© Siemens AG 2011. All rights reserved. How to create a PAT data management platform to support Continuous Production 8th Annual PAT and QbD Conference, London, February 16 th , 2011 Jan Verelst
© Siemens AG 2011. All rights reserved.
How to create a PAT data management platform to support Continuous Production8th Annual PAT and QbD Conference,London, February 16th, 2011
Jan Verelst
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 2 February 16, 2011 8th Annual PAT/QbD Conference - London
Back in time
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 3 February 16, 2011 8th Annual PAT/QbD Conference - London
What caused this disaster ?
Information :
� Hidden
� Critical
� Knowledge
Knowledge
Data
Information
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 5 February 16, 2011 8th Annual PAT/QbD Conference - London
Oral Solid Dosage today
Granulator
Dryer
Coating
Delay Delay Delay
1 to 2
months to
release
Blender
Raw Material
Tablet press
Delay
Blender
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 6 February 16, 2011 8th Annual PAT/QbD Conference - London
PAT: Key Enabler for Continuous Manufacturing and Real Time Product Release
Raw Material
Dryer
Granulator Tablet press
Blender
Qualitycheck
SIPATSIPAT
Coating
Right First Time
Real time release
In/At line check
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 7 February 16, 2011 8th Annual PAT/QbD Conference - London
Process feed
Ho
ld /
rele
ase
LabLabTemp., Speed,
Liquid addition,
Compression Force, …
Process output
LIMSLIMS
Sample
Classic controlClassic control
Process Analyzer
monitoringmonitoring
processprocess datadata
monitoringmonitoring
product product qualityquality
PATPAT
mathematicalmathematical
translationtranslation
RealReal--timetime
releasereleaseAdvanced Advanced
ControlControl
Quality build in by design
Right first time
PAT Basics – Holistic Approach
Closed loopClosed loop
controlcontrol
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 8 February 16, 2011 8th Annual PAT/QbD Conference - London
SIPAT functions
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 9 February 16, 2011 8th Annual PAT/QbD Conference - London
SIPAT Collectors
Collector = Interface to read data from external systems as input to SIPAT
Method
Standard available collectors within SIPAT
� Instrument collectors (dedicated interfaces to range of analytical instruments)
� Process collector (OPC DA Client)
� File collectors (XML, SPC, JCAMP)
� PAT Collector (results from other SIPAT Method)
� Databrowse collectors (ODBC, SQL, OLEDB)
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 10 February 16, 2011 8th Annual PAT/QbD Conference - London
Process Analyzers
Standard analyzer connection available with actual productivity pack:
� Kaiser Raman (Raman RXN1, RXN3, RXN3L, RXN4R and PhAT); Kaizer’s Analyser Control V4.2
� Thermo Fischer Antaris and Antaris II family (via Result software)
� PAA Ltd with GranuMet XP (acoustic emission measurement), Version 4
� Bruker NIR via OPUS V6.1 software
� Bruker NIR via OPUS 6.5 embedded DLL
� Malvern Insitec – Laser Diffraction PSD (RT Sizer software)
� Malvern Morphology G3
� Mettler Toledo collector for ReactIR 4000 (FT-IR) � via IC4 software
� Mettler Toledo MonARC (FT-IR) � via IC4 software
� Mettler Toledo Lasentec FBRM (PSD) � via IC4 software
� Zeiss collector MCS 500 serie (photodiode array UV/visible spectrophotometer)
� Zeiss collector MCS 600 serie (NIR)
� Brimrose Luminar 5030 and 4030 AOTF-NIR
� Expo ePAT 601 (including an Axsun NIR)
� Dr. Schleuniger tablet tester (Tandem using Dr. Schleuniger – Pharmatron - Bruker)
� ABB FTPA2000
� Bruker LancIR via embedded DLL 6.5.92
New productivity packs will be released on a frequent basis
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 12 February 16, 2011 8th Annual PAT/QbD Conference - London
SIPAT: data collection in real time
Alignment of Collectors
NIR
Process value(e.g. Temp)
Measurement
duration
Start batch
Startup off set Measurement rate
Negative Validity
Range
REFERENCE COLLECTOR
Positive Validity Range
Aggregation function :
Average
Last Value
Maximum
Middle
Minimum
First Value
Nearest to reference
