recipe driven, data visualization
TRANSCRIPT
© Copyr i gh t 2014 OSIso f t , LLC.
Presented by
Recipe driven,
data visualization
from lab to
commercial
Barry Higgins and Koen Paeshuyse
© Copyr i gh t 2014 OSIso f t , LLC.
Agenda
• Introduction
• Goal and Challenges
• S88 recipe strategy
• Recipe integration into systems
• OSIsoft PI AF/EF setup
• Data visualization
• Conclusions
2
© Copyr i gh t 2014 OSIso f t , LLC.
Introduction
3
Cardiovascular & Metabolism
Immunology Infectious Diseases & Vaccines
Neuroscience Oncology
20%
80% Internal
External
API SM* Development
* API: Active Pharmaceutical Ingredient
SM: Small Molecules
© Copyr i gh t 2014 OSIso f t , LLC.
Goal and Challenges
4
Building a Knowledge Based Organization
Common Data Warehouse
Enables leveraging of information
across systems to help drive data
driven decisions
Common Recipe Approach
Ensures that we all “speak the
same language” promoting
consistency
DPD
AD
API
Electronic
Lab Notebook
Data Infrastructure
Data Visualization
© Copyr i gh t 2014 OSIso f t , LLC.
S88 – What is it?
5
Short for ANSI/ISA-88
• ANSI is American National Standards Institute
• ISA is Instrumentation, Systems, and Automation Society
An industry standard addressing batch process control
S88 provides a consistent set of standards and terminology for batch control
© Copyr i gh t 2014 OSIso f t , LLC.
S88 – Process Model
6
Process A Process is made up of an ordered set of one or more
Process Stages
Process
Stage
A Process Stage is made up of an ordered
set of one or more Process Operations
Process
Operation
A Process Operation is made up of
an ordered set of one or more Process
Actions
Process
Actions\
Parameters ANSI/ISA-88.00.03-2003
Part 3 Figure 6
© Copyr i gh t 2014 OSIso f t , LLC.
S88 – Process Model
7
Process A Process is made up of an ordered set of one or more
Process Stages
Process
Stage
A Process Stage is made up of an ordered
set of one or more Process Operations
Process
Operation
A Process Operation is made up of
an ordered set of one or more Process
Actions
Process
Actions\
Parameters ANSI/ISA-88.00.03-2003
Part 3 Figure 6
P S
O A
Process
Process
Stage
Process
Operation
Process
Actions\
Parameters
© Copyr i gh t 2014 OSIso f t , LLC.
S88 recipe model for API SM Dev
8
ACTION
PARAMETERS
PROCESS
OPERATION
PROCESS
ACTION
Charge
Parameter
Name Target Unit Actual
Material
Quantity
Duration
pH
PROCESS
Project
PROCESS
STAGE
Product
Product
Temperature Control
Wait
Reaction (#)
Preparation (#)
Work up (#)
Agitate
© Copyr i gh t 2014 OSIso f t , LLC.
S88 recipe model for API SM Dev
9
ACTION
PARAMETERS
PROCESS
OPERATION
PROCESS
ACTION
Charge
Parameter
Name Target Unit Actual
Material
Quantity
Duration
pH
PROCESS
Project
PROCESS
STAGE
Product
Product
Temperature Control
Wait
Reaction (#)
Preparation (#)
Work up (#)
Agitate
© Copyr i gh t 2014 OSIso f t , LLC.
S88 Recipe KM* Strategy
10
Operation
Operation Operation
Operation
Operation Operation
Site Executable Recipe
Regulatory Filing = General Recipe
CONTENT Critical Parameters and Steps Structured in S88 Format
EXECUTION Data generated in S88 Format Contains additional site details
Process
Process
Stage
Process
Operation
Process
Action
VISUALIZATION Critical Parameters and Steps Context rich data
* KM: Knowledge Management
© Copyr i gh t 2014 OSIso f t , LLC.
Dataflow in the Process Chemistry lab
11
Execute experiment
Samples during experiment incorporated in same ELN experiment
LAB
LAN
SERVER
Experiment done
iC Data Center by Mettler
Output files
SN – ID Username
Start experiment on Automate Lab Reactor
View/Reprocess on Client Automatic update on server
csv file contains ALR recipe and data (to be XML format)
IDM
ALR and PAT data
iControl
OFFICE
S88 Executed Recipe Based on Approved Operatios and Actions
Analytical results
LC
KF
NMR
Initiates ELN experiment in Symyx Notebook
Scientist Chemistry Table
Experiment master data
© Copyr i gh t 2014 OSIso f t , LLC.
S88 integration in OSIsoft PI System
14
ELN
Process Analytical Technology
tools
Automated Lab
Reactor
© Copyr i gh t 2014 OSIso f t , LLC.
OSIsoft PI System Architecture
• Existing infrastructure utilized.
• Development and Production environments.
• 74 Automated Lab Reactors (ALRs) across 20 of labs.
