
Eg. Alerts
Algorithmic
Operations Support
API
Layer
How Is It Powered?
Ops Analytics
Eg.,
predictors
and alerts
Inputs
Fault
Logs
Service
Manuals
Schematics Events &
Alarms
Tools,
Spares
Work
Orders
Synergy
(Eg. Workgroups and
notifications)
Performance and
Reliability Dashboards
(Eg. Visualisation and
reporting)
Digital Tools
(Eg. Estimation image
recognition, drone
images)
Outputs
Secure Cloud Hosted
36

37. LexX Platform Features…
What Do You Get?
Intelligent Maintenance
Assistant
Configurable for Client
need or specific
maintenance use-cases
(i.e. fault management,
parts lookup)
Synergy Suite
Bring the team to context
for help, advice, or
approval
Performance and
Reliability Dashboards
Configurable dashboards
for performance
management and reliability
engineering user groups
Digital Tools
Tools that empower the
technician, such as
automation for repetitive
tasks (i.e., estimation,
image recognition, drone
images)
37

38. Ultimate
Success
Indicator:
Reduces O & M
Costs /
Improves
Machine
Availability
The Measurable Value LexX Brings to Operations & Maintenance…
Time to Troubleshoot (MTTR) improves: Technicians
get it right the first time as best practices get
consolidated; maintenance procedures get improved.
Spare Parts Utilization improves by minimizing
unnecessary use of parts.
Time on Tool improves: With less technician time
spent on the phone, searching for, retrieving and
classifying work orders, LexX improves and the
capacity of the existing workforce to improve
machine availability / Uptime. Empowers the
modern-day technician and reduces
labor/contractor costs.
Time to Productivity improves: Less
time is needed to train and job-shadow
technicians to high competency levels,
increasing the scalability and flexibility of
the workforce.
Knowledge Retention Improves.
Enhanced forecasting accuracy. Improved data
and configuration management ensures the integrity
and currency of the assets, components, spares and
resources related to maintenance. This allows the
organization to optimize its inventory through
reliable long term-planning and forecasting. 38

39. The Way Technicians Work Today How Technicians Should Work Today
Digital Transformation With LexX Enabled Maintenance
How Technicians Will Work Tomorrow
The Future of Maintenance with LexX
39

40. LexX Deployment Case Study 1: Energy Australia
Background
The Energy Australia site was Mt Piper Power Station, which
comprises of two 700MW coal-fired steam turbine generators and
can meet the energy needs of 1.18 million homes in NSW every year.
Pump Repairs could take days to weeks depending on need to
isolate, transfer pump to workshop for repairs, and parts availability.
In the meantime, operators are always applying pressure to have
the equipment brought back in to service. In some cases, a spare
could be brought in while the repairs are happening, but if there
aren't any spares then the fitters work overtime.
Objective of POV
Prove that the Lexx platform can help:
o Improve availability/up-time of generating power plants
o Demonstrate ease of use, without being intrusive
o Demonstrate availability of right information at right time and
place
40

41. Pump Repair Process
Work order comes in
•If there's an issue related
e.g. pump leaking or not
starting, technician will go in-
field.
•Based on experience, and
some ad hoc preliminary
investigation, cause may be
known.
•Some information might be
on the work order, or might
be known to the operators
e.g. temperature;
•If not found in the WO,
technician needs to check in
with operators.
If the problem can't be
fixed immediately
(which it generally
can't),
•Coordinate with operators to
have the equipment
(including components like
valves) 'isolated' for some
period
•Technician estimates how
long he thinks the equipment
needs to be isolated
(generally based on intuition)
•Depending on urgency the
isolation could be scheduled
for later in the day/week or
immediately
If the pump needs to
be removed from its
location (which it
generally does) then it
must be taken to the
workshop.
•Arrange for a crane and
rigger to be hired, and
permits granted.
•If pump is in a pit, then a
confined space permit needs
to be obtained.
•The administrative side of
things can take quite a while.
Once in the workshop,
the equipment gets
pulled down.
•Risks here that dismantling it
improperly could cause more
problems. Some root causes
don't require dismantling.
•If it's discovered that
something is broken, then
replacement parts must be
retrieved from the store,
•If there's no spares in store
they must be ordered. Also,
critical to know exactly which
part number it is..
Once these repairs are
done,
•A crane and rigger must
again be hired, permits
granted,
•Window of time arranged
with operators, to have the
equipment re-installed
•The pump is brought back in
to service.
The role of LexX here is to:
• Help the technician follow the correct procedure for identifying the root causes
• Utilising historical or OEM provided troubleshooting guides
• Identify the correct parts
• Identify the nature of issue by contextually providing to the technician
information from the work order, history, logs, manuals and any tribal
knowledge existing in the organisation
Scenarios tested with LexX:
1. Reactive maintenance: pump not starting (due to electrical
issue but thought to be mechanical)
2. Planned maintenance: check oil level and change oil as
required
3. Main belt repair
1
2 3
Understand Isolate/Fix Shift to Workshop Workshop Restore to Ops
41

42. Issue Description:
This event has occurred in the past, troubleshooting for a
sulfuric acid pump led to a no fault was found (NFF).
Since the issue was not mechanical but electrical, engineer
and the team weren't able to diagnose the problem through
isolation and disassembly.
The issue turned out to be due to an associated solenoid, for
which repairs could have been made without removing or
tearing-down the pump.
Scenarios 1: Reactive maintenance: pump not starting (due to
electrical issue but was thought to be mechanical)
Without LexX With LexX
• Turnaround time: 2
working days, 70%
reduction in
turnaround time
• Fewer NFF events
• Less time spent on
admin, more time-
on-tool
• Early intervention
• Turnaround time: 7 working
days
• Technician went too far into
mechanical troubleshooting.
42

