AI en la Industria

Manufacturera
A Simple View of an IoT Solution

Things Insights Actions

 A More Realistic View… Data storage Deployment
Enterprise integration Cloud-to-device Business process integration
commands Cold path analytics
Device recovery Solution scale Operations monitoring Cost management

Internationalization
High availability
Updating devices
Provisioning
Transport protocols devices
Device lifecycle Warm path analytics

Drivers Things Disaster recovery

Insights Hot path analytics
Actions
Device updates
Data ownership
< ---- End-to-End Security ---- >
On device analytics Data visualization
Manufacturing scale Industry and government compliance
HW certification
Device commercialization

Securing data … and why IoT needs simplifying

The basis for the fourth industrial revolution
is the availability of all relevant
information in real time by connecting all
instances involved in the value chain. The
ability to derive the optimal value-added
flow at any time from the data is also vital.
The connection of people, things and
systems creates dynamic, self-organizing,
real-time optimized value-added
connections within and across companies.
Source: PwC Industry 4.0 Survey 2016
 Azure IoT ecosystem Azure IoT platform and servicesAzure Data Platform and AI Azure D365
(to enable brownfield & greenfield scenarios)
and 3rd party providers
(SAP, etc.)

Things Insights Actions

Source: Aachen University

 Azure IoT ecosystem Azure IoT platform and servicesAzure Data Platform and AI Azure D365
(to enable brownfield & greenfield scenarios)
and 3rd party providers
(SAP, etc.)

Things Insights Actions

Source: Aachen University

Things Connectivity Data Analytics

New Business Models

All open, extensible, scalable & secure

Azure Cortana
IoT Suite Intelligence Suite
Get started in minutes and Transform data action with one
customize to meet your needs comprehensive offering
Events and Data Producers Ingest Tranform Storage Analytics Visualization

Field Gateway
Edge Analytics
and Data PowerBI
Filtering
Data Lake

IoT Hub
Machine Learning
Stream Analytics HDInsight
Web App
Azure SQL
Manufacturing IoT Real Time Data IoT Data database
Equipment

Other Data Sources Other Data Ingestion Other Data

Data Lake Analytics
Storage blob
Business Apps (ERP, MES, etc)
OnPrem Data
Store Event Hubs

Social Service
Data Factory
Fabric SQL Data
Warehouse
API App
Cloud – GIS, weather,
news, etc
 Challenges Digital Transformation Opportunities
Legacy assets See Patterns to
Current Manufacturing Predict failures
Execution Systems, PLCs, Gain Insights across
R&D
SCADAs, Plants
Supply Production Sales

Global footprint with

R&D
Supply
Marketing
Production
Services
Enable new services
varying connectivity for your customers
Data is stored in variety Smart Connected
of places products

Marketing Services

Sales

`
 User experiences
Dashboards Home Mixed Reality Interactive speech Gestures

Enterprise business processes Preconfigured Azure IoT Suite and SaaS applications
L4: Business
Planning CRM PLM ERP SCM
Remote monitoring Predictive maintenance Connected factory Microsoft IoT Central

Operations
Technology MES & Enterprise IT Business integration Analytics and artificial intelligence

L3: Mfg Ops

Mgmt, MES, Logic API BizTalk HDInsight Machine Data Lake Azure Time Cognitive Bot
SQL Server Azure Stack
CAD, PLM Apps integration Services Learning Analytics Series Insights Services Framework
Stream Analytics

L2: Supervisory Hot path analytics and application platform

Analytics

Control,
SCADA,
HMI
edge
Edge

Stream Analytics Event Hubs Service Fabric (Actors) Functions

L1: Plant
on

Control PLC, Cloud Gateway

DCS, IPC
Azure IoT Cold path analytics and storage
Device Agent

L0: Physical Hub

Equipment, Field Gateway (device
I/O, Devices,
provisioning)
Sensors
Azure IoT Gateway Data Factory DocumentDB SQL Database Data Lake Store
 Notifications

Event Frames

Asset Analytics Event Hubs / Stream Analytics Power BI

IoT Hubs

PI Asset Framework

PI Data
Archive

PI Server
 REMOTE MONITORING
Utilizarla para generar una
comunicación de dos vías con
los suscriptores

PREDICTIVE MAINTENANCE
OPTIMIZE El desarrollo de creación de
contenidos (Animaciones,
OPERATIONS efectos espaciales)

Manufacturing fast-
Securing operations
moving consumer CONNECTED FACTORY Ensuring critical
[Smart Factory] confidence and
assets availability
goods Para periodistas y consistency
editorialistas para acelerar
la velocidad de ingesta

REAL TIME PRODUCTION From reactive to

Making critical data
OPTIMIZATION predictive and
mobile and sharable
Plataforma de comunicación, preemptive
colaboración y gestión de
conocimiento
ThyssenKrupp Elevator
Global elevator manufacturing
company develops service arm

Objectives Tactics Results

ThyssenKrupp wanted to Microsoft technology • Increased elevator uptime
better compete in their enabled ThyssenKrupp to • Reduced costs for ThyssenKrupp
industry by offering monitor products via a and its customers
dramatically increased real-time dashboard and
uptime, taking preventative instruct technicians on • Developed real-time data
maintenance a step further to optimal maintenance visualization and awareness of
predictive and even activities through dynamic issues
preemptive service. predictive models.

