IoT-Enabled Predictive Maintenance

1© Cloudera, Inc. All rights reserved.
How Leading Organizations are using Data
Analytics and Machine Learning to enable IoT
IoT Enabled Predictive
Maintenance

Your Speakers for Today…
Vijay Raja
Solutions Marketing Lead, IoT
Cloudera
Chellury (Ram) Sastry
IoT Practice Lead
Tata Consultancy Services
Vinod J Nair
Solution Delivery Manager, IoT
Tata Consultancy Services

The Need for Predictive Maintenance
Unplanned outages due to
equipment failure, leading to
increased costs & lowered
productivity
Limited to no analytic modeling
capabilities to predict and prevent
issues before they impact
Unplanned Outages Fundamentally Reactive
Performing preventative
maintenance regardless of
condition can lead to
unnecessary costs and delays
Unnecessary Work

The Foundation of PredictiveMaintenance
• Sensors continuously monitor
critical attributes – temp,
pressure, vibrations/sec etc.
• Use machine learning to
detect anomalies or patterns
that are indicative of failure
• Real-time alerts
• Early intervention as soon as
initial signs of failure are
detected

The Challenge – Managing Machine Data
Data can come from a variety of “siloed” sources
• Massive volumes of intermittent data streams
• Generated from a variety of data sources
• Predominantly time-series
• Can come in streams (real-time) or batches
• Diverse data structures and schemas
• Some of it may be perishable
Combining sensor data with contextual data is the
key to value creation from IoT

Predictive Maintenance – The Data Value Chain
Data Sources Data Ingest & Storage Data Processing Machine Learning Analytics/ BI
• Diverse Data Sources
• Intermittent data streams
• Batch/ Streaming
• Diverse schema/ formats
• Real-Time Data Ingest
• Big-Data platform
• Unlimited Storage
• Low Cost/ TB
• In-memory processing
• Real-Time Processing
• Data enrichment
• Data contextualization
• Machine Learning libraries
• Unsupervised M/L
• Supervised M/L
• Iterative data modeling
• Data Visualization
• Data Science
• SQL Analytics
• Search

Cloudera Enterprise – The Data Mgmt. Platform for IoT
Sensors/ IoT
Data Sources
Enterprise Data
Sources
External Sources
BI Solutions Real-Time
Apps
Search Data Science
Workbench
SQL
Machine
Learning
Data Center
Cloud
Sensor/ IoT Data
IoT Gateway
• Data Storage
• Data Processing
• Machine Learning
• Real-time Analytics
OPERATIONS
Cloudera Manager
Cloudera Director
DATA
MANAGEMENT
Cloudera Navigator
Encrypt and KeyTrustee
Optimizer
BATCH
Sqoop
REAL-TIME
Kafka, Flume
PROCESS, ANALYZE, SERVE
UNIFIED SERVICES
RESOURCE MANAGEMENT
YARN
SECURITY
Sentry, RecordService
FILESYSTEM
HDFS
RELATIONAL
Kudu
NoSQL
HBase
STORE
INTEGRATE
BATCH
Spark, Hive, Pig
MapReduce
STREAM
Spark
SQL
Impala
SEARCH
Solr
SDK
Partners

The Cloudera Platform for IoT – Data Mgmt. Value Chain
Data Sources Data Ingest Data Storage & Processing
Serving, Analytics &
Machine Learning
ENTERPRISE DATA HUB
Apache Kafka
Stream or batch ingestion of IoT data
Apache Sqoop
Ingestion of data from relational sources
Apache Hadoop
Storage (HDFS) & deep batch processing
Apache Kudu
Storage & serving for fast changing data
Apache HBase
NoSQL data store for real time
applications
Apache Impala
MPP SQL for fast analytics
Cloudera Search
Real time searchConnected Things/ Data
Sources
Structured Data Sources Security, Scalability & Easy Management
Deployment Flexibility:
Datacenter Cloud
Apache Spark
Stream & iterative processing, ML

1Predictive Maintenance – Business Impacts
Using real-time data to
predict and prevent
breakdowns can reduce
downtime by 50 percent
50%
Repair & Replace Predict & Prevent
Predictive maintenance
could reduce maintenance
costs of factory equipment
by 10 to 40 percent.
40%
Reduced Downtime
- Source: McKinsey Analysis
Reduced Costs

IoT Enabled Predictive Maintenance – Customer Examples
Automotive Energy & Utilities Industrial Robotics
• Monitor the performance of
300,000+ trucks in real-time
• Predict failures, improve uptime
and vehicle performance
• Lower maintenance costs and
vehicle downtimes by 40%
• Continuous monitoring of
industrial-grade turbines for
hydroelectric power stations
• Gather, store and analyze noise
levels from turbines for anomaly
detection
• Predict failures in advance to
perform required maintenance
• Gather, store and analyze
sensor data from 10,000 robots
in real-time
• Zero Down Time (ZDT)
application
• Diagnostic solution predicts
potential failures and alerts the
operators in advance

TCS - Cloudera Confidential & Proprietary
TCS Sensor Data Analytics IoT Framework (SDAF)

Adopting an integrated approach with best of breed partner products to provide a strong foundation and helping
enterprises unlock compelling business value
TCS Sensor Data Analytics IoT Framework
Enterprise Data Hub powered by
Apache Hadoop
Intel Atlantic Ridge architecture
McAfee Security
Key highlights
• Certified on Cloudera
platform
• Built using open standards
• Proven scalability and
security
• Seamless and securely
connect devices
• Batch and real time
processing of data
• Deep Analytical Insights
IoT Partnership – Intel, Cloudera & TCS

