Presentation at Pro-VE'17 conference in Vicenza, Italy

1. An Approach for cloud-based
Situational Analysis for factories
providing real-time
Reconfiguration Services
Sebastian Scholze
ATB
Pro-VE`17 Sebastian Scholze, ATB-Bremen 1

2.  Manufacturing of products becomes increasingly
complex and needs to be flexible
 increasing diversity of product use & products,
 demand for more customised products,
 shorter time-to-market requirements.
 Product use and production
activities are often separated
 low efficiency,
 high costs.
 Optimisations are required to
face these problems
 Some of these optimisations can be carried out by
 adjusting production control parameters,
 reconfiguration of products
Sebastian Scholze, ATB-Bremen 2
Motivation
supplies factory deployedproducts
data analyticsFoF infrastructure
cloud
monitoring &
reconfiguration
Pro-VE`17

3. Pro-VE`17 Sebastian Scholze, ATB-Bremen 3
Motivation

4.  Methodology for reconfiguration and optimisation
of products and factories based on predictive big-
data analytics.
 address technical and organizational issues for
extension and introduction of tools and services,
 emphasizing security, privacy & trust.
 Set of tools and services to support
 predictive (big) data analytics,
 dynamic reconfiguration and optimisation,
 situational awareness technologies,
 cloud resource management.
 Cloud based secure infrastructure
 will be an add-on for existing production systems
or next generation smart factory operating system.
Sebastian Scholze, ATB-Bremen
Objectives and Expected Results
4Pro-VE`17

5.  Targets two technology challenges for factories of the
future for improving production, products and services:
 Interconnected Systems of Production Systems
 Connected Product Networks
 A key objective of the project is to develop cloud-based
analytics and reconfiguration capabilities with:
 reactive and predictive reconfiguration for production
systems and smart products;
 flexible run-time reconfiguration decisions during
production rather than pre-planned at production planning;
 real-time reconfiguration decisions for optimisation of
performance and real-time production and product
functions.
Sebastian Scholze, ATB-Bremen 5
Technical Objectives
Pro-VE`17

6. Sebastian Scholze, ATB-Bremen 6
Proposed Concept (1)
Secure Infrastructure
Big Data Platform
Connected
Product Networks
Factory
Modelling,Monitoring& ReconfigurationInterfaces
Predictive
Analytics Engine
Configuration
Quality Evaluation
Reconfiguration &
Optimisation
Engine
Cloud Resource Management
Pro-VE`17

7.  Situation Monitoring & Determination:
Services to identify the current situation
under which a product/machine is being used
or operates.
 Predictive Analytics Platform: Real-time
big data framework for manufacturing.
 Reconfiguration and Optimization
Engine: Cloud-based engine for reactive
optimization and predictive optimization,
based on past performance and analysis of
current configurations
 Security, Privacy & Trust: Protection of
both product and customer data, by a
flexible policy-enforcing scheme.
Pro-VE`17 Sebastian Scholze, ATB-Bremen 7
Proposed Concept (2)

8. Pro-VE`17 Sebastian Scholze, ATB-Bremen 8
Workflow –
Predictive Analytics Platform
Get data fromfactory&
connectedproducts
Post-processedtechnical
availabilitydata
BigDataAnalyticsPlatform
Analyse data basedon
configuration(e.g. for
predictive maintanance)
Provide analysis results
forreconfiguration
decisionsupport
Event-drivenData
IngestionServices

9.  Challenges
 Real Time data Ingestion (filter, aggregate, routing,
storage).
 Unify Real Time / Batch data processing.
 Cloud agnostic platform (On-premise / Any cloud).
 High Availability 24x7 support.
 Cost efficient platform.
 Innovations
 Cost efficient platform using dynamic resource managers.
 Improve of the so-called Lambda architecture using the
same tools for both fast and slow data.
 Develop new predictive analytics algorithms for each BC
that will run on the platform.
 New-SQL research, where strong consistency and
relational capabilities are added to highly scalable non-
relational databases.
Pro-VE`17 Sebastian Scholze, ATB-Bremen 9
Workflow –
Predictive Analytics Platform