“We wanted to go beyond the industry standard of preventative maintenance,

to offer predictive and even preemptive maintenance.”
Andreas Schierenbeck
CEO, ThyssenKrupp Elevator
Arming field technicians with actionable guidance
Objectives Tactics Results
thyssenkrupp wanted to better monitor their more They connected thousands of sensors embedded in • Reduced costs for thyssenkrupp and its customers
than 1.1 million elevators worldwide. Lack of insight their elevators to the cloud to monitor real-time • Increased reliability through predictive maintenance
led to downtime and unpredicted failures in some performance and proactively address issues with and rapid, remote diagnostic capabilities
of the world’s most famous buildings Microsoft technology

Scheduled Maintenance
• Hydraulic/Passenger Elevator
• Replace brake shoes
• Check AC motor for health

!
NEXT STOP
!

Maintenance in progress
Predictive Maintenance framework

Identify the target outcome

1 Determine what outcome you ultimately want to achieve. In this case, we want to know
how long until any given AC unit fails.

Last time an AC unit failed, it cost I want to understand how much time each
thousands of dollars and operations AC unit has left before it needs maintenance
were down for days. so we can prevent unplanned failures.

AC unit
out of order
Predictive Maintenance framework

Inventory data sources

2 Identify all potential sources of data, including the types of data and the amounts
available. The outcome you are seeking will influence what data is essential and
what’s optional.

Weather data Failure logs

Include data from a variety
100101011000
101000101101 of sources – you may be
Performance 010011001110 surprised about the places
data 101000110011 where key information can
come from.
Maintenance logs Sensor data
Predictive Maintenance framework

Capture and combine data

3 Connect all of your data to a single place and prepare it for analysis.

Lay the groundwork for

a robust predictive model
by pulling in data that
includes both expected
behavior and failure logs.
Predictive Maintenance framework

Model, test, and iterate

4 Identify unexpected patterns by developing statistical models using advanced analytics
techniques. Stank-rank models to determine which model is best at forecasting the
timing of AC unit failures.

▪ Model A Make your model

actionable by understanding
▪ Model B how much advance notice
▪ Model C
the maintenance team needs
in order to respond.
Predictive Maintenance framework

Validate model using your latest data

5 Apply your model to live, streaming data and observe how it works in real-world conditions.
Use machine learning to refine your model and ready it for full implementation.

100101011000
101000101101
010011001110
Be willing to refine your
101000110011 approach based on the
data you gather during
the real-world pilot.
Predictive Maintenance framework

Integrate into operations

6 Operationalize the model by adjusting maintenance processes, systems and resources
to act on near-real-time data. Make ongoing improvements by gaining insights from
machine learning and advanced analytics.

2 weeks until failure: Send

repair team

Strengthen your processes

3 weeks until failure: Order and procedures to take
replacement part
advantage of what you learn.
 Azure IoT Products and Strategy
• Microsoft IoT Central
• Fully managed IoT SaaS solutions
• Connect and manage devices, set rules and alerts on telemetry, connect to existing business processes
• Configuration only; requires no cloud solution development expertise.

• Azure IoT Suite

• Pre-Configured Solutions for IoT built on Azure PaaS services
• Default, customizable UX and applications that provision in minutes
• Enables full control and customization

• Services & Software for IoT

• Specialized services like IoT Hub, IoT Hub Device Provisioning Service, Time Series Insights
• General purpose services: Stream Analytics, AzureML, Data Lake, Data Lake Analytics, PowerBI, Web Apps, Logic Apps
• Open and extensible support for all popular IoT operating systems
• Windows 10 IoT Core includes out of the box support for provisioning and device management with Azure IoT Hub
• Azure IoT Device SDK – Open Source, Cross-Platform, Multi-Language
• Azure IoT Gateway SDK – Open Source, Cross-Platform, Multi-Language

• IoT Ecosystem & Partnerships

• Open Source Contributions – Eclipse Kura, Node Red, IoT Hub->Kafka connector, OPC-UA
• Connectivity Partnerships – Jasper, Ericsson, SigFox, more coming
• Security Program for Azure IoT
• Azure Certified for IoT
• Azure IoT Partner Program
• Reference Architectures
• Microsoft AppSource Marketplace for IoT