A highly customizable Big Data based IoT framework to collect, transform, store and analyze
variety of sensor data.
Highlights:
 Certified by Cloudera
 Ready to use configurable framework across Edge, Cloud and Analytics functions.
 Comprehensive Device Management & Sensor Data Stream Handling capabilities.
 End-to end Security to ensure secure transfer, storage and access to data.
 Analytics of Things – Edge Analytics, Real-time Analytics and Deep Learning at
scale.
 Built using open standards ensuring zero vendor lock-in.
 Strong collaboration and partnerships with leading technology vendors, platform
providers and IoT consortiums.
TCS
SENSOR DATA
ANALYTICS
IoT Framework

15
SDAF on ClouderaSensors
Grove,Siemens,EMX,Honeywell
Devices Communications Acquire & Analyze Smart Applications
ControllerGateway
Intel(MoonIsland),ARM
Enterprise
Intelligent
Application
Web Portals
Mobile Apps
BI Tools
Protocols
Wi-Fi,BLE,ZigBee,Modbus,NFC,3GPP,LTE
BusLayer
MQTT,CoAP,HTTP,TCP,UDP
Real-time Processing
Acquire
Deep Analytics
Failure Prediction
Anomaly Detection
Time Series Analysis
Real-time AnalyticsParse & Enrich
Cloudera Hadoop Distribution
TCS Sensor Data Analytics IoT Framework (SDAF)Partner Ecosystem
Frequent Pattern Mining
Segmentation
Demand Forecasting
Device Management
Zero Touch Device
Commissioning
Remote
Terminal GUI
Health
Monitoring
Alerts &
Notifications
SDAF
Agent

SDAF - Feature Overview
STREAMING
DATA INGESTION
DATA ANALYTICS
DEVICE DATA
ACQUISITION SECURITY
DATA
PROCESSING &
STORAGE
DEVICE
MANAGEMENT
• Zero-touch
Device
Commissioning
• Health
Monitoring
• Remote Terminal
Access
• Software and
Firmware
Upgrades
• Support for
industry standard
protocols
• Data streaming
using CoAP and
MQTT protocols
• QoS support
• Support for protocols
such as Modbus,
Bacnet, Zigbee, BLE,
etc
• High-throughput
framework
• Data Enrichment
• Complex Event
Detection
• Sliding Window
based Correlation
• Real-time Alerts
& Notification
• Data Formatting
& Filtering
• Storage into
HDFS, HBase and
Hive
• Data Governance
and Meta data
Management
• Geo-Fencing and
Geo Spatial
Correlation
• Edge Analytics
• Real-time
analytics on data
streams
• Data Profiling
framework for
time series data
• Scalable Machine
Learning library
for predictive
analytics
• Comprehensive
Edge Security
• Audit Logs
• Role based
Access control on
data
• Kerberos
integration for
authorization
• HDFS Encryption

Analytical Models
Warranty Estimation Model
Segmentation Model
Anomaly Detection
• Interesting Pattern Detection
• Peak Detection
• Outlier sub-sequences
• Time series classification
Association/Rule Mining
Telematics Model
Clickstream Analytics and
Recommendation Model
Time Series Analysis
Prediction Model
Failure Probability & Time to Failure
Frequent Pattern Mining and Apriori
• Missing Trip Identification
• Parking Location Identification
• Route Matching
• Acceleration, Braking & Cornering Scoring
• Speeding Score
• Trip Score, Daily Score and Overall Score
• Score Projection
Forecasting

Case Studies

Solution Overview:
Perform real-time monitoring of Port equipment, improve
operational efficiency & facilitate Predictive Maintenance
Technology:
TCS SDAF components for capturing data from port
equipment, store data on Amazon Web Services and
Cloudera & uses SparkML Lib for running machine
learning algorithms.
Business Outcomes:
 Increased Equipment availability
 Reduced Operational costs
Finnish Shipping and Port Equipment
Manufacturer

Largest Indian HVAC Manufacturer
Solution Overview:
Centralized monitoring of Chillers across different
locations and proactive maintenance to avoid frequent
failures and reduce maintenance cost.
Technology:
TCS SDAF for real-time data capture from Intel IoT
gateways to a centralized Cloudera platform and
SparkML Lib for running machine learning algorithms.
Device Management features to remotely monitor,
configure and manage Intel gateways.
Business Outcomes:
 New Revenue Models using Chiller Monitoring as a
Service
 Proactive customer support and service using
predictive maintenance

Demo – Integrated Maintenance Management

Visualization/ Consumption
Layer
APIGatewayTools
Data
Scientists
Business
Analysts
Asset Monitoring
Portal
Mobile Applications
Real-time
Batch
Real-time Layer Batch Layer
Broker/Server
FTP/ SFTP
DB Connectors
Data Ingestion
Load balancer
Data Management
Sematic LayerStaging Area
Raw Data
Store
Integrated
Data Layer
Complex Event Processing
Parsing
Alerts & Notifications
Data Enrichment
Data Transformations
Data Aggregation
Real-time Analytics Machine Learning
OSI/PI Server
Distributed
Queue
Maximo
PLC/SCADA
Cloud
Predictive Maintenance For Fermenter Unit in Pharmaceutical Plant

Thank you

IoT-Enabled Predictive Maintenance

Recommended

Recommended

More Related Content

What's hot

What's hot (20)

Similar to IoT-Enabled Predictive Maintenance

Similar to IoT-Enabled Predictive Maintenance (20)

More from Cloudera, Inc.

More from Cloudera, Inc. (20)

Recently uploaded

Recently uploaded (20)

IoT-Enabled Predictive Maintenance