Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
Success through collaboration
Delivering value by combining knowledge and technology that
we wouldn’t be able to achieve on our own.
Collaboration to realise IoT Solutions
1
PC-Control: Unified Universal Automation Platform
Runtime
Engineering Tools
Communi-
cation
Data
Storage
HMI IEC
61131
PLC
Safety
PLC
Safety
Motion
Library
PLC Open
Motion
DIN
66025
CNC
Robotics
Robotics
Motion
C, C++CCAT
CCATModules
Simulink®
Module
MATLAB®
Simulink®
Scientific
Automation
Measure-
ment &
Analysis
Vision
Runtime
Vision
Beckhoff Automation IO System
3
Beckhoff Automation IO System
4
▪ Increase robustness and reliability
− Protect machinery as well as staff on-site
▪ Save the environment
− Don‘t replace intact parts
▪ Save your money
− Planned and fewer downtimes
Introduction to 4.0 IIOT
5
tota
l costs
minimum
total costs
preventive reactivepredictive
▪ Time domain
− Root-mean-square (RMS)
− Excess Kurtosis
− Crest-Factor
− Signal envelope
− Analytic signal
▪ Frequency domain
− Magnitude- and Power Spectrum
− Envelope Spectrum
− Integrated- and Multi(frequency)band RMS
▪ Time-Frequency domain
− Instantaneous frequency/phase
▪ Quefrency domain
− Power Cepstrum
Condition Monitoring Toolbox | TwinCAT 3 Library
Time domain signal features
Frequency domain signal features
Time-Frequency domain signal features
Quefrencydomain signal features
Statistical evaluation
- Quantiles
- Central moments
Classification
110/240 VAC
4G data
EtherCAT P
EtherCAT and power in a single cable
IEPE Vibration sensorThermocouple
IP67 signal conditioning for temp and vibration sensors
IoT Vibration and Temp Monitoring
Controller and Gateway
▪ Frequencies to look at when aiming bearing diagnostics
− Ballpass frequency, outer race
− Ballpass frequency, inner race
− Ball roller spin frequency
− Fundamental train frequency
• (+) fixed inner ring
• (-) fixed outer ring
Condition Monitoring Toolbox | TwinCAT 3 Library
8
BRSF
BPFO
FTF
BPFI
Shaft speed [Hz]
Number of roling elements
angle of load
▪ Time domain
− Excess kurtosis
− Crest-Factor
Condition Monitoring Toolbox | TwinCAT 3 Library
bearing
Defect on outer
bearing surface
Defect on inner
bearing surface Defect on
rolling element
time
Time domain signal features
Frequency domain signal features
Time-Frequency domain signal features
Quefrencydomain signal features
Statistical evaluation
- Quantiles
- Central moments
Classification
time
T
▪ Frequency domain
− Envelope Spectrum
Condition Monitoring Toolbox | TwinCAT 3 Library
Knowledge on geometry
of bearing and rotating
speed
→ frequencies to observe
are known
frequency
1/T
2/T
3/T
e.g.
en
v. s
pectr
um
(FFT of a signal‘s envelope)
1) Measure and visualise raw data (e.g. scope view)
2) Calculate signal features (CM Lib. | signal processing
and statistics)
3) Mapping to a maintenance state (CM Lib. | classif.)
4) ISO 10816-3:2009
Application example
11
Traffic light
classification
(thresholds)
• There are lots of drilled holes on a plane (Boeing 777 wing -> 35 000
fasteners)
• Current process monitoring is to prove the process on coupons then
manual inspection
• Drilling tools are improving, monitoring systems are appearing but
often proprietary so no real opportunity to use the data
• Always looking for ‘one way assembly’
POC Use case Aerospace wing drilling MTC
Drilling Process Monitoring
Features of Interest
1. Part Thickness
The thickness of composite panels is quite variable and form tolerances mean that
shim is added to fill the gap between panels on assembly. As a consequence the
thickness of the drilled part may vary. The current process is that the depth of each
drilled hole is checked and a fastener selected long enough to fasten the parts but
not too long as to add extra weight to the wing.
2. Burr Height
Typically a burr forms on the exit side of aluminium components on drilling whose
height is quite variable. Currently the wing panel must be dismantled to inspect each
hole and deburr if necessary because there is no way of knowing the burr height in
advance. We believe that the burr formation may again be detected from the spindle
torque/power curve and alert the operator when a burr forms.
3. Tool Wear
In general we think we will see an increase in spindle power and defects as the tool
wears and becomes blunt. Currently the tool is replaced either after a set number of
holes or when the operator notices that the tool is worn. We would like to generate
an automatic alert when the tool has worn and possibly estimate the number of holes
remaining in a particular tool.