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 13 February 16, 2011 8th Annual PAT/QbD Conference - London
SIPAT: data collection in real timePrinciple
�Data alignment
�Contextualization
�Rules separating non-representativedata
SIPAT
Raw Material
Characteristics
LIMS
Process
+
Context data
SC
AD
A
Analytical
Data
NIR, PSD
Analyzer
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 14 February 16, 2011 8th Annual PAT/QbD Conference - London
SIPAT functions
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 15 February 16, 2011 8th Annual PAT/QbD Conference - London
SIPAT: data monitoring and control in real timeApplication
SIPAT brings all data (in-line, at-line, on-line and off-line) in relation for prediction
Quelle: Siemens IT Solutions and Services
Quantitative prediction
Qualitative prediction
Real Time
Prediction
Alig
nm
ent
Meta &
context data
Raw Material
Data / IPC data (Moisture, …)
Other Application data
Instrument data
(NIR, Raman, PSD, …)
Process data (pH, Temp, Pressure, …)
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 17 February 16, 2011 8th Annual PAT/QbD Conference - London
SIPAT Data Model
SIPAT MethodSIPAT MethodCollector DataCollector Data CalculationsCalculations
VisualizationsVisualizations
Info DataInfo Data
Perform Calculations
Make Predictions
� using
� Models
� Basic Calculations
� Advanced Calculations
Input from
� Analytical Instruments
� Process/Automation Level
� Raw Materials
� Other Systems
� Previous process step
� using Collectors
Static data
Semi-static data
Meta-data
Manual/automatic entry
Configure operator screens
Using Charting types
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 18 February 16, 2011 8th Annual PAT/QbD Conference - London
SIPAT: data monitoring and control in real timePrinciple
�Data alignment
�Contextualization
�Rules separating non-representativedata
�Statistical RT Models
�e.g. PLS, PCA
�Chemometrics(SIMCA, Matlab,
Unscrambler,
PharmaMV,…)
engine embedded
in SIPAT
�CQA
Aggregation of CQA over process context
- Average,
- Standard deviation
On line Monitoring
Out of control detection
Deviation tracking
AP
CR
TR
SIPAT
Raw Material
Characteristics
LIMS
Process
+
Context data
SC
AD
A
Analytical
Data
NIR, PSD
Analyzer
Critical
to
QualtyAttribute
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 19 February 16, 2011 8th Annual PAT/QbD Conference - London
SIPAT functions
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 20 February 16, 2011 8th Annual PAT/QbD Conference - London
SIPAT: data miningPrinciple
SIPAT
Analyser Data ( NIR, Raman, PSD, …)
Process Data Predicted data
Off- line lab data
(RM,process samples,…)
External application
Meta & context data Other data
- Batch number
- Trial number
- Process step
- Product
- Campaign
…. Processoptimization
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 21 February 16, 2011 8th Annual PAT/QbD Conference - London
SIPAT: data miningModel creation & process optimization
Turning data into knowledge
� Improved process understanding
� Base of QbD
Model Builder
� Integrate all relevant PAT data
� Send to Chemometrics application
� Umetrics - Simca
� Camo - Unscrambler
� Matlab
� Perceptive Engineering - PharmaMVProcess
optimization
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 22 February 16, 2011 8th Annual PAT/QbD Conference - London
SIPAT functions
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 23 February 16, 2011 8th Annual PAT/QbD Conference - London
ReportingAutomated or ad hoc
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 24 February 16, 2011 8th Annual PAT/QbD Conference - London
ReportingTable output
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 25 February 16, 2011 8th Annual PAT/QbD Conference - London
ReportingGraphical output
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 26 February 16, 2011 8th Annual PAT/QbD Conference - London
ReportingGraphical output - spectrum
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 27 February 16, 2011 8th Annual PAT/QbD Conference - London
ReportingGraphical output – batch comparison
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 28 February 16, 2011 8th Annual PAT/QbD Conference - London
Why SIPAT ?
