1 of 18

Application of FLEDGE at Neuman Aluminium: An Industrial Use Case

Sebastian Kropatschek, Thorsten Steuer

dunkelblau (RGB): 35/51/100

blaugrau (RGB): 115/125/156

2 of 18

Agenda

Neuman Aluminium

Roadmap

Transaction Manager

Demo

Learnings

Austrian Center for Digital Production

Architecture Variants

Application of Fledge

State Management

Feedback Develop and Debug FLEDGE Plug-Ins

Application Scenario

Mapping Vision to Architecture

3 of 18

Center for Digital Production -Research

Digital Engineering

Adaptive Manufacturing & Smart Factories

Process-Based Manufacturing Orchestration

Data Integration and Analytics for Digital Production

Sustainable Production Systems

Merge Design & Manuf.

Data Driven Design

Digital Twin Fidelity

„Wear-aware“ CAM

Reconfigurability

Process Adaptability

Predictability

Shop Floor OS

Orchestration & Data

Process based PLM

Deployability (HMI)

Data Contextualisation

Interoperability

Advanced Analytics (AI)

PMV based Analytics

Knowledge Graphs

Flexible Safety

Fail-safe Manufacturing

Circular Economy

Production as a Service

4 of 18

Neuman Aluminium

Sheet Metal Processing

The Neuman Aluminium Group is your global partner for high-quality aluminium solutions.

5 of 18

Application Scenario

Scenario:

Multiple Production Facilities in Europe
Controlled by a Centralized MES
MES Submits a Production Order to a Facility
Machine Processes the Order and Notifies MES of Current Production Status
MES Sends Data to a Subsequent Machine
Once the Connection to the MES is Lost, Production will Continue as long new Input is Required

The centralized nature of the system creates a critical dependency on the network connection!

Sheet Metal Processing

6 of 18

Architecture Variants

Device Services

Core Services

Appl. Services

Support Services

7 of 18

Roadmap for connected Production Systems (Neuman)

Open-source software stack to deliver device, core, application and supporting services for production equipment.

Self-healing, scalable, up-gradable, flexible, platform independent and offline capable edge devices.

Knowledge or/and ML driven smart edge devices to semi- autonomously control machine, forward data and adjust production equipment.

8 of 18

Mapping Vision to Architecture

Fledge

Data Streaming Service

Infrastructure

Orchestration

& Distribution

Cloud

REST

OPC-UA

MQTT

Edge Device N

Edge Device 1

Transaction Manager

Rule Engine

Semantic Search

ML Model Training

Data Analysis

ML Application

Plug-Ins:

System Monitoring

Low-Code Platform

9 of 18

Fledge as Transaction Manager

Fledge...

... is open source.

… delivers data to different cloud services.

... collects data from any sensor.

… aggregates, combines and organizes data.

… transforms and filters data.

… buffers data & resends after connection is reestablished.

… is highly performant and resource efficient.

10 of 18

Current Project Status

EDGE Device

Grafana + Kibana

Node-RED

API + UI

State

[{

qty: { cnt1: 112 , cnt2: 122 }

state: (Hand, Automatik , Teilautomatik)

prematerialNr: 2345345

handlingUnit: 2343234

job : {ID: P43000, product: S5445, …}

quality: [{Part: 3222, state: ok}]

….

}]

State Management

Actions

Machine UI triggers action to manipulate production state.

Fledge stores the machine and production state.

Fledge forwards data to next machine or/and cloud services.

States

EDGE Device

Grafana + Kibana

Node-RED

State Management

Siemens S7 Controller

State

[{

qty: { cnt1: 112 , cnt2: 122 }

state: (Hand, Automatik , Teilautomatik)

prematerialNr: 2345345

handlingUnit: 2343234

job : {ID: P43000, product: S5445, …}

quality: [{Part: 3222, state: ok}]

….

}]

Actions

States

11 of 18

S7 South / North Plug-in

New Fledge South S7 Plug-In

The plug-in is used to read data from a Siemens S7 PLC.

New Fledge North S7 Plug-In

The plug-in is used to write data to a Siemens S7 PLC

Features:

Read Various PLC Data Types
Arrays
Objects/ Structs (UDTs)
Array of Objects
Optimized Reading of Data as Blocks

Output:

JSON Object /

Escaped String

Flat Variable List

Features:

Write Various PLC Data Types
Verification: Write + Read
Add Static Datapoint to Asset
Limited Bool Support because of Snap7 -Python Library

12 of 18

Fledge State Management

Redux Data Flow

State Management

Redux Principles:

Single Source of Truth: The state of your whole application is stored in an object tree within a single store.

State is Read-only: The only way to change the state is to emit an action, an object describing what happened.

Changes are made with pure functions: A reducer is a central place where state modification takes place. Reducer is a function which takes state and action as arguments, and returns a newly updated state.

13 of 18

Notification Service for State Management

New Fledge Rule Plug-In

The rule is used to detect if a data point is different from its previously received value within an asset.