15
J&J Business LAN
Beerse J&J Automation VLAN
Pilot Plant
DeltaV 7.x
PI Base
1K
Automation Firewall
(VLANs) Failover
(Phase II)
Primary PI Interface
Node (Beerse)
Enterprise PI Server
(Primary)
Ø Windows Server 2003
Ø Terminal Services
Ø PI Enterprise 2010
Ø Interface Monitoring (PI ACE).
Janssen Beerse
PI System Architecture
TQS Integration Ltd
Rev 0.4
Sandbox PI Server
Ø Windows Server 2003
Ø Sharepoint Server (MOSS)
Ø PI WebParts
Ø RtReports
Enterprise PI Server
(Secondary)
Ø Windows Server 2003
Ø Terminal Services
Ø PI Enterprise 2010
Ø Interface Monitoring (ACE)
PI Web Parts Server
Ø Windows Server 2003
Ø Sharepoint Server (MOSS)
Ø PI tWebParts
Ø RtReports
Secondary PI Interface
Node (Beerse)
Sandbox PI Interface
Node (Beerse)
Higi Automation Isolation Network
Failover
(Phase II)
Secondary PI
Interface Node
(Higi)
Primary PI
Interface Node
(Higi)
Outside Clean Area
Coating Machine (PP-015)
Fixed Wire Devices
A
Coating Machine (PP-016)
Top Granulator (PP-002)
Livepoint Server
Ø Windows Server 2008
Ø Livepoint
PD Plant – DCS PD Plant - CRS
S7 PLC’s (16 No)
Win CC / PI
Reporting Nodes
PI ENT
5KReporting Nodes
PCS7
SIPCIP
10 Devices
PD - Standalone
WB0
1
RH0
2
Miniplant
PCS7
PST Plant
S7 PLC’s
Win CC / PI
PI ENT
5K
Development
Labs
ALR
Mettler ALR
© Copyr i gh t 2014 OSIso f t , LLC.
PI Event Frames 1 – ALR Data
• CSV generated from ALR, containing both continuous and event data.
• PI UFL Interface developed to write data to tags.
• PI System configured to generate PI Event Frames
17
UFL EF GEN
© Copyr i gh t 2014 OSIso f t , LLC.
PI Event Frames 2 – ELN Data
• XML output from ELN exported every night.
• Stored Procedures run via PI RDBMS.
18
RDBMS
EF GEN
© Copyr i gh t 2014 OSIso f t , LLC.
PI Event Frames 3 – ELN and ALR Data
• Challenge – ELN and ALR data are not time
linked.
• Goal – Create a single event frame that contains
and combines the ELN and ALR actions and all
its data for analysis
19
© Copyr i gh t 2014 OSIso f t , LLC.
PI Event Frames 3 – Bringing it all together
• Data and action
timestamps.
• ELN Timestamps adjusted.
• Stored procedure call
writes to PI System.
• Entire experiment is
captured in one event.
20
© Copyr i gh t 2014 OSIso f t , LLC.
Visualizing the data
• Ability to view, analyze and share experiment data away from the lab.
• Web based tools preferred.
• PI Coresight live point.
21
© Copyr i gh t 2014 OSIso f t , LLC.
PI System lab data integration overview future
22
Pilot and
Commercial
Data
iControl,
touch panel
PAT
ELN
PI Interfaces
Analytical
Data
© Copyr i gh t 2014 OSIso f t , LLC. 23
Internal External External External External
Source Independent Cloud-based Data Presentation Layer
? ? ? ? ?
• Very disparate systems landscape – now
including MES, ELN, Manual Data Entry…
• Ability to share contextualized process data with
external partners is critical.
• Require capability to capture and aggregate data
for visualisation, reporting & analysis
Value of the OSIsoft PI System: Future
© Copyr i gh t 2014 OSIso f t , LLC.
• Visualizing the data outside of
the lab.
• By applying the S88 structure,
lab to commercial comparisons
are possible
Solution Results and Benefits
Recipe driven, data visualization
from lab to commercial
Business Challenge
• Streamline and implement
the S88 recipe strategy in the
different variety on lab
systems we have in place
• Combining the data from
multiple data sources
• OSIsoft PI System as our
global Data Infrastructure is
the key enabler towards this
recipe concept.
• PI Asset Framework and
PI Event Frames components
to bring the data together
• Enabling scientists to generate and capture
consistent process data in a standard recipe context
to generate information and knowledge from
experiments.
• Enabling efficient transfer from the lab through our
pilot plants, commercial operations and ultimately to
our patients, capturing and building knowledge
throughout.
24
OSI PI
© Copyr i gh t 2014 OSIso f t , LLC.
Acknowledgments
Ryan Bass, Nick Dani, Adam Fermier, Alison
Harkins, Jim Kenyon, Mike McGorry, Luc Moens,
Terry Murphy, Chris Nichols, Gaby Wevers
…and many others
25
© Copyr i gh t 2014 OSIso f t , LLC.
Thanks
• Barry Higgins – [email protected]
• Koen Paeshuyse – [email protected]
• Janssen PDMS
26
© Copyr i gh t 2014 OSIso f t , LLC. 27
Questions
Please wait for
the microphone
before asking
your questions
State your
name &
company