4. Good/Bad Hole
More generally we would like to explore the possibility of using machine learning to
flag bad holes to the operator so that they do not have to inspect each hole but are
directed to the holes in need of attention.
POC Use case Aerospace wing drilling MTC
Data Capture
Analogue output +/- 10V
• Spindle power (8ms)
• ~1200 data points per hole
Beckhoff IPC IOT Gateway
OPC UA
• Spindle power (8ms)
• Timestamp (8ms)
• Filename
MQTT
• Spindle power (8ms)
• Timestamp (8ms)
• Filename
Spreadsheet
• Hole position
• Part thickness
• Hole quality
• Gap
• Exit Burr
Manual Measurements
Drilling Process Monitoring
Data Capture – IoT Gateway
Remotely
accessed IoT
Gateway
ENABLING GLOBAL IOT CONNECTIVITY
Secure Private Networks using Cellular Networks
ENABLING GLOBAL IOT CONNECTIVITY
SECURITY
• The Arkessa Core Network creates a SECURE PRIVATE NETWORK between device and
cloud.
• Redundancy in the Arkessa Core Network ensures HIGH-AVAILABILITY.
• Public Internet access is RESTRICTED unless required by the application.
Enterprise Cloud
Public Internet
SECURE PRIVATE NETWORK
ENABLING GLOBAL IOT CONNECTIVITY
SECURE PRIVATE NETWORKING
OVER CELLULAR NETWORKS
GGSNSGSN RADIUS DATA
ARKESSA NETWORK
Arkessa AAA Services for fixed IP
allocation and 2nd level authentication
ENTERPRISE NETWORK
ResilientEncrypted
Links
Internet
Network to Network
Interconnect
Private APN
Radius
SIM APN
GEA/UEA/UIA2 GTP VRF
IPSec
VRF or Access Control List
IPSec / VLAN
Private IP Address LOCAL AND ROAMING
MOBILE NETWORKS(~600 networks, ~200 countries)
ENABLING GLOBAL IOT CONNECTIVITY
MASSIVE BROADBAND AUTOMATION & SAFETY
Metering Smart Cities
Agriculture &
Environment
Wearables &
Healthcare
Transportation
Digital
Surveillance
Drones
Retail &
Hospitality
Robotics &
Machine Vision
Smart GridAV & EV
Process and Factory
Automation
Low-cost & Low-Power
Low Data volumes
Massive deployment density
High Data volumes
Data streaming
Real-time
Ultra-reliability and low-latency
Wireless Time Sensitive Networking
Positioning
ONE NETWORK FOR ALL INDUSTRIES
< Commercial deployment and scaling Early pilots >
CELLULAR
2G
3G
4G LTE
LTE Cat-NB1
LTE Cat-M1
5G
TM
© 2019 InVMA Limited All Rights Reserved
INTRODUCING INVMA AND ASSETMINDER
Rob Dinsmore – Sales Manager
April 2019
Need resized image
Fundamentally the PHYSICAL and DIGITAL worlds have converged
22© 2019 InVMA All Rights Reserved
ASSETMINDER
Receive instructions
and updates from you
• Firmware updates
• Control services remedy
Tell you when the asset isn’t
being used properly• Warranty claims
• Health and safety
Tell you when there
maybe a failure• Predict failure
• Reduce risk of further damage
Tell you when they need servicing
• Revenue
• Minimise downtime
Tell you their location
• Theft
• Violations
• Audit
To provide adequate service to your customers or to make sure your assets are working for
you, they need to be able to:
23© 2019 InVMA All Rights Reserved
ASSETMINDER
• Telematics monitoring the location and status of vehicles, equipment and people is common
• Using data from assets/products in the field to diagnose problems remotely is a reality
• Companies are getting value from this data and are changing the way they work.
• BUT with many solutions:-
Narrow
Closed
Rigid
£!
They answer one question such as where your asset is or what it is doing. They work well with one type of asset such as a vehicle, a generator or a building but never all of your assets. They might just work with one manufacturer.
You can’t easily connect them to your current processes through your existing Help Desk, Service Management, Product Lifecycle Management, Customer Relationship Management, Enterprise Resource Planning, or other systems.
They are built to solve one problem and can’t grow with your needs and ideas for business improvement. Small changes require customisation rather than configuration.
The hardware and software used is relatively expensive and a one-size fits all with a reliance on one technology to satisfy all of your needs.
