Sergio Gusmeroli from the Politecnico di Milano
June 18
The BDVA Smart Manufacturing Industry group is going to present in this webinar, challenges and opportunities related to the adoption of Big Data-driven solutions in the manufacturing domain. Following the grand scenarios defined by EFFRA (Smart Factory, Smart Supply Chain, Smart Product), the discussion will make use of real cases and pilots from research projects to elaborate on the topic.
BDVe Webinar Series - Politecnico di Milano - Big Data in the Smart Manufacturing Industry
1. The role of Data Sovereignty in European
Commission communication “Towards a common
European Data Space”
Sergio Gusmeroli, Politecnico di Milano
2. 2
The Regulatory Context: Sharing Private Sector Data
Models to B2B Data Exchange
a) An Open Data approach: The data in question are made available
by the data supplier to an in principle open range of (re-)users with
as few restrictions as possible and against either no or very limited
remuneration.
b) Data monetization on a data marketplace: Data monetization or
trading can take place through a data marketplace as an
intermediary on the basis of bilateral contracts against
remuneration. Suitable when either (1) there are limited risks of
illicit use of the data in question, (2) the data supplier has grounds
to trusts the (re-)user, or (3) the data supplier has technical
mechanisms to prevent or identify illicit use.
c) Data exchange in a closed platform: Data exchange may take
place in a closed platform, either set up by one core player in a data
sharing environment or by an independent intermediary. The data
in this case may be supplied against monetary remuneration or
against added-value services, provided e.g. inside the platform.
3. Open Data: the vision of a FIWARE for INDUSTRY DataLab
I4.0LAB @ POLIMI
Open Data in Manufacturing
a) Open Data Models for SI entities (such as
robots, AGVs, machines, conveyors) with
Data in Motion and Data at Rest
b) Network of open Didactic Factories
producing and sharing their data thanks to
standard protocols and data formats (e.g.
OPC-UA, MQTT, ROS, AQMP).
c) Data Transformation techniques for non-
public Data such as Aggregation, Filtering,
Anonymization, Pseudonymiz.
d) One stop shop for search-discovery-
selection, Distributed Repositories for Data
Storage (iSpaces)
e) Ecosystem of Innovators testing and
experimenting their solutions on open data
(Data-AI Community)
4. Data Marketplaces: FIWARE enabled Data Platform
Access to
Competencies
COLLABORATION PLATFORM
for DIGITAL INNOVATION HUBS
Ideation
Platform
Wikis &
Fora
Search
Engine
Human Skills CV
Manager
Partner Search
and Selection
Access to
Technology
IT Assets & OSS
Catalogue
Reference
Architectures
Applications
Marketplace
Access to
Experiments
Industrial
Experiments
Best Practice
Success Story
KPIs Lessons
Learned
Access to
Knowledge
Maturity
Model
Training &
Formation
Brownfield
Integration
Access to
Market
Ideas
Incubation
Business
Acceleration
Capital &
Funding
Tangible Assets
Manager
Living Lab
Innovation
5. Trusted B2B Data Exchange: Sovereignty on Access
Broker
App
Store
Data
Source
Connector
Data Provider Data Consumer
Dataset(s) transferred from
Provider to Consumer
Metadata Description of
Datasets/Provider/Consumer
Application for specific data
manipulation
Data exchange (active)
App download
Metadata exchange
Data exchange (inactive)
Connector Data
Sink
Connector
Meta
Meta
Meta
Meta
Meta
…
App
Data
Meta
App
App
App
App
Data
Meta
6. Trusted B2B Data Sharing: Sovereignty on Usage
Data
Models
Ontologies
7. IDSA-BDVA Towards a EU Data Sharing Space
http://www.bdva.eu/node/1277
• Create the conditions for the
development of a trusted
European data sharing
framework
• Incorporate data sharing at the
core of the data lifecycle to
enable greater access to data.
• Provide supportive measures
for European businesses to
safely embrace new
technologies, practices and
policies.
• Assemble a European-wide
digital skills strategy to equip
the workforce for the new data
economy.
9. Big Data Value Spaces for Competitiveness of European
Connected Smart Factories 4.0
G. Petrali (WHIR)
Whirlpool Use Case
10. TIER 1 SUPPLY BASE
TIER 1 SUPPLY BASE
Raw material
Change in Tier 2
Wrong Instructions
SPARE PART CENTER
Parts delivery
PREDICTION
TOOL
Prod. Orders Request
and wharehouse
management
suggestion
Planning and
Quality suggestion
Warning alert,
training and
medium term planning
PRODUCTION FACTORY
Whirlpool Pilot : Cause and Effect Diagram
11. Whirlpool Pilot Architecture
STATISTICAL DEMAND
FORECAST GENERATION
Spare Parts Consumption History
INVENTORY & SUPPLY
PLAN GENERATION
PROCUREMENT
PRODUCTION
DISTRIBUTION
AS IS
NEW
PREDICTION
TOOL
(FABRIC CARE)
(in parallel with the current
statistical demand forecast
generation)
Pre-elaboration 1:
Factory Test Data
Pre-elaboration 2:
Sell-In / Demand Forecast
Pre-elaboration 3:
Service Order Confirmations
Pre-elaboration 4:
Smart Appliances
Output 1 :
Spare Parts Demand Forecast in Qty
Output 2 :
Spare Parts Demand Forecast in
# of Order Lines
Output 3 :
Monitoring Tools &
Demand Forecast Analysis
Output 4 :
Quality Reports for :
- Factory and Product Engineers
- Market and Service Partners Training
- Smart Appliance Predictive Maintenance Approach
TO BE
Pre-elaboration 5:
Service Incident Rate
12. From PoC to the Large Scale
Data extractors ready and
deployed
Large scale data
available and
loaded
Forecasting models
tuning
New forecasting
reports and price
analysis
Operational KPIs
monitoring
activation
Environment setup
and activation for the
large scale
Forecasting analysis from monthly
to weekly view
TeraLab Big Data
environment certification
and activation
KPIs calculation: demand
forecast error, human effort
for planning
Integration with other information
that can influence the forecasting,
extention to other EMEA countries
13. BC expansion: IDS allow extending the benefit of
forecasting tool to all Value Chain
TIER 1 SUPPLY BASE
Third party
Field Service