New Fledge Notification Plug-In

The notification is used to create a new asset based on data previously received from the rule plugin and the assets already collected by Fledge.

Features:

Configurable Asset Name to Create
Can be Configured to Choose which Assets and Data Points Should be Used to Create the New Asset.
Can be Configured to Rename Data Points
Can be Configured to Authenticate against the FLEDGE REST API.

ACTION

STATE

triggers

Features:

Configurable to Monitor Multiple Assets
Can be Configured to Rename Data Points

14 of 18

Developed and Relevant Fledge Plug-Ins

Provided Plug-Ins by FLEDGE

South

fledge-south-http https://github.com/fledge-iot/fledge-south-http.git

fledge-south-opcua https://github.com/fledge-iot/fledge-south-opcua.git

fledge-south-modbustcp https://github.com/fledge-iot/fledge-south-modbustcp.git

fledge-south-mqtt https://github.com/fledge-iot/fledge-south-mqtt.git

North

fledge-north-kafka https://github.com/fledge-iot/fledge-north-kafka.git

fledge-north-opcua https://github.com/fledge-iot/fledge-north-opcua.git

fledge-north-http-c https://github.com/fledge-iot/fledge-north-http-c.git

fledge-north-http https://github.com/fledge-iot/fledge-north-http.git

Notification

fledge-notify-asset https://github.com/fledge-iot/fledge-notify-asset.git

fledge-rule-average https://github.com/fledge-iot/fledge-rule-average.git

fledge-rule-simple-expression https://github.com/fledge-iot/fledge-rule-simple-expression.git

fledge-notify-mqtt https://github.com/fledge-iot/fledge-notify-mqtt.git

Filter

fledge-filter-asset https://github.com/fledge-iot/fledge-filter-asset

fledge-filter-delta https://github.com/fledge-iot/fledge-filter-delta.git

fledge-filter-expression https://github.com/fledge-iot/fledge-filter-expression.git

fledge-filter-change https://github.com/fledge-iot/fledge-filter-change.git

Developed by ACDP

Transaction Manager

fledge-south-s7-python https://github.com/kropatschek/fledge-south-s7-python.git
fledge-north-s7-python https://github.com/kropatschek/fledge-north-s7-python
fledge-north-opcuaclient https://github.com/kropatschek/fledge-north-opcuaclient

State Management

fledge-rule-delta https://github.com/kropatschek/fledge-rule-delta.git
fledge-notify-customasset https://github.com/kropatschek/fledge-notify-customasset.git

15 of 18

Demo

EDGE UI -> Container vorzeigen Microk8s| Sebastian
Node-REd for debugging. Kommt die Nachricht rein| Sebastian
Grafana System monitoring| Sebastian
Kibana LOG monitoring| Sebastian
Fledge

Workflow

South Service S7 support for structs| Sebastian

Show data in Edge UI direct read from database

Action Channels daten von MES| Sebastian

Jobstart und job strop in Edge UI

MES for jobstart trigger to start produciton| Sebastian
Notification Service| Thorsten

Hadnlingunit configuration quality data for next machine production id and program identifier for washing machine for booking in Kafka to MES.
Produce configuration

Represents the quantiy of the produces parts what parts are produced

North to Kafka| Thorsten

For cloud and MES

North Nodered| Thorsten

For debugging or procees data or include business logic via low code platform

North to-plc S7| Thorsten

To trigger controller of the produciton machine

Show how to trigger job from UI and that fledge shows receiving and sending to PLC| Thorsten

16 of 18

Learnings

Performance

We tested FLEDGE with sampling rates of 50 ms.
FLEDGE is very resource efficient, runs on a Raspberry Pi.

Flexibility

The current approach supports a variety of architecture concepts.
The services and plug-ins of FLEDGE can be combined to cover a big variety of use cases.
Supports multiple protocols south and north.

Extendability

Plug-Ins can relatively easy be developed after an initial training period.
Other plug-ins can be used as templates.

Usability

Documentation provides great support for users.
Community is helpful and open to new contributors.

Open Challenges

Better open-source tool support for monitoring and orchestration of edge devices.
Improvement of FLEDGE Documentation for Developers.

17 of 18

Bottom Line

Trust open-source and start to adapt and improve FLEDGE!!!

18 of 18

Feedback Develop and Debug FLEDGE Plug-Ins

Documentation for developers should be extended.
Logger cannot be stopped manually from the GUI to look at specific error messages.�All Plug-Ins must be stopped sometimes to catch the error message.
Only single log pages can be viewed and the search only works page wise.
For debugging North Services Node-RED is a useful tool.
Debugging North Services can be tricky since sometimes after reconfiguring North Services the Service runs into unpredictable states.
Cannot set debug level via GUI for python plug-ins.
There is no option available to clean the log if it is full.
Importing other fledge configurations is not supported and exporting only the configuration is not supported.
Increase community for testing FLEDGE or develop more unit tests.
Provide more sample code for plug-ins.
It is difficult to overview what classes can be used and how to achieve a certain goal.