10. Pro-VE`17 Sebastian Scholze, ATB-Bremen 10
Workflow – Situation Model and
Determination Services
Get data fromfactory&
connectedproducts
Identifycurrent situation
SituationDeterminationServices
Refine identified
situations byreasoning
techniques
Forward determined
situations
Event-drivenData
IngestionServices
Big-Data Analytics
Platform

11.  Challenges:
 Modelling of Context
 Interpreting data from various existing
systems
 Innovations
 Situation monitoring services, identifying
product use patterns
 Situation determination services,
supporting reconfiguration of
factories/products/services
 Providing new insights into product use
patterns and product environmental
impact patterns.
Pro-VE`17 Sebastian Scholze, ATB-Bremen 11
Workflow – Situation Model and
Determination Services

12. Pro-VE`17 Sebastian Scholze, ATB-Bremen 12
Workflow – Optimisation &
Reconfiguration Engine
Identification of
product/production
processes and
reconfiguration points
Identification and linking
of optimisation metrics,
constraints and
monitoring data
Optimisation&ReconfigurationEngine
Search-based
optimisation
Issuing of a
reconfigurationSituation
Determination
Services
Big-Data Analytics
Platform
Event-driven Data
Ingestion Services
Reconfiguratio
n Interfaces

13.  Challenges
 Context-driven optimisation
● Situation Monitoring/Determination
● Analytics
 Timely response to process change
 Innovations
 Best solver can be context-determined
 Multiscale optimisation
● Solvers can be:
 Federated (cloud-based)
 Hardware-accelerated (FPGAs/GPUs)
Pro-VE`17 Sebastian Scholze, ATB-Bremen 13
Workflow – Optimisation &
Reconfiguration Engine

14. Pro-VE`17 Sebastian Scholze, ATB-Bremen 14
Conceptual Architecture
SAFIRE
Manufacturer/ Factory
Optimisation&
Reconfiguration
Engine
Situation
Determination
Services
Predictive
Analytics
Engine
Reconfiguration
Quality
Evaluation
Services
Event-driven Data Ingestion & Situation Monitoring Services
Reconfiguration
Interfaces
Secure SAFIRE infrastructure
Connected ProductNetwork

15.  Electrolux
 Products connected to a cloud-based
system can be optimised through a
reconfiguration process i.e. Cloud-
driven product optimisation
 OAS
 Optimise production processes and
preventive maintenance activities
through reconfiguration of processes
based on Big Data analysis in the Cloud
 ONA
 Improvements in Adaptive Machining,
Part Quality and Sustainability, based
on analysing machine usage behaviour
compared to nominal machine usage
behaviour
Pro-VE`17 Sebastian Scholze, ATB-Bremen 15
Case Studies

16.  Adaptive machining capability.
 Manufacturing quality improvement through a
reconfiguration process.
 Sustainability improvements during machine use
with reconfiguration capability
 Reconfiguration process generates new
knowledge for new optimized Machine/Product
Design
 New business opportunities around
reconfiguration services.
Pro-VE`17 Sebastian Scholze, ATB-Bremen 16
Expected Impacts

17.  Approach for optimisation and
reconfiguration opportunities was
presented, which is based on
 Predictive analytics
 Situational awareness
 Four building blocks are presented that
are typical examples of new
reconfiguration functionalities needed for
factories of the future.
 Situation Monitoring & Determination
 Predictive Analytics Platform
 Reconfiguration and Optimization Engine
 Security, Privacy & Trust
Pro-VE`17 Sebastian Scholze, ATB-Bremen 17
Conclusions

18. Cloud-based Situational Analysis for Factories
providing Real-time Reconfiguration Services
More info: www.safire-factories.org
Pro-VE`17 Sebastian Scholze, ATB-Bremen 18