SIPAT
Increase process understanding / manage variability
Bring all data (in-line, at-line, on-line and off-line) in relation
Provide on-line predictions on product/process quality
Shift from lab testing towards in process testing
1 common interface for all PAT tools
Close the gap between R&D and Manufacturing
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 29 February 16, 2011 8th Annual PAT/QbD Conference - London
Project context
GEA Pharma Systems
� Process Know-How
� Supply specific equipment +
instrumentation to handle a specific
production operation
� Focus on equipment specific
optimization
� Equipment automation specification
(PLC, HMI, drive, instrumentation,
...)
� Equipment validation
� Equipment specific 21 CFR Part 11
functionality
� Worldwide Packaged Unit Support
Siemens� Industry know How
� Standard platform
� Scalable solutions
� Easy integration
� GMP supporting functions
� Cost effective solutions
� Long term Support (investment protection)
� Worldwide Product Support
End-User
� Product Know-How
� Specify required machinery to fulfill a specific production operation
� Equipment must comply with the own
specific norms and standards
� Focus on the optimized
functioning of the
complete production line
� Easy integration of
the control level into
Plant-IT
End-user
Supplier
OEM
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 30 February 16, 2011 8th Annual PAT/QbD Conference - London
PAT Instruments
NDC - Moisture
NIR (J&M LHP)
Moisture+ C. uniformity
Malvern
Particle Size
At line analyzer
Weight, Hardness,Thickness, Assay
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 31 February 16, 2011 8th Annual PAT/QbD Conference - London
MES LIMS
SCADA
PLC / Motion Control
PROCESS
ANALYZERS
SIPAT
Line
SITE
(2)
(5)
(1) (3)
(4)
(6)
High Level System Architecture
�Control Analyzers
�Multivariate Data handling
�On-line monitoring of the Critical to Quality Attributes (CQA)
�Deviation Tracking CQA
�Out of Control Detection & input to APC
� Supervisory & Control
� Alarm handling
� Line Recipe handling
� Trending Univariate Data
� Deviation Tracking Critical Process Parameters
�Real Time Releasereporting
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 32 February 16, 2011 8th Annual PAT/QbD Conference - London
Traditional PAT in solid dosage
•
Drying
control
Process
parameters
Milling Blending
Lubrication
control
Process
parameters
Compression
control
Process
parameters
Coating
control
Process
parameters
Dispense
& Blend
control
Process
parameters
� API
� Excipients
Wet
Granulation
control
Process
parameters
Liquid addition
Lubricant excipient
Coating solution
Moisture
Bulk Physical
Defects
Test against
specifications
Bulk Physical
Defects
Content
Uniformity
Packaging
control
Process
parameters
Particle size
Sampling &
Off-line analysis
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 33 February 16, 2011 8th Annual PAT/QbD Conference - London
End product quality predictions
•
Lubricant excipient
Drying
control
Process
parameters
Blending /
Lubrication
control
Process
parameters
Compression
control
Process
parameters
Coating
control
Process
parameters� Api
� Excipients
Dispense
& Blend
control
Process
parameters
Liquid addition
Wet
Granulation
control
Process
parameters
Coating solution
Moisture
(NIR)
Assay
Dissolution
Disintegration
Packaging
control
Process
parameters
Particle size
(Malvern)Weight
Hardness
Thickness
Input material
characteristics
Visual
Inspection
Coating
thickness
Content
Uniformity
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 34 February 16, 2011 8th Annual PAT/QbD Conference - London
Lot 1 Lot 2
Lot 1 Lot 2
Lot 1 Lot 2