Traditional System Landscape
Logistics SuppliersExternal
Production Quality Inventory
Business Systems
(ERP, CAD, PLM)
Level
4
Operational Control
(MES, MOM)
Level
3
Process Monitoring
(HMI, SCADA, Test Stands)
Level
2
Control Level & Data Acquisition
(PLC, CNC, Autoclave)
Level
1,0
An Industrial Innovation Platform
complements and brings
• Unified visibility across systems, assets, devices and people
• Wrap and extend existing systems
• Actionable and comprehensive operational intelligence
• Real-time, social, mobile
• Rapid application developmentMainten. Energy
USE NEW TECHNOLOGY TO WRAP AND EXTEND
25© 2019 InVMA All Rights Reserved
ASSETMINDER
The Broad, Open and Flexible Solution
• Set up multiple Geo-fence Triggers• Locate your asset and its velocity• Use the location to trigger process
• Set your parameters• Configure your dashboard• Configure rules on asset and
system
• Alarms and configurable workflow• Analytics and Machine Learning• Email, text, vmail, tweet
• Job assignment• Time to next service counters• Knowledge base and service blog• Auditable compliance trail
Built on the secure, robust and flexible ThingWorx Platform from PTC
• Remote access and triggers• Software updates• Remote diagnosis
LOCATE
MONITORCONTROL
SERVICE INFORM
INTEGRATE
CAFM, ERP, CRM, PLM,
CAD, SLM, CPM, BI
26© 2019 InVMA All Rights Reserved
ASSETMINDER
The Broad, Open and Flexible Solution
27© 2019 InVMA All Rights Reserved
ASSETMINDER
Deliver Value and Scalability With Low Risk
Project Hurdles
Integration
Business Logic
User Interface
Database
Security
Protocol Translation
Hardware
Process Change
People Change
InVMA & AssetMinder
Out of the Box
Drag and drop
Pre-built “Mash-Ups”
Out of the Box
Out of the Box
Out of the Box
Approved Suppliers
Design Method
Change Method
Outcome
Faster
Lower Cost
Lower Risk
More Robust
Thank you
Start trace
2mm above
surface
Stop trace at
end of
countersink
Drill tip cutting
composite
Drill tip cutting
aluminium
Drill tip in air Start forming
countersink
Drilling Process Monitoring
Drilling Process
Part thickness
Burr height
Tool wear
Good / bad hole
Drilling Process Monitoring
Setting up ThingWorx
Drilling Process Monitoring
Setting up ThingWorx
Drilling Process Monitoring
Setting up ThingWorx
ThingTemplate
Drilling Process Monitoring
Setting up ThingWorx
ThingShape
Property definition
• ‘payload’ property used to listen for the MQTT payload
• On change, code snippet is run (next slide)
• ‘spindlePower’ is persisted (saved to db) as the result of the code snippet.
• Defining this property declares it as a property to be stored in db
Drilling Process Monitoring
Setting up ThingWorx
Thing
• MQTT connection from Template
• Data management functionality (parsing and posting) comes from Shape
• Properties also from Shape
• Very easy to hook up/initialize new drill Things, may just need to change MQTT topic
Drilling Process Monitoring
Setting up ThingWorx
...this part is TBC
There is still the question of how to associate the post-processing measurements
Drilling Process Monitoring
Analysis
During Processing
Start trace
2mm above
surface
Stop trace at
end of
countersink
Drill tip cutting
composite
Drill tip cutting
aluminium
Drill tip in air Start forming
countersink
Drilling Process Monitoring
Analysis
Post Processing
• After drilling we need to check the quality of the hole, therefore we take post-drill-time measurements
• Measurements taken manually…
• This is important so that we have “ground truth” data to use for Machine Learning
Drilling Process Monitoring
Analysis
Start trace
2mm above
surface
Stop trace at
end of
countersink
Drill tip cutting
composite
Drill tip cutting
aluminium
Drill tip in air Start forming
countersinkAnalysis Approach:
• Extract features from torque
data
• Attempt:
• pattern based recognition
approaches to predict key
measurements
• Simpler “deterministic”
measurements (e.g. time
where torque is below ‘x’
for gap test and
thicknesses)
• Will include ‘environment’
features, such as drill bit age
• Finally predict: Part thickness,
burr heights, gaps + thickness,
countersink height
Thank you
1. Identify business problem that causes pain. (unreliability, poor efficiency, variable quality)
2. Identify simple project to prove value that can be scaled if successful
3. Select sensors and interfaces required to collect data
4. Select method of communication to cloud (wired network, 4G…)
5. Choose application software that provides value
6. Choose cloud provider that will host application software
Conclusion