43. Issue Description
Work order comes in
requesting a check of the oil
level for a pump.
In some work orders, all the
instructions are already
broken down step-by-step,
but level of detail is
inconsistent (depends on
who wrote the work order).
When changing oil, you
need the exact oil type for
each pump.
If the wrong oil is used then
the damage can be
catastrophic, requiring the
pump to be trashed.
Test Scenarios 2: Planned maintenance: check oil level and change
oil as required
Without LexX: Cycle time: 1 hour
• There is a list with pump/oil
information on the computer;
technician would need to look
up the pump and then the oil.
• However, the software system
they use is very difficult to use
and technicians tend to avoid it
as much as possible.
• Technician has stuck a paper
version of the list to his locker,
• If that doesn't cover it, he asks
a colleague with a better
memory for which pump takes
which oil.
Results with LexX: Cycle time: 10 minutes, 80 %
Reduction
• Gets to pump and oil in an instant
• "Takes the guesswork out of it."
• No paper on his locker
• No consultation needed
• With LexX technician can entrust the job to an
apprentice, focus elsewhere.
43

44. Test Scenarios 3: Main Belt Repair
Issue Description
Apprentice provides support to
experienced technician for a repair
to the main belt and had a six-hour
window for the job.
As it was his first time doing the job,
there was a bit of anxiety and
uncertainty about what needed to
be done.
Experienced technician knew certain
things ahead of time based on
experience, like the dimensions of
the bolts, what tools to use, what
type of oil/grease etc.
Apprentice wouldn't have known
any of this stuff, so there was no
chance he could have made the
repair on his own.
Anxiety, uncertainty, lack of
independence
Results:
Using the LexX platform, apprentice says he has been able to
prep for the repair ahead of time.
He is able to look at photos, diagrams, identify which tools he
would need, and be able to be more on-the-ball and proactive.
He has been able to take the initiative rather than waiting for
instruction from the experienced technician.
The benefits highlight by the apprentice were, that having LexX
platform will make him:
· Confident, proactive, "backing himself";
· Less reliant on senior staff;
· Spend less time on training; and writing weekly
report.
Quote: “Great training tool, to review procedures, schematics etc. in own time, and
in preparation for jobs. Liked the 'Google' style of searching because it would save
time trying to find the right information”.
44

45. Background
Lexx Technologies was the winner of a US Energy Provider’s Global Innovation
Award in 2019. From this, a 12-week Pilot commenced to validate a set of field
troubleshooting and performance management objectives, ahead of wider
deployment. This Pilot took place at the Energy Provider’s Wind Farm located in
Illinois.
Objectives of the Pilot
The Energy Provider’s key initiatives:
Provide a user-friendly troubleshooter for desktop and mobile for
technicians in the field
Provide a troubleshooter configured for support, targeting fault
resolution time
Enable Algorithmic Work Order Classification by resolution type and
auto-correction functionality
Enable reports addressing performance analysis questions and spare and tools
consumption trends
Results: Estimated Efficiency Savings
From the use of Lexx during the Pilot:
40% Troubleshooting Time Reduction
40% Time-To-Productivity Savings
10% Part Replacement Prevention
LexX Deployment Case Study 2: US Wind Farm
45

46. Lexx was able to ingest a
large amount and array
of data (including OEM
manuals, work
instructions schematics,
SAP data, work orders,
notifications, fault codes,
OPs data, ROCC fault
logs, events & alarms,
inventory/spares, and
SCADA)
Data Ingestion
Lexx improved the quality of data, andtherefore
performance analysis
Lexx was able to link various datasets accuratelyand
efficiently to support performanceanalysis
Lexx automated mundane taskssuchascorrecting,
improving quality and classifying rawdata
Lexx captured accurate data from troubleshooting to
assist performance analysis
Lexx
empowered
and reduced
the workload
of lead
technicians
Showed that
Lexx can link
troubleshooti
ng and
performance
analysis
Auto correct and
auto classification
returned high
accuracy, with some
results better than
human
interpretations
Provided a single
searchable
interface to
disconnected
datasets linked
to faults and
turbines
The learning
feature
enabled
better capture
of turbine
behaviour and
technician
behaviour
The Value | Key Findings
Overall, Lexx enabled the availability and efficiency of critical equipment
The Technician’s Response
84% agreed Lexx improved reliance on data,
takingthe guesswork out of troubleshooting
86% agreed Lexx increased the reliance on digital
tools while on the tower
90% agreed Lexx improved quality of repairs and
allowed better utilisation of individual skills of
technicians
46

47. LexX can be used in the following
maintenance & operational areas:
Troubleshooting;
Planned, preventive and
breakdown maintenance;
Performance diagnostics and
analysis;
Compliance & reporting;
Issue and problem management;
Asset life cycle management;
Training and simulation;
And publication management
…and even more
How Technology Can Be Used…
The user interface is
customised around
individual Client
requirements and
processes
Demo link:
https://www.youtube.com/channel/UCdOmj3fGvKiGrvpCyxicrkA
47

48. Thank You
CONTACT DETAILS
E
M 7208408051
W www.vittiai.com
shwetank@vittiai.com
www.vittiai.com
Imagine AI Innovation
http://www.agilemumbai.com/