F1C1 F1C2 F1C3 F1C4 F1C5 F1C6 F2C1 F2C2 F2C3 F2C4 F2C5 F2C6 F3C1 F3C2 F3C3 F3C4 F3C5 F3C6 F4C1 F4C2 F4C3 F4C4 F4C5
F1C1 F1C2 F1C3 F1C4 F1C5 F1C6 F2C1 F2C2 F2C3 F2C4 F2C5 F2C6 F3C1 F3C2 F3C3 F3C4 F3C5 F3C6 F4C1 F4C2 F4C3 F4C4
F1C1 F1C2 F1C3 F1C4 F1C5 F1C6 F2C1 F2C2 F2C3 F2C4 F2C5 F2C6 F3C1 F3C2 F3C3 F3C4 F3C5 F3C6 F4C1 F4C2 F4C3
F1C1 F1C2 F1C3 F1C4 F1C5 F1C6 F2C1 F2C2 F2C3 F2C4 F2C5 F2C6 F3C1 F3C2 F3C3 F3C4 F3C5 F3C6 F4C1 F4C2
API
Excipient 1
Excipient 2
Granulation
Ingredients
Drum 1 Drum 2 Drum 3Output
Time
Drying
Millling/Blending
Compression
Coating
Material Residence Time distribution Product Plug tracking
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 35 February 16, 2011 8th Annual PAT/QbD Conference - London
Lot 1 Lot 2
Lot 1 Lot 2
Lot 1 Lot 2
F1C1 F1C2 F1C3 F1C4 F1C5 F1C6 F2C1 F2C2 F2C3 F2C4 F2C5 F2C6 F3C1 F3C2 F3C3 F3C4 F3C5 F3C6 F4C1 F4C2 F4C3 F4C4 F4C5
F1C1 F1C2 F1C3 F1C4 F1C5 F1C6 F2C1 F2C2 F2C3 F2C4 F2C5 F2C6 F3C1 F3C2 F3C3 F3C4 F3C5 F3C6 F4C1 F4C2 F4C3 F4C4
F1C1 F1C2 F1C3 F1C4 F1C5 F1C6 F2C1 F2C2 F2C3 F2C4 F2C5 F2C6 F3C1 F3C2 F3C3 F3C4 F3C5 F3C6 F4C1 F4C2 F4C3
F1C1 F1C2 F1C3 F1C4 F1C5 F1C6 F2C1 F2C2 F2C3 F2C4 F2C5 F2C6 F3C1 F3C2 F3C3 F3C4 F3C5 F3C6 F4C1 F4C2
API
Excipient 1
Excipient 2
Granulation
Ingredients
Drum 1 Drum 2 Drum 3Output
Time
Drying
Millling/Blending
Compression
Coating
Material Residence Time distribution Real Time Release floating window
Release window
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 36 February 16, 2011 8th Annual PAT/QbD Conference - London
SIPAT: data monitoring and control in real timePrinciple
�Data alignment
�Contextualization
�Rules separating non-representativedata
�Statistical Real Time Models
�e.g. PLS, PCA
�Chemometrics
(SIMCA, Matlab,
Unscrambler,
PharmaMV,…)
engine embedded in SIPAT
�CQA
Aggregation of CQA over process context
- Average,
- Standard deviation
On line Monitoring
Out of control detection
Deviation tracking
AP
CR
TR
SIPAT
Raw Material
Characteristics
LIMS
Process
+
Context data
SC
AD
A
Analytical
Data
NIR, PSD
Analyzer
Critical
to
QualtyAttribute
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 37 February 16, 2011 8th Annual PAT/QbD Conference - London
37
Data Driven Manufacturing
• SiPAT allows to measure inline the CQA’s, real time
• Increased number of measurements
Test case of 50hrs run @25kg/hr=1250kg material
- Moisture measured 960 times
- Particles size distribution
measurements: 960 times
- Content uniformity: 100 times
- Tablet weight: each tablet (1,8 Mio)
�Enhanced product security
�Data correlation across units
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 38 February 16, 2011 8th Annual PAT/QbD Conference - London
Results of the trials
Main findings of the feasibility study:
� Reduced process development time
� “Unlimited” number of reference batches as a basis for model creation
� Real Time Release enabling
� Quality information available on tablet level rather than on batch level
� Increased process understanding
� Through Real Time Quality Assurance
� Time-based process: scalability improvement
� Savings in construction, space, energy, maintenance
� Increase of yield and equipment efficiency
� Reduction of scrap, waste & rework
� Reduction of raw material
� Reduction of human interference (cost, safety)
� …
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 39 February 16, 2011 8th Annual PAT/QbD Conference - London
Article available on request
© Siemens AG 2011. All rights reserved.
Industry Sector
Page 40 February 16, 2011 8th Annual PAT/QbD Conference - London
Better process understanding…leads to a more efficient result !
© Siemens AG 2011. All rights reserved.
Jan Verelst
Siemens
Industrial IT
+32 475 432 560
Thank you for your attention!